Report No. 24221 -GU Guatemala Poverty in Guatemala Februayra 20, 2003 Poverty Reduction and Economic Management Unit Human Development Sector Management Unit Latin America and the Caribbean Region ,Document of the Wtorld Bank CURRENCY EQUIVALENTS US$1 = GTQ7.70 (July 1, 2000) US$ 1= GTQ7.78 (Current) FISCAL YEAR December 31 - January I ACRONYMS AND ABBREVIATIONS ANACAFE = National Coffee Association AP = All poor ARI = Acute respiratory infections BNI = Basic Needs Indicators CA = Central America CONRED = Coordinadora Nacional para la Reducci6n de Desastres DHS = Demographic and Health Survey ENCOVI = Encuesta Nacional de Condiciones de Vida (Living Standards Measurement Survey) FPL = Full poverty line GUAPA = Guatemala Poverty Assessment Program (World Bank) HDI = Human Development Index HIPC = Highly Indebted Poor Countries IGSS = Instituto Guatemalteco de Seguridad Social INCAP = Instituto de Nutricion de Centro Ame'rica y Panamd INE = National Statistics Institute; Instituto Nacional de Estadistica - Guatemala KAI, KA2 = Kaqchiqel communities in QPES study KII, K12 = K'iche communities in QPES study Li, L2 = Ladino communities in QPES study LAC = Latin America and the Caribbean Region MI, M2 = Mam communities in QPES study MAGA = Ministerio de Agricultura, Ganaderfa y Alimentaci6n MDG = Millennium Development Goals MECOVI = Program for the Improvement of Surveys and Measurement of Living Conditions in Latin America and the Caribbean MINEDUC = Ministry of Education MINUGUA = United Nations Verification Mission - Misi6n de Verificaci6n de las Naciones Unidas en Guatemala MSPAS = Ministry of Public Health and Social Assistance NP = Non-poor PRONADE = Programa Nacional de Autogesti6n para el Desarollo Educativo PRSP/ERP = Poverty Reduction Strategy Paper PTA = Parent-teachers association (school committees) QEI, QE2 = Q'eqchi communities in QPES study QPES = Qualitative Poverty and Exclusion Study (ECEP) SA = Social Assistance SAS = Secretaria de Asuntos Sociales de la Municipalidad de Guatemala SBS = Secretaria de Bienestar Social de la Presidencia de la Republica SEGEPLAN = Secretaria de Planificaci6n y Programaci6n Si = Social Insurance SIAS = Sistema Integral de Atenci6n de Salud SIAF = Integrated Financial Management System SOSEP = Secretaria de Obras Sociales de la Esposa del Presidente SP = Social Protection UNDP = United Nations Development Program URL = Universidad Rafael Landfvar VAT = Value Added Tax WHO = World Health Organization XPL = Extreme poverty line XP = Extreme poor Vice President David de Ferranti Country Director Jane Armitage LCSPR Director Ernesto May LCSHD Director Ana-Maria Arriagada Sector Manager-Poverty Norman Hicks Sector Manager-Social Protection Christopher Chamberlin Sector Leader Helena Ribe Resident RepresentativeTask Manager Eduardo Somensatto Task Manager Kathy Lindert Guatemala Poverty Assessment TABLE OF CONTENTS MAIN REPORT EXECUTIVE SUMMARY ....................................i Chapter 1 - Introduction .....................................1 Context for Report: the GUAPA Program .....................................1 Objectives of Report .................... . . . . ... . . . 2 Analytical Framework ............... . . ; ;,.2 Information Sources: Integrating Qualitative and Quantitative Data ...............................................3 Overview of the Report ........................... . , , .4 PART 1 - THE MAGNITUDE AND CAUSES OF POVERTY ............................................................7 Chapter 2 - "The Problem:" Poverty and Social Indicators .........................................................................7 Monetary Indicators of Poverty and Inequality in Guatemala .........................................................7 Non-Monetary Indicators of Poverty and Living Conditions in Guatemala .................................. 13 Perceptions of Poverty and Welfare in Guatemala ........................................................................ 16 Chapter 3 - The Determinants of Poverty and Inequality: Endowments and Household Characteristics.. 23 The Determinants of Poverty ............................................................................. 23 The Sources of Inequality ............................................................................ 25 Chapter 4 - Historical and Contextual Factors . 30 Diversity, Ethnicity and Isolation ................................. 30 Exclusionary Forces in Guatemala's Historical Pattern of Development ...................................... 33 Guatemala's 36-Year Civil War: Significant Costs for Long-Term Development ....................... 36 The 1996 Peace Accords: Towards a More Inclusive Course of Development ............................ 37 PART 2 - KEY CHALLENGE: BUILDING OPPORTUNITIES AND ASSETS ............................. 42 Chapter 5 - Growth and Poverty .................................................................. 42 Poverty and Growth Over Time .................................................................. 42 Poverty and Growth in the Future: Targets and Projections .......................................................... 45 Sunmnary of Key Issues and Priorities .................................................................. 47 Chapter 6 - Livelihoods, Labor Markets, and Rural Poverty .................................................................. 49 Incomes and Inequality .................................................................. 49 Labor Markets ................................... ; 50 Rural Poverty and Livelihoods ................................... 54 Summary of Key Issues and Priorities ................................... 59 Chapter 7 - Education and Poverty ................................... 62 Sectoral Overview ................................... 62 Educational Stock, Coverage and Equity ................................... 63 Internal Efficiency ................................... 65 Disparities in the Quality of Education ................................... 67 Barriers to Enrollment and Attainment ................................... 69 The Returns to Education ................................... 73 Public Spending and Equity .......................................................... 74 Summary of Key Issues and Priorities ........................................................... 77 Chapter 8 - Health, Malnutrition and Poverty ........................................................... 81 Main Health Challenges .......................................................... 81 A Focus on Malnutrition: A Red Flag! .......................................................... 83 Access to Health Care ........................................................... 86 Supply-Side Factors .......................................................... 88 Demand-Side Factors .......................................................... 91 Summary of Key Issues and Priorities .......................................................... 95 Chapter 9 - Basic Utility Services and Poverty .......................................................... 98 The Benefits of Basic Services .......................................................... 98 Sectoral Overview ........................................................... 98 Access and Equity .......................................................... 99 Barriers to Improved Access: Supply vs. Demand-Side .......................................................... 102 Achieving Universal Coverage: Time and Costs .......................................................... 103 Costs and Subsidies for Basic Services .......................................................... 103 Summary of Key Issues and Priorities .......................................................... 105 Chapter 10 - Transport, Poverty and Isolation .......................................................... 107 Access to Roads .......................................................... 107 Road Quality and Dependability .......................................................... 108 Public Transport ...........................................................111 The Effects of Isolation: Limited Access to Services and Opportunities .................................... 111 Summary of Key Issues and Priorities .......................................................... 113 PART 3 - KEY CHALLENGE: REDUCING VULNERABILITY .................................................. 114 Chapter 11 - Vulnerability and Vulnerable Groups .......................................................... 114 Shocks as a Source of Vulnerability in Guatemala ........................................................... 115 Vulnerable Groups in Guatemala .......................................................... 123 Summary of Key Issues and Priorities .......................................................... 128 Chapter 12 - Social Protection - Private Transfer and Poverty .......................................................... 130 Overview of SP Programs: Types, Magnitudes and Spending .................................................... 130 Coverage of SP Programs and Private Transfers .......................................................... 135 Distributional Incidence (Targeting Outcomes) of SP Programs ................................. ............... 138 Overall Effectiveness of SP Programs ........................................................... 140 Impact on Poverty and Inequality .......................................................... 142 Cost-Benefit Analysis .......................................................... 142 Summary of Key Issues and Priorities ........................................................... 143 PART 4 - KEY CHALLENGE: EMPOWERMENT .......................................................... 146 Chapter 13 - Building Institutions and Empowering Communities ......................................................... 146 Public Sector Management .......................................................... 146 Governance .......................................................... 150 Community Participation and Social Capital ........................................................... 153 Role of Other Actors .......................................................... 157 Summary of Key Issues and Priorities .......................................................... 158 PART 5 - CONCLUSIONS AND RECOMMENDATIONS ............................................. 161 Chapter 14 - Priority Actions to Reduce Poverty ................................................. 161 A Broad Agenda for Poverty Reduction in Guatemala ................................................. 162 Some Progress and Key Issues ................................................. 163 Priority Actions for Poverty Reduction in Guatemala ................................................. 164 Priority Target Groups for Poverty Reduction in Guatemala ................................................. 168 Monitoring Poverty Reduction Efforts ................................. 172 REFERENCES FOR MAIN REPORT .............................. 175 TEXT TABLES 2.1 Poverty in Guatemala, 2000 - Poverty Indicators by Welfare Measure. 8 2.2 Poverty Patterns in Guatemala .1 2.3 Inequality (Gini coefficients) .13 2.4. Trends in the Human Development Index, 1975-97 .15 2.5 Perceptions of Welfare from 10 Rural Villages, QPES .17 2.6 Perceptions of Poverty from 10 Rural Villages, QPES .18 2.7 Summary Report Card for Monetary and Non-Monetary Indicators of Poverty and .21 Living Conditions 3.1 The Correlates of Poverty .26 3.2 Decomposition of Inequality in Guatemala .28 4.1 Language Ability by Socio-Economic Group .31 4.2 Performance of Select Peace Monitoring Indicators - A Snapshot for the Years 2000, 2001 . 39 5.1 Average Real Growth Rates, 1950s-1990s .43 5.2 Main Macroeconomic Indicators, 1997-2001 .44 5.3 Pro-Poor Growth? The Sectoral Pattern of Employment and Annual Growth (in %) .45 5.4 GDP Growth and Poverty Reduction .46 5.5 GDP Growth and Non-Monetary Indicators .46 5.6 Meeting Poverty Strategy and Millennium Targets: Hit or Miss? .47 6.1 Income Sources, by Consumption Quintiles .50 6.2 Rural Poverty by Land Status and Main Source of Income .54 6.3 Types of Crops Produced, by Type of Household .58 7.1 Net Enrollment Rates, by Level and Group .65 7.2 The Direct Cost of Schooling .73 7.3 Returns to Education, by Gender .74 7.4 Public Expenditures on Education .75 7.5 Distributional Incidence of Public Spending on Education and Educational Support .77 (Demand-Side) Programs 8.1 International Comparison of Various Health Indicators, 1999 .82 8.2 Decomposition of Coverage Deficit - Access to Health Facilities .88 8.3 Public Spending on Health, 1995-2000 .89 8.4 Distributional Incidence of Public Spending, by Facility .90 8.5 Use of Different Health Facilities by People with Type of Insurance Coverage .90 8.6 Share of Individuals Covered by Health Insurance ..................................... 91 8.7 Utilization of Health Services by Children with Diarrhea or ARI .92 8.8 Utilization of Health Services by Individuals ................................ 92 8.9 Prenatal Care and Treatment During Delivery, Pregnant Women .93 8.10 Average Cost of Prenatal and Delivery Care, by Type of Facility .93 9.1 International Comparisons Coverage of Basic Services (Percentage of Households with. 99 Access) 9.2 Coverage of Basic Services, by Area and Quintile .................................... 100 9.3 Probability that an Unserved Household was Connected .101 9.4 Coverage Gap for Modern Utilities ................................................................. 102 9.5 How far Away is Universal Coverage? ..................... ............................................ 103 10.1 Access to Roads by Type of Road ................................................................... 109 11.1 Shocks Reported in the QPES: Types of Shocks and Impacts .................................................... 117 11.2 Ranking the Severity of Shocks ................................................................. 119 11.3 Shocks for the Rich and Poor9 ................................................................. 120 11.4 Main Coping Strategies, by Wealth Quintiles ................................................................. 120 11.5 Main Coping Strategies and Formal Assistance in 10 Rural Villages ......................................... 121 11.6 Profile of Poverty and Vulnerability ................................................................. 124 12.1 Public Spending on Social Protection, 2000 ................................................................. 131 12.2 Overview of Public Social Protection Programs ................................................................. 134 12.3 Coverage of Social Protection Programs and Private Transfers .................................................. 136 12.4 Absolute Target Incidence of Social Protection Programs and Private Transfers ....................... 139 12.5 Relative Incidence of Social Protection Programs and Private Transfers (the "importance" ....... 140 of the transfers) 13.1 Public Spending, 1996-2001 ................................................................. 147 13.2 Distribution Incidence of Public Spending, by Sector and Socio-Economic Group ................... 148 13.3 Composite Governance Indicators, International Comparisons, 2001 ........................................ 150 13.4 Social Capital and External Assistance (Percentage of Total Communities in .......................... 157 ENCOVI Sample) 14.1 Priority Target Groups ................................................................. 170 14.2 Menu of Options and Key Actions for Poverty Reduction .......................................................... 173 TEXT BOXES 1.1 The MECOVI Program: Filling a Critical Information GAP ..........................................................1 1.2 Analytical Framework ..................................................................3 2.1 Other Studies and Surveys of Poverty ..................................................................9 2.2 Linguistic Approximations of "Welfare" .................................................................. 16 2.3 Definitions of Welfare: Examples from QPES Villages ................................................................ 17 2.4 Definitions of Poverty: Examples from QPES Villages ................................................................ 18 4.1 Who is "Indigenous"? A Historical Perspective ................................................................. 32 4.2 Life in a Finca Village: the Story of KA1 (QPES) ................................................................. 34 4.3 Rebuilding After the Violence of the 1980s: the Story of Ki 1 (QPES) ........................................ 37 6.1 Migration and Livelihoods ................................................................. 51 6.2 Rural Credit ................................................................. 55 6.3 Land Redistribution Programs ................................................................. 56 7.1 The PRONADE Program ................................................................. 63 7.2 Schooling in Guatemala: Children's Perspectives ................................................................. 68 7.3 Language, Culture and Education ................................................................. 69 8.1 Health and Ethnicity: The need for a Culturally Sensitive Approach ............................................ 94 10.1 Classification of Roads in the ENCOVI ........................... 107 11.1 Disaster Management Programs in Guatemala ................................................... 117 11.2 Catastrophic Consequences of Hurricane Mitch: the Case of LI (QPES) ................................... 118 12.1 Access to Market-Based Risk Mitigation Mechanisms .............................. ..................... 136 13.1 Internal and External Conflicts: the Case of QEI (QPES) ................................................... 153 13.2 Strong Village Bonds: the Case of MI (QPES) ................................................... 154 14.1 Poverty Map ................................................... 172 TEXT FIGURES 2.1 Tentative Trends in Poverty Over Time (1989-2000) ....................................................9 2.2 Poverty Rates by Department ................................................... 12 2.3 Inequality: % of Total Consumption Received by each Quintile .................................................. 13 2.4 Inequality - "Jumps": Ratios Across Quintiles (for consumption) ............................................... 13 iv 2.5 Perceptions of Changes in Household and Community Welfare Over Time ................................ 19 5.1 Structure of Growth, 1965-2000 ........................... 45 5.2 Structure of Economy, 1965-2000 ........................... 45 6.1 Employment Diversity in the Urban Informal Sector, % of Informal Sector Workers ................. 51 6.2 Employment Diversity in the Rural Informal Sector, % of Informal Sector Workers .................. 51 6.3 Evolution of Real Monthly Wages, by Sector ................................................................... 53 7.1 Poverty, Illiteracy, and Net Primary Enrollment Deficit by Region ............................................. 64 7.2 Poverty, Illiteracy, and Net Primary Enrollment Deficit by Ethnicity .......................................... 64 7.3 Improvements in Literacy, Primary Enrollment Over Time ......................................................... 64 7.4 Educational Attainment (Stock): Improvements and Closing of Gender, Poverty . ...................... 64 Ethnic Gaps Over Time 7.5 Late Initial Entry in Primary School - % of Students Aged. 7-12 Enrolling in First Grade ........... 65 at Age 8+ 7.6 Primary School Repetition Rates by Grade, Poverty Group .......................................................... 67 7.7 Secondary School Repetition Rates by Grade, Poverty Group ...................................................... 67 7.8 Supply- vs. Demand-Side Barriers to School Enrollment ............................................................. 70 7.9 Reasons for not Enrolling in Primary School - % of Children Aged 7-12 not Enrolled ................ 72 7.10 Reasons for not Enrolling in Secondary School - % of Children Aged 13-19 not Enrolled .......... 72 7.11 Incidence of Public Spending on Primary School: International Comparison .............................. 75 7.12 Incidence of Public Spending on Secondary School: International Comparison .......................... 77 7.13 Incidence of Public Spending on University: International Comparison ...................................... 77 8.1 Population Structure by Poverty Group ................................................................... 81 8.2 Health Indicators by Quintile - % of Children < 6 ................................................................... 82 8.3 Poverty and Malnutrition by Ethnicity, % of Individuals Below the Full Poverty Line ............. 83 and % of Children < 5 who are Stunted (HAZ) 8.4 Poverty and Malnutrition by Region, % of Individuals Below the Full Poverty Line and % of ... 84 Children < 5 who are Stunted 8.5 Reasons for not Seeking Health Care Treatment when Needed, % of those who Needed . ......... 87 % of those who did not Seek Treatment when Ill and Believed Treatment was Necessary 8.6 Distributional Incidence of Public Spending on Health, International Comparison ...................... 89 9.1 Social Fund Investment in Rural Infrastructure since 1993 ........................................................... 99 9.2 Coverage of Basic Services by Quintile ........................... 100 9.3 Decomposing the Coverage Deficit: Supply- or Demand-Side Constraints? .............................. 102 9.4 Household Spending on Basic Services ................................................................ 104 9.5 Target Incidence of Electricity Subsidies ................................................................ 104 10.1 Lack of Access to Paved and Motorable (paved + unpaved) Roads, by Quintile ........................ 107 10.2 Poverty and Road Access, by Region ................................................................ 108 10.3 Percentage of Population Poor, by Municipio, 2000 Roads ........................................................ 110 10.4 Access to Public Transport, % of Households in PSU Sample with Access ............................... 111 10.5 Access to Health Service: Travel Times with and without Access to Motorable Roads .............111 10.6 Access to Opportunities: Travel Times with and without Access to Motorable Roads ............... 112 Rural Areas 10.7 Access to Institutions: Travel with and without Access to Motorable Roads, Rural Areas ........ 112 11.1 Incidence of Reported Shocks During the Year 2000 ................................................................ 116 11.2 Classification of Poverty and Vulnerability: Transient vs. Chronic? .......................................... 126 11.3 Risk by Main Age Group, Leading Indicators of Risks, Uncovered Poor/Vulnerable and ......... 127 Suggested Interventions 12.1 Average Benefit Levels and Dispersion among Social Protection Programs .............................. 133 12.2 Average Levels and Dispersion Among Private Transfers .......................................................... 133 12.3 Coverage of SA Programs by Poverty Group ................................................................ 135 12.4 Duplications and Gaps in Social Risk Management Arrangements ............................................ 137 12.5 Effectiveness of Social Protection and Private Transfers in Reducing Poverty .......................... 141 12.6 Effectiveness of Social Assistance in Reducing Poverty ............................................................. 142 v 13.1 Estimated Share of Social Fund Spending Received by the Poor ............................................... 148 13.2 Country Policy and Institutional Assessment (CPIA), 1977-98: LAC and Guatemala ............... 150 13.3 Social Capital Participation by Gender ................................................................. 157 13.4 Participation in Organizations, by Poverty Group ................................................................. 157 METHODOLOGY AND DATA ANNEXES 1. Measuring Consumption Using the ENCOVI 2000 2. Measuring Income Using the ENCOVI 2000 3. Measuring Poverty Using the ENCOVI 2000 (Poverty Lines) 4. Statistical Appendix 5. Qualitative Poverty and Exclusion Study (QPES): Overview of 10 rural villages 6. Supply Versus Demand-Side Constraints to Coverage of Education, Health and Basic Utility Services: Cluster Methodology TECHNICAL BACKGROUND PAPERS Available on the web at http://www.worldbank.org/guatemalapovertv 1. Livelihoods, Labor Markets, and Rural Poverty, by Renos Vakis 2. Education Sector Review, by Maria Elena Anderson 3. Education and Poverty LSMS Analysis, by John Edwards 4. Perceptions of Education (Qualitative Study), by Martha Rodriguez 5. Health and Poverty, by Michele Gragnolati and Alessandra Marini 6. Malnutrition and Poverty, by Alessandra Marini and Michele Gragnolati 7. Poverty and Modem Utility Services, by Vivien Foster and Caridad Araujo 8. Transport and Poverty: A Profile Using Data from the ENCOVI 2000, by Jyotsna Puri 9. Vulnerability and Vulnerable Groups: A Quantitative and Qualitative Assessment, by Emil Tesliuc and Kathy Lindert 10. Social Protection, Private Transfers and Poverty, by Emil Tesliuc and Kathy Lindert 11. Exclusion and Poverty in Guatemala's Rural Villages: The Challenge of Tackling Cumulative Barriers, by Carine Clert and Ana-Maria Ibanez 12. Social Capital in Guatemala: A Mixed Methods Analysis, by Ana-Maria Ibafiez and Michael Woolcock 13. Qualitative Poverty and Exclusion Study (QPES), Main Report, by COWI Consulting vi ACKNOWLEDGEMENTS This report is part of a collaborative multi-year program of analytical work and technical assistance that signals the commitment of the World Bank to poverty reduction. The program seeks to (a) conduct a thorough, multi-dimensional analysis of poverty and exclusion in Guatemala using both quantitative and qualitative data; (b) provide relevant policy analysis, advice, and tools on the incidence and impact of public programs, which could serve as inputs into the formulation and implementation of the Government's poverty reduction strategies; (c) contribute to the World Bank's assistance program, including the development of the next Country Assistance Strategy (CAS) and the design of a new operations in a variety of sectors; and (d) foster institutional development and capacity building in our counterpart agencies for greater ownership and sustainability of the analysis and results. The program has been highly collaborative, spanning several years (1998-present) and involving continuous cooperation and numerous missions (over 30 in total). The main phases and activities supported by the program have included: (a) support to the MECOVI' Program to develop an integrated system of household surveys and build capacity for survey design, implementation and analysis, including the production and analysis of the Encuesta Nacional de Condiciones de Vida (ENCOVI 2000) (1998-present); (b) qualitative data collection2 (field work conducted in 2000); (c) training in poverty analysis (initially in 1999-2000 with data from the ENIGFAM, then in 2001 with data from the ENCOVI); (d) training in, and construction of, poverty maps (occurring in 2000-2001); (e) support to the formulation of the Government's Poverty Reduction Strategy (from mid-2000-early 2002); and finally (f) analysis and preparation of the present report (from mid-2000 to 2002). This program is collectively called the Guatemala Poverty Assessment Program (GUAPA Program). The main counterpart agencies for the GUAPA Program have been the Instituto Nacional de Estad(stica - Guatemala ([NE) and the General Planning Secretariat under the Presidency (SEGEPLAN) in collaboration with the University of Rafael Landivar (URL). The World Bank team extends a heartfelt thanks to the teams from each of these agencies for their fruitful collaboration and partnerships. An earlier draft of this report was discussed with representatives of the Government (SEGEPLAN, INE, Ministry of Finance, Ministry of Communications, Secretarfa Ejecutiva de la Presidencia, the Secretarfa de. Bienestar Social, the Congreso de la Repuiblica, FIS, FOGUAVI, INFOM-UNEPAR, and the Ministry of Health), the international community (UNDP, JICA, Costa Rican Embassy, Canadian Embassy, Japanese Embassy, JICA, KfW, and FAO), and the academic community (URL, ASIES, CIEN, FUNDAZUCAR, FLACSO, and IDC) during a series of consultations held in Guatemala in June 2002. We are very grateful for their participation, suggestions, and comments, which we have taken into account in the current draft. The World Bank's task team included Kathy Lindert (Task Manager), Carlos Sobrado, Renos Vakis, Diane Steele, Caridad Araujo, Carine Clert, John Edwards, Emily Gustafsson-Wright, Vivien Foster, Ana-Marfa Ibanez, Michele Gragnolati, Alessandra Marini, Jyotsna Puri, Gloria Rubio Soto, Emil Tesliuc, Jean-Philippe Tre, Quentin Wodon, Michael Woolcock. Carlos Becerra and Martha Rodriguez provided valuable inputs in Guatemala. Lerick Kebeck provided task assistance and managed the production of the report. The Spanish version was translated by LTS translation, with editing and review kindly provided by Lorena Cohan. Shelton Davis, Peter Lanjouw, Jose Roberto Lopez Calix, and Tomas Rosada Villamar served as Peer Reviewers and Advisors throughout the process. Michael Walton, Alberto Valdes, Indermit Gill, Quentin Wodon, Jeni Klugman, and Jehan Arulpragasam provided additional advice. The team also benefited from the guidance of the management team: Guillermo Perry (Chief Economist), Ana-Marfa Arriagada (Director and Sector Manager), Donna Dowsett-Coirolo and Jane Armitage (Country Directors), Suzana Augusto (Country Officer), Ian Bannon and Felipe Jaramillo (Lead Economists), Norman Hicks (Sector Manager), Helena Ribe (Sector Leader), Eduardo Somensatto (Resident Representative), and Chris Chamberlin (Sector Manager). Finally, the team benefited from the advice of two "Quality Enhancement Review Panels," one on vulnerability analysis consisting of Lant Pritchett, Lynne Sherburne-Benz, Michael Woolcock, Tamar Manuelyan Atinc, and the other transport and poverty, Christina Malmberg-Calvo, Dominique van de Walle, Robin Carruthers, and Hanan Jacoby along with Guillermo Ruan, our partner on the transport side. In addition, we benefited from the inputs and support of the Multi-Sector Team Learning Initiative, particularly from the assistance of our "coach," Jennifer Sara. Funding for the study was generously provided by the World Bank (including from the Institutional Development Fund), the Government of Denmark, the Government of Japan, and the Government of the Netherlands. Funding for the MECOVI Program was generously provided by the World Bank, the IDB, CEPAL, USAID, the Soros Foundation, UNDP, UNICEF and the ILO. 'Program for the Improvement of Surveys and Measurement of Uving Conditions in Latin America and the Caribbean (MECOVI). Sponsors of the program include the Inter-American Development Bank (IDB), the World Bank (IBRD), and the Economic Committee for Latin America and the Caribbean (CEPAL), as well as USAID, the Soros Foundation, UNDP, UNICEF, and the [LO. 2 This initiative, called the "Qualitative Poverty and Exclusion Study" (QPES) was generously sponsored by the Danish Govemment. The work was carried out by local consultants under a contract with COWI consulting. GUATEMALA POVERTY ASSESSMENT EXECUTIVE SUMMARY Objectives of Report. This report is part of a sources, which are both valuable in their own right but collaborative multi-year program of analytical yield important synergies when used together. The work and technical assistance (the Guatemala primary source of quantitative information is the Poverty Assessment Program, or "GUAPA" Living Standards Measurement Survey (Encuesta program). The poverty assessment report itself has Nacional de Condiciones de Vida, ENCOVI 2000), three main objectives. The first is to conduct an which was conducted by the Instituto Nacional de in-depth, multi-dimensional analysis of poverty Estadistica - Guatemala (INE) under the auspices of building on the framework of the World Bank's the MECOVI Program.' The study also draws on the World Development Report (WDR) for results of a Qualitative Poverty and Exclusion Study 2000/2001 using both quantitative and qualitative (QPES) covering 10 rural villages of different data. The second is to examine the impact of ethnicities that were also included in the ENCOVI government policies and spending on the poor. sample. The QPES was designed to complement the The third is to use the empirical findings to ENCOVI with in-depth information of perceptions of identify options and priorities for poverty poverty, vulnerability, social capital, education, public reduction in the future. Policy options are services, and gender roles. outlined not only in general, but for the specific themes and sectors covered. It is hoped that the Collaborative Approach. The approach embraced by report will make an empirical contribution to the GUAPA/MECOVI program is as important as this improving the Government's anti-poverty policies report itself. It is based on a highly collaborative and strategies (broadly as well as for specific process between the Government, the World Bank, sectors). The report is also expected to contribute and other donors2 designed to build local capacity for, to the definition of the country assistance strategy and ownership of, the analysis of poverty. The (CAS) and lending operations of the World Bank process involved continuous cooperation, technical and other donors, so as to make these assistance and training with three main counterpart interventions more effective in the future. This agencies: INE, the planning secretariat (SEGEPLAN), type of study is conducted by the World Bank in and the University of Rafael Landivar (URL). The its client countries on a regular basis with the process has already generated several outputs, objective of assessing the poverty situation of the including the ENCOVI and its database; a new and country in question. improved Poverty Map designed to serve as a policy tool for targeting public spending and interventions Analytical Framework. The GUAPA embraces (constructed by the SEGEPLAN-INE-URL technical the conceptual framework developed in the World team with World Bank technical support); the training Development Report (WDR) 2000/2001 that and capacity building of numerous staff in these stresses the multi-dimensionality of poverty. agencies; and technical support to the Government's Poverty is associated with (a) a low level and poverty reduction strategy (PRSP). productivity of assets, which constrain opportunities; (b) exclusion from institutional, GUATEMALA'S POVERTY PROBLEM IS SERIOUS social and political spheres; and (c) vulnerability to risks and shocks. Reducing poverty requires Poverty in Guatemala is high and deep. In 2000, concerted efforts on all of these inter-connected over half of all Guatemalans - 56% or about 6.4 dimensions: expanding the opportunities of the million people - lived in poverty. About 16% lived in poor, empowering them, and improving the extreme poverty. Available evidence suggests that security of their well-being. poverty in Guatemala is higher than in other Central American countries, despite its mid-range ranking Data. An innovative aspect of the GUAPA is its using per capita GDP (US$3,630 in PPP terms). The combination of qualitative and quantitative data costs of reducing poverty are also high. Given i average consumption levels of the poor, it is countryside. Three quarters of all rural residents live estimated that the minimum annual cost to in poverty and one quarter live in extreme poverty. eradicate poverty is equal to about 8% of GDP. Poverty is also significantly higher among the To put this in context, total Government spending indigenous (76% are poor) as compared with the non- in 2000 was about 13%. Moreover, these costs indigenous population (41% are poor). are hypothetical: they represent the cost of bringing all the poor up to the poverty line, While pockets of poverty pepper the country, there excluding the inevitable administrative costs or is also a significant "poverty belt" in the Northern leakages to the non-poor associated with virtually and North-Western Regions. Poverty in Guatemala all poverty-alleviation schemes. is a national problem, with pockets of poverty spread throughout the country. Nonetheless, poverty is Although poverty has fallen over the past significantly lower in the Metropolitan region around decade, it has increased in recent years. the capital, and much higher in the North and Poverty is estimated to have fallen from about Northwest Regions, as well as the Department of San 62% in 1989 to 56% in 2000.3 This drop is Marcos, which were largely affected by the country's slightly slower than what would have been three-decades long civil war. predicted given Guatemala's growth rates, suggesting that growth has not been particularly Inequality is also quite high. With Gini indices for "pro-poor." This pattern arises largely because consumption and income of 48 and 57 respectively, growth in the rural sectors - where the poor are Guatemala ranks among the more unequal countries of largely concentrated - has been slower than in the world. The population is characterized by a large other areas. Moreover, projections suggest that "low-income" majority and a small high-income poverty has actually increased slightly in 2001 minority, with the top quintile accounting for 54% of and 2002, due to a series of economic shocks. total consumption. There are significant inequities These projections are based on estimated growth across ethnic groups. Although the indigenous rates which are estimated to have fallen in those represent 43% of the population, they claim less than a years (with slightly negative per capita growth quarter of total income and consumption. rates once population growth is taken into account). Malnutrition rates among Guatemalan children are abysmally high - among the worst in the world. Poverty and vulnerability are mainly chronic, Some 44% of children under five are stunted.4 There not transient. While 56% of Guatemala's is a strong correlation between poverty and population lived in poverty in 2000, the majority malnutrition, as four fifths of malnourished children of these (79%) were chronically poor, whereas are poor. Moreover, malnutrition is declining more only a fifth were transient poor. Likewise, while slowly in Guatemala than in other countries. 64% of the population could be considered vulnerable to poverty, the majority of these are Guatemala also ranks poorly for health outcomes. vulnerable due to low overall expected Guatemala ranks among the worst in LAC for life consumption rather than high volatility of expectancy, infant mortality and maternal mortality. consumption. The chronic nature of poverty and The patterns of health indicators also suggest worse vulnerability highlights the importance of building conditions for the poor, rural, and indigenous the assets of the poor, rather than focusing populations. Though health outcomes have improved primarily on the expansion of public safety nets or over the past 20 years, Guatemala's progress has been social insurance. Nonetheless, some public slower than the low-income countries of Bolivia, transfers (social assistance) could indeed be Nicaragua and Honduras. desirable to alleviate the poverty and suffering of the extreme poor, particularly when linked to Although Guatemala's performance in education participation in health and education activities. still lags, with important biases against the poor, progress has been made. With an illiteracy rate of Poverty is predominantly rural, and higher 31%, only Nicaragua and Haiti rank worse in LAC. among the indigenous. Over 81% of the poor Likewise, educational attainment is extremely low (4.3 and 93% of the extreme poor live in the years on average). Nonetheless, Guatemala has made ii progress in improving the educational stock. In terms of context, two key features play an Current primary enrollment has also expanded, important role in shaping the profile of poverty: particularly since the signing of the Peace geographic isolation and ethnic exclusion. First, Accords in 1996. Nonetheless, coverage is still while Guatemala is a physically diverse country, low and biased towards the non-poor. geographic isolation -- due to its complex topography and an inadequate road network - limits opportunities, Progress is also evident for basic utility constrains social networks, and fosters vulnerability. services, though important gaps and disparities Overall, 13% of households lack any form of adequate remain. Overall, about 70% of Guatemalan motorable road access, with an even higher degree of households have piped water and electricity. physical isolation among the poorest quintiles and the Almost 90% have some kind of basic sanitation, indigenous.5 Second, while Guatemala's population is though fewer than half have sewerage. About rich in cultural and linguistic diversity, this diversity 20% subscribe to either a fixed line and/or a has historically been accompanied by conflict, cellular telephone service. Expansions in exclusion and a dualistic social and economic coverage have accelerated since the signing of the structure. In a population of over 11 million, about Peace Accords in 1996, with a targeted expansion half the population is indigenous, including some 23 for disadvantaged groups. Nonetheless, important ethno-linguistic groups. Internationally, countries disparities remain, with significant coverage gaps with significant indigenous populations tend to have for the poor, particularly in rural areas. higher overall poverty rates. Within these countries, the indigenous tend to be poorer than the non- WHILE POVERTY IS LARGELY DETERMINED BY indigenous population due to historically exclusionary HOUSEHOLD CHARACTERISTICS AND ASSETS... forces. In this regard, Guatemala is no exception. Poverty is clearly associated with lower levels In fact, past policies greatly contributed to an or productivity of key assets, including labor, exclusionary pattern of development in Guatemala, education, physical assets (including basic utility particularly for land, labor and education. All of these services, land, and housing), and social capital. spheres were intertwined with each other, and with the Geographic location and household size are also development of coffee, Guatemala's primary export found to be important correlates of poverty. crop. Policies such as massive land expropriations, Disparities in assets also constitute the main forced labor, and exclusion from the education system sources of inequality, with education accounting (as part of a broader political strategy), all sought to for over half of all inequality in Guatemala. The promote economic growth, but to the exclusion and cross-sectional relationship between these assets detriment of the indigenous population. Women were and poverty is similar to that found in other also excluded from these spheres. As a result of these countries. policies - and despite having almost twice the per capita GDP of Honduras, Nicaragua, or Bolivia -- ... HISTORICAL AND CONTEXTUAL FACTORS Guatemala already ranked behind these countries by ALSO sHAPE GUATEMALA'S POVERTY PROFILE 1960 on several key social indicators. It also lagged significantly behind the collective group of History and context matter. A combination of comparable "lower-middle income countries." historical and contextual factors have fundamentally influenced Guatemala's The 36-year civil war further imposed costs on the performance regarding the levels of the Guatemala's development. Somewhat endowments and characteristic determinants of paradoxically, Guatemala managed to maintain poverty observed today. They also provide hints reasonable growth rates in the early phases of the war, about key challenges and potential levers to though growth did fall during the peak of the conflict reduce poverty tomorrow. This report does not in the 1980s. In addition to a large loss of life, the war purport to conduct an exhaustive review of these had serious short- and long-run impacts on historical and contextual factors; rather it seeks to Guatemala's development, for the overall economy simply highlight the importance of these factors in (lost jobs, productivity, output), human capital (and influencing poverty in Guatemala. hence long-run growth), and for life at the village level. THE PEACE ACCORDS SIGNALED A SIir However, changing the course of history in such a TOWARDS A MORE INCLUSIVE DEVELOPMENT short time span is not easy in any country. The PATH hierarchical relations, attitudes, and institutional forces that have pervaded for centuries do not disappear over The Peace Accords represented a turning point night. Furthermore, recent events (including for Guatemala's development path, paving the Hurricane Mitch and political instability) have delayed way for a transformation to a more prosperous and the implementation of the Peace Agenda. inclusive nation. Key areas of emphasis related to economic development and poverty reduction Moreover, despite progress, households do not include: a focus on human development, goals for perceive significant improvements in living productive and sustainable development, a conditions. While communities in the ENCOVI do program for the modernization of the democratic perceive progress - and attribute it to improvements in state, and strengthening and promoting basic services - households are decidedly more participation. The rights of the indigenous and pessimistic about changes in their welfare since the women were also highlighted as cross-cutting Peace Accords. They attribute these perceptions to themes throughout the accords, in an attempt to economic factors, such as a lack of increases in reverse the historical exclusion of these groups. incomes and opportunities (factors that directly affect "their wallets"). Moreover, given the importance of improving living conditions to lasting peace, poverty Many of the challenges for poverty reduction reduction has taken center stage on the current coincide largely with the remaining actions on the social policy agenda. In particular, the Peace Agenda. In particular, select development- Government recently outlined its poverty related targets supported by the Peace Accords have reduction strategy in an important policy not been met, especially those involving outcomes in document "Estrategia de Reducci6n de la health, education and economic growth as well as Pobreza" (ERP) 6 presented at the Consultative fundamental institutional reforms. The lack of Group meetings in February 2002. General progress for these key development outcomes reflects principles emphasized in the ERP include: a rural the need for poverty reduction and improvements in focus, using the poverty map for targeting; living conditions, which are crucial for lasting peace. efficient and transparent public spending; This overlapping between the Peace Agenda and the decentralization; and participation. Key action poverty agenda highlights remaining priority areas include: (a) promoting growth with equity; challenges in several key areas: (a) creating (b) investing in human capital (emphasizing opportunities; (b) reducing vulnerability; and health, education and food security); and (c) improving institutions and empowering (c) investing in physical capital (particularly water communities. and sanitation, rural roads, electricity, and rural development). Cross-cutting issues in the ERP KEY CHALLENGIE: BUILDING OPPORTUN1TIES AND include multiculturalism and inter-culturality, ASSETS gender equity, and vulnerability. Despite Guatemala's historically reasonable PROGRESS HAS OCCURRED ... BUT CHALLENGES economic growth rates, current growth is neither REMAIN sufficiently fast nor oriented towards the poor. Guatemala has historically enjoyed relative In the six years since the signing of the Peace macroeconomic stability and reasonable growth Accords, Guatemala has taken important steps (averaging 3.9% over the period from 1950-2000). on this new development path, with progress in Nonetheless, growth did not favor the poor because public sector management, public revenues and the economy did not generate enough low-skilled jobs. spending, and improvements in the coverage and Agriculture, which employs the majority of the poor, equity of education and basic services. experienced below-average growth rates over the past Importantly, these steps signal that progress is 20 years. In addition, other sectors did not grow fast possible, despite the magnitude of the challenge of enough to offer enough employment opportunities for changing the course of history. the poor. Reflecting these trends, the estimated iv decline in poverty over the past decade has been low share of women participate in the labor market, slightly slower than what would have been those that do are highly concentrated in the informal expected with neutral growth. Moreover, growth sector, which generates lower incomes. Finally, the has fallen in recent years, and may have caused a indigenous face considerable wage discrimination, slight increase in poverty. even after controlling for human capital and job differences. Women also face wage discrimination, As a result, faster growth and interventions to but the wage gap between men and women is smaller reach the poor are necessary to reach the than the one between the indigenous and non- Millennium Development Goals (MDGs). indigenous. Given current growth projections, the record for meeting such targets is likely to be mixed. Agriculture is unlikely to serve as a major vehicle Targets for most social indicators established by for poverty reduction. While the poor are highly the Government's poverty reduction strategy dependent on agriculture (subsistence farming and (ERP)7 should be met by 2005. Nonetheless, agricultural jobs), agriculture is not likely to be a extreme poverty is not expected to fall as dynamic source of new employment opportunities and ambitiously as anticipated under the ERP due to will probably continue to shrink as a share of GDP. slower overall growth in 2001 and 2002. Agriculture has faced declining growth rates over the Moreover, given projected growth rates, it does past several decades. Within agriculture, traditional not seem likely that Guatemala will meet most of crops, such as coffee, which tend to employ a the more ambitious targets for health and significant number of workers, are contracting in the education established under the international face of a structural terms-of-trade decline. The MDGs (Table 5.6). As a result, faster growth and production of non-traditional crops has expanded interventions to boost the assets of the poor are significantly, but not enough to replace the earnings clearly needed to improve living conditions and employment opportunities lost by the coffee crisis. enough to meet these goals. As discussed below, Data from the ENCOVI suggest that relatively few further work is needed to define key actions for a households overall are involved in the production of pro-poor growth strategy. non-traditional exports, and most are non-poor.8 Poor farming households are primarily involved in the The poor are constrained in terms of production of subsistence crops. Similarly, while very oPPortunities and livelihoods. The poor are few households report having received technical highly dependent on agricultural income (which assistance (only 3% overall), over 70% of public accounts for about half of the total income of the technical assistance was reported by non-poor poorest quintile, as compared with just 3% for the households. As such, although the poor will likely to top quintile). Poorer households are fairly continue to depend on agriculture as an important homogeneous in their occupations, dividing their source of income, it is unlikely that agriculture will labor primarily between agriculture, self- provide the solution to the poverty problem or that employment, and blue collar jobs (mainly in the many people will escape poverty via agriculture. informal sector), which all yield significantly relatively lower incomes. Some 87% of the rural Although land is an important asset, its ownership poor depend on agriculture, either as small-scale is highly inequitable in Guatemala. Moreover, the subsistence farmers or agricultural day laborers. holdings of the poor tend to be: (a) quite small, often Indeed, poverty rates among these groups are providing below-subsistence incomes; (b) untitled; (c) significantly higher than among those whose main poorly located (geographically isolated); and (d) of source of income comes from non-agricultural poor quality. Market-based land reform efforts are employment. promising and should continue to be pursued, although high costs and design issues remain to be addressed. Similarly, women and the indigenous face both constrained opportunities and discrimination Non-farm employment opportunities could provide on the labor market The indigenous appear a route out of poverty in rural areas. Multivariate limited to lower-paying jobs, primarily in regressions signal a strongly negative correlation agriculture. Non-Spanish speakers also face between non-farm activities and poverty. Indeed, considerably lower incomes. While a relatively almost half of the rural non-poor are landless, v typically working in a variety of jobs (e.g., as self- of attending school, are the main constraints to employed entrepreneurs in commerce or increased coverage at the primary level. Secondary manufacturing or in non-farm salaried jobs). enrollment is constrained by both supply- and Nonetheless, a variety of barriers constrain access demand-side constraints, particularly the direct costs by the poor to non-farm opportunities, particularly of attending school and opportunity costs (work and geographic location, inadequate infrastructure and domestic duties). The targeting of public spending on education. Additional analytical work should be education is neutral at best, and highly regressive at conducted to further define such a strategy. the secondary and university levels. PRONADE is the exception: it is highly targeted to the poor (with only Indeed, geographic location is highly correlated 8% of beneficiaries being non-poor). Existing with poverty and employment opportunities. demand-side programs - such as scholarships and the The spatial proximity to larger cities offers school transport subsidy - are highly regressive, with considerable advantages, with higher poverty rates most benefits going to the non-poor. Other programs and fewer employment opportunities in smaller (e.g., school feeding and the bolsa de utiles program) municipalities. are slightly better targeted, but mainly benefit the middle quintiles of the population. Improving education is central to both the Peace Agenda and the poverty agenda. Indeed, While Guatemala has made some progress in the education is a crucial determinant of poverty, health sector, significant challenges remain for inequality, and earnings. It also greatly influences improving health outcomes. Progress has mainly health outcomes, malnutrition and fertility rates. centered on sectoral reforms (e.g., some decentralization and deconcentration). In spite of There have been significant improvements in these, key health outcomes - malnutrition, infant education, particularly since the signing of the mortality, maternal mortality, and morbidity - are not Peace Accords in 1996. Notably, (a) the sector improving as fast as they should, and Guatemala has undergone important institutional and remains among the worst performers in LAC. Health structural reformns (including some outcomes are worse among the poor, the indigenous, decentralization and deconcentration); (b) public and rural residents, suggesting a need for better spending on education has increased significantly targeted interventions. A significant share of the since 1996, with the bulk going to the primary population lacks access to affordable health services, level; (c) literacy and educational attainment are particularly the poor and rural and indigenous increasing over time, with important reductions in residents. disparities between genders, ethnicities, and the poor versus the non-poor; (d) coverage has A combination of supply- and demand-side factors accelerated at all levels since the Peace Accords, appears to be blocking improved health access. On particularly the primary level, and the expansion the supply side, services are fragmented; insurance has been well-targeted to the poor (largely coverage is minimal; inefficiencies in public funding through the PRONADE program); and (e) official are generated by use of highly-subsidized public statistics on internal efficiency suggest facilities by the few who are insured (virtually improvements. exclusively the non-poor); and even when facilities are available, they often lack medicines, doctors or staff. Nonetheless, important challenges remain. Public spending on health has not increased Significant coverage gaps and disparities remain, sufficiently and public spending is not well targeted to particularly for the poor, girls, rural, and the poor. On the demand side, economic barriers indigenous children. Very few poor make it to the (direct costs of health care) present the main constraint secondary level, and inequalities in earnings are to improved access. Although public health care is largely generated at this level. Despite progress, highly subsidized, private health care is relatively indicators of internal efficiency suggest serious expensive. As such, in situations in which only structural deficiencies in the educational system. private services are available, disadvantaged groups Moreover, the low returns to primary school lack access due to economic constraints. Cultural suggest shortcomings in the quality of schooling. barriers further constrain access of the indigenous Demand-side factors, particularly the 'direct costs population to health care. vi There has been significant progress in KEY CHALLENGE: REDUCING VULNERABILITY BY expanding the equitable provision of basic BUILDING ASSETS utility services since the signing of the Peace Accords in 1996. Notably: (a) sectoral reforms The lack of adequate assets makes the poor have improved competition and efficiency; (b) the vulnerable to shocks. Despite the lack of any major volume of resources channeled towards the macro shocks in 2000, households in Guatemala report expansion of rural service provision has increased a high incidence of shocks that year, and most substantially through a variety of new and existing experienced multiple shocks with varying duration of institutional mechanisms; (c) overall coverage of impact. The effects of shocks are multi-dimensional, basic services has accelerated considerably since affecting not only income, wealth and consumption, 1996; and (d) this expansion has been well- but also community assets, the psychological and targeted, with new connections going social well-being of individuals, families and disproportionately to traditionally disadvantaged communities, health and education. The poor are groups. more exposed to natural disasters and agriculture- related shocks. They also have lower resilience to Nonetheless, significant coverage gaps and shocks than the non-poor. The cost of shocks is disparities in access remain. A significant share significant. Economic shocks have larger and more of those without access to basic services live in severe impacts than other types of shocks. Possible communities where the services are present but do sources of vulnerability in the future include: not connect due to demand-side barriers, such as (a) worsening terms-of-trade and job loss (e.g., those the direct costs of connecting to and using associated with the crisis in the coffee sector); and services. As such, demand-side interventions, not (b) natural disasters. All are likely to have lasting and just the physical provision of infrastructure will be severe impacts on the poor. needed. Given existing rates of expansion, it will take more than eight years to reach universal Certain sub-groups of the population are coverage for all services except sanitation, and the inherently or structurally vulnerable due to special total cost of meeting universality is estimated at circumstances. Specifically, key vulnerable groups between US$1-1.5 billion. In addition, energy include young children, who are vulnerable to subsidies (under the "tarifa sociar') are extremely malnutrition and lack of development; school-aged poorly targeted, benefiting primarily the non-poor. children, who are vulnerable due to lack of Finally, the quality of piped water services is poor educational opportunities and child labor; the working (non-potable and irregular). poor, particularly those in agriculture, due to low Guatemala especially needs to extend the road earings and susceptibility to natural shocks; poor network and public transport to the poor, households lacking basic services; seasonal migrants particularly in rural areas. Rural residents and and their families; and poor, rural households living in the poor are relatively more isolated in terms of areas prone to natural disasters. road and transport access. There is a significant Faced with shocks, Guatemalan households tend to inverse correlation between access to the rely primarily on their own assets, with little motorable road network. Year-round access is Government assistance. The main coping strategies also crucial, with road closures from rains and include reduced consumption or self-help. Few landslides further cutting off access to households report reeiving any formal govemental oppornities and services. The poor also lack hor non-governmental assstance in the face of shocks. access to public transportation, which appears to be correlated with a lackofadequateroads The poor are less equipped than the non-poor to fight shocks, and are more likely to reduce consumption Accords appear to have favored the non-poor and (regrettably, of basic staples) or use existing assets urban residentso The ENCOVI shows that (particularly labor). The non-poor are more likely uranrsiets heECOIshw ta than the poor to use market-based insurancee inadequate road access significantly constrains the access of the poor and rural residents to health mechansms. services, opportunities and institutions, further indeed, existing public social protection9 programs exacerbating their isolation. are poorly targeted and inefficient. Social insurance vii is virtually exclusively limited to (a small share Guatemala) to develop a national program to promote of) the non-poor. Public social assistance transparency and reduce corruption. programs are scattered across many agencies, with At the community level, social capital is lmited and many gaps and duplications in coverage, concentrated among the more privileged groups in Moreover, they are generally regressive. Private society. Communities have an important role to play transfers (such as international and domestic in promoting their own development, particularly in remittances, charity and donations) are an light of Guatemala's weak public sector. Social importance source of income, accounting for capital can offer significant benefits to community almost half of all transfers received by households welfare, including managing local public goods, in Guatemala. Nonetheless, the current coping with shocks, and leveraging external distribution of transfers reveals that they do not go assistance. Nonetheless, the ENCOVI and QPES to the poorest groups (and in fact are regressive), suggest that social capital in Guatemala is mainly and hence do not compensate for inadequacies in concentrated in strong horizontal, within-village the public social safety net. connections, with weaker bridges to other KEY CHALLENGE: BUILDING INSTITUTIONS communities or links to formal institutions. This ANTD EMPOWERING COMMUNITEES pattern reflects the physical isolation of many communities and decades of civil war and exclusion. Despite some progress, a weak public sector Moreover, the ENCOVI and QPES suggest that social has hampered Guatemala's efforts to improve capital appears to be concentrated among the more living conditions and promote a more inclusive privileged groups, with women, the poor, and the society. In addition, it has affected the menu of uneducated significantly less likely to participate at options for reducing poverty and the ways in the community level. The recent passage of three which these interventions are carried out. Though laws" on decentralization and participation is an there has been some progress in public finances important step towards creating a legal framework for and public expenditure management, significant the empowerment of local communities. The challenges remain, including: (a) the weak tax implementation of these laws should seek to reverse base (about 10% of GDP in 2001), which the traditional exclusion of women, the poor, and significantly constrains public spending (about uneducated from community-level participation. 13% of GDP in 2001); (b) public expenditure Other actors - particularly the private sector, management needs additional strengthening, with NGOs, and religious groups - are active players in better links to policy, planning, and priorities; (c) the fight against poverty. Given the limited size and public spending is poorly targeted; (d) public scope of the public sector, partnerships should be sector accountability and responsiveness needs to sought with these actors to help advance the poverty- be improved; (e) the civil service is weak, reduction agenda. hampered by inefficient hiring practices; and (f) the Government is highly centralized, A BROAD AGENDA TO ]REDUCE POVERTY especially given the heterogeneity of Guatemala's In developing a broad agenda for poverty reduction, population and communities.'0 three caveats are important to emphasize at the outset: Governance is also weak. Good governance is important for poverty reduction, and has been o First, while there is no single "blueprint" for linked to higher incomes, lower infant mortality poverty reduction, there are some key levers and higher literacy. Guatemala scores poorly on that take central stage for national efforts to most governance indicators, particularly those for reduce poverty, and these are the emphasis of this corruption, the rule of law and the justice system, report. Nonetheless, efforts should be made to and political instability, all of which damage the tailor this broad agenda to local conditions, climate for growth and investment. The particularly in a country as heterogeneous as Government has recently undertaken a number of Guatemala. To this end, efforts are underway in important initiatives designed to improve Guatemala to develop not only a national poverty governance, including issuing a Letter of Intent reduction strategy, but also localized strategies at (Carta de Intenciones del Gobierno de the department and municipal levels.'2 viii * Second, the policy discussion is aimed size and capabilities of Guatemala's public sector. primarily at the perspective of policy Moreover, a recurring theme that arises in the makers and the role of the public sector; analysis is the fact that the poor, particularly the hence it emphasizes interventions that both (a) rural poor, women and the indigenous, are not would have a substantial impact on poverty; able to fully participate in, or benefit from, the and (b) merit the use of public resources in a overall economic system. Therefore, improving market-oriented economy. Nonetheless, other employment and earnings opportunities is actors - such as the private sector and other essential, and this depends largely on the actions facets of civil society (e.g., communities, of the private sector. The pattern of growth needs NGOs, religious organizations), also have an to be made more "pro-poor," with an emphasis on important role to play in reducing poverty. building opportunities for the rural poor, women The private sector, in particular, will provide and the indigenous. This will depend on the central arena for economic growth and complementary policies in two other key areas: productive activities, which are crucial for building the assets (education, infrastructure, land poverty reduction. Other facets of civil and physical capital) of the poor (particularly the society (e.g., communities, NGOs, religious rural poor), as well as improving institutions and organizations) are clear partners in this the investment climate. poverty reduction agenda, and will play crucial roles in prioritization and * Building the assets of the poor. Given the implementation of public sector actions, as chronic nature of poverty in Guatemala, existing well as the provision of other services and disparities, and linkages to the other key areas, interventions that are beyond the scope of the this is arguably the most important area for public sector. poverty reduction. Key assets include: education, health, basic utility services (particularly water * Third, poverty reduction is a multi- and sanitation), rural roads, and land and physical dimensional and long-term process. There capital. is no single magic bullet to reduce poverty. Rather, efforts should be made to attack the * Reducing vulnerability. Again, the central path poverty problem from a multitude of angles, for reducing vulnerability is to build the assets of including those to foster opportunity, build the poor, since most vulnerability in Guatemala is assets, reduce vulnerability, and improve associated with low expected earnings (due to institutions and empower communities. weak assets) rather than high volatility of Moreover, poverty reduction does not occur consumption. Nonetheless, disaster management over night. Implementation of key actions to is important, given the poor's exposure to natural reduce poverty takes time, and often the and agriculture-related shocks. Moreover, much impact of such actions occurs over an even could be done to improve the efficiency and longer time frame (for example, into effectiveness of existing social protection subsequent political cycles - or even programs. Many of these - such as scholarships, subsequent generations). That said, the school feeding - could also play a role in building sooner actions are undertaken, the more the assets of the poor by easing demand-side quickly the inter-generational cycle of poverty constraints to improved coverage. Efforts should can be broken. be made to consolidate and improve existing programs, particularly with respect to their Broadly speaking, a concerted strategy sfiould targeting. Strategic priorities should seek to be adopted to reduce poverty in Guatemala by maintain the current focus of social protection building opportunities and assets, reducing programs on children, given their inherent vulnerability, improving institutions and vulnerabilities and prospects for long-term empowering communities. A broad agenda of transmission of poverty. actions in these areas is outlined in Table A. Improving institutions and empowering * Building opportunities. Economic growth is communities. Weaknesses in the public sector crucial, particularly given the relatively small and poor governance strongly shape the menu of ix feasible options and effectiveness of poverty investment and productive activities. Yet the reduction efforts. They also influence the actions of the public sector in this supporting role overall climate for investment and economic are crucial. In particular, priority actions include: growth. As such, improvements in this area * Maintaining macroeconomic stability; are deemed to be of high priority, consistent * Enforcing a tight fiscal position, with a careful with the strategic emphasis on "modernization plan for strengthening tax collection and of the state" in the Peace Accords. The role redirecting public spending towards the social of communities in promoting their own sectors so as to build assets that are crucial to development is also important, as both growth and poverty reduction; acknowledged in the Peace Accords, and * Fostering a climate that is conducive to poverty, reduction efforts should seek to private investment and growth, including partner with communities in determining improvements in governance and public sector priorities. Nonetheless, explicit efforts should management; and be made to reach out to groups typically * Promoting growth with special emphasis on excluded from community decision-making sectors that are likely to generate substantial (namely, the poor, women, and the employment for the poor. Additional uneducated). Partnerships should likewise be analytical work is needed to define a more sought with private-sector and NGOs to comprehensive pro-growth strategy. extend and improve service delivery. Nonetheless, while a thorough sectoral analysis of growth is beyond the scope of this These priorities are consistent with the overall study, available data do suggest certain levers principles of the Peace Agenda. Indeed, that would have stronger impacts on poverty reducing poverty and improving living conditions reduction than others for urban and rural is central to lasting peace in Guatemala. areas: PRIORITY ACTIONS FOR POVERTY REDUCTION o In urban areas, this requires policies to support labor-intensive sectors, Certain actions stand out as top prioritv. particularly micro-, small- and medium- Within this broad agenda, actions should be enterprises (MSMEs), as well as further prioritized using the following criteria: (a) education and technical training. likely poverty impact; (b) political, institutional, o In rural areas, this means developing non- and administrative feasibility; (c) economic agricultural activities that are better feasibility and costs; and (d) their need and remunerated and have better long-term justification for public sector resources. Such prospects than taditional agriculture. prioritization will likely require further dialogue Key tevtion toasupport growthrin and analysis (e.g., institutional assessments, non-farm acvtioes include: (a) increasing costing of actions, public expenditure analysis). and improving the targeting of As a first cut, certain actions should be considered nvestments i education and technical as top priority, based on a cursory review of such taining; (b) increasing investments in criteria: transport and basic infrastructure, which (1) Promoting economic growth. Guatemala are crucial for the diversification, growth must raise its rate of economic growth if it is and inclusion of the poor in the rural to make significant progress in reducing economy and with facilitating the poverty and achieving key development and adjustment to the coffee crisis; and peace targets. This is true internationally, but (c) policies that promote micro-, small- particularly relevant for Guatemala, given the and medium-enterprises (MSMEs), a limited scope for public sector action and segment of the private sector that tends to redistribution. In this context, the main generate a lot of employment. While engine of growth is likely to come from the agriculture is unlikely to generate enough private sector, with the public sector playing a additional employment opportunities to supporting role affecting growth mainly reduce poverty on a large scale in the insofar as it stimulates private-sector medium term, it will continue to be an important source of incomes for the poor x (at least in the short run). In this as to improve internal efficiency and the context, diversification efforts should returns to education, particularly at the focus on non-traditional products with primary level; and better demand and price prospects * Investing in early childhood development to than traditional export crops. Policies promote: (a) improved child nutrition at an should also continue to facilitate early age, since nutritional status is a productivity improvements (such as significant factor in determining enrollment technical assistance), so as to boost and attainment and since nutritional the earnings of those who remain in deficiencies emerge at a young age; and (b) agriculture. Investments in early educational opportunities, including infrastructure (e.g., rural roads to links between traditional schooling and pre- improve marketing opportunities and primary schooling. education to improve farm- 3 management practices) will likewise (3) Investing in health, with an emphasis on be important. expanding access and usage using both supply- (2) Investing in education, with priority and demand-side interventions. Again, both actions to improve quality and access to theory and empirical analysis using the ENCOVI pre-primary and primary education. Both point to important linkages between health and theory and empirical analysis of the ENCOVI productivity (economic growth), vulnerability demonstrate the crucial role of education in (health shocks), and poverty. Guatemala's health emon omicagrowth;crucialroe rfeducaning outcomes have lagged significantly behind those promoting economic growth; reducing o poverty and malnutrition; reducing in other countries as well as the targets set by the vulnerability by making the labor force more Peace Accords and the MDGs. A significant agile and able to adjust to shocksh and share of the population still lacks access to health reducing inequality, social disparities and facilities - or fails to use them when available - exclusion. Since Guatemala is still a due to a mix of supply- and demand-side "primary" country on average (with average constraints. As such, priority actions should seek attainment of 4.3 years) and since the poor in to improve health outcomes by: particular fail to complete primary school, * Expanding access to affordable health care investments should still focus on expanding using both supply- and demand-side and improving primary education, at least in interventions. Such interventions should be the medium-term. In the longer-run, as a targeted to the poor and priority groups (for larger share of the poor complete primary example,using thevpoverty map); school, efforts should emphasize expanding * Emphasizing preventative care, infectious and access to secondary school. As such, priority parasitic diseases, reproductive health, key actions should focus on: outcomes (mortality, malnutrition); and Increasing access to primary education, * Expanding access to potable (not just piped) largely through demand-side water and improved sanitation to complement interventions, since supply-side the basic health care package. constraints are no longer binding for most of the population. However, as supply- (4) Integrating actions to reduce malnutrition into side. gaps are filled, the Government the basic health-care package. The high and should consider easing eligibility criteria stagnant rates of malnutrition in Guatemala so as to allow poor communities that require the highest attention. Their lasting effects already have schools to be eligible for the also result in inter-generational transmission of PRONADE-type community-based poverty. Reducing malnutrition should be school-management model. To target this designated as a top priority. Malnutrition expansion, the poverty map could be used interventions should be integrated into the to identify eligible schools and preserve MSPAS basic health care package and provided at PRONADE'n s exemplary targeting record; the community level through outreach workers, so * Improving the quality of education, as to improve their effectiveness and reach and curricl and prfance sdards so foster the integration of malnutrition as a key curlculum and perforTnance standards so v xi concern into the health system. The target improvements in governance and the population for these schemes should be pre- effectiveness and credibility of the public school children (particularly those under 24 sector; months of age) and mothers (including o Improving the targeting of public spending pregnant and lactating women). Priority particularly for investments in education, actions include: health, basic utility services, and transfers; o Promotion of proper health, hygiene, and o Improving public expenditure management, feeding practices; with stronger links to policy, planning and o Growth monitoring of pregnant women priorities; and children under aged two; o Expanding and building on recent initiatives o Micronutrient supplementation to fight corruption (e.g., adopting an anti- (particularly for iron); and corruption charter); o Deworming treatments and oral o Improving incentives for better service rehydration therapy. delivery (e.g., implementing the recently- passed laws on decentralization,'3 local (5) Reducing isolation and improving "control social," and service "report cards" communications by investing in rural and client satisfaction surveys); and transport and roads. Many communities in o Improving the rule of law and the justice Guatemala are still relatively isolated due to a system. lack of road access. Empirical analysis using the ENCOVI has demonstrated the effects of PRiORiTY TARGET GROUPS FOR POVERTY isolation on opportunities, productivity, REDUCTION iN GUATEMALA vulnerability (shocks), and access to services. The broad agenda for poverty reduction can Expanded rural transport helps build the become even more effective by focusing efforts on assets of the poor, promote economic growth key priority groups. For example, while economic and opportunity, reduce vulnerability, and kypirt rus o xml,wieeooi empower communities. Priority actions in growth is needed in general, growth that provides this area include focusing on improving and opportunities for the rural poor will be even more thipareincthe network of motorable roads in effective in reducing poverty. While building assets expanding the network of moth road of the poor in general is essential, priority is needed to rural areas, particularly those with untapped tackle the issues of malnutrition and the relative economic potential and a high concentration disparities against poor women and indigenous of poor people. residents. (6) Improving governance and the effectiveness of the public sector. Actions As suclh, thpe Government should prioritize among are needed to reduce corruption, improve poverty groups, according to tde prevalence of transparency, improve public expenditure poverty, specific risks, andl demographic mangemnt,and better target existing circumstanuces. Specifically, the analysis reveals management, and Such act wing several priority groups that should be emphasized in resources to the poor. Such actions will have pvryrdcinefrs a oradmlorse multiple benefits, including: (a) making the poverty reduction efforts: (a) poor and malnourished most of existing scarce resources and children; (b) poor women and girls; (c) poor improving service delivery; (b) fostering a indigenous households; (d) the rural poor; and climate that is more conducive to economic (e) specific geographic areas (Table B). Clearly, these growth; (c) assuring that public resources groups can have considerable overlaps. For example, reach the poor (needed for impact); and (d) a poor or malnourished indigenous girl living in rural improving the credibility of government and areas in the North or North-Western parts of the its ability to increase revenues in the future country would probably qualify for just about any (without such improvements, the Government anti-poverty intervention. will face continued resistance to tax increases). Priority actions include: o lPoor and malnourished children. The o Improving the tax base and increasing developmental status of children renders them puli smpending,the which will dend n ing extremely vulnerable to the risks of living in public spending, which will depend on xii an impoverished environment. Youth disadvantages, reflecting the historical pattern (particularly early childhood) is the point of exclusion and decades of conflict. Poverty in the life cycle when physical, cognitive, is higher among the indigenous. Indigenous and psycho-social development occurs at children also suffer higher rates of its most accelerated pace and is most malnutrition and less access to education, susceptible to abnormal development which affect their earnings ability in the from poverty conditions. As such, future. The indigenous also have less access childhood poverty also increases the to health and basic utility services. They are likelihood of inter-generational further constrained in employment transmission of poverty. Children in opportunities (particularly those who don't Guatemala are particularly disadvantaged speak Spanish), and face considerable wage and vulnerable: they have_ relatively discrimination (even after taking into account higher rates of poverty; close to half are disparities in endowments). Finally, they also malnourished; infant mortality is report perceptions of discrimination by public alarmingly high; a significant share of officials and service providers. children fail to enroll in school, thereby missing crucial opportunities for social * Rural poor. Poverty is higher in rural areas, and cognitive development; and child and even higher among specific rural sub- labor is common, particularly among poor groups, including small land-holders, children, further compromising their agricultural day laborers, and seasonal migrant chances of attending school. In this agricultural workers. The rural poor context, poverty reduction efforts should (particularly these sub-groups) have relatively confer top priority to poor and limited access to services and infrastructure malnourished children as a key target (education, health, utilities, transport, group. markets). They also have limited employment and earnings opportunities, particularly those Poor girls and women. Girls and women living in more geographically isolated areas face cumulative disadvantages in and smaller municipalities. They also face Guatemala, reflecting historically lower returns to their labor and are. rarely exclusionary policies (e.g., in land and covered by formal labor and IGSS benefits. education) and a general culture of Finally, they are quite susceptible to shocks, machismo. They face limited access to particularly natural disasters, agricultural- education (with fewer girls attending related shocks, and recent economic shocks school even when schools are available), (such as the coffee crisis which has worsened constrained employment opportunities, the terms-of-trade for producers and caused explicit wage discrimination (even after job loss for day laborers). The emphasis of taking into account differences in the ERP on rural areas is thus correct and endowments), and traditional exclusion should be maintained. from land ownership. Women are also at risk for health shocks, with Guatemala * Specific geographic areas. While poverty is recording extremely high levels of clearly a national problem in Guatemala, maternal mortality. Furthermore, women poverty is significantly higher in the "poverty participate significantly less in belt" in the Northern and North-Western community decision-making (limited regions as well as the departments of San social capital networks). Yet women's Marcos. The poverty map helps further roles are crucial in promoting long-term pinpoint specific municipalities with higher development, with a strong influence, for incidence of poverty. The ERP's inclusion of example, on the nutritional status of the poverty map as a key tool is thus children. appropriate. * Poor indigenous households. The Given budget constraints, certain activities should indigenous likewise suffer cumulative be actively targeted to the poor. As a "rule of xiii thumb," incremental increases in public spending (such as poor or malnourished children, girls or on areas such as education, health, basic utility women). Given the traditional exclusion of certain services, core communication links, social groups from community decision-making, however, assistance transfers, or employment schemes care should be taken to ensure that these patterns are should be explicitly targeted to the poor in order not repeated. Fifth, a unified proxy-means database to better integrate them into the economy and (such as those in Costa Rica or Colombia) could be improve social indicators. Decisions regarding developed for programs targeted to individuals, the allocation of investments in other services, though this could require significant administrative such as more intensive infrastructure, institutional capabilities. support, or banking services, should generally follow indicators of economic potential, which MONrrORING POVERTY REDUCTION EFFORTS could also be combined with targeting criteria Monitoring of both poverty and poverty reduction (e.g., the poverty map). Ideally, a strategy to interventions is necessary, and adequate resources promote pro-poor growth and reduce poverty should be made available for this task. First, the would focus on areas that have both a large MECOVI program seeks to develop an integrated concentration of poor people, but also a strong system of household surveys to track living conditions potential for future economic activity, and provide data for the evaluation of the impact of interventions. The system will build on the ENCOVI The Government can use a variety of tools to 2000, and should execute similar surveys every 3-5 better target programs to priority groups. . better target programs to priortygroups. years. In addition, LNE is currently developing an Improved use of limited public resources is employment and incomes survey that would be crucial for poverty reduction efforts. Ensuring executed on a more regular basis, to fill crucial gaps in such resources are channeled to key poverty Guatemala's information base. Finally, the upcoming groups (Table B) is a first step in improving the Population Census will provide additional information effectiveness of public spending and poverty for the monitoring of poverty, including an reduction efforts. The Goverment has at its opportunity to update the poverty map (combined with disposition several potentially potent tools for data from the ENCOVI), as well as infrastructure targeting its poverty reduction efforts to these priority groups. First, the poverty map, recently constructed by SEGEPLAN-INE-URL, can be extremely useful (alone or with other targeting m i aon ns to ru poverty ind the to tools) in ensuring that resources get channeled to including: further elaborating the ERP (fleshing out municipalities with high concentrations of the dis fortsei ectorand devElP ing dp t poor. Second, efforts have been made to develop .an mun cipal-l evelp yion strate e and other geographic-based maps and databases, such d evel of mon ion iniators frt as a extnsiv roa netork nvenory/ap, developing a system of monitoring indicators for the as an extensive roadnetwork inventory/ma targets set by the ERP. Ideally, a goal-based poverty vulnerability maps, conflict maps, municipal-level trgedt setrathe w dinlve a syste thatret databases oneuainadhalhsrie,ec reduction strategy would involve a system that relates databases on educatioa he ate etc. actions and exteral conditions to progress in reaching The upcoming census will help update many of tegas noprtn vlainmcaim n these~~~~~ ~~~ mas A.nfe egpi nonto the goals, incorporating evaluation mechanisms and thesemaps.lA unifned ghepogertymaph winformation feedback loops. The development of. this type of system could combine the poverty map with these sse hudceryb oriae ih a fot other maps and databases to better target specific st gherld (e.g. wt tedCVi prora) to gather data (e.g., with the MECOVI prograrm); interventions to the poor. Thiird, certain services (b) efforts to monitor the targets set by the Peace - such as health posts and community health Accords and the MDGs, (c) the SIAF, which is centers - are self-targeted to the poor. Other developing performance monitoring indicators for programs could be channeled through these public expenditure management; and (d) the various facilities to take advantage of this inherent self- executing agencies (e.g., sectoral ministries). targeting (and perhaps even promote use of these Adequate financial and technical resources should be facilities). Fourth, community-based targeting made available to the concered agencies for the could be used (perhaps after broader program allocations are made using the poverty map) to posesof stent select specific individuals eligible for programs xiv Table A - Menu of Otions and Key Actions for Poverty Reduction MAIN CONSTRAINTS I MAIN RECOMMENDATIONS Ke IssuesIPriorit Ke Actions & Time Period for A a BUILDING OPPORTUNITIES AND LIVELIHOODS: Priority overall, especially in rural areas * Growth has slowed and isn't very "pro- ... * Maintaining macroeconomic stability, with a careful plan for allocating public poor." Economic growth is crucial for expenditures and strengthening tax collection; ACT: on-going, IMP: ST, MT reducing poverty and building opportunities, ... * Improving the climate for growth, including governance and public sector particularly given the relatively small size . management; ACT: ST, MT, IMP: MT, LT and limited capabilities of Guatemala's . * Improving regulation and supervision of financial sector; ST public sector. ... * Promoting growth with emphasis in sectors that are likely to generate * Households do not perceive employment, such as non-agricultural sectors, via education and training, improvements, largely due to constrained transport, basic infrastructure, and support to SMEs.; ACT: ST, MT; IMP: LT opportunities and limited eamings * Reducing transactions costs in accessing markets (e.g., with road access, basic * Limited opportunities and earnings for services); ACT: ST, MT; IMP: MT, LT the poor, particularly the rural poor, women, . * Creating mechanisms to discourage labor-market discrimination for women and the indigenous: and the indigenous; ACT: MT; IMP: LT o Discrimination for women, indigenous * Expanding land titling and land markets programs; establishing financial o Low profitability in agriculture institutions in rural areas; ACT: MT, IMP: MT o Constrained entry for non-farm * Expanding seasonal employment creation programs (such as existing food-for- opportunities work programs) to provide opportunities for the rural poor; ACT: MT; IMP: .__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ M T BUILDING THE ASSETS OF THE POOR - EDUCATION: Priority for poor overall, especially for girls, indigenous, rural * Disparities, gaps in access: .. * Continuing increases in public spending on education, particularly at primary o Pre-primary: all poor, esp. rural and pre-primary levels; ACT: ST, MT; IMP: MT, LT o Primary: poor, esp. girls, indigenous ... * Expanding coverage, especially for girls and indigenous. Expansion should be o Secondary: all poor implemented via decentralized PRONADE program using poverty map to * Demand-side constraints (both primary replace supply-side restrictions as targeting mechanism; ACT: ST, MT; IMP: and secondary) MT, LT * Supply-side constraints (mainly at *. * Lowering official age of entry for primary school from 7 to 6; ACT: ST; IMP: secondary) MT * Internal efficiency, quality *.. * Reviewing and improving quality, curriculum and performance standards, * Weak targeting of pubic spending, particularly at grades 1, 7, and 10 (transition years); ACT: ST, MT; IMP: MT, education programs LT .. * Promoting, expanding, consolidating and improving demand-side programs, with emphasis on girls and indigenous children (e.g., scholarships, school feeding, bolsa de utiles); ACT: ST, MT; IMP: MT, LT Soo * Increasing investments in early childhood development; ACT: ST, MT; IMP: MT; LT ... * Using poverty map and other mechanisms, to better target public spending and demand-side programs (e.g., scholarships, school feeding, bolsa de utiles); ACT: ST; IMP: MT BUILDING THE ASSETS OF THE POOR - HEALTH: Priority for poor overall, especially for girls, indigenous, rural * Health outcomes - malnutrition, infant and ... * Increasing public spending and expanding access to health care combined with matemal mortality, and morbidity - are better targeting (via poverty maps and health posts/community centers); ACT: inadequate and not improving fast enough ST, MT; IMP: MT * Public spending inadequate and not well ... * Emphasizing preventative care, infectious and parasitic diseases, reproductive targeted health, ey outcomes (mortality, malnutrition); ACT: ST, MT; IMP: MT * Significant share of population lacks .. * Conducting a critical review of existing malnutrition interventions; ACT: ST; access to affordable health care, IMP: ST particularly the rural poor and indigenous ... * Implementing specific interventions for malnutrition as a top priority: * Supply-side constraints, including community-based information and behavioral change programs; growth fragmented services, minimal insurance monitoring for pregnant women and children under age two; micro-nutrient coverage, waste in public spending, lack of supplements. ACT: ST, MT; IMP: MT, LT medicines, doctors, staff . * Focusing on demand-side interventions (e.g., conditional transfers) that could * Demand-side constraints, including cost be channeled through self-targeted health posts/community centers; ACT: ST, and cultural barriers .. MT; IMP: MT * Promoting culturally-sensitive health care practices; ACT: ST, MT; IMP: MT * Conducting full review of supply-side issues; ACT: ST; IMP: MT * Developing monitoring system for health outcomes, including better and more regular measurement of infant and maternal mortality; ACT: ST; IMP: MT * Adopting measures to improve efficiency and quality of services delivered (see Chapter 8); ACT: MT, LT; IMP: LT .. * Facilitating increased awareness of family planning options so as to reduce Guatemala's high population growth rates, which constrain per capita income growth; ACT: ST, MT; IMP: LT *-v = top priority; *- = medium priority; * priority; ST = one year period; MT = 1-3 years; LT = more than 3 years; ACT = period for implementation of actions; IMP = period needed for impact on poverty XV Table A,Cont'd - Menu of Options and Key Actions for Poverty Reduction MAIN CONSTRAINTS MAIN RECOMMENDATIONS Key Issues I Priority Key Actions & Time Period for Actions and Impact . BUILDING THE ASSETS OF THE POOR - BASIC SERVICES: Priority for poor overall, especiaUy for rural, indigenous * Significant coverage gaps and disparities, .. Maintaining and, if possible, increasing resources for expansion of services; especially among rural poor and indigenous ACT: ST, MT; IMP: MT * Demand-side factors (connections costs) oe * Targeting service expansion to poor (particularly rural) using poverty map * Supply-side constraints (not available) combined with geographic information on coverage gaps; ACT: ST; IMP: MT * Energy subsidies poorly targeted 0o * Developing strategy for demand-side constraints; ACT: ST; IMP: MT * Quality of water is poor (not potable, 0 * Eliminating "tarifa social" energy subsidy and using resources to fund new irregular) connections instead; ACT: ST, but gradually; IMP: MT 0 0 * Allowing water tariffs to rise to a level that allows water utilities to become financially sustainable and improve the quality of service offered; ACT: ST but gradually; IMP: ST-MT 00 * Encouraging social funds and other providers to consider measures to improve quality of water; ACT: ST; IMP: ST * Complementing water and sanitation programs with measures to improve household hygiene and water treatment practices; ACT: ST; IMP: ST BUILDING THE ASSETS OF THE POOR - TRANSPORT: Priority for rural poor * Geographic isolation for rural poor, due to 0. * Focusing public spending on transport on rural areas; ACT: ST; IMP: ST, MT limited road network and public transport 0o0 0 Expanding and improving motorable road network in rural areas, particularly services by improving existing roads (including dirt roads); ACT: ST, MT; IMP: ST, * Road quality and closures limit year-round MT access o * Targeting expansion and rehabilitation using combination of poverty map with * Road improvements have favored non- road maps; ACT: ST, MT; IMP: ST, MT poor, urban areas o Inadequate road access significantly constrains access of rural poor to health services, opportunities, institutions REDUCING VULNERABILITY: Priority for alI poorivulnerable, particularly rural and specific vulnerable groups * Lack of assets makes poor vulnerable to see * Building assets of poor and key vulnerable groups (see Table B); ACT: ST, shocks, particularly natural disasters and MT; IMP: MT, LT agriculture-related shocks 00 e Expanding and improving disaster management relief; ACT: ST, MT; IMP: * Key sources of future vulnerability: (a) MT coffee crisis; (b) lost remittances from 0 ° Introducing catastrophic insurance schemes; ACT: MT; IMP: MT global slowdown; (c) natural disasters 00 ° Improving targeting of social protection programs; ACT: ST, MT; IMP: MT * Certain sub-groups are particularly o o Eliminating energy subsidy and school transport subsidy; ACT: ST; IMP: ST vulnerable due to special circumstances 00 0 Consolidating and improving scholarships and school feeding programs; ACT: 0 Faced with shocks, households rely on ST, MT; IMP: ST, MT own assets with little formal assistance 00 0 Improving targeting of bolsa de utiles program; ACT: ST, MT; IMP: ST, MT * Existing social protection programs are poorly targeted and inefficient IMPROVING INSTITUTIONS AND EMPOWERING COMMUNITIES: Priority for aU poor e Weak public sector hampers poverty 00 0 Improving tax base and tax collection; ACT: ST, MT; IMP: ST, MT reduction efforts 0oo o Increasing and improving targeting of public spending; ACT: ST, MT; IMP: o Weak tax base, limited public ST, MT spending 000 ° Improving public expenditure management, with stronger links to policy, o Public exp. management needs planning, and priorities; ACT: ST, MT; IMP: ST, MT strengthening 00 0 Strengthening the civil service; ACT: MT; IMP: MT o Public spending poorly targeted 00 0 Improving incentives for better service delivery (e.g., implementing recently- o Weak civil service passed laws on decentralization, local "control social," and service "report o Overly centralized 000 cards" and client satisfaction surveys); ACT: MT, LT; IMP: MT, LT • Governance weak, constrains growth and o Expanding and building on recent initiatives to fight corruption (e.g., an anti- poverty reduction efforts: corruption, lack G corruption charter) and making it a top priority. ACT: ST, MT; IMP: ST, MT of rule of law, inadequate justice system, 0 * Improving rule of law, justice system; ACT: ST, MT; IMP: ST, MT political instability. e Promoting community-based development but with explicit outreach programs * Social capital limited, concentrated to ensure participation of excluded groups (women, poor, uneducated) in among privileged community-decision making; ACT: ST, MIT; IMP: ST, MT o Limited networks outside villages * Partnering with private sector, NGOs to extend services; ACT: ST, MT; IMP: o Community participation limited for ST, MT women, poor, uneducated *o= top priority; *. = medium priority; o = priority; ST = one year period; MT = 1-3 years; LT = more than 3 years; ACT = period for implementation of actions; IMP = period needed for impact on poverty XVI Table B - Priority Target Groups Priority Target Groups Key Constraints/Challenges Possible Targeting Tools Poor and malnourished * Poverty * Poverty map combined with information children, especially pre-school * Malnutrition (stunting) on malnutrition and educational (age 0-6) and primary-aged * Not enrolled in school enrollment children (7-13) * Child labor * Self-targeting via health posts and * Vulnerable phase of life cycle community health centers (e.g., a growth * Inter-generational transmission of poverty monitoring program channeled through these facilities) * Community-based targeting * Proxy means testing Poor women and girls * Historical pattern of exclusion * Poverty map Por woenandgrls * Less access to education * Gender-based targeting (e.g., programs * Constrained in employment and earnings that restrict eligibility to girls, such as opportunities . scholarships) * Face wage discrimination * Community-based targeting * Face discriminatory attitudes (culture of * Proxy means testing machismo) * Excluded from participating in community decision making (social capital) Poor indigenous * Historical pattern of exclusion * Poverty map combined with language * Higher poverty and malnutrition map * Less access to education, health services * Proxy means testing * Less coverage by basic utility services * Constrained in employment and earnings opportunities * Face wage discrimination * Face discrimination in treatment by public officials and other service providers The rural poor, particularly * Higher poverty and malnutrition * Poverty map small land-holders, agricultural * Less access to education, health services * Vulnerability maps (e.g., natural day laborers, seasonal migrant * Less coverage by basic utility services disasters) agricultural workers * Geographic constraints (isolation, roads, small * Proxy means testing, with certain proxies municipalities) emphasized (e.g., land holdings, * Constrained employment and earnings electricity connections, etc.) opportunities * Migration maps (that could be developed * Low returns, limited coverage of labor and from census data) showing municipalities IGSS benefits with significant concentrations of * Susceptible to shocks seasonal migrants Specific geographic areas, * Higher poverty rates, malnutrition * Poverty map, combined with other asset- especifyic g heographicares . Lower access to basic services specific maps/info. (e.g., gaps in coverage belt" (Norte, Nor-Occidente, * Geographic isolation, limited road network of roads, education, health services, San Marcos) utilities, etc.) P:\GENERAL\LCSHD\GUAPA\GREY COVER ENGLISH\GUAPA Executive Summary.doc February 11, 2003 9:13 AM ' The "Program for the Improvement of Surveys and Measurement of Living Conditions in Latin America and the Caribbean" (MECOV), is sponsored by the Inter-American Development Bank (IDB), the World Bank (IBRD) under the GUAPA program, and the Economic Committee for Latin America and the Caribbean (CEPAL) with additional funding from a number of other donors, including USAID, the Soros Foundation, UNDP, UNICEF, and the LO. 2In particular, the IDB, UNDP, USALD, UNICEF, the Soros Foundation and the ILO. 3Comparing estimates of poverty over time in Guatemala is complex due to large differences in survey and measurement methodologies (see Chapter 2). The estimates for 1989 are adjusted to make them more methodologically comparable to those for 2000. 4According to the Height-for-Age (HAZ) measure. xvii 5 This statistic refers to households in the ENCOVI questionnaire for which community-level data were available, see Chapter 10 for details. 6SEGEPLAN (November 2001). 7Estrategia de Reducci6n de la Pobreza (November 2001). 5 Though it is possible that some were poor before getting involved in the production of non-traditional exports. The single year nature of data coUlected in the ENCOVI do not allow for such analysis. 9 An adequate social protection system is an important element of a comprehensive strategy to reduce poverty and vulnerability. Social protection (SP) has been traditionally defined as "a set of public measures aimed at providing income security for individuals." (Holzmann and Jorgensen 2000). The final goal of public social protection policies is to increase the welfare of the population, and to that end, these schemes have generally included social assistance (SA) and social insurance (SI) programs. Social assistance programs are generally designed to help individuals or households cope with chronic poverty or transient declines in income that would cause them to live in a situation of poverty or worsening poverty. As such, they help alleviate poverty and reduce vulnerability to poverty. SA programs as a whole make up what is commonly referred to as "the social safety net," and include programs such as transfers (in cash or kind), subsidies, and workfare. Social insurance schemes include publicly-provided or mandated insurance for unemployment, old age (pensions), disability, survivorship, sickness, and so forth, which are designed to help mitigate income risks. Private transfers can complement public social protection interventions. '° Though there is a current move towards greater decentralization, with the passage of three recent laws by the Congress (the Ley de los Consejos de Desarrollo Urbano y Rural, the Codigo Municipal, and the Ley General de Descentralizaci6n). " The Ley de los Consejos de Desarrollo Urbano y Rural, the Codigo Municipal, and the Ley General de Descentralizaci6n. 12 These efforts are being led by SEGEPLAN under the ERP initiative. 13 Specifically, the Ley de los Consejos de Desarrollo Urbano y Rural, the Codigo Municipal, and the Ley General de Descentralizaci6n. xviii Chapter 1: Introduction CONTEXT FOR REPORT: THE GUAPA PROGRAM This report is part of a broader, multi-year program of analytical work and technical assistance that signals the commitment of the World Bank to poverty reduction and seeks to: (a) contribute to filling the crucial information gaps on poverty and living conditions; (b) deliver timely outputs on a regular and on-going basis in response to client requests and data availability; and (c) provide longer-term partnering and collaboration on poverty analysis and strategy. The four main interconnected "prongs" of the program include: (a) "GUAPA collaborative," which seeks to foster institutional development and capacity building in counterpart agencies for greater ownership and sustainability of the analysis and results. As such, the program adopts a collaborative approach, providing technical assistance and hands-on training to counterpart- agencies for poverty measurement, data collection, analysis, and policy and strategy formulation;' (b) "GUAPA analytical," which seeks to conduct a thorough, multi-dimensional analysis of poverty building on the framework of the World Bank's World Development Report for 2000/2001 using both quantitative and qualitative data (the main product is this present report); (c) "GUAPA policy," which seeks to contribute to the design of the Bank's upcoming Country Assistance Strategy (CAS),2 as well as the upcoming Country Economic Memorandum (CEM), the government's poverty reduction policies and strategies, and the poverty-effectiveness of interventions and policies in Guatemala; (d) "GUAPA operational," which seeks to forge linkages to lending operations currently under preparation to improve their poverty focus, such as those in social protection, education, and transport. This program is collectively called the Guatemala Poverty Assessment Program (GUAPA Program), which is intricately linked to the World Bank's support for the MECOVI Program in Guatemala (see Box 1.1). The main counterpart agencies for the GUAPA Program have been the Instituto Nacional de Estadistica - Guatemala (INE) and the General Planning Secretariat under the Presidency (SEGEPLAN) in collaboration with the University of Rafael Landivar (URL). Beyond this report, some of the outputs already generated with the support of the GUAPA/MECOVI program include the Living Standards Measurement Survey (Encuesta Nacional de Condiciones de Vida, ENCOVI 2000) and its database (see below), the poverty map (see Box 14.1 in Chapter 14), a profile of poverty prepared by the multi-agency technical team from INE- SEGEPLAN-URL, as well as the training and capacity building of numerous staff in these agencies. The work has also provided technical support to the Government's poverty reduction strategy (PRSP).4 Box 1.1 - The MECOVI Program: Filling a Critical Information Gap Guatemala does not have a strong tradition in collecting statistics on the living standards of its population. This largely reflects the three-decade long civil war which, until recently, made topics such as "poverty," "living conditions," and "social equity" taboo in government circles. As a result, Guatemala has an excessively large gap in terms of recent, systematic, and comprehensive data on poverty, social indicators, and living conditions. To fill these gaps and generate comprehensive, systematic, and integrated household survey data on living conditions, the Government of Guatemala requested to be incorporated in the "Program for the Improvement of Surveys and Measurement of Living Conditions in Latin America and the Caribbean" (MECOVI), sponsored by the Inter-American Development Bank (IDB), the World Bank (IBRD), and the Economic Committee for Latin America and the Caribbean (CEPAL). The MECOVI-Guatemala program also draws on financial inputs from a number of other donors, including USAID, the Soros Foundation, UNDP, UNICEF, and the ILO. The objectives of this program involve (a) developing an integrated system of household surveys, including two Living Standards Measurement Surveys (ENCOVI 2000 and 2003) and labor and income surveys in interim years (2001 and 2002); and (b) building national capacity for designing, implementing, and analyzing household survey data. One of the elements of the GUAPA Program involves providing both financial and technical support to this MECOVI Program. .1 OBJECTIVES OF REPORT This poverty assessment report has three main objectives. The first is to conduct an in-depth, multi- dimensional analysis of poverty building on the framework of the World Bank's World Development Report (WDR) for 2000/2001 using both quantitative and qualitative data. The second is to examine the impact of government policies and spending on the poor in key sectors. The third is to use the empirical findings to identify options and priorities for poyerty reduction in the future. Policy options are outlined not only in general, but for the specific themes and sectors covered. It is hoped that the report will make an empirical contribution to improving the Government's anti-poverty policies and strategies (broadly as well as for specific sectors). The report is also expected to contribute to the definition of the country assistance strategy (CAS) and lending operations of the World Bank and other donors, so as to make these interventions more effective in the future. Poverty Assessments are studies conducted by the World Bank in its client countries on a regular basis (usually every 3-5 years). They are not intended to be a critique of a particular government administration; rather, they assess the poverty situation of the country in question. They provide a basis for a collaborative approach to poverty reduction by country officials and the World Bank, and help to establish the agenda of issues for policy dialogue. ANALYTICAL FRAMEWORK Poverty is a multi-dimensional phenomenon in terms of its definition, measurement, manifestations, causes, and solutions. The analytical framework adopted for this report embraces this multi-faceted view, which was formalized in the recent World Development Report 2000. The poor are poor because they lack the resources to attain basic necessities - food, shelter, clothing, and acceptable levels of health and education. The poor are also particularly vulnerable to the impact of adverse shocks, with limited assets to be able to cope with them. Finally, a sense of voicelessness and powerlessness, particularly with regards to their representation and interaction with institutions, also characterizes the manifestation of poverty. An important aspect of this framework is the inter-action between these factors, which tend to reinforce each other. This analytical framework can be simplified into the following three inter-related facets of poverty (see Box 1.2):6 e Opportunity. Opportunity (or lack of it) to generate incomes and attain basic necessities is central to the manifestation of poverty. Economic growth is essential for expanding economic opportunities for poor people. Growth in turn depends on the functioning of markets, the policy environment, institutions, and initial conditions (such as geography, social fragmentation, initial incomes). The ability of growth to influence the opportunities of the poor depends not only on the pace of growth but also on the pattern of growth in the economy (e.g., favoring or disfavoring sectors which employ substantial segments of the poor population). The poor also rely on a portfolio of assets in order to forge opportunity, including human capital (their own labor, education, and health), physical assets (basic services, housing and land), financial assets (savings and credit), and social capital (horizontal and vertical connections, informal and formal organizations, etc.). Generally, the poor suffer from an unequal distribution of these assets. * Empowerment. The social and political context in which the poor operate affects their daily lives. A lack of political and social voice, discrimination, social barriers, conflict, inappropriate treatment by public officials, alienation from service providers, language barriers, and a lack of information about services and rights all affect the ability of the poor to achieve their potential and create a sense of voicelessness and powerlessness. Strengthening democratic processes, enforcing the rule of law, promoting participation in political processes and decision making, and removing social barriers are all important steps to empower the poor so that they can expand their opportunities and reach their 2 potential. These contextual factors are particularly important in Guatemala, where a history of conflict, exclusion and discrimination have constrained the mobility of the poor for generations. * Vulnerability. Vulnerability comes from the notion that certain groups in society are more vulnerable to shocks that threaten their livelihood and/or survival. Other groups are so vulnerable that they live in a chronic state of impoverishment where their livelihood remains a constant state of risk (e.g., street children). The sources of risk may be natural (e.g., hurricanes) or the result of human activity (e.g., job loss, conflicts, or violence). Low levels of assets (see above) make poor people especially vulnerable to the impact of negative shocks. Strengthening the assets of the poor enhances their ability to manage risks and weather shocks. Risk management interventions - such as social insurance or social assistance - can also help the poor mitigate and cope with shocks. Box 1.2- ANALYTICAL FRAMEWORK Growth ssets Institutions Economic Risks Governance I l l s l l |~~~~~~~~~ Natural Disasters| LEconomic Policy Human Capital I Social Socio-Political T ~~~~~FragmentationRsk Physical Assets (e.g., ethnic) Institutions Basic Services Housing Land Social Capital Health Risks Social I Fragmentation Financial Assets Initial Conditions 1 Social Capital (e.g., geography) | Source: K. indert, October 2000 based on WDR 2000 INFORMATION SOURCES: INTEGRATING QUANTITATIVE AND QUALITATIVE DATA An innovative aspect of the GUAPA is its combination of qualitative and quantitative research methods, which are both valuable in their own right but yield important synergies when used together. Quantitative Data on Living Conditions. Quantitative household and community surveys are suitable for gathering information that can be quantified, tabulated and analyzed statistically. They have the advantage of providing concrete data on a range of generalizable indicators for a statistically representative sample of the popAulation. 3 The main source of quantitative information is the Living Standards Measurement Survey (ENCOVI- 2000), which was conducted under the MECOVI Program (see Box 1.1).7 The ENCOVI provides a unique opportunity for poverty analysis, providing a comprehensive snapshot of living conditions for a representative sample of the Guatemalan population. The survey was executed by INE, with field work carried out during the period from July - December 2000. It covers a sample of 7,276 households and is statistically representative at the national level and for a number of strata including: (a) urban and rural areas; (b) eight regions8 (and urban and rural areas in these regions); and (d) ethnic groups according to the language classifications in the 1994 Census including: Ladinos (Spanish speakers), K'iche, Kaqchiqel, Mam, Q'eqchi, other Maya, and "other indigenous." The ENCOVI is based on the LSMS survey methodology which combines an integrated set of questionnaires to collect data on household living standards with extensive quality control features. The household questionnaire gathers information needed to generate monetary measures of poverty (using consumption and/or income). It also collects information on key assets and other living standards indicators, including: labor and migration, housing, land, basic services, transport, health, education, financial assets, household enterprises, agriculture, fertility and nutritional status. The Guatemala ENCOVI also includes a number of non-standard modules designed to collect information on the social dimensions of poverty, including: social capital, participation and citizenship, crime and violence, risks and shocks, and time use. The community questionnaire seeks responses from community focus groups about infrastructure and basic services (which allow for control variables), labor market conditions, product markets, formal social capital (e.g., community organizations), and a variety of qualitative issues such as perceptions of services, living standards, crime and violence, and key issues facing the community. A price questionnaire also collects necessary information for constructing detailed spatial and temporal price indices. Qualitative Data. Qualitative instruments, by contrast, are useful for gathering information on the influence of motives, attitudes and preferences on economic behavior, on perceptions, and on the barriers and opportunities that determine poverty and mobility. They are not intended to be statistically representative or reflect measures of central tendency. Rather, they yield information that is primarily descriptive but can broaden the field of inquiry to include questions, issues and factors in the quantitative instruments which are otherwise likely to be missed. A Qualitative Study of Poverty and Exclusion (QPES) provides the main source of qualitative information for GUAPA.9 The objectives of the QPES are to gather information on perceptions and the nature of constraints to and opportunities for economic mobility so as to better understand the dynamic processes that perpetuate or reduce poverty and exclusion. The QPES collected data in ten rural communities that are also included in the ENCOVI. The configuration of these villages seeks to examine perceptions of poverty and exclusion for a number of ethnicities; as such, the sample includes two villages from each of the following ethnic groups: Mam, K'iche, Q'eqchi, Kaqchiqel, and Ladino (non-indigenous). For the purposes of protecting the anonymity of respondents, each village is given a code name (e.g., MI, M2, Kl, K2, etc.). The field work covered a number of themes including: perceptions of poverty and welfare; perceptions of risk, shocks and vulnerability; social capital; user perceptions of public programs; community perceptions of education; and gender roles and issues. The main research instruments included: community focus groups, direct interviews, social mapFing, and observation. Annex 5 presents a summary of the main findings of each of the ten QPES villages.' OvERVEW OF THE REPORT This report is divided into five sections. Part 1 examines the magnitude and causes of poverty in Guatemala. Chapter 2 finds that poverty rates and inequality are among the highest in the LAC region. Moreover, Guatemala ranks among the worst in the region for various social indicators. Malnutrition rates in particular are abysmally high." Chapter 3 shows that poverty is determined largely by household 4 characteristics and endowments, including key assets such as human capital, physical assets, and social capital. Chapter 4 contends that other factors - both historical and contextual - have also fundamentally shaped Guatemala's performance regarding the levels of these endowments and characteristics. In particular, various forces in Guatemala's historically exclusionary pattern of development have put Guatemala at a relative disadvantage in terms of its development and key social indicators. Moreover, the long civil war imposed further costs on Guatemala's development. Chapter 4 then contends that the Peace Accords signaled a fundamental shift in Guatemala's pattern of development, paving the way to a more prosperous and inclusive nation. In the six years since the Peace Accords, Guatemala has taken important steps on this new development path, with progress in public sector management, public revenues and spending, and improvements in the coverage and equity of education and basic services. These steps signal that progress is possible, despite the magnitude of the challenge of changing the course of history. Nonetheless, challenges remain. Many of the key challenges for poverty reduction coincide largely with the remaining actions on the Peace Agenda. In particular, progress on the Peace Accords is deficient for key development targets, especially for those involving health and education outcomes, and economic growth. These outcomes reflect the need for poverty reduction and improvements in living conditions, which are crucial for lasting peace. They highlight remaining priority challenges in several key areas - which reveal the significant overlaps between the Peace Agenda and the Poverty agenda: (a) creating opportunities; (b) reducing vulnerability; and (c) improving institutions and empowering communities. Part 2 examines the first of these key challenges, specifically, the challenge of building opportunities and assets. Specifically, Chapter 5 examines the relationship between poverty and economic growth in Guatemala. Chapter 6 builds on this macro-economic context to further examine the livelihoods and earnings opportunities of the poor from a more "micro," household-level perspective, with a focus on rural livelihoods. The poor also rely on a portfolio of assets in order to forge opportunity, including education (Chapter 7), health (Chapter 8), basic utility services (Chapter 9), land and financial assets (Chapter 6), and access to transport (Chapter 10). Generally, the poor suffer from an unequal distribution of these assets. Part 3 examines the second challenge area, namely the challenge of reducing vulnerability and improving social protection. In Chapter 11, shocks, particularly natural disasters and economic shocks, are found to have a considerable impact on the poor. Poverty and vulnerability, however, are found to be primarily chronic - rather than transient - in nature, highlighting the importance of building the assets of the poor. Chapter 12 finds that the use of existing resources devoted to social protection could be greatly improved by eliminating certain programs and consolidating, improving the targeting and expanding the coverage of others. Part 4 examines empowerment as the third key challenge. Chapter 13 contends that these institutional forces are crucial for both the Peace Agenda and the Poverty agenda, influencing the both menu of options available to the Government in future efforts to reduce poverty and the way in which these options are carried out. Indeed, poverty and economic growth are not only driven by economic processes, but also by interacting economic, social and political forces. In particular, they depend on the accountability and responsiveness of state institutions. Social institutions - kinship systems, community organizations, and informal networks - also greatly affect poverty outcomes, helping communities manage public goods, cope with risks and shocks, and leverage external assistance. In light of the importance of these factors for both poverty and the Peace Agenda, this chapter reviews key institutional challenges in the areas of (a) public sector management; (b) governance; and (c) community participation and social capital. The chapter also considers the role of other important actors in development, namely the private sector, NGOs, and religious organizations. Part 5 (Chapter 14) identifies a menu of priority actions for poverty reduction efforts in Guatemala, building on the in-depth empirical analysis offered by the combination of quantitative and qualitative data from the ENCOVI and the QPES, as presented in preceding chapters. The chapter contends that these 5 priority actions are consistent with the overall principles of Peace Agenda. Indeed, reducing poverty and improving living conditions is central to lasting peace in Guatemala. ' Two grants from the World Bank's Institutional Development Fund (IDF) supported this approach, one to INE for the development of capacity for an integrated household survey system (TF027318) and one to SEGEPLAN for institutional strengthening for poverty analysis and policy and strategy formulation (TF027423). 2 'elinkages between the GUAPA and the CAS were greatly strengthened by support from a joint GUAPA-CAS Multi-Sector Team Learning Grant. In particular, this grant supported two retreats (to date) to discuss the findings of the GUAPA and their implications for the CAS, including the majority of the World Bank's Guatemala country team. 3 The MECOVI-Guatemala program also draws on financial inputs from a number of other donors, including USAID, the Soros Foundation, UNDP, UNICEF, and the ILO. 4SEGEPLAN: Estrategia de Reducci6n de la Pobreza (ERP), November 2001. 5 Poverty assessments are regular studies conducted by the World Bank in its client countries. They provide the basis for a collaborative approach to poverty reduction by country officials and the Bank and help to establish the agenda of issues for the policy dialogue. They are required for all client countries and are conducted on a regular basis (usually every few years). They are not intended to be a critique of a particular govemment administration; rather they assess the poverty situation as it pertains to the country in question. 6 This analytical framework is largely based on the framework set forth in the World Development Report 2000. 7Data from the ENCOVI 2000 are available to the public from by contacting NE at inedifusion@intelnet.net.gt. Documentation and information on the survey is available from the World Bank at http:H/www.worldbank.orgllsms,lsmshome.html a Guatemala, Norte (covering the departments of Baja Verapaz, Alta Verapaz), Nor-Occidental (covering Huehuetenango and Quiche), Sur- Occidental (covering Solola, Totonicapan, Suchitepequez, Quetzeltenango, Retalhuleu, and San Marcos), Central (covering Chimaltenango, Escuintla, Sacatepequez), Nor-oriental (covering Izabal, Zacapa, Chiqimula, El Progreso), Sur-oriental (covering Santa Rosa, Jalapa, Jutiapa) and Petdn. 9 Funding was provided by the Govemment of Denmark (1F039498). Field work was conducted in July and October 2000. ° The full report is also available on the web at: www.worldbank,orvAlacpoverty "Particularly for stunting (height-for-age), see Chapters 2 and 8. 6 PART 1: THE MAGNITUDE AND CAUSES OF POVERTY Part I of the report identifies the magnitude of poverty in Guatemala and examines its causes. Reflecting the multi-dimensionality of poverty, Chapter 2 examines the poverty "problem" using an array of monetary and social indicators, as well as perceptions of poverty identified by Guatemalan communities and households themselves. Likewise, echoing its multiple facets, poverty has many causes. In general, poverty is determined by key household endowments and characteristics. These are analyzed in Chapter 3. Yet historical forces and contextual factors also play a crucial role in shaping patterns of poverty. These factors are discussed in Chapter 4. Chapter 2: "The Problem:" Poverty and Social Indicators This chapter identifies the magnitude of the poverty problem in Guatemala. A systematic analysis of poverty requires some measure of welfare. Ideally, such a measure would (a) capture the multi-dimensional aspects of poverty; (b) be observable and measurable in a consistent way across households, space and time; and (c) be objective. Since no single measure fully captures all such features, living conditions should be monitored over time using a battery of indicators rather than with a single measure. This report considers both monetary and non-monetary indicators of well-being. The chapter starts with an analysis of poverty and inequality using monetary measures. Non-monetary indicators of poverty and living conditions - including malnutrition, health, education, basic services, composite indices, and perceptions -- are then explored. The chapter ends with a summary "report card" of key monetary and non-monetary indicators, for cross-sectoral comparison and easy reference. The main findings of this chapter are that the magnitude of the poverty problem is quite large. In fact, poverty rates and inequality in Guatemala are among the highest in the LAC region. Moreover, Guatemala ranks the worst in the region - and among the worst in the world - for malnutrition.' It also performs poorly for indicators of education, health, and basic utility service coverage, though it has made some progress in expanding the coverage of education and utility services. Finally, the chapter reveals interesting patterns about perceptions of changes in poverty and welfare over time. Specifically, while communities do perceive progress - and attribute this to improvements in basic services - households are decidedly more pessimistic about changes in their welfare since the Peace Accords. These more negative perceptions are attributed to economic factors, such as a lack of improvements in incomes and opportunities. MONETARY INDICATORS OF POVERTY AND INEQUALITY IN GUATEMALA2 Monetary Measures of Poverty: Consumption and Poverty Lines Typical monetary measures of welfare include income and consumption. This study primarily uses consumption because (a) it is fairly comprehensive;3 (b) consumption data tend to be more reliable than income data due to incomplete measurement, underreporting, and seasonality of income; (c) it tends to fluctuate less than income (which can even go to zero in certain months due to seasonality), making it a better indicator of living standards; and (d) consumption is less subjective than basic needs indices (BNI), which rely on some form of subjective weighting across their components. Using data from the ENCOVI 2000, two poverty lines were calculated: an extreme poverty line and a full poverty line.4 The extreme poverty line is defined as yearly cost of a "food basket" that provides the minimum daily caloric requirement of 2,172.5 The selected "food basket" is based on the average consumption patterns observed in the ENCOVI 2000 for the entire population.6 The annual cost of this minimum caloric requirement yields an extreme poverty line of Q.1,912. Below this level of consumption (or income), individuals cannot maintain the minimum level of caloric consumption even if all resources were allocated to food. The full poverty line is defined as the extreme poverty line (the cost of food that 7 satisfies the minimum caloric requirement) plus an allowance for non-food items. This allowance is calculated as the average non-food budget share for the population whose food consumption was around the extreme poverty line (Q. 1,912).7 It is assumed that, since these individuals barely meet the minimum caloric requirements, whatever share of total consumption they actually allocate to non-food must be essential. The analysis found the non-food share for this group to be 56%. This method yields a full poverty line of Q.4,3 19,8 below which individuals would be considered poor.9 Poverty in Guatemala: Levels, Trends and Patterns Poverty in Guatemala is very high. In 2000, over half of all Guatemalans - 56% or about 6.4 million people - lived in poverty (Table 2.1).'° About 16% lived in extreme poverty. International comparisons of poverty are always difficult due to various methodological differences (welfare measures, poverty lines, survey samples). Nonetheless, available evidence does suggest that poverty in Guatemala is higher than in other Central American countries, despite its mid-range ranking using per capita GDP.' It is also generally deeper (P1 measures) and more severe (P2 measures). The relatively12 lower incidence of extreme poverty in Guatemala is a reflection of the food consumption patterns of the population. As discussed above, the extreme poverty line is based entirely on the cost of food items. The Guatemalan diet derives a large share (46%) of calories from corn and corn products, which are also the cheapest source of calories (Q.0.00057/calorie). Comparisons with other extreme poverty lines are valid in the sense that they measure the capability of satisfying a minimum caloric requirement, but also taking into account the ability of the population to satisfy this requirement in the most efficient way.'3 As such, by taking into account the observed "cost efficiency" of the minimum caloric intake, the extreme poverty classification is a test that only a small share (15%) of the population fail to meet. Table 2.1 - Poverty in Guatemala, 2000 Poverty Indicators by Velfare Measure AU Poor (below EPLl Extreme Poor (below XPL) GNI Per % Poora | Depth' I Severity' % Poora ' Depth5 | Severityc Capita, PPP Using Consumption 56.2% 22.6 11.7 15.7% 3.7 1.3 $3,630 Using Income 65.6% 35.1 25.9 31.9% 15.1 22.2 $3,630 Sources: GNI (gross national income) per capita estimates for 1999 in PPP US$, from World Bank, World Development Indicators 2001. Poverty estimates calculated by INE-SEGEPLAN-URL with technical assistance from World Bank using the ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. a. Incidence of poverty or headcount index (% of population whose total consumption or incomne falls below poverty line, FPL or XPL). AU poor includes extreme poor (throughout study). b. The Poverty Depth Index (PI) represents the amount needed to bring all poor individuals up to the poverty line (FPL or XPL), expressed as a percent of the poverty line taking into account the share of the poor population in the national population. c. The Poverty Severity Index (P2) is a derivation of P1 that takes into account the distribution of total consumption among the poor. In other words, it is a measure of the degree of inequality among the population below the poverty line. The costs of reducing poverty in Guatemala are high. Given average consumption levels of the poor, it is estimated that the minimum annual cost of eradicating poverty represents about Qz. 11,121 million, or 8% of GDP (as compared with 5% in Panama and 17% in Nicaragua). Moreover, these costs are hypothetical annual costs. They are calculated as the cost of bringing all poor individuals up to the poverty line, excluding the inevitable administrative costs or leakages to the non-poor associated with virtually all poverty-alleviation schemes. These administrative costs and leakages can be quite high, as discussed in Chapter 12. To put this in context, total Government spending in Guatemala in 2000 was about 13% -- and spending on public investments was only 4%. Total public spending on health and education amounted to 3.6% of GDP. Public spending on social protection (assistance and insurance) was 1.8% and with another 1.7% by social funds. Clearly, the task of eliminating - or even reducing - poverty represents quite a challenge. Poverty appears to have fallen over the past decade. Comparing estimates of poverty over time Figure 2.1 - Tentative Trends in Poverty over in Guatemala is complex due to large differences in Time (1989-2000) survey and measurement methodologies (Box 2.1). 100% Until recently, the 1989 Socio-Economic Survey l0% 55 6%. provided the only nationally-representative measures 60% 56% of welfare and poverty. Using this survey, poverty 400% _,-, in 1989 was estimated at about 75% of the 20% population.'7 These estimates, however, were based 0% _ . u I n S, on a very simple income question (with only five 1989 1989 2000 (cons.) categories). To get towards better comparability, the Original Adjusted ENCOVI 2000 repeated these same exact questions (five categories, as well as other income questions for a more complete measurement). The results suggest that in 2000, the "1989 comparable income aggregate" was about 18% lower on average than the consumption aggregate. As such, poverty rates for 1989 were likely overestimated. Adjusting for these differences, poverty that year was probably around 62%. As such, poverty is estimated to have fallen by about six percentage points from 1989 to 2000 (Figure 2.1). These results are roughly consistent with what would be expected given past growth rates (and growth- poverty elasticities, see Chapter 5). Box 2.1 - Other Studies and Surveys of Poverty Three surveys allow for the calculation of poverty rates at the national level over the past decade (a) the 1989 Encuesta Nacional Sociodemografica (ENS); (b) the 1998-99 Encuesta Nacional de Ingresos y Gastos Familiares (ENIGFAM); and (c) the Encuesta Nacional Sobre las Condiciones de Vida (ENCOVI) carried out in 2000. While it is tempting to compare the results of these different surveys, differences in welfare measures (consumption vs. income; different ways to measure both), poverty lines, and samples often render such comparisons meaningless. Using these surveys, various studies of poverty in Guatemala have yielded the following results. Poverty Estimates for 1989. Using data from the 1989 ENS, CEPAL (1991) and the World Bank (1995) both estimated poverty to be quite high (80% and 75% respectively). The iricome variable collected in the 1989 ENS, however, was highly under- estimated as it was based on very restricted set of questions on income. Subsequent analyses by UNDP (2000) and the World Bank have attempted to correct for this over-estimation, yielding very consistent estimates (63% and 62% respectively).'4 Poverty Estimates for 1988-89. Using income as a welfare measure and data from the 1988-90 ENIGFAM, UNDP (2000) estimates poverty at 56.7% and extreme poverty at 26.7%. Comparisons between the results of the ENIGFAM and other surveys are complex, however, due to the fact that the sample for the ENIGFAM was only 20% rural and the survey questionnaire was not designed for the analysis of poverty but for the calculation of weights for consumer price indices (there are substantial differences in measurement of welfare and poverty lines between the ENIGFAM and other surveys). SEGEPLAN (November 2001) reports similar, but slightly lower results in its Estrategia de Reducci6n de la Pobreza (ERP): 54.3% for full poverty and 22.8% for extreme poverty. 16 Poverty Estimates for 2000. Using data from the 2000 ENCOVI and consumption as a welfare indicator, a multi-agency team consisting of INE, SEGEPLAN, and URL, estimated poverty at 56.2% and extreme poverty at 15.7%. These are the figures adopted in the present study. The ENCOVI 2000 offers many advantages over previous surveys in terms of poverty measurement and analysis, including a nationally representative sample with extensive rural coverage, detailed questions on consumption and income, and a broad range of topics covering the multi-dimensions of poverty and living conditions. As part of the MECOVI Program, INE expects to repeat comparable ENCOVI surveys every few years to allow for comparable monitoring of poverty and living conditions. Consistencies in Patterns and Trends Across Studies and Surveys. Despite difficulties in comparison, the various studies and surveys yield fairly consistent patterns. Over, time, poverty appears to have fallen from 62-63% in 1989 to between 54-56% by the end of the 1990s. Moreover, all studies suggest similar patterns, such as higher poverty and worse overall living conditions in rural areas and among the indigenous. The consistency of these trends and patterns is more important for policy than the exact number or rate of poverty in any given year. Nonetheless, with negative growth in GDP per capita, projections suggest that poverty may have increased slightly since 2000. As discussed in Chapter 5, Guatemala's economic growth rates have slowed substantially in recent years, reflecting a series of economic shocks. In fact, taking into account population growth rates, growth has actually declined in per capita terms. As such, poverty has likely increased. Taking into account these trends in the pace of growth, projections using the ENCOVI suggest that poverty likely rose slightly from 56.2% in 2000 to 56.6% in 2001, with extreme poverty rising from 15.7% to 16.0% (see Chapter 5). About two-thirds of all Guatemalan children live in poverty. Due to higher fertility rates among the poor, a large share of children live in poverty. In fact, 68% of children under six (about 1.7 million) and 63% of all children under 18 (about 3.8 million) live below the poverty line. In contrast, 45% of senior citizens over age 60 live in poverty.'8 This lower share of elderly in poverty (compared with 56% of the overall national population) could reflect lower life expectancy among the poor than the average or non-poor population. Poverty is not higher among households headed by women. Only a small share of households (14%) in Guatemala report having a woman as household head. Contrary to popular belief, households headed by men have slightly higher poverty rates than those headed by women (Table 2.2). Controlling for other factors, multivariate regressions show no significant difference between male and female household heads in determining poverty status (see Chapter 3). Poverty is predominantly rural and extreme poverty is almost exclusively rural. A disproportionate share of the poor and extreme poor live in rural areas in comparison with the share of rural residents in the national population (Table 2.2). Over 81% of the poor and 93% of the extreme poor live in the countryside. Three quarters of all rural residents fall below the full poverty line and one quarter live in extreme poverty. Poverty is significantly higher among the indigenous, but with differences between indigenous groups. As discussed in Chapter 4, Guatemala is a multi-ethnic society, including some 23 different ethno-linguistic indigenous groups (21 of which are Mayan). Although the indigenous represent about 43%19 of the national population, they account for 58% of the poor and 72% of the extreme poor. Over three-quarters of the indigenous population live in poverty, as compared with 41% of the non-indigenous. Poverty is also deeper and more severe among the indigenous (see Annex 4 for details on these indices). The ENCOVI 2000 also reveals that there are important differences in poverty rates between indigenous groups. The largest indigenous groups are the K'iche, the Kaqchiqel, the Mam, and the Q'eqchi. Among these, the Mam and Q'eqchi have the highest poverty rates. Pockets of poverty are present throughout the country, but there is also a significant "poverty belt" in the Northern and North-Western Regions. Poverty in Guatemala is a national problem, with pockets of poverty spread across the country (see Figure 2.2).2o Regionally, the ENCOVI shows that poverty is significantly lower in the Metropolitan region of Guatemala City. It is moderately high in the Nororiente and Central Regions of the country (Table 2.2). It is much higher in the Norte and Nor-Occidente Regions as well as the Department of San Marcos (which collectively comprise a "poverty belt"), which were largely affected by the country's three-decades long civil war (see Chapter 4). 10 Table 2.2 - Poverty Patterns in Guatemala % of National Incidence of Poverty Contribution to National Poverty Population Headcount Index (% of pop.) (% of AP or XP) All Poor Extreme Poor All Poor Extreme Poor Total Guatemala 100.0 56.2 15.7 100.0 100.0 By Geographic Area Urban 38.6 27.1 2.8 18.6 6.9 Rural 61.4 74.5 23.8 81.4 93.1 By Ethnicity Non-Indigenous 57.6 41.4 7.7 42.4 28.3 Indigenous 42.6 76.1 26.5 57.6 71.7 Kaqchiqel 8.9 62.6 13.6 9.9 7.7 K'iche 9.4 64.4 19.1 10.8 11.5 Q'eqchi 6.5 83.5 38.0 9.6 15.6 Mam 8.3 89.7 34.2 13.2 18.0 Other Indigenous 9.5 83.6 31.3 14.1 19.0 By Region Metropolitana 21.7 18.0 0.6 6.9 0.9 Norte 8.1 84.0 39.1 12.1 20.1 Nororiente 8.2 51.8 8.9 7.6 4.7 Suroriente 8.8 68.6 20.1 10.7 11.3 Central 10.7 51.7 8.7 9.8 6.0 Suroccidente 26.5 64.0- 17.0 30.1 28.6 Noroccidente 12.9 82.1 31.5 18.8 25.9 Peten 3.3 68.0 12.9 4.0 2.7 By Gender of Household Head Male 85.3 57.6 12.3 87.5 90.8 Female 14.7 47.8 8.4 12.5 9.2 Source: Poverty estimates calculated by INE-SEGEPLAN-URL with technical assistance from World Bank using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Metropolitana mainly covers Guatemala City (and Departmnent); Norte includes the Departments of Baja Verapaz and Alta Verapaz; Noroccidente covers the Departments of Huehetenango and Quichd; Suroccidente includes the Departments of SololA, Totonicapan, Suchitepequez, Quetzaltenango, Ratalhuleu, and San Marcos; the Central Region covers the Departments of Chimaltenango, Escuintla, and Sacatapequez; Nororiente covers the Departments of Izabal, Zacapa, Chiquimula, and El Progreso; Suroriente includes the Departments of Santa Rosa, Jalapa, and Jutiapa; and Peten covers Peten. Figure 2.2 - Poverty Rates by Department2' ...... = 1-3 Estimat~~~zxes base on te Poet Map Y* 35 60 SEPA-EULmneCnu14 -NI --A - 1 eg .......... . . . . . . . . . . . . . . . . . . . - - - . . . . . . 12B " & " ......... =e ffi~~~~. . . . . . . . . . . 1 § : 3 Se .. .. ... ..S: W ~ ~ ~ ~ ~ ~~ ~...... .......... 9 | g E------------- S | B ... .. ... ... .. . . . . . . . . . . - . . . . . . . . . . . . . . . . . . . . . . . . San~~~~~~~~~~ Marc" Qu etzaften an E E I 0 Q _ . .~~~~~~~~~~~~~~~ - - .5 ~~~~~~~~~~~~~~~~~~~1 - E 5 __ - ;, -,~~~~~~~~~~~~~~3 - 4 r EStiate base ontePvryMa 5-6 -15- 90 12 Inequality in Guatemala: Levels and Patterns22 International evidence suggests that inequality is a handicap to longer-term poverty reduction for two reasons.23 First, greater income inequality leads to lower investment in physical and human capital and hence slower economic growth - which translates into higher poverty. Second, cross-country evidence suggests that higher inequality results in a lower rate of poverty reduction at any given growth rate. Guatemala is among the more unequal countries in Table 2.3 - Inequality (Gini coefficients) the world. The Gini indices using consumption and Inequality (GINIs) income for Guatemala are 48 and 57 respectively. Using Consumption as Welfare Measure This is more unequal than most other LAC countries, Nicaragua (1998) 48 which as a whole has higher inequality than other Panama (1997) 49 regions in the world (Table 2.3).24 LAC median 46 Using Income as Welfare Measure The population distribution in Guatemala is Guatemala (2000) 57 Nicaragua (1998) 54 characterized by a large "low-income" majority Panama (1997) 60 and a very small "high-income" minority. As Honduras (1996) 55 shown in Figures 2.3 and 2.4, income shares climb El Salvador (1998) 52 slowly, with a fairly flat distribution, across the lower Costa Rica (1998) 46 quintiles. They then jump up dramatically in the top LAC median 55 quintile, which accounts for 54% of total consumption (almost three times higher than the next highest quintile and 11 times higher than average consumption in the bottom quintile). There are significant inequities across ethnic groups and geographic areas. Although the indigenous represent 43% of the population, they claim less than a quarter of total income and consumption in the country, with the non-indigenous accounting for 75%. Nonetheless, inequality as measured by Gini coefficients is higher among (within) the non-indigenous population than among the indigenous. Likewise, whereas rural residents account for almost two thirds of the nation's population, they claim only about a third of total income and consumption. Inequality within urban areas, however, is higher than among rural residents. Fgure 2.3 - tnequalltr % of totat consumption Flgumr 2.4 - Inquatlty 'Junpas: Ratios acrsas qufft les (for received by each quintUle consumption) 50% 543 2.8 40%4 30%~~~~~~~~~~~~~~ 28 -_% 12 1g | X 1 10%0 03 5 _ _ _ _ _ _ _ _ _ _ .l 02 03a02/t 03/02 04/03 05/04 NON-MONETARY INDICATORS OF POVERTY AND LIVING CONDITIONS IN GUATEMALA To complement monetary measures of poverty and welfare, this section examines patterns and trends in non- monetary indicators of poverty for a fuller understanding of the multi-dimensionality of poverty. Some non- monetary indicators considered here are topic- or sector-specific, malnutrition (measured by anthropometric data), health, education, and basic services. Others are composite indices, such as the Human Development Index proposed by the UNDP (which is actually a mixed indicator including income, health and education indices).25 13 Malnutrition:26 A Red Flag! Extremely Poor Performance, Little Progress Malnutrition among Guatemalan children is extremely high - among the worst in the world. Guatemala has among the worst performances in the world - and the worst in LAC -- in terms of child growth attainment, with an overall stunting rate of 44% of all children under five.27 As discussed in Chapter 8, there is a strong correlation between poverty and child malnutrition, as four fifths of malnourished children in Guatemala are poor (height-for-age). Malnutrition is much higher among poor children than non- poor children (64% of extreme poor and 53% of all poor children versus 27% of non-poor children). Malnutrition is also higher among rural and indigenous children than their urban or non-indigenous counterparts, as discussed in more detail in Chapter 8. Moreover, malnutrition is declining slower in Guatemala than in other countries. Guatemala has made some progress in reducing malnutrition, with stunting rates falling from 59% in 1987 to 44% in 2000. However, the yearly percentage reduction (1.7%) has been the slowest in the LAC region. This lack of adequate progress paradoxically contrasts with significant progress in other areas (basic services, health, education), as discussed in subsequent chapters. Health:28 Poor Performance, with Slow Progress Guatemala also ranks poorly for health indicators. Life expectancy at birth (65 years) is the lowest in Central America, and far lower than the average for LAC countries (70) or lower-middle income countries (69).29 Infant mortality (40-45 per thousand)30 is also the highest in Central America, and far higher than the average for LAC (30) or lower-middle income countries (32).3' Only Bolivia and Haiti perform worse for life expectancy or infant mortality in LAC; Guatemala does worse than other low-income countries, such as Nicaragua and Honduras. The patterns of health indicators also suggest worse conditions for the poor, rural, and indigenous populations (see Chapter 8). Though Guatemala has improved health outcomes over the past 20 years, its progress has been slower than the low-income countries of Bolivia, Nicaragua and Honduras. Education:32 Poor Performance With Biases Against the Poor, but With Some Progress Guatemalan literacy is not only below average in Latin Ameirica, it isfar lower. With an illiteracy rate of 31% in 2000, only Nicaragua and Haiti rank worse. Illiteracy among women, the poor, indigenous, and rural residents is particularly high (see Chapter 7). Despite this poor performance, Guatemala has seen improvements over time, with a slight quickening of the pace since the signing of the Peace Accords in 1996 (see Chapter 7). Although educational attainment is quite low in Guatemala, with significant gender, ethnfic and poverty gaps, there has been some progress over time. Guatemala is still a "primary" country, with an average educational stock of 4.3 years (for those aged 14+). Attainment is even lower among women, the poor, and the indigenous, and these gaps have been narrowing over time (see Chapter 7). Guatemala has made progress in improving primary enrollment, but coverage is still low and biased towards the non-poor. In the early 1970s, primary schools enrolled just over half3 of the target population. Net enrollment rates increased dramatically in only one generation, to 79% in 2000.34 Importantly, progress has been significantly faster in the years since the signing of the Peace Accords in 1996. Notwithstanding its commendable progress, primary coverage is still low by international standards. Guatemala's net enrollment is the lowest in Central America, and lags significantly behind the averages for LAC (91%) and lower- middle income countries (99%). As with other indicators, enrollment is lower among girls, the poor, indigenous, and rural children (Chapter 7). Coverage at the pre-primary, secondary, and superior levels are even lower and more biased against disadvantaged groups than at the primary level (see Chapter 7). 14 Basic Services: Important Progress in Improving Coverage and Reducing Inequities35 Overall coverage for basic services in Guatemala is typical for Central America. Overall, about 70% of Guatemalan households have water36 and electricity. Almost 90% have some kind of basic sanitation, though fewer than half have sewerage. About 20% subscribe to either a fixed line and/or a cellular telephone service. Around 17% of Guatemalan households do not have access to any kind of modem network utility service. While overall coverage rates are average for Central America, they lag slightly the average for Latin America and other lower-middle income countries. Coverage has accelerated considerably in recent years, with a targeted expansion for disadvantaged groups. Coverage indices for electricity, water, and sanitation increased far faster following the Peace Accords than in the years leading up to the Accords (see Chapter 9 for details). Moreover, data suggest that this expansion has been well targeted and that the poor, rural, and indigenous households have seen their probability of receiving service more than double following the Peace Accords, increasing more than for other segments of society. Even this substantial improvement, however, has not been enough to offset their historic disadvantage. As such, despite relative progress, in absolute terms, these groups still remain the least likely to receive services and a significant coverage gap remains (as discussed in more detail in Chapter 9). Composite Measure: The Human Development Index (HDI) Guatemala has made progress in raising the level of the HDI, but this level remains below what would have been expected given the level of GDP per capita of the country. The UNDP's Human Development Index (HDI) is a composite indicator of well-being that combines a weighted sum of three indices covering life expectancy, educational attainment, and per capita income (the higher the score the better). Because per capita income is included in the HDI, this index is a mixed indicator rather than a purely non-monetary indicator of well-being. On the positive side, Guatemala has improved its HDI since 1975 (Table 2.4). On the negative side, however, in absolute terms, the 1999 performance is well below the average for Central American countries. Moreover, despite a higher level of economic development, Guatemala's human development performance is below that of Latin America's Highly Indebted Poor Countries (HIPC). Finally, Guatemala is the only country in Table 2.4 for which the HDI ranking is far worse than its GDP ranking (a difference of 16 positions in the rankings). This suggests that, despite progress over time, the country still lags significantly behind in human development. Table 2.4: Trends in the Human Development Index, 1975-99 PRSP (HIPC) countries in LAC Non-HIPC Central America countries . HO BO GUY NI All CR ES GUA PA All HDI index (higher number is better) 1975 0.517 0.512 0.678 0.569 0.569 0.745 0.585 0.505 0.711 0.637 1980 0.565 0.546 0.681 0.580 0.593 0.769 0.584 0.541 0.730 0.656 1985 0.596 0.572 0.670 0.588 0.607 0.770 0.604 0.554 0.745 0.668 1990 0.614 0.596 0.676 0.596 0.621 0.789 0.642 0.577 0.746 0.689 1995 0.627 0.628 0.699 0.618 0.643 0.807 0.681 0.608 0.769 0.716 1999 0.634 0.648 0.704 0.635 0.655 0.821 0.701 0.626 0.784 0.733 1999 HDI and GDP ranking (lower number is better) GDPranking 112 111 93 113 107 47 86 92 67 62 HDI ranking 107 104 93 106 103 41 95 108 52 74 GDP-HDI ranking +5 +7 0 +7 +4 +6 -9 -16 15 -12 Source: UNDP (2001). 15 PERCEPTIONS OF POVERTY AND WELFARE IN GUATEMALA This section considers subjective perceptions of Guatemalan households and communities regarding definitions of poverty and welfare, changes over time, the main causes of these changes, and priorities for the future. Defining Poverty and Welfare Multi-Dimensional Perceptions of Welfare. The Qualitative Poverty and Exclusion Study (QPES) gathered perceptions of the concept of welfare from focus groups in 10 rural villages. Despite several interesting cultural and linguistic discussions (Box 2.2), there is considerable consensus across these villages as to the meaning of welfare and, to most, the concept is multi-dimensional. First, all study villages identify material aspects of welfare (Table 2.5). These include basic material necessities, such as food and clothing. Villagers also identify incomes Box 2.2 - Linguistic Approximations of "Welfare" (salaries) and health in this rubric In some of the indigenous communities in the QPES, focus group n participants discussed words that encompass the concept of welfare or well- noting the links between health and being. Some examples: physical conditions needed for working Maakd chik tink'aduxla - Not having to worry about anything (Q'eqchi and generating income. Participants in village, QEl) all study communities also identify Sa tatwanq - To live well (QEI) access to - and ownership of - land and Sahaqinch-ool - To be content (QEI) Wanq inna 'aj - To have somewhere to live (QEI) housing as key material aspects of Quicotemal - To be content: "Sundays we go to the market and I'm content welfare. Ownership was highlighted in to go to the market... There we buy our salt, sugar, and other things" particular by residents of fincas (KAI); (K'ich6 village, K12) who live in homes and work land that is Q'ino - In general, this term can mean "rich." It is also used for the name owned bytepatto,u oof a tropical fruit (jocote) which is "sabrosa" (delicious) (Mam village, M2) owned by the plantation, but do not own these assets themselves. Second, all villages identify access to public and communal services as being important aspects of welfare, including health services, water, education, roads, credit and technical assistance. The specific services vary somewhat, sometimes reflecting what the villages currently lack. Third, several villages equate welfare with having an education. In these cases, the concept of education transcends the merely productive, income generating aspects, with villagers noting the empowering importance of education for acquiring skills, talents, and especially knowledge to "permit them to think clearly so as to make correct decisions." Finally, participants in several villages note the social, psychological and spiritual aspects of welfare (Box 2.3). Most villages note the importance of having good social relations within the family and community as contributing to well-being. Examples include having a family life without conflicts, abuse (principally towards women or children), or alcoholism; and harmonious communal relationships (friendships, trust, mutual respect). Some note the importance of spiritual relations, emotional balance and happiness, and freedom (libertad). 16 Table 2.5 - Perceptions of Welfare from 10 Rural Villages, QPES Q'e,chi Mam K'iche Kaqcbiqel Ladino QEl QE2 Ml M2 Kll K12 KAI KA2 Ll L2 Material/Physical Dimensions _ _ Basic material necessities (e.g., food, clothing) X X X X X X X X X X Access and ownership of assets (land, housing) X X X X X X X X X X Public and Community Services X X X X X X X Having an Education (knowledge) X X X X Social, Psychological, Spiritual Dimensions I Having good social relations (family, communal) X X X X - X X X Spiritual relations, emotional balance X _ X X X Freedom X _ X Source: Qualitative Poverty and Exclusion Study (QPES), COWI/World.Bank (2001). Box 2.3 - Definitions of Welfare: Examples from QPES Villages Material/Physical Definitions: --"What we think about is our work, our milpa (com). We all have this necessity. We are happy because we have enough to eat from our milpa. Well-being means having com and beans." (K12) --Welfare is "having money to buy what we need to live well...food... not suffering from hunger" (M2) --"Having food... having clothing... having money... living well... having a job." (KA2) --"Having a parcel of land would be adequate... Having our own land and house for our family." (KAI, finca village) --"Having land... Having a house with two stories." (QEI) Public/Community Services: --'To have water" (K12); 'To have enough water" (M2) --'To have midwives in the community; to have health centers, hopsitals... to have the possibility of credit/borrowing" (Ml) --To have "a good road (un buen camino)." (KA2) Having an Education: --"That we have knowledge" (QEI) --"When children study its to improve their futures." (QE2) --"To know how to read and write... To have an education." (KA2) Social/Emotional/Spiritual Dimensions: --"Welfare is that husbands don't hit their women... that we don't fight, my husband doesn't fight or hit me" (KAI) --"Peace... that there are no problems... that there are no deceptions... that thee is peace in the family and the community... that there is no alcoholism... alcoholism creates mistrust." (QEI) --"Welfare is to look for God... although my house might not be good, I look for God... Only God can help us." (KAI) --"To feel good with God... Although we don't have anything, yet we are good with God, we participate in church and we have God in our hearts... to be at peace... to be content." (L2) --"To have freedom for people small and large. To have freedom to look for work in other communities... That the woman has the freedom to participate." (LI) Perceptions of Poverty. Much like with welfare, the QPES villages all point out the multi-dimensionality of poverty. First, all study villages equate poverty with a lack of basic material necessities, such as food and clothing (Table 2.6). Most also associate it with a lack of access or ownership of assets (land, housing). These results are very similar to those found in an earlier study of perceptions of poverty by von Hoegen and Palma in which 89% of all interviewed (560/627) identified that the absolute priority was food as a productive input ("No tener dinero para comer").37 A significant share of those interviewed in this earlier study also identified poverty not only with hunger (food), but also with housing ("No tener lote ni vivienda propia"). Second, the majority of villages link poverty to a inadequate public or community services and particularly education. The von Hoegen and Palma study also found a large share of those interviewed attributing poverty to a lack of education and training (368/627 interviews). Third, poverty is associated with a social, cultural, and spiritual dimensions in many of the study communities. 17 Box 2.4 - Definitions of Poverty: Examples from QPES Villages Lack of Material Necessities/Physical Assets --"Not having food... not having work... not inheriting (land)... not having a house or having a house in bad condition." (LI) --Poverty is "when there's not enough to eat... when there is no money to buy medicines." (KA2) --"Not having anything... not having work, there is no land." (L2) --Poverty is "little land; little harvests (milpa)... due to a lack of inheritance." (MI) Inadequate Public/Community Services: --"Not having potable water in all the houses." (M2) --"Death for not bringing the sick quickly to the health post or hospital." (LI, where there are no health services) Inadequate Education: --"Not having an education. The school offers only 6 years... the lack of training." (M2) --Poverty is "the lack of education in the community... Many are still illiterate, they can't communicate with everyone else." (KIt) Social/Emotional/Spiritual Dimensions: --"If we were united, we could look for markets and export our products... The rich are always united, we the poor are always disorganized... the problem is that we are not united." (Kil) --Poverty is "misunderstanding in marriage... when the woman is good and the man is bad or the reverse.. having a disordered life and not looking for God." (KAI) --"The ancestors used natural medicines; curing with natural medicine is being lost, now only with doctors and in health centers. The woman is poor because she doesn't worry about this.... --Poverty is "a lack of community organization... there are no leaders." (MI) Table 2.6 - Perceptions of Poverty from 10 Rural Villages, QPES Q'e chi Maim K'iche Ka chiael La ino QOE QE2 MI M2 KI I K12 KAI KA2 LI L2 Material/Physical Dimensions Lack of basic material necessities (e.g., food, clothing) X X X X X X X X X X Lack of access/ownership of assets (land, housing) X X X X X X X X Inadequate Public and Community Services X X X X - X X X - Inadequate Education X X X X X X Social, Psychological, Spiritual Dimensions Lack of good social relations (family, communal) X X X X X Lack of spiritual relations X - X X X Lack of freedom, cultural identity X =_==_ Source: Qualitative Poverty and Exclusion Study (QPES), COWtIWorld Bank (2001). Perceptions of Changes in Poverty and Welfare Over Time: Possible Paradox?38 The ENCOVI 2000 collected data on perceptions of changes in welfare and poverty over the "past five years"(period from 1995-2000). This corresponds with the period following the Peace Accords, and hence offers a unique opportunity to gauge perceptions of progress over this crucial period. Such data were collected in both the household and community questionnaires. Households and communities report strikingly different perceptions of changes in welfare over time, suggesting a possible disconnect in household and community welfare models. While there is a clear perception of improvements in welfare at the community level, such gains are not perceived at the household level. Perceptions of community-level welfare seem to be based more on the provision of "public goods," such as basic services and education, which indeed have seen improvements since the signing of the Peace Accords (as discussed above). Such progress is not acknowledged at much with respect to household welfare. Perceptions of changes in household welfare seem to be driven primarily by economic factors that more directly affect the "wallets" of consumers - employment, incomes, and prices. 18 Communities do perceive that community welfare and living conditions have improved over the past five years. Over half of all communities sampled in the ENCOVI 200039 perceive welfare has improved at the community level Figure 2.5- Perceptions of Changes in Household over the past five years (1995-2000), which and Community Welfare over Time, ENCOVI 2000 spans the period covering the Peace Accords 0 (Figure 2.5). Only 10% perceive that 0 community welfare has worsened, with the 0 60 51 57 remainder perceiving no change in living - 50 34 0 4037 *Worse conditions. These results are almost identical 20 2- 1 Bette for urban and rural communities alike. E 120 j] 0 SBet Interestingly, with a higher share of indigenous 8 o __ communities surveyed report improvements in 0 ,e?' 9 ' welfare than non-indigenous. In particular, a 'q& ROO large share of K'iche (72%) and Q'eqchi (63%) communities (which were indeed affected by the war) report perceived improvements in welfare during the period following the Peace Accords. Communities mainly attribute these gains to improvements in public services and education. Over half of those perceiving improvements in community welfare attribute this progress to improvements in public services, such as water, sanitation, energy, and transport. The next strongest factor is education (13%). These perceptions are indeed consistent with the gains observed for the extension of public services and education (as discussed above and in Chapters 9 and 7 below). In contrast, perceptions of changes in household welfare are decidedly more pessimistic than communities. Interestingly, the same community focus groups that ranked community living standards as having improved were much less positive in their ranking of changes in households welfare over time.40 The majority of community focus groups in the ENCOVI sample perceive no changes in household welfare, while only a third indicated that household-level welfare had improved over the past five years, with 20% saying household welfare had worsened. Household perceptions of changes in their own welfare are consistently pessimistic (Figure 2.5). The bulk of poor households perceive that living conditions remained stagnant over the past five years, whereas perceptions of non-poor households were spread more across the spectrum (with a larger share perceiving both improvements and a worsening of living conditions). This more pessimistic view of household welfare is attributed mainly to factors that more directly affect consumption and incomes. Households and communities are quite consistent in their reporting of the factors that cause changes in welfare and poverty at the household level. Both link household welfare to factors that directly affect the purchasing power of consumers. The labor market - unemployment, employment opportunities, and incomes - is by far the top explanatory variable for household welfare and poverty in the view of both households and communities. Prices (cost of living) are the second most cited factor (by both communities and households). Education and corruption/bad government are also mentioned as factors affecting poverty, but by a much smaller share of households. It seems that despite observed progress in basic services and education, Guatemalans are more concerned with factors that more directly affect the "wallets" of consumers in basing their views of household welfare. These findings are consistent with those of the QPES, in which virtually all communities equated welfare with the satisfaction of material necessities, such as food and clothing, as well as the ownership of productive physical assets, such as land and housing (as discussed above). They are also consistent with the findings of the von Hoegen and Palma (1999) qualitative survey of perceptions of poverty in which low incomes, the cost of living, a lack of land, and a lack of employment opportunities were ranked by interviewees as the top four causes of poverty. 19 Nonetheless, when asked what are the priority problems for communities, households seem to revert to a "public goods" view of state intervention. Despite their emphasis on "livelihoods" as key factors affecting household welfare, very few households (less than 3%) identify labor market factors (employment opportunities, unemployment, incomes and salaries) as priority problems for resolution in their communities. Instead, public services - such as water, energy, telecom, and transport - were the main priorities cited by households (50%), followed by health (11%), social problems (8%, including violence, alcoholism, family problems), and education (5%). 20 Table 2.7 - Summ Report Card for Monetary and Non-Monetary Indicators of Poverty and Living Conditions Level (most recent International Comparisons Changes Over Time rty g Worse for: year) I_I Poor Rural Indigenous By Re ion Poverty (2000) 56% Worst in Central America Improving as expected not applicable Worse Worse National problem (CA), below LAC average with growth rural: 75% indig: 76% Worse for: Norte I _________________ __________I_____________ I___________________ __________________ [urban: 27% non-indig.: 41% Noroccidente Health and Malnutrition Malnutrition (2000) 44% Worst in LAC, among worst Little progress (slower Worse Worse Worse Same as poverty children 0-5 in world than other countries) poor: 53% rural: 51% indig.: 58% .________________ ________________________ ____________________ non-poor: 27% urban: 33% non-indig: 33% Infant mortality 40-45 depending on Among worst in LAC, worst Slow progress (slower Mixed results, worse No Worse Central, Sur-occidente, (1998/99) source in CA than other countries) in lower "wealth" indig. 56% Nor-oriente (finca quintiles non-indig. 44% zones) Diarrhea, 31.3% n.a. Worsening Worse Worse Worse Norte children 0-5 poor: 33% rural: 35% indig.: 35% (much worse) (2000) non-poor. 27% urban: 25% non-indig.: 27% ARI 47.9% n.a. Worsening Not strong pattem Worse Not significant Norte, Nor-oriente children 0-5 rural: 51% (2000) urban: 41% Dengue (1999) 3.3 per 10,000 n.a Improving n.a. n.a. n.a. n.a. Cholera (1999) 18.7 per 100,000 Education Illiteracy 31% illiterate Worst in LAC except Haiti Improving over time, Worse Worse Worse Same as poverty adults aged 14+ 22% men and Nicaragua espec. since Peace poor: 46% rural: 42% indig. 49% (2000) 39% women Accords non-poor: 17% urban: 16% non-ind.: 20% ind. women: 62% Educational 4.3 years n.a. Improving over time Worse Worse Worse Same as poverty Attainment 4.9 men poor: 2.4 yrs rural: 2.7 yrs indig: 2.5 years adults aged 14+ 3.8 women non-poor: 6.4 yrs urban: 6.4 yrs non-ind: 5.5 yrs (2000) ind.women: 1.8 Net enrollment - 79% Lowest in CA Improving, especially Worse Worse Same as poverty Primary 81% boys Below LAC and lower- since Peace Accords poor: 78% rural: 75% indig: 75% (2000) 76% girls middle income averages non-poor: 90% urban: 85% non-ind.: 84% ind g ids: 67% Basic Services (Coverage all for 2000) Piped water in 69% Average for CA, below LAC Improving, especially Worse but Worse Worse Worse in Norte and dwelling or yard and lower-middle income since Peace Accords improving rural: 53% indig: 62% Peten averages poor: 56% urban: 87% non-indig.: 73% non-poor 79% Sanitation (latrines, 87% Average for CA, below LAC Improving, especially Worse but Worse Worse Worse in Peten, septic tanks, and lower-middle income since Peace Accords improving rural: 79% indig: 84% Suroriente, Nororiente sewerage) averages poor: 79% urban: 98% non-indig.: 88% non-poor: 94% Electricity 73% Average for CA Improving, especially Worse, but Worse Worse Worse in Norte, since Peace Accords improving rural: 56% indig: 61% Peten, Nor-occidente, poor: 54% urban: 95% non-indig: 81% Nor-oriente non-poor: 89% Telecommunications fixed phone: 15% Average for CA Improving, especially Worse: Worse: Worse: Worse in Norte, Peten, cell phone: 10% since Peace Accords poor: 1% rural: 3% indig: 4% Nor-occidente, Sur- non-poor: 27% urban: 31% non-indig: 22% oriente Sources: World Bank calculations using ENCOVI 2000 - Instituto Nacional de Estadistica - Guatemala unless otherwise noted. Other sources: intemational comparisons mostiy come from World Bank World Development Indicators 2001; infant mortality indicators from DHS 1998/99; denguetcholera from Ministry of Health. All numbers are cited and explained in more detail (along with sources) in the technical papers (Annexes) to this report. 21 ' Particularly for stunting, see below and in Chapter 8. 2 Most of these results come from collaborative work conducted jointly by INE, SEGEPLAN, and URL with technical assistance provided by the World Bank under the GUAPA/MECOVI programs using data from the Encuesta de Condiciones de Vida (ENCOVI 2000), Instituto Nacional de Estadfstica - Guatemala. Detailed tables are included in the Statistical Appendix (Annex 4). 3In terms of comprehensivity, consumption covers different sources of consumption (purchased and non-purchased including consumption of own-produced products, autoconsumo). It also provides wide coverage of the multiple dimensions of welfare, including basic material necessities (such as food, clothing), the current value of physical assets (such as land, housing), and the consumption of basic services (e.g., water, energy), health and education. Other measures, particularly basic needs indices, consider only a fraction of these components (e.g., excluding basic material items such as food and clothing). These poverty lines were calculated using data from the ENCOVI 2000 by an inter-agency technical team from [NE, SEGEPLAN, and URL with technical assistance provided by the World Bank under the GUAPA and MECOVI programs. 5 This minimum average daily caloric requirement was estimated by the Instituto de Nutricion de Centro Amirtica y Panamd (INCAP), representing a weighted average based on the assumption of moderate activity, taking into account the actual age and gender distribution of the Guatemalan population in accordance with official population projections. 6 The reference group excludes the bottom and top 2% of the population to avoid extreme values. 7Households with a yearly per capita food consumption within 5% of the extreme poverty line value. 8 Q. 1,912 (44%) for food, and Q. 2,407 (56%) for non-food items. 9An advantage of this method for calculating the extreme and full poverty lines (in contrast to those methods that use the cost of a pre-detemined "normative" basket of food and non-food items) is that it does not impose assumptions about the consumption preferences and pattems of the population. These patterns are empirically observed for the full Guatemalan population (rural and urban alike) - rather than imposed by the analyst. '° Poverty rates were estimated using data from the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala by a multi-agency technical team from INE, SEGEPLAN and URL with technical assistance from the World Bank's GUAPA team. See Annex 4 for additional results on poverty pattems as well as sensitivity analysis. " For example, poverty rates in Nicaragua and Panama were 47.9% and 37.3% respectively (using consumption as a measure of welfare). 12 Relative to the overall incidence of poverty. 13 In other words, sometimes differences in poverty rates reflect not only the economic conditions of the country, but also the endowments and the customs of the population. This characteristic does not diminish the validity of the methodology, but rather enhances it because, since the objective is to measure welfare, it takes into account not only what people have (endowments), but also how they use it (e.g., consuming the relatively cheaper com). 14 The adjustments by the World Bank take advantage of the fact that, for purposes of comparison only, the ENCOVI 2000 repeated the same set of income questions from the 1989 ENS (along with an expanded set for better measurement). An analysis of the "1989 comparable" income aggregate using these comparable questions shows that income was underestimated by about 18%. Hence poverty rates were probably about 18% lower than originally estimated for 1989. 15 The results reported by SEGEPLAN (November 2001) were based on joint work by SEGEPLAN-INE-URL to construct the poverty map by combining data from the 1994 Census and the 1998-99 ENIGFAM. They are based on projected consumption estimates. 16 The results reported by SEGEPLAN (November 2001) were based on joint work by SEGEPLAN-INE-URL to construct the poverty map by combining data from the 1994 Census and the 1998-99 ENIGFAM. They are based on projected consumption estimates. " World Bank (1995). 18 All comparisons presented in this report that are based on ENCOVI 2000 data are statistically significant at the 90 percent level or more. 19 This estimate is from the Instintuto Nacional de Estad(stca - Guatemala using the expanded sample of the ENCOVI 2000 and updated sample weights based on projections from the 1994 census. Other estimates put the indigenous population as high as 60% (CELADE, 1994). 20 This map will be updated using data from the ENCOVI rather than the ENIGFAM for the final version of this report. 21 This map will be updated using data from the ENCOVI rather than the ENIGFAM for the final version of this report. 22This section is based primarily on the results of the ENCOVI 2000 - Instituto Nacional de Estadtsrica. Detailed tables can be found in the Statistical Appendix, Annex 4. 23 See Deininger and Squire (1997), Ravallion and Chen (1997), and Ravallion (mimeo, February 12, 1998) for a survey of cross-country evidence. 24 As for poverty, inferring distributional changes for income over time is a challenging task due to differences in survey designs and methodological issues. Nonetheless, using the "1989 comparable income aggregate," the Gini coefficient for 2000 is 61, compared with 60 for 1989, suggesting that income inequality has slightly increased during the last decade. Source: World Bank calculations using data from the ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala 25 Other parts of this report do attempt to examine further aspects of living conditions, including crime and violence, justice, isolation and transport, and empowerment, reflecting the multi-dimensionality of poverty and welfare. 26 See GUAPA Technical Paper 6 (Marini and Gragnolati, 2002) for details. 27 These stunting rates (height for age) are consistent for various recent surveys (ENCOVI 2000: 44%; DHS 1998: 46%). See Chapter 8 for additional details on malnutrition in Guatemala. 28 See GUAPA Technical Paper S (Gragnolati and Marini, 2002) and Chapter 8 for more details. 29 Source: World Bank (2001b). 30 Estimates of infant mortality for Guatemala vary. The 1998/99 Demographic and Health Survey (DHS) puts the average at 45. The World Bank World Development Indicators (2001), based on official health statistics, puts the estimate for that same year at 40. " World Bank (200 lb). 32 See Chapter 7 and GUAPA Technical Papers 2, 3, and 4 for details. " World Bank WDI 2001. 34 Source: World Bank calculations using ENCOVI 2000 - Instituto Nacional de Estadfstica Guatemala. 35 Refer to matrix at end of chapter + Working Paper!!!! 36 Defined as piped water to dwelling or yard. Note that piped water does not guarantee potable water in Guatemala. 37 von Hoegen and Palma (1999). 38 All statements about perceptions of changes in welfare and poverty over time in this section are based on the results of the ENCOVI 2000. See Annex 4 for details. 39 The ENCOVI community interviews were carried out as focus group discussions with a variety of community members. 40 Both households and communities in the ENCOVI were asked about their perceptions of changes in household welfare. Perceptions of changes in community welfare were only collected in the community survey. 22 Chapter 3: The Determinants of Poverty and Inequality: Endowments and Household Characteristics This chapter analyzes the relative importance of key household characteristics and disparities in assets as correlates of poverty and inequality in a multi-variate setting. Three key conclusions emerge from the analysis: * The determninants of poverty and inequality overlap substantially - and most involve crucial policy levers. These same factors also play a role in determining malnutrition, as discussed in Chapter 8. * Disparities in key assets - human capital, physical assets, and social capital- are indeed strongly correlated with poverty and inequality. * These factors are similar to the correlates of poverty and inequality in other countries. In other words, the relationship between key household characteristics and assets and poverty is similar across countries. Nonetheless, the level of endowments held by Guatemalan households is determined by complex historical processes and contextual factors, which shape the poor performance observed in Guatemala today. These historical processes and contextual factors are discussed in Chapter 4. THE DETERMINANTS OF POVERTY This section analyzes the correlates of consumption in a multi-variate setting. The analysis is useful, first, to verify the relative role of various household characteristics and endowments in determining poverty status, and second, to assess the potential impact that policy-inducing changes in these factors are likely to have on poverty, holding all other factors constant. The correlates of consumption were examined for the nation as a whole, as well as for urban and rural areas.' A positive coefficient on a variable in the regressions signals that that variable is associated with higher consumption2 (and hence lower chances of being poor). The results are presented in Table 3.1 (below). It is important to note the limitations of this analysis at the outset. First, the analysis does not capture the dynamic impact of certain causes of poverty over time. For example, historical processes of exclusion cannot be directly captured (though they are revealed to some extent in the level of endowments observed). These factors are nonetheless important, and as such are discussed in Chapter 4. In addition, the impact of changes in economic growth - most certainly a key determinant of poverty - cannot be assessed using this static, cross-section model. The issue of growth and poverty is discussed in Chapter 5. Second, the analysis is limited by the variables available at the household level from the ENCOVI 2000. Other factors - such as social conditions or physical conditions (e.g., variations in climate or access to markets) - could not be included due to a lack of data at this level. Finally, although theory holds that many of the variables included in the analysis do indeed contribute to (cause) poverty (or poverty reduction), the statistical relationships should be interpreted as correlates not as determinants, since in some instances, causality can run both ways. Poverty is clearly associated with lower levels of key assets, including labor, education, physical assets, social capital and infrastructure characteristics for the municipality. Geographic location and household size are also found to be important correlates of poverty. Key findings are summarized below: * Education,3 particularly higher levels, can serve as an elevator for upward economic mobility. As discussed in Chapters 2 and 7, educational disparities are quite high in Guatemala. The higher the educational attainment in the household, the higher the household consumption, and hence the lower the chances a household lives in poverty. Overall, having someone in the household who has completed primary schooling raises consumption by 23%4 (for an average yearly return in total consumption of 3.8% per year of schooling completed). The gains are even stronger with higher 23 levels of education. Having a member of the household who has completed secondary schooling raises consumption by 46% (for an average yearly return in total consumption of 7.5% per year of schooling completed). Having more than one member with an education also significantly increases consumption, and hence reduces the likelihood that a household will be poor. The impact of a second member in the household with formal education is especially strong for completed secondary and university studies (40% and 66%.) 'Working idn agriculture, blue collar jobs or as a domestic servant is strongly correlated with poverty. Households in which the highest income-earner6 works in agriculture have significantly lower consumption levels (and hence are more likely to be poor) than those depending on work in other sectors (e.g., construction, commerce, transport or services). Likewise, those households in which the highest income-earner has a blue collar or domestic servant job have significantly lower consumption levels than those in other types of jobs. Interestingly, unemployment (no work) and under-employment do not have strong effects on consumption. Rather it is the sector (agriculture) or type of job (blue collar, domestic servant) that drives lower consumption levels and poverty status. These results are not uncommon. Unemployment is rarely associated with poverty in countries with little or no safety net (e.g., unemployment benefits). The poor simply cannot afford to be unemployed. They generally work, but in lower paying sectors and jobs - as found in the present analysis. Larger agricultural land holdings7 are associated with higher consumption in rural areas.8 Possessing and working plots of land between 1.5-3.5 hectares raises rural consumption levels from 4.3% to 10.2% total (by 2.9% per hectare9) over those with no land. Similarly, working plots between 3.5-10.5 hectares raises rural consumption levels up to 23.1% (by 2.2% per hectare) over those with little or no land. The returns to larger plots are significantly lower on a per hectare basis but the total values are higher: up to 30% for plots between 10.5-50 hectares (0.6% per hectare), and a 0.3% per hectare for larger than 50 hectares plots. (see Table 3.1). Small plots (less than 1.5 hectares) are not associated with lower (or higher) consumption levels. O Households in municipalities with more access to basic utility services are significantly less likely to be poor. As discussed in Chapters 2 and 9, disparities in access to basic utility services are quite large. In urban municipalities, the level of access to sewage connections is associated with consumption levels that are up to 12% higher than those without access.'° Municipal electricity connections is associated with higher consumption levels in both urban (32% higher) and rural (23% higher) areas. The municipality variables can be interpreted as a proxy for local endowments and represent the effect of local characteristics not captured by other variables." a Participation in "bridging" social capital is correlated with higher levels of consumption. As discussed in Chapter 13, bridging social capital refers to the bridges (horizontal connections) that people make with a broader group of people who have comparable economic status, education levels and political power. School groups, professional associations, groups that manage community-level public goods, and social groups are examples of bridging social capital.'2 Households with members who participate in bridging activities and organizations tend to have higher levels of consumption. The direction of causality is not clear, however, since households could be better off due to such connections, or they could participate more because they are better off. Nonetheless, the correlation remains between bridging connections and higher. economic status. 3 Even after accounting for disparities in assets and endowments, geographic location is a major factor in explaining poverty in Guatemala. Households located in any region are more likely to be poor than those residing in the Metropolitan Region. Those residing in the Sur-Oriente, Norte, and Peten regions have significantly lower levels of consumption than those in the capital region (42%, 24 36%, and 32% respectively). Interestingly, after controlling for other assets and characteristics (such as ethnicity, language ability, agricultural employment, education, region, and municipal infrastructure), residents of rural areas are not significantly more likely to be poor than their urban counterparts. * Spanish speakers and non-indigenous households are less likely to be poor." Even after accounting for other differences, such as education, ethnically indigenous households and those who do not speak Spanish have lower levels of consumption and are thus more likely to be poor. Spanish-speaking ability raises consumption levels by about 9%, indigenous households tend to have consumption levels that are 17% lower than otherwise similar non-indigenous households. Still, additional analysis suggests that the main source of welfare differences between indigenous and non- indigenous households is due to low asset holdings, as opposed to differences in the marginal returns to these assets. 14 * Larger households tend to be poorer, particularly those with many young children. Overall, each additional child under six years old lowers total consumption by 23% (higher in rural areas than urban); each additional member from age 7-24 lowers total consumption by 17%. The magnitude of these values suggests that increased awareness and use of family planning methods could have a significant effect on reducing poverty. Interestingly, despite the numerous disadvantages faced by women in Guatemala (e.g., access to education, labor returns, as discussed in other chapters), households headed by women are not inherently more likely to be poor than those headed by men once other factors are taken into account. * Shocks - such as natural disasters and economic crises - are also associated with higher poverty. These are analyzed using additional multivariate regressions in Chapter 11. THE SOURCES OF INEQUALrry The sources of inequality in Guatemala overlap substantially with the correlates of poverty. As with poverty, disparities in assets and endowments - labor, education, housing, land, and basic services - explain a large share of inequality (Table 3.2 below). Other sources of inequality include geographic location and household size and composition.'5 Three quarters of Guatemala's consumption inequality can be explained by disparities in education, certain services, labor (agricultural work), land, ethnicity, geography, and household size.'6 Education constitutes the single most important determinant of inequality, accounting for over half of all inequality in Guatemala. This reflects the large disparities in access to education and the large wage differentials associated with various levels of educational attainment (as discussed in more detail in Chapter 7). The long-term nature of many of these variables is significant: as international experience has shown, it is very difficult to reduce inequality. Many of these factors are somewhat structural (e.g., geography), difficult for Governments to change (land distribution, sectoral employment opportunities), or affect the distribution of consumption mainly in the long run (e.g., education and fertility). As such, inequality is a particularly "stubborn" variable, quite resistant to change over time.'7 25 TabRe 3.1 - The Correllates of Poverty Dependent Variable Natural Log of Yearly per capita consumption All rban Rural Urban (Rural excluded) NS n.a n.a. Z North (Metropolitan excluded) -0.358 -0.352 -0.326 0 Northeast Region (Metropolitan excluded) -0.216 -0.209 -0.159 Southeast Region (Metropolitan excluded) -0.415 .-0.386 -0.395 Central Region (Metropolitan excluded) -0.246 -0.260 -0.195 Southwest Region (Metropolitan excluded) -0.287 -0.309 -0.239 Northwest Region (Metropolitan excluded) -0.281 -0.344 -0.219 Peten Region (Metropolitan excluded) -0.321 -0.281 -0.321 LANG. Spanish speaking household (ability to speak)c 0.088 0.091 0.098 ETHNICrIY Indigenous Household (self identification)' -0.168 -0.193 -0.147 SIZE # of HH Members 0 to 6 years old -0.228 -0.200 -0.235 # of HH Members 7 to 24 years old -0.173 -0.178 -0.169 # of HH Members 25 to 59 years old -0.148 -0.195 -0.111 # of HH Members 60 or more years old -0.094 -0.147 -0.081 # of HH Members 0 to 6 years old squared 0.023 0.015 0.023 # of HH Members 7 to 24 years old squared 0.012 0.011 0.012 # of HH Members 25 to 59 years old squared 0.009 0.015 NS 9 I# of HH Members > 60 years old squared NS NS NS OTHERS Female Household Head (versus male) NS NS NS Age of Household Head 0.013 0.016 0.008 Age of Household Head Squared -0.0001 -0.0002 -0.0001 % of males in the household 0.001 NS 0.001 Primary incomplete (Maximum in household) 0.116 NS 0.092 F Primary complete (Maximum in household) 0.233 0.230 0.191 Secondary incomplete (Maximum in household) 0.280 0.275 0.272 O i Secondary complete (Maximum in household) 0.456 0.426 0.452 University (Maximum in household) 0.472 0.502 0.299 Primary incomplete 0.044 NS 0.084 a z ca Primary complete NS L NS 0.115 O Secondary incomplete 0.220 NS 0.324 W Secondary complete 0.404 0.297 0.467 _ ~~~~~~Uni versity 0.658 0.3 0.636 I st. No work 0.150 0.262 NS I st. Mining/manuf/energy (agr. excluded)d NS NS NS 0 d - 1st. Construction (agr. excluded) 0.110 0.134 0.121 3 Ist, Commerce (agr. excluded) 0.128 0.171 0.106 0 1st. Transport (agr. excluded) 0.121 0.139 0.183 0 Z- Ist. Services (agr. excluded)" 0.127 0.175 0.104 z 2nd. No work' -0.058 NS -0.074 5 2nd. Mining/manuf/energy (agr. excluded)' NS NS NS O 2nd. Construction (agr. excluded)' NS NS NS 2nd. Commerce (agr. excluded)' NS NS NS 2nd. Transport (agr. excluded) eNS NS NS 22nd. Services (agr. excluded)' NS NS NS 26 Dependent Variable = Natural Log of Yearly per capita consumption All Urban Rural 1 st. White collar (Public excluded) d NS NS NS 1st. Blue collar (Public excluded) -0.152 -0.143 NS 5 1st. Domestic (Public excluded) d -0.120 NS NS O I st. Self employed (Public excluded) NS NS NS Ist. Unpaid (Public excluded)d NS 0.245 NS 2nd. White collar (Public excluded) ' NS NS -0.185 2nd. Blue collar (Public excluded)' NS NS -0.232 2nd. Domestic (Public excluded) e NS NS -0.235 2nd. Self employed (Public excluded) e NS NS -0.222 2nd. Unpaid (Public excluded) -0.203 NS -0.361 1 st.: >12 & <=20 hours/week b NS NS NS Z Ist.: >20 & <=40 hours/week b NS NS NS 1 st.: > 40 hours/week b 0.055 NS 0.064 2nd.: >12 & <=20 hours/weeke NS NS NS 2nd.: >20 & <=40 hours/week' NS NS NS 2nd.: > 40 hours/week e -0.042 NS -0.047 HECTARIES OF 0.5-1.5 ha worked (0-0.5 ha excluded) NS NS NS AGRICULTURAL LAND 1.5-3.5 ha worked (0-0.5 ha excluded) 0.025 NS 0.029 WORKED a 3.5-10.5 ha worked (0-0.5 ha excluded) 0.016 NS 0.022 10.5-50 ha worked (0-0.5 ha excluded) 0.005 0.012 0.006 >50 ha worked (0-0.5 ha excluded) NS NS 0.003 HOME House owned and paid (rented excluded) NS 0.111 NS OWNERSHIP House owned and not paid (rented excluded) 0.114 0.186 NS Other House ownership (rented excluded) NS -0.085 NS MUNICIPIO LEVEL Sewage connection 0.120 0.123 NS BASIC UTILITY Electricity connection 0.242 0.324 0.232 SERVICES b SOCIAL CAPITAL Participate Organizations - Bridging 0.049 0.082 NS Participate - Bridging activities 0.081 0.081 0.074 Participate - Linking activities NS 0.057 NS Number of Organizations - Cluster NS -0.005 NS Constant 8.621 8.572 8.776 Number of observations 7,275 3,423 3,852 Number of strata 8 8 8 Number of PSUs 737 353 384 Population size 11,383,521 4,395,934 6,987,587 F 86.85 48.19 35.23 Prob > F 0 0 0 R-squared 0.73 0.69 0.57 NS: Not significant at P.c=10%; _: Significant at P<=10%; all others: significant at p<=5% a. The variable used was the reported number of hectares worked for each category (no dummy.) b. Basic service data are taken from the mean values at the municipal level in the census to avoid simultaneity problems. c. For the majority of household members d. 1' = characteristics of the person with the highest labor income in the household. e. 2'd = characteristics of the person with the second highest labor income in the household. Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Variables excluded due to lack of significance: no partner (wife, husband, et.), participation in any community group, perceived exclusion (from education services, health services, connection to drinking water, training, access to credit, technical assistance, public transport, insurance, sewage connection, road improvement, and justice), and if household perceived exclusion: human capital, social security and physical assets. 27 Table 3.2 - Decomposition of Inequality in Guatemala % of inequality explained Variable Groups Sub-Groups by: National All Guatemala 100.0% 1 . Areas (Urban/Rural) 334% 2 8 Regions 31.5% 3 Number of Members aged 0-24 27.7% 4 Number of Members aged 0-6 25.3% 5 Household Size (regardless of age) 23.8% 6 HH composition size Number of Members aged 7-24 12.6% 7 Number of males in household 1.8% 8 Gender of Household Head (male or female) 0.4% 9 Number of Members aged 25-59 0.2% 10 Number of Members aged 60+ 0.1% 11 Average education for those aged 13 and older' 52.5% 12 Average education for all members a 52.5% 13 Average education for those aged 18-59 a 47.1% 14 Education Higher education level in the household a 46.3% 15 Education of household head a 43.9% 16 iteracy (household average) b 31,6% 17 Literacy of household head 16.9% 18 elephone connection 42.6% 19 rainage 34.6% 20 HH services iped water inside house 24.1% 21 Electricity connection 21.3% 22 Home ownership (Owned & paid, Own & not fully paid, renting, other) 8.4% 23 Stratum (pre-sample classification based on area characteristics)c 32.4% 24 Sector (Agric., Mining, Manufacturing, Utilities, Construction, etc.) 28.8% 25 Work in agriculture d 24.9% 26 Labor & Credit Job type (Public, White Collar, Blue Collar, Domestic, Self Employment) d 16.0% 27 Formal sector work (versus informal sector) d 9.8% 28 CREDIT: Household has more than Q. 1000 in loans 4.8% 29 Work in the public sector (vs. outside public sector) 3.6% Etnicity: specific self classification (K'iche,Q'eqchi,Kaqchiquel,Mamn, 30 ther Maya, Non-Indig.) 23.4% 31 Ethnicity & Languag: self classification (indigenous versus non-indigenous) 213% 32 thc &LnugLanguage: Spanish speaking ability (Majority of HH members) 17.4% Language: Specific language spoken (Majority of HH members: K'iche, 33 Q'ueqchi, Kaqchiguel, Mam or other indigenous) 10.4% 34 Land Hectares of agricultural land worked (0,1,2-3,4-6,7-9,>9) 16.2% 35 Hectares of agricultural land owned (0,1,2-3,4-6,7-9,>9) 9.0% TOTAL: % of inequality explained by variables 1, 4,11,18, 25,31,34 75.9% Defined by the years of education in groups of: 0, 1-3, 4-6,7-9,10-12, 13-17 and 18 and more years. "In groups of 0%, 1-25%, 26-50%, 51-75%, md 75-100%. ' The stratum classification is based on the average household characteristics for each Primary Sampling Unit from the latest Census ata.d Of the highest labor income earner in the household. Source: World Bank calculations using data from the ENCOVI 2000, Insnttuo Nacional de Estad(stica - Guatemala. Based on a decomposition of entropy Measures of Inequality (Theil indices). Figures presented for the E-0 measure of inequality. Those for E- I and E-2 yield very similar nagnitudes and results. 'The dependent variable is the log of per capita consumption. 28 The analysis uses the maximum educational attainment of the members with highest and second highest education in the household. Since the education of these members (usually adults) generally precedes their current economic status, it could validly be considered as having a causative influence on poverty status and consumption (whereas the educational levels of young dependents in the household may be low because poverty prevents them from affording an education). 4Compared to households in which none of the members has any formal education. See Chapter 6 for additional information on labor, incomes and livelihoods. 6 This includes only labor income (since other sources of income can not be directly attributed to specific household members.) ' The variable used was "holdings of agricultural land worked." a See Chapter 6 for additional information on land holdings and poverty. 9 The total percentage is the result of multiplying the 2.9% by the plot size for the group (1.5 and 3.5 hectares.) '° The 12% refers to households in municipalities with 100% access compared to households in municipalities with 0% access. 'This variables are included to take into consideration the "locational" or fixed effects into the model. 12 Bridging social capital contrasts with (a) bonding social capital, which refers to more narrow horizontal ties connecting family members, neighbors, close friends and business associates; and (b) linking social capital, which constitutes the vertical ties people have with formal organizations, institutions or people in positions of influence. See Chapter 4 and Technical Paper 12 (IbAfiez, Lindert and Woolcock, 2002) for more information on the different types of social capital. 13A household is classified indigenous when the majority of its members identify themselves (self-identification) as such. A household is classified as Spanish-speaking when the majority of its members can speak Spanish (regardless of where or how often.) 14 By using the Oaxaca-Blinder decomposition, the difference in consumption levels between indigenous and non-indigenous households was decomposed in terms of differences due to asset holdings and in those due to the retums to the assets. The exercise reveled that about 70% of the differential is attributed to asset holding differentials. 15 The analysis reveals that gender is not a relatively important attribute in explaining between-household inequality (or poverty) in Guatemala. The unimportance of gender is consistent with intemational evidence for developing countries (see Ferreira and Litchfield, 1997). It is important to note, however, that inequality within households is not captured by ENCOVI data, and intra-household inequality between genders could indeed be significant. In fact, other factors (such as unequal access to education for girls) suggests that it is. . 16 ft is important to note that, with the exception of this "combined" analysis (analyzing the different variables together), the decomposition of inequality using each variable is not additive. " Even in Chile, which has aggressively pursued structural policies to promote growth, and reduce poverty and inequality for some time (trade policies, education reform, targeted public investments, etc.), the Gini index of inequality has only fallen by 1.5 percentage points in seven years. See Ferreira and Utchfield (January 1997) and World Bank (June 10, 1997). 29 Chapter 4: Historical and Contextual Factors The last chapter showed that household endowments, assets, and characteristics play an important role in "determining" poverty in Guatemala. From a technical point of view, the "cross-sectional" relationships between these variables and consumption (the "betas"' in the regressions) are quite similar to those found in other countries.! Yet Guatemala's high degree of poverty and exclusion did not emerge over night. This chapter argues that other factors - both contextual and historical - have fundamentally shaped Guatemala's performance regarding the levels of the endowments and characteristic determinants of poverty observed today (the levels of the "x's" used in the regressions). This report does not purport to conduct an exhaustive review of these historical and contextual factors; rather it seeks to simply highlight the importance of these factors in influencing poverty in Guatemala. Specifically, this chapter examines the key contextual and historical factors and their role in shaping Guatemala's poverty profile today, highlighting key challenges for tomorrow. It begins with an overview of two predominant contextual factors in the make-up of Guatemala as a country: rich cultural diversity and geographic isolation. With this context in mind, Section 2 then reviews key exclusionary forces in Guatemala's historical pattern of development: (a) massive land expropriations from the indigenous population which have resulted in one of the most unequal distributions of land in the world; (b) forced labor policies that exploited indigenous labor from the 1800s through the middle of 1900s; and (c) human capital accumulation, which has historically suffered as from exclusionary education policies, as part of a broader political strategy, and as an outcome of land and labor policies. These forces all put Guatemala at a historical disadvantage in terms of its development and key social indicators. Moreover, the long civil war imposed further costs on Guatemala's development, as reviewed in Section 3. Section 4 then asserts that the Peace Accords, signed in 1996, not only yielded a formal end to the armed conflict, but signaled a fundamental shift in Guatemala's pattern of development, paving the way for a transformation to a more prosperous and inclusive nation. In the six years since the Peace Accords, Guatemala has taken important steps on this new development path, with progress in public sector management, public revenues and spending, and improvements in the coverage and equity of education and basic services. Importantly, these steps signal that progress is possible, despite the magnitude of the challenge of changing the course of history. Nonetheless, challenges remain. This is to be expected. Changing the course of history in such a short time span is not easy in any country. The hierarchical relations, attitudes, and institutional forces that have pervaded for centuries do not disappear over night. Moreover, events such as Hurricane Mitch and political transitions have delayed the implementation of the Peace Agenda. Section 4 contends that many of the remaining challenges for the Peace Agenda coincide largely with priorities for poverty reduction. In particular, progress on the Peace Accords is deficient for key development targets, especially for those involving health and education outcomes, and economic growth. These outcomes reflect the need for poverty reduction and improvements in living conditions, which are crucial for lasting peace. They highlight remaining priority challenges in several key areas - which reveal the significant overlaps between the Peace Agenda and the challenge of poverty reduction: (a) creating opportunities; (b) reducing vulnerability; and (c) improving institutions and empowering communities. These challenges are discussed in subsequent chapters. DIVERSITY, ETHNICITY AND ISOLATION Guatemala is a physically diverse country, with many isolated areas. The country is divided into numerous distinct geographic zones, including forested highlands in the west, fertile lowland coasts, and tropical forests in Peten. Two-thirds of the country is mountainous and volcanic. Most of the population lives on land between 900 and 2,500 meters above sea level. Unfortunately, this natural resource diversity is being 30 increasingly threatened by erosion, deforestation, and population pressures on the land. The country is not physically united and many villages are fairly isolated, with long inter-village distances, due to an inadequate road network (see Chapter 10). About 13% of households in the ENCOVI 2000 sample did not have access to motorable roads;2 this figure reaches close to 20% in the Nor-Occidente, Nor-Oriente, and Norte regions, which are also among the poorest. Such physical isolation is also higher among the poorest quintiles and the indigenous, than the non-indigenous (see Chapter 10 for details). Village access is further complicated by impassability from road closures (due to landslides, mudslides, flooding, etc), as discussed in Chapter 10. Table 4.1 - Language Ability by Socio-Econonic Group Language spoken Spanish Indigenous Monolingual Spanish Monolingual Indigenous Bilingual All Guatemala 89 35 65 ,11 24 Area Urban 97 19 81 2 17 Rural 84 44 56 16 28 Gender Male 91 35 65 9 26 Female 87 35 65 13 22 Poverty Extreme poor 68 66 34 32 33 All poor 82 60 50 17 33 Non-poor 98 15 85 2 12 Ethnic group Non-indigenous 100 1 99 0 1 Indigenous 74 84 16 27 57 Indigenous K'iche 79 75 25 21 54 Q'eqchi 37 86 4 63 33 Kaqchiquel 91 75 25 9 65 Mam 87 81 19 13 68 Indigenous male 79 82 18 21 61 Indigenous female 68 85 15 32 53 Age group (for indigenous only) 7-13 68 77 23 32 45 14-18 77 78 22 23 55 19-25 76 82 18 24 58 26-40 76 84 16 24 60 40+ 67 89 11 33 56 Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadtstica - Guatemala. Mirroring the physical diversity of the country, Guatemala's population is rich in cultural and linguistic diversity. In a population of over 11 million, about half the population is indigenous, including some 23 ethno-linguistic groups, 21 of which are Mayan.3 The largest indigenous groups include the K'iche (22% of the indigenous population), the Kaqchiqel (21%), the Mam (19%), and the Q'eqchi (15%).4 The Mayan population - descendents of the great civilization that created the magnificent pyramids and ceremonial centers of Mexico and Central America - live primarily in hundreds of small, rural communities scattered throughout the western and central highlands.5 Taking language as a cultural asset, the ENCOVI 2000 reveals that some 84% of the indigenous population (self-identified ethnicity, see Box 4.1), speaks an indigenous language (Table 4.1). This share differs by specific indigenous group, with a higher share of the Mam and Q'eqchi speaking Indigenous languages than the K'iche or Kaqchikel. An inter-generational loss of cultural heritage also seems evident, since a smaller share of indigenous children and youth speak the Indigenous languages than the older generations. Moreover, the loss is particularly strong with inter-ethnic marriage: while 90% of those between 7-25 years who boast "full Mayan linguistic lineage" (4 Mayan-speaking grandparents and 2 Mayan-speaking parents) speak indigenous languages, only 41% speak the language if only 3 grandparents are Mayan speakers (even if both parents are), and only 18% speak indigenous languages if only one parent does (even if all four grandparents do). Taking Spanish-speaking ability as an indicator of economic access (e.g., job opportunities), the ENCOVI indicates that, while 89% of the overall population speaks Spanish, only 31 three-quarters of the indigenous population do. This share differs significantly by specific ethnicity and gender, with only a third of the Q'eqchi population and two-thirds of indigenous women reporting that they speak Spanish (Table 4.1). Taking bilingualism (ability to speak both Spanish and indigenous languages) as an indicator of possession of both cultural and economic assets, however, the indigenous are clearly better endowed (Table 4.1). Box 4.1 - Who is "Indigenous"? A Historical Perspective The history of how many Indigenous people there are in Guatemala is as much an account of efforts by a dominant political group to set itself apart from the general population as it is a description of demographic change. Guatemalan census data have always distinguished between indigenous and non-indigenous people, but the criteria for classification evolved with the maturation of national identity. The Indian/non-Indian dichotomy was at first unequivocal. When European conquerors first established themselves in the territory., the word "indigenous" was an adjective that meant native-bom. However, Spanish law granted full citizenship to anyone bom of Spanish parents. This was true even if they were locally bom and if only one parent was Spanish. As this new group of locally- bom descendents of Europeans grew, its members came to be known as "Ladinos." During the colonial period, Indians were subject to a head tax, which provided an economic incentive for classification as Ladinos. Zilberman de Lujan (1995) notes that throughout the colonial period the "Poblacicln de Castas" -later called "Ladinos" -grew in proportion to the Indian population because indigenous people wished to evade the tribute. Table BI shows a fairly steady but very uneven decline in the percentage of "Indians" reported. For instance, between 1893 and 1921 there was almost no change. In both years classification was left up to the interviewer, who was told to "discretely write down the person's race" without asking the subject, for it was believed that asking resulted in "erroneously recording" individuals as Ladinos. Another 30 years later, the recorded percent Indigenous dropped by 11 points when the census instructed the census taker to "use as a basis for classification whatever was the local social perception of-that individual." Another 11 point drop resulted in only 14 years, when the local-standards methodology was "cross-checked" with "objective" criteria, including language spoken, dress, whether shoes were worn and whether the person ate bread or com tortillas. In 1973 and 1981, crosschecking was dropped and the census taker was allowed to decide ethnicity, except in the case of maids, when the patron's opinion was to be recorded. The result was that between 1964 and 1973 the proportion recorded as indigenous rose slightly for the first time in Guatemala's history. By 1994 data collection is imbued with the spirit of peace-process democracy and pluralism. The census report notes that "the individual's right to self-identification with an ethnic group was respected. For this reason [ethnicity] was obtained by means of a direct question and not through simple observation." While the ENCOVI 2000 includes numerous possible indicators of ethnicity, including language, self-identification, and language of ancestors, the definition used here is "self-identification" (consistent with the census). % of Population by Ethnicity, 1893-2000 Ladinos Indigenous Ladinos Indigenous 1893 Census 35.3% 64.7% 1973 Census 56.2% 43.8% 1921 Census 35.2% 64.8% 1981 Census 58.1% 41.9% 1950 Census 46.4% 53.6% 1994 Census 58.3% 41.7% 1964 Census 57.8% 42.2% 2000 ENCOVI 57.5% 42.6% Source: Adapted from GUAPA Technical Paper 3, Edwards (2002). Unfortunately, Guatemala's diversity has historically been accompanied by conflict, exclusion, and a dualistic social and economic structure. Internationally, countries with significant indigenous populations tend to have higher poverty rates. Within these countries, the indigenous tend to be poorer than the non- indigenous population due to historically exclusionary forces.7 In this regard, Guatemala is no exception. Indeed, inequality between ethnicities is a pervasive feature in Guatemala. As discussed in Chapter 2, while the indigenous make up about 43% of the population, they account for less than a quarter of total income and consumption. In contrast, economic and political resources remain concentrated among the economic elite of predominantly European descent and the Ladino population. The linkages between these groups have historically been weakened by decades of exclusion and conflict, as discussed below. Unlike its topography or.people, Guatemala's economy is not so diverse. As discussed in Chapters 5 and 6, agriculture still dominates, accounting for a quarter of GDP,8 and employs 36% of all workers.9 The 32 economy is still largely based on exports of coffee and sugar, notwithstanding some success in promoting non- traditional exports. Despite a fall in international prices, coffee still accounts for over 25% of export earnings. Sugar, bananas, and cardamom follow as the principal cash crops. The main export crops all require large inputs of seasonal labor for harvest. Subsistence agriculture traditionally revolves around the production of corn and black beans. Maquila (free-trade assembly and re-export zones), mining, energy, commerce, and services have all grown fairly rapidly in the past decade. Tourism strengthened in the 1990s and now exceeds coffee or sugar as the main source of foreign exchange, though its success depends largely on political stability and security. Remittances are also an important source of income (see Chapters 5 and 6). EXCLUSIONARY FORCES IN GUATEMALA'S HISTORICAL PATTERN OF DEVELOPMENT In fact, past policies greatly contributed to an exclusionary pattern of development in Guatemala, particularly for land, labor and education. All of these spheres were intertwined with each other. They are all also intimately connected with the development of coffee, Guatemala's primary export crop. Such policies sought to promote economic growth, but the indigenous population was excluded from benefiting from this growth. Women were also excluded from these spheres. Land The historical practice of expropriating land from the indigenous gained momentum with the development of agricultural exports, particularly coffee. The indigenous have consistently been divested of their. land in Guatemala since the conquest and colonial times. Land privatization and expropriation accelerated in the late 1800s, however, with significant changes in the legal environment determining property rights,'° which coincided with the spread of coffee. Coffee production depended on secure private property rights, and it was only then that coffee developed rapidly. These laws encouraged privatization by simplifying the conversion of communally-held indigenous lands (ejidos)," into individually titled holdings. A central aim of land privatization and consolidation was the formation of large plantations (fincas), with the creation of a class of large landowners at the expense of indigenous cultivators, to take advantage of the expanding world market for coffee.'2 Since the ideal terrain for coffee occurs between 800-1,500 meters of altitude, the indigenous peoples who had been cultivating these lands were compelled to located to higher and less fertile grounds for their subsistence cultivation.'3 Nearly one million acres of land were privatized between 1871, when land privatization decrees were initiated, and 1873. alone.'4 With the shift towards larger plantations, some 3,600 persons received plots averaging 450 hectares each during the period between 1896 and 1921.15 Such consolidation has continued until recent times, as diversification of exports brought new expropriations from peasants. As discussed above and in Chapter 5, Guatemala's economy has diversified substantially in recent decades, though coffee still dominates. During the country's brief democratic period, from 1944-1954, the Government introduced a series of agrarian reform measures, including laws to protect communal lands. The reforms were aborted in with the military coup in 1954, however, delaying any significant land reforms until the 1990s. Land consolidation continued during this period. While communal lands accounted for 12% of agricultural land in 1950, this share had dropped to 4.8% by 1964 and only 1.1% in 1979 (the year of the last, agricultural census).'6 Between 1950 and 1970, the number of farm families - most of them indigenous- possessing parcels of land too small to provide subsistence incomes increased by 37% and the number of landless peasants increased to about one-fourth of the rural workforce.'7 Estimates from 1979 indicate that less than 2% of the population owned at least 65% of the land and less than 1% of all farms were over 2,500 hectares in size and accounted for over 20% of the land, while over 78% of all farms were under 3.5 hectares and accounted for slightly more than 10% of the land.18 With an estimated Gini coefficient for the distribution of land of 85 in 1979, Guatemala's land inequality is among the most skewed in the developing world.'9 Women have also been consistently denied the right to hold land, both legally and then by tradition. According to INTA, only 8% of land appropriations between 1954 and 1996 went to women.20 Indeed, tradition still bars women from inheriting land in most of the QPES study villages. 33 Box 4.2- Life in a Finca Village: the Story of KAl (QPES) "Los que nos tratan mral son los encargados de lafmnca...Aquino hay que hablar de bienestar, sino de pobreza. Aquf todos somos pobres. Son muchos nuestros deseos, pero nada tenemos. .No nos vamos porque no tenemos a donde ir" - Villagers of KA 1, QPES, July 2000. The 200-some villagers of KAI live and work permanently on a coffee plantation (finca).2' Most were born on the finca, their relatives having migrated there in previous generations. The villagers of KAI own almost zero physical assets; the finca owns the houses and small plots they are allowed to use for subsistence production (mainly of corn). When their spouses pass away, widows are not allowed to maintain use of these plots or houses, unless they move in with relatives who are actively employed on the finca. The houses are in terrible condition, with incomplete wooden walls and metal roofs. The villagers lack access to most basic services, such as piped water, sanitation, or energy. The only access to the finca comes from an unpaved road, which gets flooded during heavy rains and becomes impassable. Financial capital is virtually non-existent, as the villagers have little opportunity to acquire surplus or borrow. Almost all adults are illiterate, and the single-teacher school only offers through grade three. Child labor prevents many from attending, particularly the boys. There is no health clinic or pharmacy, only a traveling midwife. Workers on the finca receive benefits according to their labor classification. Permanent workers ("rancheros" or "03") do receive benefits. However, many "permanent residents" - particularly union members - have been classified as "volunteers" ("06") and have much more tenuous job security, facing periodic suspensions in an apparent strategy by the finca to avoid paying labor benefits and discourage union affiliation. Their classification as volunteers (same grouping as temporary migrant workers, or cuadrillas) derives from the fact that they seek work on other fincas during periods of suspension. Women on the fincas do not receive labor benefits, despite contributing to the production process, both directly (especially during the harvest) and indirectly by maintaining the workers via household chores and cooking. Social capital and capacity to engage in collective action is virtually non-existent. The villagers do not feel they belong to the community. Vertical authoritarian relations between the finca administrators and the villagers dictate daily life and discourage the formation of horizontal connections within the village. While some workers do maintain ties to the union, this has been a bit of a mixed blessing due to their reclassification as volunteers (06). In addition to a lack of these basic assets, the villagers of KA I seem to have lost their cultural identity, identifying themselves and their language not as the Mayans (Kaqchiqel and k'iche) that they are, but simply as "natives" of the finca who speak "lengua". Children only speak Spanish, though they do understand the Indigenous languages. Few residents wear traditional Indigenous clothing (traje). On top of their extreme poverty, the villagers of KAI have been subject to several significant shocks, with lasting effects, including the earthquake of 1976 (their homes have still not been repaired), repeated large-scale labor shocks (including vengeful dismissals for union affiliation), and other recurring natural disasters (including flooding and landslides). They have little hope for the future, noting that "everything stays the same on the finca, without water, without lighting, without drainage or latrines" and that "they have no where else to go." Children's aspirations do not go beyond the life they know, with boys indicating that when they grow up they will "work on the coffee harvest and clean the hillside" and girls answering they will "clean the house." Labor Guatemala's economy essentially grew on the backs of Indigenous workers who suffered numerous forms of mandatory forced labor, again connected to the expansion of the coffee sector. Indigenous labor in Guatemala has historically been viewed as an exploitable asset. In fact, land policies of the late 1800s had the additional goal of reducing land available to subsistence Indigenous to create a low-wage labor force. In the late 1800s, insufficient cheap labor was a barrier to the expansion of coffee. The expropriation of Indigenous communal lands helped create rural unemployment by forcing families into marginal areas or leaving them without access to sufficient land.22 Such conditions were precisely the prerequisites for several types of forced labor: * Mandamiento. In 1877, the state instituted its infamous mandamiento forced-labor system in which villages (pueblos) were required to supply coffee plantations with work crews of up to sixty people for periods of fifteen to thirty days.23 * Forced Work on Roads. In 1873, the contribuci6n de caminos decreed that all able male citizens were obliged to provide free labor on public projects to build roads or pay a commutation fee.24 Supposedly, this free work requirement applied to all, but in practice; only the indigenous population was forced to perform it.25 34 o Debt Servitude. Indentured labor was also common. Under this system, advances were provided to workers in anticipation of a certain amount of work; debts were then deducted from the worker's harvest or in cash. Such debts commonly built up to levels high enough that the workers were essentially "owned" by the finca owners (patrones).26 Debts were monitored by local public authorities who were authorized to arrest any defaulters, as evidenced by debt-recordings in the personal workbooks required for all indigenous peasants.27 o Vagrancy Law. The 1934 ley contra vagancia obliged landless peasants to work at least 150 days per year on plantations. Proof of service was required in the workers' personal workbooks, and the system was monitored and enforced by the State itself.28 In conjunction with mass land expropriations, an increasing share of indigenous were forced into these conditions. These forced labor laws remained in effect until the middle of the 2Oe century. During the country's brief democratic experiment (from 1944-1954), the Government enacted the first Labor Code in 1947, providing for minimum wages and recognizing the freedoms of organization, finca owners were still allowed to pay up to 35% of workers' salaries in food, a practice that even appeared in the Constitution of 1985. An important step occurred with the Agrarian Reform Act of 1952 that prohibited all forms of servitude and slavery. This law was again abandoned, however, in 1954, after the military coup, when a decree permitted landowners to reintroduce the semi-feudal colono system under which landowners could avail themselves of a cheap labor force by providing subsistence plots on their plantations in exchange for labor during the harvest season,29 a practice that continues today (see Box 4.2). Finally, the Constitution of 1985 reestablished most modem labor rights,30 though as shown in Chapter 6, many are not enforced (e.g., labor benefits, minimum wages). Education and Human Capital Cross-country evidence3t suggests that Guatemala's human capital accumulation - and hence long-term economic growth - suffered as a result of land and labor policies. In Guatemala, like El Salvador, land and labor policies favored the development of the coffee sector via large plantations. Such an approach reflects the fairly close ties between the landed elite and both the conservative and liberal parties in the 1800s. The subsistence-wage economy that resulted provided little incentive for workers or firms to invest in human capital. In contrast, the economies of Costa Rica and Colombia benefited substantially from policies that protected and promoted the development of coffee based primarily on small-holder production. Small-holders have stronger incentives to accumulate human capital because they are more likely to reap its returns. This structure fostered faster human capital generation, and hence a stronger base for economic growth in the long- run. Relatively lower human capital accumulation among the indigenous is also an outcome of a historically exclusionary educational system - which reflected a broader strategy of political exclusion. Education in Guatemala was traditionally reserved for "citizens," a status not fully extended to women or the Indigenous until 1945 32 The liberal government of 1871 instituted a substantial education reform, shifting responsibility for providing education from the Church to the state, and making education free and mandatory. The policy was not implemented, however, as the state lacked the necessary resources (teachers or funds). Nonetheless, some indigenous education centers and secondary schools for women were established during this period. In practice, the education system was ethnically segregated: on the one hand, the state promoted a paternalistic approach to "civilize" and assimilate the indigenous as agricultural producers; on the other hand, the education system continued to exclude the indigenous, being largely reserved for ladino males. As such, virtually the entire indigenous population (and most women) remained illiterate. Miteracy was then used as a pretext for ineligibility for voting, making the educational policy part of a broader strategy of political exclusion. With the temporary introduction of a more democratic state in 1945 came education reform, including a literacy campaign, the creation of the Instituto Indigenista Nacional, and a policy to officialize indigenous languages.33 Progress was made, and overall literacy rose from 19% in 1930 to 29% in 1950,34 and to 37% by 1964. Further policy changes were legislated in 1985 under the National Education Law, which emphasized 35 decentralization and participation of indigenous communities in the education system. This was the first time that a law asserted that the education system should be take into account Guatemala's ethnic diversity.35 Overall Development By outbreak of the civil war Guatemala was already far behind in its development, particularly its social indicators. Despite having almost twice the per capita GDP of Honduras, Nicaragua or Bolivia, by 1960, Guatemala ranked worse than these countries on several key social indicators. In particular, Guatemala had the lowest gross primary enrollment rate in LAC, the second shortest life expectancy (second only to Bolivia), and infant mortality rates only slightly lower than Honduras, Nicaragua or Bolivia.36 Guatemala also lagged significantly behind the collective group of comparable "lower-middle income countries" on these social indicators in the 1960s. GUATEMALA'S 36-YEAR CIVIL WAR: SIGNIFICANT COSTS FOR LONG-TERM DEVELOPMENT Guatemala's was one of the longest of Latin America's recent civil wars. Spanning 36 years (from 1960- 1996), the war underwent several phases, only later involving the indigenous on a large scale. The first wave broke out in 1960 with a group of army officers revolting against government corruption. Initially, the movement was centered in the eastem region and involved primarily the non-indigenous population. When the revolt failed, the officers fled to the countryside and launched a war against the Government.37 The second wave broke out in the 1970s, this time in the western highlands, with some Indigenous communities becoming active participants. During the final and longest phase, from the late 1970s into the 1980s, social tensions exploded into a full scale civil war with active indigenous participation (involving close to half a million Mayans during that period).38 Although Indigenous communities had been undergoing important social changes for more than a century,39 it was not until this later phase that they became a source of mobilization and support for the country's guerrilla movements, and hence a perceived threat to the country's powerful economic and military elites.40 The guerrilla military offensive reached its height in 1980-81, gaining 6,000- 8,000 armed fighters and 250,000-500,000 active collaborators. In response, the military launched a major counterinsurgency effort that reached genocidal proportions in the early 1980s, executing scorched-earth warfare tactics, mandatory paramilitary "civilian self-defense patrols"(PACs), forced resettlement camps, and the militarization of the entire administrative apparatus of the country.4' Guatemala's war was also one of the bloodiest in LAC. In total, more than 200,000 people (over 2% of the population) were killed or "disappeared" and another million (10%) were displaced. Over 600 villages were completely destroyed, their residents massacred.42 The results of the investigation by the Historical Clarification Commnission indicate that 93% of the actions of violence were attributable to agents of the state (particularly the military), and that 53% of the victims were Mayan, 11% were Ladino, and 30% were of unregistered ethnicity. Although Guatemala recorded reasonable economic performance during the early phases of the conflict, the conflict imposed significant costs on the economy and village life. Somewhat paradoxically, despite the war, Guatemala managed to maintain reasonable growth rates during the 1960s and 1970s, though they fell significantly during the 1980s.43 Nonetheless, in addition to the loss of life, the war had serious short- and long-run impacts on Guatemala's development, for the overall economy and life at the village level (Box 4.3). While it is impossible to quantify the full range of impacts, some are directly quantifiable. The Historical Clarification Commission report estimates that, during the 1980s alone, the costs of the war were equivalent to 15 months of production in Guatemala, or 121% of GDP in 1990. The majority of these costs arise from the loss of productive potential and abandoned economic activities due to death, disappearance, or forced displacement. Destruction of physical capital, including private, community and infrastructure assets was also costly (estimated at 6% of GDP in 1990). The engagement the potential workforce in military - rather than productive - capacity also further reduced Guatemala's output. UNDP (2000) estimates that the PACs diverted a significant share of the economically active population, with a cost of $3,000 million in 1990 or 39% of GDP 36 for that year. Using time-series models and average growth-poverty elasticities, Lopez (October 2001) estimates that if the armed conflict had not occurred, per capita GDP in 2000 would have been about 40% higher and poverty would have been about 12 percentage points lower. Social indicators lagged even further behind growth, with Guatemala remaining among the worst performers for education and even falling in the rankings behind Honduras and Nicaragua for infant mortality during the period from 1975-1990. Policies and strategies to reduce poverty were essentially non-existent. The mere use of the tern "poverty" was considered "taboo" in official circles until the 1990s, since such egalitarian concepts were associated with leftist insurgents." Other impacts include the repression of participation of civil society organizations was discouraged and the (often deliberate) destruction of vast areas of the highlands (e.g., via burning of forests).45 Box 4.3 - Rebuilding After the Violence of the 1980s: the Story of KIl (QPES) "Ella ha pasado penas.. el esposo desapareci6, ya nunca los vimos...A mi esposo lo llevaron y como mi hijo tenia 10 meses en el pecho, mam6 la tristeza y el nino murio... Ya no pudimos trabajar...ya no vimos la cosecha de maz_.. Afecto a los ninos, ya no fueron a la escuela... Ya no se comunicaban entre vecinos yfamiliares... Un mont6n de gente se fue a la capital por miedo." - Villagers of Kll describing the violence of the 1980s, QPES 2000 Although the QPES teams did not explicitly seek out villages that had been subject to substantial violence during the civil war, several happened to fall into the sample. The K'iche village of KIl is one of them. The shock was apparently severe, with numerous families losing relatives (killed or missing), girls being raped, and houses being looted. A huge share of the population fled to the capital in fear, retuming only in the 1990s. In addition to the obvious psychological trauma suffered by the victims and their families, agricultural production halted, children didn't attend school, and communication between neighbors and families was severed out of fear and lack of trust. "Envious" villagers apparently "wrongfully" accused other residents of being guerrilla members. People were even afraid to talk inside their homes for fear of being listened to and reported by their neighbors. Even today, they still complain of susto (post-traumatic stress syndrome) as a lasting effect. Nonetheless, the community seems to have rebuilt itself substantially since the conflict ended. In terms of village organization and social capital, it boasts numerous committees (for development, water, electricity, stoves, PTA, a cooperative, and a farmers association), though some allege misuse of funds by the water committee. The villagers have fairly strong links to extemal bodies and are receiving assistance from the municipality, various bilateral agencies, social funds, NGOs, and ministries. On the other hand, there are some religious conflicts and divisions in the community, women are not active in community decision making, the illiterate seem to be excluded ("no one takes them into account") and links between poor and wealthier families are scarce. Land inequality is high, and land conflicts have arisen due to unclear borders between properties. In terms of assets, the community is well-endowed, with extensive road access, water, latrines, electricity, and substantial communal infrastructure (market, mills, 17 churches, sports fields). The village also boasts relatively extensive school coverage, with pre-primary, primary, and secondary schools. Health services are missing, however, and they only have midwives. Major sources of livelihoods are fairly well- diversified, including agriculture (com and fruits) and non-farm activities (artisan crafts, textiles, commercialization). THE 1996 PEACE ACCORDS: TOWARDS A MORE INCLUSIVE COURSE OF DEVELOPMENT The Peace Accords: A Turning Point for Guatemala's Development Path The Peace Accords aimed not only to formally end the armed conflict, but to reverse the country's historically exclusionary pattern of development. Recent developments in Guatemala have been shaped in large part by the signing of the Peace Accords in December 1996. The four main areas of the agreements involved (a) resettlement, re-incorporation, and reconciliation issues; (b) an integral human development program; (c) goals for productive and sustainable development; and (d) a program for the modernization of the democratic state, including a strengthening of the capacity of participation and consultations of the distinct segments of civil society. Three cross-cutting themes were also emphasized throughout the accords: the rights of indigenous communities, commitments regarding the rights and position of women, and a strengthening of social participation.46 Importantly, the main themes on the Peace Agenda were maintained throughout the protracted negotiations process, despite numerous changes in Government, three different peace commissions, and various changes in the military. The endurance of these main themes throughout the peace process bears testimony to their importance as priorities for the country. 37 The agenda for economic development and the reduction of poverty and exclusion were the focus of two main accords. In particular, the Accord on the Identity and Rights of Indigenous Peoples, signed in March 1995, proposed to formally define Guatemala as a multi-ethnic, multi-cultural and multi-lingual nation and recognized that the identities of the indigenous peoples are fundamental to the construction of national unity. The accord contained many provisions to overcome the historic exclusion and exploitation suffered by indigenous peoples, including: proposals for anti-discrimination legislation; the protection of cultural rights; educational reforms (including decentralization and the promotion of multilingual and multicultural education); recognition of the traditional forms of organization and communal land ownership; and the creation of several joint commissions (comisiones paritarias) to guide these reforms.49 Unfortunately, poor voter turnout at a referendum in 1999 resulted in a rejecting of a package of constitutional reforms necessary for full implementation of this accord. The Socioeconomic and Agrarian Issues (SEA) Accord, signed in May 1996, established the overall development agenda, with stronger social orientation and a general goal of closing the huge gap between rich and poor. The accord covered a range of development areas, including: growth, tax revenues, public spending, education, and health, each with specific monitorable targets (see below). It also emphasized broader citizen participation in development and for the decentralization of development projects through "development councils" (consejos de desarrollo). With respect to land, rather than endorsing full-fledged land reform, the Accords focused instead on (a) market-assisted land redistribution; (b) the creation of a Land Fund, from which land would be acquired by the Government and made available to landless peasants; and (c) the implementation of a national land survey and land registry. Conspicuously absent from the SEA Accord are specific actions and targets regarding labor markets and job creation (see Chapter 6).48 Progress Since the Peace Accords In the six years since the signing of the Peace Accords, progress has been made in a number of important areas. In addition to bringing formal end to the war, reducing the size of the armed forces and creating a new civilian police force (PNC), progress has been achieved in several. areas on the economic development side of the Peace Agenda, including: * Improving public financial management. As discussed in Chapter 13, progress has been made with the introduction and implementation of the Integrated Financial Management System (SLAF) since 1998. * Increasing tax revenues. In August 2001, the Government increased the value added tax rate (VAT) from 10% to 12% in order to increase revenues as a key part of the Peace Agenda. Efforts have also been made to improve the efficiency of the tax collection agency to clamp down on tax evasion. As a result, tax revenues have increased, rising from 8.7% of GDP in 1996 to an estimated 9.8% in 2001, with total revenues rising from 9.2% to 11.1%. * Increasing public spending, particularly for certain sectors. Overall public spending has increased from 10.4% of GDP in 1996 to 13.4% in 2000.49 Notably, spending on education and basic services (particularly via the social funds) has increased significantly, as discussed in Chapters 7 and 9. Such resources are crucial for Guatemala to be able to invest in improved living conditions. * Improving educational coverage. As discussed in Chapter 7, Guatemala has made progress towards improving educational coverage and narrowing disparities between genders, ethnicities, and poverty groups. Importantly, such progress has accelerated since the signing of the Peace Accords in 1996. * Improving the coverage of basic utilit services. As discussed in Chapter 9, Guatemala has likewise made significant advances in extending the coverage of basic utility services (water, sanitation, electricity) and reducing disparities in access to these services. 38 Progress has also occurred in other areas. For example, for land, numerous entities have been created and initiatives launched, though their reach has been limited to only a few thousand households (see Chapter 6). These steps signal that progress is possible. While Guatemala has not achieved the full set of targets set up by the Peace Accords, it is important to recognize that progress has been made. Given Guatemala's long history of exclusion and conflict, such progress is noteworthy since it essentially constitutes the first steps on a new, more inclusive development path. Table 4.2 - Performance of Select Peace Monitoring Indicators A Snapshot for the Years 2000, 2001 Obiective/Target I Base 1995 Target 2000 Actual 2000 Estimated 2001 Economic and Fiscal rgets Growth Rate (%) 6.0 6.0 33a .1.9g Tax Revenues (% GDP) 7.6 12.0 96a 9.8a Health Spending (% GDP)b 0.9 1.3 1.1 1.1 Preventive Care (% of health budget)' 38 >50 52 n.a. Education Spending (% of GDP)b 1.6 2.5 2.5 2.8 Judicial/Public Min. Spending (% of GDP)b 0.2 0.3 0.5 0.6 Military Spending (% of GDP)' 1.0 0.7 0.7 0.9 Investment in Rural Development (mn Q)C n.a. 50 >300 n.a. Investment in Rural Infrastructure (mn Q.)c n.a. 300 >300 n.a. Social Targets Literacy Rate (%) 64.2 d 70.0 6 _8.9e n.a. Primary Education Coverage three years - gross rate 84 c 100 125 n.a. three years - net rate' 69c 100 72e_. n.a. Infant Mortality (deaths per 1000 live births) 40 c 20 40-45s . n.a. Maternal Mortality Rate (per 100,000 giving birth) 97 c 2709 48.5 -1990 - 270 g n.a. Vaccination Coverage Polio c 80 85 88 n.a. Measles' 83 95 n.a. Sources: a. World Bank macroeconomic database. b. SIAF/Ministry of Education (Feb. 12, 2002), based on executed ("devengada") spending figures. c, World Bank (February 2000). d. INE-1994 Census. e. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala f. DHS 1998-99. g. World Health Organization. h. World Development Indicators (2001). i. Note that the Peace Accords targets specified these coverage indicators as being for the first three years, rather than the standard six. Note that the difference between gross and net enrollment rates is primarily the enrollment of over-aged students, aged 10+. Shading signifies targets that are under-achieving. Remaining Challenges: Important Overlaps in the Peace Agenda and the Poverty agenda Nonetheless, significant challenges remain. Changing the course of history hasn't been easy. The hierarchical relations, attitudes, and institutional forces that have pervaded for centuries will not disappear over night. Moreover, recent events, including reconstruction efforts required after Hurricane Mitch struck in 1998 and instability caused by the political transition in 1999-2000, have delayed the implementation of the Peace Agenda. As such, close to 200 actions on the agenda had to be rescheduled and are now programmed to be implemented over the next three years.50 Many of the key challenges for poverty reduction coincide largely with the remaining actions on the Peace Agenda. In particular, key development-related targets supported by the Peace Accords have not been met, especially those involving outcomes in health, education and economic growth as well as fundamental institutional reforms (Table 4.2). The lack of progress for these key development outcomes reflects the need for poverty reduction and improvements in living conditions, which are crucial for lasting peace. This overlapping between the Peace Agenda and the poverty agenda highlights remaining priority challenges in several key areas: 39 * Building Opportunities and Assets. The most recent peace monitoring report by the UN Secretary General and MINUGUA5' notes uneven progress in the areas of human and sustainable development. Improving the climate for growth and opportunities and building key assets (such as human capital, access to basic utility services and transport, and land and physical capital) are crucial for meeting targets in this area. These same factors present significant challenges for poverty reduction, as discussed in Chapters 5-10. * Reducing Vulnerability. Improving security (both personal and economic) is crucial for both lasting peace and poverty reduction. Indeed, both poverty and instability can be greatly worsened by the onslaught of shocks (natural, political, economic). The challenges of reducing vulnerability and improving social protection are discussed in Chapters 11 and 12. * Improving Institutions and Empowering Communities. The most recent peace monitoring report also highlights key challenges for modernizing the state and promoting community participation as a means to empower Guatemalan people and bring about a more inclusive society.52 Such challenges are also crucial for poverty reduction and include: (a) strengthening public sector management; (b) improving governance (particularly corruption and the rule of law); and (c) fostering more inclusive participation at the community level, as discussed in Chapter 13. These institutional factors are important in their own right, but also have a strong influence on the menu of options available to the Government in the other areas in their efforts to reduce poverty. 'For a cross-country review in LAC, see Wodon (2000). See also various poverty assessments. 2 Motorable roads defined as paved, gravel or unpaved roads. Excludes dirt roads, tracks and paths. Estimates are for households in the ENCOVI sample for which community-level information was collected. See GUAPA Technical Paper 8 (Puri, 2002) for additional details. 3Tovar G63mez (May 1998). 4Population estimates by ethnicity based on World Bank calculations using the ENCOVI 2000 (expanded sample), Instituto Nacional de Estadtstica - Guatenala. 5 Davis (1988) and FLACSO (November2001). 6 Indeed, the language abilities of parents and grandparents were significant determinants of Mayan-speaking abilities in multi-variate regressions. See GUAPA Technical Paper 3 (Edwards, 2002) for details. 7Psacharopoulos and Patrinos (1994). 8 World Bank macroeconomic database. 9 For the population aged 15+. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadstica, Guatenmala. '5A Presidential Decree of 1873 provided for the sale of national lands in individual lots of between 45 and 225 hectares each. Plant (1995). Communal rights to ejidos were not entirely abolished, however, for two reasons: (a) it was in the interest of plantation owners unable or unwilling to support a full-time labor force that the Indigenous retain access to some means of subsistence production; and (b) because such lands provided an opportunity for taxation for the Govemment. Grandin (2000) and UNDP (2000). 12 Although coffee is produced by both large and small farmers in Guatemala (with over 30,000 producers), the bulk of production is from the larger producers. Plant (1995). This model of coffee development based mainly on large plantations is similar to that of El Salvador's, but contrasts with relatively more efficient small-holder production practices in Costa Rica and Colombia. In the late 1800s, close to 80% of coffee in Guatemala was produced on farms greater than 50 hectares, as compared with 58% in El Salvador, 38% in Costa Rica, and only 14% in Colombia. Nugent and Robinson (August 2000). 3 Plant (1995). '4 Nugent and Robinson (August 2000). 'Plant (1995). 16 UNDP (2000). '7 Davis (1988). 8 Plant (1995). 19 This compares with land Gini coefficients in other Central American and LAC countries of: 82 for El Salvador, 81 for Panama, 80 for Costa Rica and Nicaragua, 77 for Honduras and Bolivia, and 61 for Mexico. Source: Deininger and Orlinto (2000). The ENCOVI 2000 yields a Gini coefficient of 79 for land, though this survey was designed as a household survey and not an agricultural or land census. 2 UNDP (2000). 21 ANACAFE (2001) estimates that some 300,000 people live and work permanently on coffee fincas in Guatemala. 22 McCreery (1976). 23 This new law was actually a rationalization and expansion of state labor obligations in effect since the 1 830s. Gandin (2000). Workers were forcibly recruited unless they could demonstrate from their personal workbooks (libreto de jomaleros) that such service had been recently performed. McCreery (1983). 24 The requirement was originally for three days of free labor, but this was extended to two weeks per year in 1910. Grandin (2000). " UNDP (2000). 26 Grandin (2000). Nugent and Robinson (August 2000). UNDP (2000). "Plant (1995). 28 Plant (1995) and UNDP (2000). 40 29 Plant (1995). 30 UNDP (2000). 3' Analysis conducted by Nugent and Robinson (August 2000). '2 UNDP (2000). 33 UNDP (2000). 34 By 1930, literacy had already reached 67% in Costa Rica and 52% in Colombia. It was lower in El Salvador, at 27%, but still higher than in Guatemala. Nugent and Robinson (August 2000). 35 UNDP (2000). 36 World Bank World Development Indicators Database using GDP per capita in constant 1995 US$ (Guatemala = $928, compared with $513, $638, and $827 for Honduras, Nicaragua, and Bolivia respectively); life expectancy at birth (Guatemala = 46 compared with 43 for Bolivia, 47 for both Honduras and Nicaragua; and 56 for LAC overall); gross primary enrollment rates (Guatemala = 45%, compared with 64, 66, and 67 for Bolivia, Nicaragua, and Honduras respectively); and infant mortality (Guatemala = 130 per 1000 live births, compared with 166, 143, and 139 for Bolivia, Honduras, and Nicaragua respectively, 128 for lower-middle income countries, and 105 for LAC). 37 World Bank and the Carter Center (July 1997). 38 Davis (1988) and Jonas (2000). 39 Davis (1988) describes a widespread "sociological awakening" of Guatemala's indigenous population during this period. The roots of this sociological mobilization can be traced to the political changes (e.g., the formation of political parties, trade unions, and peasant leagues) that took place in Guatemala in the period of popular democracy and social reformism from 1945 to 1954. In the 1970s, opposition political parties, especially the Christian Democratic (CD) party also began to organize and gain influence in the Guatemalan countryside, some even sponsoring successful indigenous candidates for office. 40 Davis (1988). 41 Jonas (2000). 42 Comisi6n para el Esclarecimiento Hist6rico (CEH, Truth Commission) (1999). 43 Growth averaged 5.4% p.a. in the 1960s and 5.5 p.a. in the 1970s. It then fell to 0.9% p.a. in the 1980s. Sources: IMF and Banco de Guatemala. 44 This point was repeatedly emphasized at a workshop on poverty organized by SEGEPLAN in October 2000. 45 Davis (1988) and Jonas (2000). 46 MINUGUA (June 2001). 47 Jonas (2000). 4s Jonas (2000). 49 Source: SlAF/Ministry of Finances (Feb. 12, 2002). 50 Comisi6n de Acompafiamiento del Cumplimiento de los Acuerdos de Paz (December 2000). 5' MIINUGUA (June 2001). 52 MINGUA (June 2001). 41 PART 2: KEY CHALLENGE: BUILDING OPPORTUNITIES AND ASSETS Chapters 3 and 4 both emphasize the importance of assets and opportunities as contemporary and historical determinants of poverty. Indeed, opportunity (or lack of it) to generate incomes and attain basic necessities is central to the manifestation of poverty. Promoting growth and development is also crucial for meeting the targets set by the Peace Accords, as discussed in Chapter 4. The next few chapters seek to examine more deeply the issues relating to poverty and opportunity with a view of informing policy. Specifically, Chapter 5 examines the relationship between poverty and economic growth in Guatemala from a "macro" perspective. Chapter 6 builds on this macro-economic context to further examine the livelihoods and earnings opportunities of the poor at the household level ("micro" perspective), with a focus on rural livelihoods. The poor also rely on a portfolio of assets in order to forge opportunity, including education (Chapter 7), health (Chapter 8), basic utility services (Chapter 9), land and financial assets (Chapter 6), and access to transport (Chapter 10). Generally, the poor suffer from an unequal distribution of these assets. Chapter 5: Growth and Poverty Economic growth is essential for expanding economic opportunities for poor people. As countries become richer, the incidence of poverty tends to fall. Other indicators of well-being, such as average levels of education and health, tend to improve as well. For these reasons, economic growth is a powerful force for poverty reduction.' The ability of growth to influence the opportunities of the poor depends not only on the pace of growth but also on the pattern of growth in the economy (e.g., favoring or disfavoring sectors which employ substantial segments of the poor population). Growth itself is also endogenous, depending on a range of factors, including, inter alia, human capital development, success in building more effective, transparent and inclusive institutions (which affect investment levels and allocations), functioning markets, and so forth. This chapter examines the relationship between poverty and growth in Guatemala. It first reviews the pace and pattern of growth over time, suggesting potential effects on poverty. The chapter then assesses the potential pace of future poverty reduction and improvements in social indicators, taking into account the expected future rate of growth. Such projections are compared to the goals set by the Government's poverty reduction strategy (ERP) and the international millennium development goals (MDGs). While a full analysis of the sources of growth is beyond the scope of this study, the chapter offers some suggestions as to key priorities for promoting pro-poor growth in the future. POVERTY AND GROWTH OVER TIME Poverty and the Pace of Growth For much of its recent history, Guatemala has enjoyed relative macro-economic stability and reasonable growth. Inflation has generally been held back and growth averaged about 3.9% over the period from the 1950s through the 1990s (Table 5.1). The exception occurred during the 1980s, when growth rates slowed to about 1%, in conjunction with the intensification of the civil war and a demnise in the international economic environment. Growth attempted a rebound during the 1990s, averaging 4%, which is slightly higher than the average for Latin America (3.4% p.a.). Nonetheless, with one of the highest 2 population growth rates in the region (2.6% for the period from 1980-99), per capita growth rates were significantly lower (and even negative during the 1980s), averaging 1.3% p.a. over the past 50 years. 42 Table 5.1 - Average Real Growth Rates, 1950s-1990s 1950s 1960s 1970s 1980s 1990s Average annual change (%) 3.7 5.4 5.5 0.9 4.0 Source: IMF (April 2001) / Banco de Guatemala. Simulations using the ENCOVI yield an elasticity of poverty to GDP per capita of -0.99%. This estimate is similar to what was estimated using the 1998-99 ENIGFAM and growth-poverty elasticities in other LAC countries.3 It is important to note, however, that the elasticity was simulated using single year data and assumes that inequality remains constant and that everyone benefits from growth equally (i.e., if growth is 1% on average, all individuals see their total consumption increase by 1%, and poverty falls by close to 1%). In fact, while survey data suggest that poverty has fallen over the past decade, this decline was slightly lower than what would have be expected given observed growth rates during that period. As discussed in Chapter 2, poverty appears to have fallen from 62% in 1989 to about 56% in 2000. However, given average per capita growth rates of 1.4% p.a. over that same period, poverty would have been expected to have fallen to 53% by 2000 (using the simulated elasticities described above). The difference between the observed reduction in poverty from 1989-2000 (6 percentage points) and the predicted reduction in poverty (9 percentage points) could arise due to measurement issues (e.g., a lack of comparability between the 1989 and 2000 estimates, as discussed in Chapter 2). More seriously, they could also signal that the pattern of growth was not neutral (as assumed), but rather favored the non-poor. Indeed, the sectoral patterns of employment and growth suggest that growth was not neutral or pro-poor, as discussed below. Despite longer-term progress in reducing poverty, simulations suggest that poverty may have increased slightly in 2001 due to the recent economic downturn. A number of recent macro-economic shocks have undermined both economic expansion and poverty reduction. In particular, the global economic contraction led by the US recession is expected to negatively influence key sectors of Guatemala's economy (Table 5.2), including exports and tourism as well as private investment and inflows, particularly remittances. Problems in Guatemala's financial sector have also damaged overall confidence and could further put the brakes on investment (though recent legal initiatives could improve the situation). Finally, the coffee sector (Guatemala's main export) is experiencing a structural crisis, facing the lowest prices in several decades (see Chapter 6). In light of these economic shocks, the real growth rate is estimated to have slowed to 1.9% for 2001, implying a contraction in growth per capita of -0.8%. As such, poverty may have increased slightly from 56.2% in 2000 to 56.6% in 2001, with extreme poverty rising from 15.7% to 16.0%. 43 Table 5.2 - Main Macroeconomic Indicators, 1997-2001 Actual Estimate 1997 1998 1999 2000 2001 Annual real growth rates (1958 prices): GDP at market prices 4.4% 5.0% 3.8% 3.3% 1.9% Population Growth 2.67% 2.68% 2.68% 2.68% 2.69% GDP per capita 1.7% 2.3% 1.2% 0.6% -0.8% Consumption 4.5% 4.9% 3.8% 3.5% 2.6% Exports (GNFS) 7.4% 1.0% 5.5% 4.8% 1.1% Imports (GNFS) 19.2% 23.2% 1.4% 2.6% 2.7% Gross domestic investment 18.4% 38.9% -2.4% 0.9% -0.3% GDP Levels Nominal GDP (US$ millions) 17,768 19,306 18,225 18,988 20,671 GDP per capita (US$) 1,689 1,788 1,644 1,668 1,768 Trade (US$ millions) Exports of GNFS 3,175 3,455 3,466 3,865 3,858 Imports of GNFS 4,188 5,028 5,010 5,358 5,452 Trade balance -1,013 -1,573 -1,544 -1,494 -1,594 Current account balance (US$ millions) -686 -1,037 -1,010 -913 -922 Current account as % of GDP -3.9% -5.4% -5.5% -4.8% -4.5% Change in net international reserves (US$ millions) -232 -243 -121 -674 -344 Investment balances, as % of GDP Gross domestic investment 13.7% 17.4% 17.4% 16.8% 16.0% o/w, Public Sector investment 3.7% 4.5% 5.7% 4.5% 4.1% Incremental capital-output ratio (ICOR) 2.1 2.1 3.6 3.9 6.6 Price and exchange rate indicators CPI inflation (average) 9.2% 6.6% 5.2% 6.0% 7.0% Annual average exchange rate (Quetzales/ US$) 6.1 6.4 7.4 7.8 7.8 Government finances (% of GDP) Total revenues, of which 9.4% 9.7% 10.5% 10.5% 11.1% Tax revenues 8.8% 8.7% 9.3% 9.6% 9.8% Total expenditures, of which 10.1% 11.9% 13.3% 12.4% 13.6% Consumption 3.6% 4.1% 4.4% 4.8% 5.3% Deficit(-)/Surplus(+), after grants -0.8% -2.3% -2.9% -1.9% -2.5% External debt Public Debt / GDP 14.7% 15.0% 18.1% 17.6% 18.6% Source: Banco de Guatemala, World Bank macroeconomic database. Projections by World Bank- Poverty and the Pattern of Growth: Pro-Poor? Growth did not favor the poor because the economy did not generate enough low-skilled jobs. Agriculture, which employs the majority of the poor, experienced below-average growth rates over the past 20 years (Table 5.3). In addition, other sectors did not grow fast enough to offer enough employment opportunities for the poor. As a result, growth was not pro-poor because the economy. did not generate enough low skilled jobs (both agricultural and non-agricultural) to absorb labor surpluses. 44 Table 5.3 - Pro-Poor Growth? The Sectoral Pattern of Employment and Annual Growth (in %) Sectoral growth Contribution to Employment Distribution (annual, in %)" GDP (in %)8 (% of people in: )b 1980-1989 1990-2000 For 2000 Poor Non-poor Total GDP at market prices 1.0 4.0 100 100 100 100 Agriculture 1.1 2.9 24 55 17 36 Industry 0.3 3.9 21 18 24 21 Construction -1.1 4.3 2 6 6 6 Gas, electricity, water 4.5 9.3 4 0.2 1.0 0.4 Mining and quarrying 4.0 11.2 1 0.2 1.0 0.3 Manufacturing 0.2 2.6 14 12 16 14 Services 1.2 4.4 50 27 59 43 Transportation 2.6 5.8 10 3 4 3 Trade/Commerce/Tourism -0.2 4.0 26 15 29 22 Banking/Financial 2.4 6.8 6 1 5 3 Public administration 4.5 4.7 8 2 8 5 Other 1.0 3.3 5 n.a. n.a. n.a. Sources: a. World Bank macroeconomic database/Banco de Guatemala; b. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Figure 5.1 - Structure of Growth, 1965-2000 Figure 5.2 - Structure of Economy, 1965-2000 5000.0- 100% 3000.0 I J - 2 2000.0l 2000.00%- 0 % 0.0 9 9 1965 1971 1977 1983 1989 1995 | Agriculture * Construction O Utilities [OAgriculture EConstruction 0Utilities 0 Mining a Manufacturing 0 Transport 3 Mining a Manufacturing TTransport ElTrade QDwellings USBanking @ Trade 0 Dwellings * Banking U Public admin. * Public admin. POVERTY AND GROWTH IN THE FUTURE: TARGETS AND PROJECTIONS Understanding the potential impact of growth on poverty can be useful for projecting poverty outcomes and other welfare indicators. As discussed above, simulations using data from the ENCOVI suggest that a 1% increase in per capita income could result in approximately a 0.99% decline in the headcount index of poverty. This relationship between poverty and growth assumes that inequality remains constant and that the pattern of growth is neutral.4 Using the expected projections for growth, the projected evolution of overall and extreme poverty is presented in Table 5.4 for the period from 2000 to 2020. Overall, poverty is projected to decrease from 56.2% in 2000 to 44.4% by the year 2020. Likewise, elasticities of social indicators to growth and urbanization from a worldwide panel can be used to forecast how growth would affect these other indicators of well-being.5 Similar to poverty projections, growth 45 yields a significant improvement over time for all indicators (Table 5.5). For example, infant mortality is expected to fall from 40 deaths per 1,000 births in the year 2000 to 26 in the year 2020. Universal coverage of water and sanitation is also expected to be reached during that period. Table 5.4 - GDP Growth and Poverty Reduction 2000 2001 2002 2003 2004 2005 2010 2015 2020 Assumptions Real annual growth rates n.a. 1.9%. 2.3% 3.3% 4.0% 4.0% 4.0% 4.0% 4.0% Population growth n.a. 2.7% 2.7% 2.7% 2.7% 2.7% 2.7% 2.7% 2.7% Real annual per capita growth rates n.a. -0.8% -0.4% 0.6% 1.3% 1.3% 1.3% 1.3% 1.3% Poverty Projections Extreme poverty rates 15.7% 16.0% 16.2% 15.9% 15.5% 15.1% 12.8% 10.4% 8.6% Overall poverty rates 56.2% 56.6% 56.8% 56.5% 55.8% 55.0% 51.6% 47.8% 44.4% Note: Projections based on growth elasticity to poverty of -1. Source: World Bank simulations using data from the ENCOVI 2000, Instituto Nacional de Estadrstica - Guatenala. Table 5.5 - GDP Growth and Non-Monetary Indicators 2000 (base) 2001 2002 2003 2004 2005 2010 2015 2020 Health Indicators Infant Mortality 40.7 39.9 39.0 38.2 37.4 36.6 33.0 29.5 26.3 Under-five Mortality 53.8 52.7 51.6 50.5 49.4 48.4 43.3 38.7 34.3 Life Expectancy 65.3 65.7 66.1 66.5 66.8 67.2 69.1 71.1 73.2 Malnutrition (stunting) 44.0 43.4 42.9 42.3 41.8 41.3 38.7 36.4 34.1 Education Indicators Illiteracy Rate 31.0 30.0 29.1 28.3 27.4 26.6 23.3 20.1 17.4 Net Primary Enrollment 79.0 79.2 79.4 79.6 79.7 79.9 80.6 81.5 82.4 Net Secondary Enrollment 25.0 25.5 26.1 26.6 27.2 27.8 30.5 34.0 37.9 Gross Primary Enrollment 99.0 99.4 99.9 100.4 100.8 101.2 102.7 104.7 106.5 Infrastructure Indicators Access to Piped Water 69.0 70.9 72.8 74.6 76.4 78.1 82.7 92.0 100.0 Access to Sanitation 87.0 89.0 90.9 93.0 95.0 97.1 100.0 100.0 100.0 Telephone Mainlines 15.0 16.2 17.6 19.0 20.5 22.1 30.4 44.5 65.4 Sources: Malnutrition, all education and infrastructure indicators: ENCOVI 2000, Infant mortality: DHS 1998/1999, Under five mortality: WDI 2000 for 1999, life expectancy: WDI 2001 for 1999. Simulations by World Bank using SlhSIP software developed by Wodon et. al. based on a worldwide panel of elasticities, average growth and per capita growth rates of 4.0% and 1.3% p.a., population growth rates of 2.7% p.a. and UN projections for urbanization. Given current growth projections, the record for meeting the targets set by the Government's poverty strategy and the Millennium Development Goals (MDGs) is likely to be mixed (Table 5.6.). Targets for most social indicators established by the Government's poverty reduction strategy (ERP)6 should be met by 2005. Nonetheless, extreme poverty is not expected to fall as ambitiously as anticipated under the ERP due to slower overall growth in 2001 and 2002. Moreover, given projected growth rates, it does not seem likely that Guatemala will meet most of the more ambitious targets for health and education established under the international MDGs (Table 5.6). Faster growth and interventions to boost the assets of the poor are clearly needed to improve living conditions enough to meet these goals. 46 Table 5.6 - Meeting Poverty Str tegy and Millennium Targets: Hit or Miss? ERP Targets MDG Targets Model Projections (Table 5.5) 2005 2015 2005 2015 Extreme Poverty Reduce by Reduce by 50% Reduced by half Reduced from 3 percentage points of the 1990 level a % point 15.7% in 2000 to (by 2% p.a.) 10.4% (by 2% p.a.) Health Indicators Malnutrition (hunger, None Reduce by 50% by 41% 36% measured as stunting) year 2015 (from 59% to 30%) Infant Mortality Reduce to 35 per 1000 Reduce by 66% of Reduced from 40 Reduced from 40 the 1990 level in 2000 to 37 per in 2000 to 30 per (by 2.6% p.a) 1000 1000 (by 1.7% p.a.) Under-five Mortality Reduce to 48 per 1000 Reduce by 66% of 48 per 1000 Reduced from 54 the 1990 level in 2000 to 39 per (by 2.6% p.a.) 1000 (by 1.9% p.a.) Life Expectancy Increase to 67 years n.a 67 years 71 years Education Indicators Illiteracy Rate Reduce to 20% n.a 26.6% 20.1% Net Primary Enrollment Reduce to 88% Universal 79.9% 81.5% Infrastructure Indicators Access to Piped Water Increase coverage to 60% Universal Overall coverage Overall coverage in rural areas 78.1% 92.0% Access to Sanitation Increase coverage to 60% Universal Overall coverage Overall coverage in rural areas 97.1% 100.0% Model projections based on SIMSIP software developed by the World Bank. SUMMARY OF KEY ISSUES AND PRIoR1TIEs The above discussion yields the following main messages: o Economic growth is necessary for reducing poverty and improving living conditions. This is true internationally, but particularly relevant for Guatemala, given the relatively small size and limited capabilities of Guatemala's public sector (see Chapter 13). e Despite Guatemala's historically reasonable growth rates, current growth patterns are neither sufficiently fast nor disproportionately oriented towards the poor. Guatemala's growth rates will not likely be sufficient to reach the international MDGs. Moreover, the poor do not seem to be benefiting from the existing pattern of growth. o As such, steps should be taken to improve both the pace of growth and make it more pro- poor. In this context, the main engine of growth is likely to come from the private sector, with the public sector playing a supporting role affecting growth mainly insofar as it stimulates private- sector investment and productive activities. Yet the actions of the public sector in this supporting role are crucial. While a full analysis of the sources and engines of growth in Guatemala is beyond the scope of this paper, several areas appear to be important priorities: o Maintaining macroeconomic stability; o Enforcing a tight fiscal position, with a careful plan for strengthening tax collection and redirecting public spending towards the social sectors so as to build assets that are crucial to both growth and poverty reduction; 47 o Fostering a climate that is conducive to private investment and growth, including improvements in governance and public sector management (see Chapter 13), as well as the regulation and supervision of the financial system. Without such improvements or a resolution of underlying social conflicts, private investment will remain depressed or will be channeled into low-productivity areas (e.g., linked to corruption), further hampering Guatemala's prospects for growth; and o Promoting growth with special emphasis on sectors that are likely to generate substantial employment for the poor in both rural and urban areas. Additional analytical work is needed to define a more comprehensive pro-growth strategy. Nonetheless, while a thorough sectoral analysis of growth is beyond the scope of this study, available data do suggest certain levers that would have stronger impacts on poverty reduction than others for urban and rural areas: * In urban areas, this requires policies to support labor-intensive sectors, particularly micro-, small- and medium-enterprises (MSMEs), as well as education and technical training. * In rural areas, this means developing non-agricultural activities that are better remunerated and have better long-term prospects than traditional agriculture. As discussed in Chapter 6, key interventions to support growth in non-farm activities include: (a) increasing and improving the targeting of investments in education and technical training; (b) increasing investments in transport and basic infrastructure, which are crucial for the diversification, growth and inclusion of the poor in the rural economy and with facilitating the adjustment to the coffee crisis; and (c) policies that promote micro-, small- and medium-enterprises (MSMEs), a segment of the private sector that tends to generate a lot of employment. While agriculture is unlikely to generate enough additional employment opportunities to reduce poverty on a large scale in the medium term, it will continue to be an important source of incomes for the poor (at least in the short run). In this context, diversification efforts should focus on non-traditional products with better demand and price prospects than traditional export crops (as discussed in Chapter 6). Policies should also continue to facilitate productivity improvements (such as technical assistance), so as to boost the earnings of those who remain in agriculture. Investments in infrastructure (e.g., rural roads to improve marketing opportunities and education to improve farm-management practices) will likewise be important. ' World Bank (2001e). 2World Bank World Development Indicators 2001. 3 See Wodon et. a]. (2000). 4As discussed above, this assumption might not hold, since it appears that the pattern of growth has not been neutral, but rather has favored the non-poor. 5 The worldwide panel of elasticities for social indicators and simulation software (SIMSIP) was developed by Wodon and Ryan (2000). The model used here assumes an annual growth rate of 4% (1.3% per capita taking into account the population growth rate of 2.7%), urbanization as projected by the UN. In addition, the expected projections also assume constant inequality and neutral growth. Estrategia de Reducci6n de la Pobreza (November 2001). .48 Chapter 6: Livelihoods, Labor Markets, and Rural Poverty "Poverty is not having anything...not having work, there is no land. " Ladino Villager, L2 (QPES) While the last chapter discussed the general importance of economic growth in reducing poverty from a "macro" perspective, this chapter attempts to examine more closely the livelihoods of the poor from a more "micro" (household-level) perspective, with a view to informing policy-makers of potential paths for promoting opportunities for the poor. Indeed, incomes, livelihoods and opportunities are among the top concerns ranked by Guatemalan households.' In fact, despite progress in other areas, a significant share of households do not perceive improvements in welfare since the Peace Accords, and they attribute this primarily to a lack of opportunities (factors that directly affect "their wallets," as discussed in Chapter 2). As such, improving livelihoods, particularly for the rural poor, presents one of the main challenges for both the poverty and peace agendas in Guatemala. In this context, this chapter begins with a review of income sources in Guatemala. Since labor is the main productive asset for the poor, the chapter then examines the constraints faced by the poor in generating incomes, and how such constraints may exclude them from participating in the overall economic system. In addition, given that these constraints are directly connected to a lack of access to credit, insurance and opportunities, it is important evaluate how these failures relate to vulnerability and exclusion of people or specific groups. Finally, since poverty in Guatemala is highly concentrated in rural areas, the chapter analyzes issues pertaining to rural livelihoods, including land, agriculture and non-farm opportunities. The chapter is mainly based on an analysis of data from the ENCOVI,2 though most of the findings are also confirmed in the QPES. INCOMES AND INEQUALITY Incomes in Guatemala are unequal both in their distribution and in their sources. As discussed in Chapter 2, income inequality in Guatemala is quite high, with a Gini coefficient of 57. While the poorest quintile of the population receives only 3% of total income in Guatemala, the top quintile captures 62%. Disparities are also evident in the sources of income (Table 6.1). Specifically: * The poor are largely dependent on agricultural income, reflecting their predominantly rural location. Agriculture accounts for about half of the total income of the poorest quintile, as compared with just 3% for the top quintile. As discussed below, the poor, particularly the extreme poor, seem to lack opportunities outside the agricultural sector, which generates relatively low incomes. Moreover, agriculture (particularly day-labor jobs) offers few labor benefits, such as job security or pensions. * In contrast, the non-poor have access to a much more diversified set of employment opportunities, with significant employment in non-agricultural sectors, such as services. As discussed in Chapter 5, growth rates have been significantly higher in these sectors than in agriculture over the past 20 years. * Transfers, both public and private, are important sources of income for the poor. Private transfers in the form of remittances constitute more than 20% of per capita income for households that receive them. In addition, international remittances (especially from the United States and Mexico) are on average twice as large as domestic (see Box 6.1). Simulations of the effects of current adverse shocks in the coffee industry (which is already affecting domestic remittances via the decreases in seasonal employment) and the global economic slowdown (especially in the United States), suggest a sharp decline in per capita income and a slight increase in both overall and extreme poverty (as discussed in Chapter 5). On the other hand, while public transfers 49 represent a higher share of income for the poor (Table 6.1), they are regressive in their absolute levels. These issues stress the need for policies that help affected populations cope with and adjust to shocks (as discussed in Chapters 11 and 12). They also suggest that improving targeting mechanisms while implementing policies for the poor and other vulnerable groups is essential (as discussed in Chapter 11). Table 6.1 - Income Sources, by Consumption Quintiles Consumption quintiles 1 2 3 4 5 Total Income per capita (Q) 1,429 2,408 3,487 5,064 15,503 5,578 Labor income (%) 77 78 77 76 70 73 Agricultural 49 38 24 14 3 13 Salaries 30 18 11 6 1 6 Formal sector 13 9 6 4 1 3 Informal sector 17 9 5 2 0 3 Net income from production 19 20 13 8 2 7 Non- Agricultural 28 40 53 62 67 60 Salaries 17 25 39 47 46 42 Formal sector 8 13 26 36 42 35 Informal sector 9 12 13 11 4 7 Own business 11 15 14 15 21 18 Formal sector 1 1 1 1 7 4 Informal sector 10 14 13 14 14 14 Non-labor income (%) 22 22 24 26 30 27 Return to capital' 10 8 10 11 16 14 Donations, gifts 11 12 12 10 6 8 Remtttances 3 4 5 5 4 4 Private 1 2 1 2 1 1 Public 7 6 6 3 1 3 Pensions, indemnizaciones 1 1 1 3 5 3 Other' 0 1 1 2 3 2 Percentages may not add up to 100 due to rounding. ' As interest received was negligible, the return to capital includes: income from rental of equipment, rental of property and the interest received. b For example, inheritance or lottery winnings. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadtstica - Guatemala. LABOR MARKETS Labor Market Participation Participation in the labor market depends significantly on gender and education levels. The Guatemalan labor force consists of about four million people (with an additional half a million children that are employed between the ages 7-14). Participation is high among men (89%) and moderate for women (44%). Poor men are more likely to participate in the labor force compared to non-poor men, while for women the opposite is true. In fact, more educated women (who tend to be non-poor) seem to self- select into the labor market, which is not surprising given the higher retums to education for women, as discussed in Chapter 7. Finally, open unemployment is very low but underemployment (based on hours worked) affects about a third of the working population, which is consistent with a number of hypotheses such as exclusionary practices or lack of employment opportunities. Informality Labor markets are characterized by a high degree of informality. Informality seems to be growing in Guatemala. Based on available data, it is estimated that an annual average of some 6,400 forrnal sector jobs were lost during the 1990s.3 Indeed, the formal sector seems to have been incapable of absorbing the 50 growing labor supply, with surplus workers being pushed into the informal sector.4 The ENCOVI 2000 shows that more than two thirds of the employed and three quarters of the working poor are engaged in the informal sector. Women are also more likely to work in the informal sector than men (with 71% of female workers engaged in the informal sector, as compared with 62% of male workers). Importantly, the informal sector is very dynamic and heterogeneous in Guatemala, ranging from small scale farmers to textile workers and merchants (Figures 6.1 and 6.2).5 The structure of the informal sector also differs significantly between rural and urban areas. This diversity implies a variety of paths out of poverty. For example, the negative correlation between non-agricultural informal work and poverty, both in rural and urban areas, highlights the importance of non-agricultural employment opportunities for poverty reduction (as discussed in more detail below). Fing 6.1 - Snplnn, Divnity in the tkin Inftl Seo, % of Infomanl Sno Figum 6.2 - En,ploynfm Diversity in the Ron.1 Sectr, %of lnfown Sector Woken. ENCOVI 20OO Wokenn, ENCOVI 2000 1 2 3 4 5~~~~~~~~~~~~~~~~~~~~ octon quiimoten3 4 COAgriouwee UMonofnctena Cctnane OCoEnMnnitty Dotb | t- _ _ . |as~~~~~~~~~~~~~~~~~~~~~~~~~AgnictInoee Moot,v actio Oonnare OCornnnooy Uottee Box 6.1 - Migration and Livelihoods Migration has long been a domninant feature of Guatemalan life, though migration patterns have shifted over time. Seasonal Migration. During colonial times, indigenous people migrated to the south coast to work in the production of indigo and salt.6 Accelerated land expropriations, forced labor mandates and the emergence of coffee in the late 1800s resulted in massive seasonal, and some permanent, migration of the indigenous from the northwest highlands to satisfy labor shortages in the southem coffee region, as discussed in Chapter 4. Today, between 400,000 and 800,000 people migrate temporarily each year.7 More than two thirds of seasonal migration occurs within Guatemala. Seasonal migration is more common among poor and rural residents, the indigenous, bilingual individuals and among men, according to data from the ENCOVI 2000. Indeed, poverty rates among temporary migrants are high (75% as compared with 55% of permanent migrants and 56% of the overall population). While a lack of income-eaming opportunities is the main motive for migration according to migrant households in the ENCOVI, migration is also used as a coping mechanism in response to crises (such as the civil war or Hurricane Mitch).8 The crisis in the coffee sector and emerging problems in sugar production will definitely affect an already vulnerable population that relies extensively on these activities for their incomes. International Migration. Intemational migration has grown from about 50,000 people in the 1980's to more than a million people in late 1990s.9 Destinations vary, but the main recipients are the United States and Mexico. Overall, 9% of all households receive intemational remittances, though a higher share of the non-poor (12%) receive them than the poor (6%).l5 The QPES suggests that remuneration (both in-kind and in cash) and working conditions are better in thefincas in Mexico than those in Guatemala. Moreover, it is apparent that as community members migrate abroad, their remittances contribute substantially to those families that receive them. Indeed, there are obvious differences in the houses of those with family members living in the US and those withouL It seems that intemational migrafion introduces a certain degree of within-village inequality at the community level due to remittances and goods sent back "home." Permanent Migration. The ENCOVI shows that permanent migrants are more likely to be non-poor living in urban areas and having completed primary education. They are also predominantly non-indigenous. On the one hand this may reflect limitations faced by the indigenous in finding permanent opportunities and relocating prior to the signing of the Peace Accords, such that permanent relocation was contained among the better off, namely, the non-indigenous. On the other hand, the negative correlation between permanent migration and poverty may also suggest permanent rural-to-urban migration as a possible path out of poverty. While it is impossible to test this hypothesis with ENCOVI data, it may be the case that by migrating, these people managed to take advantage of better opportunities, assets and services in urban areas such that most of them are not poor today. 51 Labor Policies Active labor market interventions are not effective instruments for reducing poverty. In fact, existing labor market policies do not benefit the poor since the do not reach them. The dominance of informal sector jobs among the poor puts poor workers out of reach of such policies." In fact, data from the ENCOVI show that, while about 40% of salaried workers earn less than legal minimum, this share rises to 60% of poor workers.'2 Moreover, only 16% of the poor receive other labor benefits (such as the thirteen month salary bonus). Indeed, villagers in the QPES community KA1 describe the schemes used by employers to avoid paying labor benefits to the finca workers (including dismissing them regularly throughout the year so as to maintain them on non-permanent contracting status and hence avoid paying the benefits). Since active labor policies introduce distortions into the labor market, these policies can in fact be counterproductive. As discussed below, policies should instead focus on enabling the poor to participate in the labor market, which requires investments in education, technical training, and infrastructure (such as roads). Such investments will help break down barriers to opportunities and improve renumeration. Opportunities The poor have fewer opportunities for higher-paid jobs and a limited ability to diversify income sources. The non-poor are twice as likely to work in a higher paying jobs in the public sector or white- collar occupations as the poor. Moreover, the poor are three times as likely to work in agriculture as the non-poor. In fact, poorer households appear to be rather constrained in employment opportunities, dividing their labor between self-employment, blue-collar jobs, and agriculture, which all yield significantly lower incomes than other sectors. In addition, household employment portfolios reveal a pattern whereby poorer households are more homogenous in their occupations (e.g. most of the individuals are in agricultural), while for non-poor households there is more diversity among the occupations of working members. While some of this disparity is explained by differences in human capital accumulation, this unequal distribution in opportunities and income-generating possibilities is also attributable to the lack of spatial integration for these groups. Non-Spanish speakers and indigenous populations face similar constraints in employment opportunities. Geographic location is correlated with poverty and employment opportunities. The spatial proximity to a bigger city may offer a number of advantages to a household such as employment opportunities but also access to services and infrastructure not available in a smaller community. Combining data from the 1994 census (which allows for the construction of municipal population sizes) with those from the ENCOVI reveals that location is central to employment opportunities. First, 75% of the households that reside in small municipalities (less than 10,000 people) are rural, whereas 40% of those living in municipalities with more than 30,000 people are rural. Second, poverty rates are significantly higher among households in smaller municipalities. Third, the share of non-farm income is higher for those households residing in larger municipalities implying that non-farm employment opportunities (that yield higher incomes) are more likely to be available in these areas. In fact in the rural areas, the share of non- farm self-employed income for households living in larger municipalities is almost twice the income of those from smaller municipalities. Therefore, if municipal size is a proxy for opportunities and infrastructure, these correlations imply that integrating households with markets and the rest of the economy is key for allowing the poor to access opportunities. Wages and Returns to Labor Real wages have followed a divergent trend between sectors. The evolution of real monthly wages by industry over the last decade reveals two interesting patterns.'3 First, while in the early 1990s, real monthly wage growth was relatively equal among the different industries, dispersion in wages has increased in the late 1990s, as has been observed for many other countries in LAC. The increase in wage inequality could be a response to reforms'4 and trade liberalization, which shifted relative wages in favor of higher skilled 52 jobs (and perhaps has driven the apparent increase in overall inequality, as discussed in Chapter 2). Second, agricultural wages have been increasing much slower than in other sectors since the mid-1990s. Nonetheless, despite lagging levels, wages in agriculture seem to have increased in real terms during the 1990s, probably explaining a good part of the improvements in poverty. Figure 6.3 - Evolution of Real Monthly Wages, by Sector 3000 .~~~~~~~~~~~ Mninq Construction Transport 2000 - 2000 / Commerce zD _ Com~~~~~~~~~~~~~~~~~~~munity Manufacturing sic Services 1000- =000 = ~ LAg riculture 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Ariusted for prbces Year Source: ILO (labost ilo.org) The lowest labor earnings occur in agricultural occupations, rural areas, the informal sector and among marginal groups such as the poor and indigenous. Real hourly wages average Q7.3.15 Wages are more than twice as high in urban as compared with rural areas and also for the non-indigenous as compared with the indigenous. In addition, the average real wage of Q3.3 in agriculture is almost five times smaller than the Q15.8 wage for financial services. The average hourly wage in the informal sector is less than half of that in formal occupations in the private sector. Wages decrease dramatically for poorer individuals and increase with higher levels of education. Similar patterns are also observed for earnings among the self-employed. Men in lower-skilled occupations receive more than women while there is more wage equality among higher skilled jobs. Discrimination in labor markets is often reflected both in hiring practices but also in the earnings differentials between different groups such as men and women. In Guatemala, wages for men are up to 50% higher than for women in jobs like manufacturing and commerce. This wage differential is smaller and even negligible in the public sector or white-collar occupations, where typically the educational attainments are higher. Yet, as the analysis shows below, wage differentials cannot be fully explained by educational attainments alone, implying that there is a high degree of discrimination. Ability to speak Spanish is correlated with higher earnings. Men and women who speak Spanish earn more than 30% more than those who do not, even when other factors are taken into account using regression analysis. This is also true for bilingual speakers, indicating the role of language ability. Yet, as shown below, neither educational attainments nor language ability are enough to explain earnings gaps between indigenous and non-indigenous, implying the existence of labor market discrimination. Wage discrimination is high for indigenous groups. While human capital endowments explain some of the variation in earnings, wage discrimination (based on the unexplained part of the determinants of 53 earnings) is prevalent for the indigenous. 16 The average wage gap between indigenous and non-indigenous workers is 50%. Controlling for differences in human capital endowments, experience, sector of employment and other characteristics, a significant share of this gap can be attributed to wage discrimination (95% for men and 35% for women) against indigenous workers. Similarly, close to half of the wage gap between men and women can be attributed to wage discrimination (controlling for other factors), though the average wage gap between men and women is small (1.3%). Child Labor Child labor is common in Guatemala. About half a million children between the ages of 7 and 14 are employed in Guatemala, with a third of them working in plantations (mainly coffee and sugar fincas). Most of these children come from poor households (75%) and reside in rural areas (80%). Boys are more likely to work in agriculture, while girls generally work in both agriculture and commerce. They receive significantly lower wages than adults and work about 30 hours a week, seriously inhibiting the ability to attend school (see Chapter 7). RURAL POVERTY AND LIVELIHOODS As discussed in Chapter 2, rural areas have disproportionately higher rates of poverty than urban areas. This section examine the livelihoods of the rural poor, arguing that agriculture and land reform are unlikely to provide significant policy levers to reduce poverty. Subsistence farming and traditional exports such as coffee, in particular, are unlikely to be significant sources of employment growth for the rural poor. Non- traditional exports could serve as a potential source of growth, but their reach has been limited in scope, particularly for the poor. Rather, data from the ENCOVI suggest that non-farm opportunities are likely to drive rural growth and employment opportunities. As such, removing constraints faced by the poor to engage in such occupations is crucial. Agriculture, Land and Rural Poverty The vast majority of the rural poor are subsistence farmers or agricultural day laborers. Some 87% of the rural poor depend on agriculture, either as small-scale subsistence farmers or agricultural laborers (Table 6.2). Indeed poverty rates among these groups are significantly higher than those whose main source of income comes from non-agricultural sources. As discussed in Chapter 3, dependence on agriculture is significantly related to lower consumption levels (even after controlling for other differences). Access to land is inversely correlated with poverty. Land is not only necessary for agricultural production, but can also serve as collateral, allowing households to obtain credit for input purchases or diversify their income portfolio by engaging in other activities Table 6.2 - Rural Poverl t by Land Status 'and Main Source of InJcome l Noenetheless land ownershtipvins % of rural % of rural poor Poverty rate Nonetheless, land ownership inmPo ulation (PO) Guatemala is highly skewed and All rural 100 100 75 unequally distributed, as L-l hecars 32 35 82 discussed in Chapter 4. The poor 1-2 hectares 10 12 87 tend to have smaller plots and 2-5 hectares 7 7 80 5-15 hectames 3 3 74 poverty rates are higher among >15 hectares 2 2 74 those with smaller plots (Table Tenants 19 20 80 6.2). Indeed, larger holdings are Landless Households' Agricultural day laborer-s 8 8 76 significantly correlated with Non-agricultural workers 19 13 49 higher consumption in rural areas a. Landless households are divided into those that derive most of their income from agriculture (even controlling for other and those that derive most of their income from non-agricultural sources Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - factors, as discussed in Chapter Guatemala. 54 3). Moreover, the relative geographic isolation of the poor suggests that their land is poorly located, further depressing the values of this asset. The plots owned by the poor are also likely to be of lower quality than those owned by the non-poor. In addition, poor landowners are less likely to have land titles or use land markets. Land titles allow farmers access credit by using their land as collateral for credit (see Box 6.2). They also offer security and even provide incentives for the farmer to be more productive. Overall, only 41% of landowner households possess formal title to their land. Only a third of the poor have land titles compared to almost half of the non-poor. Larger landowners are also more likely to have titles. Moreover, the non-poor use land markets more than the poor. In particular, more than half of the land owned by non-poor households was purchased on the market as compared with only a third of the land owned by extreme poor households (the remainder being acquired by inheritance). Box 6.2 - Rural Credit Credit can provide an important input to both agricultural and non-farm profitability. Credit allows households to acquire land and other inputs. It also enables households to diversify their income sources by engaging in other activities besides farming. Nonetheless, access to credit in rural Guatemala is minimal. Only 13% of the rural households applied and received any kind of a loan (15% among the non-poor and 13% among the poor). Just 10% of landless households applied and received a loan as opposed to 15% for landowner households, indicating the importance of collateral (as discussed below). Loan sizes were significantly higher for the non-poor than the poor. Lack of collateral is one of the most frequent reasons households refer to for not applying for a loan. The ENCOVI suggests that about 40% of rural households wanted to receive credit but chose not to apply. Among those who believed they needed credit, the top three reasons for not applying for a loan were: (a) lack of collateral; (b) too expensive; and (c) fear of ineligibility for a loan. In the case of collateral, it is not necessarily the lack of assets per se but the lack of formal ownership of the assets that may prevent households to access credit. For example, lack of land titles (discussed above) may disable farmers from obtaining credit even though they do own the land. Interestingly, it appears that credit institutions do exist in rural areas. However, the fact that that people feel that loans are expensive may be a signal of a non-competitive credit market (for example, a few local informal money lenders). In addition, a lack of information may lead to people with the wrong impressions about their ability to qualify for a loan as well as the costs and risks associated with lending. As such, addressing these issues of information and supply is important. In fact, a recent study on rural financial services in Guatemala recognizes the short supply of formal lending institutions as an important impediment to obtaining credit in rural areas.'7 Indeed, only half of the households that reported receiving loans in the ENCOVI survey obtained them from formal lenders. Furthermore, the poor and landless are more likely to receive credit from informal lenders, corroborating the findings that formal lenders require land titles as collateral and the fact that the poor face constraints in accessing credit from formal institutions. Nonetheless, interventions in credit markets should be complemented by investments in rural infrastructure, education and information access in order to raise living standards in rural areas Market-based land reform efforts are promising and should continue to be pursued. However, existing pilot land programs have proved difficult and slow due to their high costs and other design issues (Box 6.3). Further, low agricultural returns in traditional crops require that beneficiaries of these programs look for better prospects in non-traditional alternatives. However, the staggering costs of providing the bulk of the rural poor with good quality parcels and complementary inputs suggest that this is not a feasible option to eradicate mass poverty in the rural sector. Similarly, agriculture overall is unlikely to provide a significant route out of poverty for the bulk of the rural poor. Agriculture has faced declining growth rates over the past several decades, contracting as a share of GDP (see Chapter 5). Within agriculture, traditional-crops such as coffee, which tend to employ a significant number of workers, are contracting in the face of a structural terms-of-trade crisis, as discussed below. Non-traditional crops have increased significantly, but not enough to replace the earnings and employment opportunities lost by the coffee crisis, as discussed below. As such, although the poor will 55 likely to continue to depend on agriculture as an important source of income, it is unlikely that agriculture will provide the solution to the poverty problem or that many people will escape poverty via agriculture. Box 6.3 - Land Redistribution Programs in Guatemala Land redistribution is a sensitive topic, both politically and culturally. Since the 1980s, a number of land programs have been at work helping farmers access land in Guatemala. Programs like the "Fundaci6n Guatemalteca para el Desarrollo-Fundaci6n del Centavo" (FUNDACEN), the "Fondo para la Reinserci6n Laboral y Productiva de la Poblaci6n Repatriada" (FORELAP), and the "Fondo Nacional de Tierras" (FONATIERRA) have benefited more than 7,000 households. As part of the Peace Accords in 1996, the Guatemalan Government established the Comisi6n Institucional para el Desarrollo y Fortalecimiento de la Propiedad sobre la Tierra (PROTIERRA) whose responsibilities include: (i) a cadastral-based land registry; (ii) a land fund to promote market-driven land reform; (iii) land conflict resolution mechanisms and free legal services with special attention to land access and land traditional management by rural communities; (iv) a national geographic information system; (v) a comprehensive land tax system; (vi) agricultural development; and (vii) rural investment programs. The idea of the land fund is to allow poor households to access land via a credit subsidy. The program gives a loan to qualified households that allows them to buy land and inputs. In addition, the participants receive technical assistance. The loan lasts for 4 years and it costs 5% in annual interest. Up to today, about 5,000 households have participated in the program. However, this and previous programs have been criticized for being slow and the overall target of 335,000 (in the case of the land fund) to be unfeasible. In addition, critics mention the lack of incentives to repay the loans as repossessing the land is hard to implement, preferential treatment in the way that land is allocated, the low interest rates charged and the program's possible overcrowding of other land projects as important problems. Currently, both the government and local agencies are considering alternative market-based mechanisms to facilitate land access. For example, rental markets are likely to be cheaper to implement. In addition, title programs (such as the "Registro General de la Propiedad" (RGP) or pilot programs such as the "Catastro" project may not only allow households to access credit but also enable them to participate in rental programs. Finally, land leasing programs with the option to buy the land are also considered. Sources: The land fund program, (1998); Caffera (1999) Traditional Agricultural Exports: Vulnerability and the Coffee Crisis The worldwide structural change in the coffee industry is seriously affecting Guatemala. Coffee has always played an important role for the Guatemalan economy. It is the most important export of the country with receipts of more then $570 million in 2000 (20% of total export earnings), making Guatemala the fifth largest coffee exporter in the world.18 However, the recent entry of a number of new producers (particularly Vietnam), as well as overproduction in Latin American countries (e.g. Brazil) have severely depressed international coffee prices, resulting in significantly lower revenues for coffee producers in Guatemala. Indeed, the National Coffee Association (ANACAFE) estimates that export volumes in 2001 fell by 1 million bags, to 5.3 million, and receipts by 50% to less than $300 million. As the downward trend in international prices was caused by structural changes on the worldwide coffee market, these changes are likely to be permanent (rather than a temporary price shock). Coffee production provides a significant source of income for many rural households. The ENCOVI reveals that 11% of rural households produce coffee. About 140,000 rural households receive incomes from coffee production, of which more than 75% are poor.19 On average, coffee producing households received Q4,526 in coffee sales. Non-poor coffee-producing households received almost five times more in coffee sales than poor households. Coffee income comprises about 25% the total income per capita, irrespective of poverty status (for coffee producing households only). As expected, most coffee producers are landowners. 56 Many more people depend on coffee production from the demand of agricultural labor. According ANACAFE, there are an estimated 200,000 people permanently employed in the coffee industry. This figure increases to more than 500,000 during coffee harvest season. Most laborers (jornaleros) in the coffee sector are seasonal migrants from poor households that depend on the coffee sector to augment their incomes, as discussed in Box 6.1. Indeed, poverty rates among temporary migrants are high (75% as compared with 55% of permanent migrants and 56% of the overall population). The crisis in the coffee industry will thus affect a significant share of the rural population. Lower revenues are likely to push some coffee producers to dramatically decrease their demand for labor or force them completely out of business. According to the Ministry of Agriculture more than 40,000 coffee production related jobs are expected to be lost in 2002 (ANACAFE puts this figure at 60,000). As most of these jobs are expected to be low-end jobs, the effect on the poor is likely to be significant. Faced with this crisis, a number policies should be considered. First, for those producers who will be able to survive the crisis as coffee farmers, assistance should be provided to expand Guatemala's presence in well-paid "niche" markets (e.g., specialty, organic, "fair trade" coffees) by improving quality, helping extend long-term contracts and establishing contacts with purchasing firms in developed countries. Second, for those who are forced to exit the sector, assistance should be provided to help with diversification in agriculture and non-agriculture, including: developing a large portfolio of substitute crops, livestock activities, service sector activities (e.g., commerce, restaurants, tourism, cultural activities), and light manufacturing opportunities (e.g., handicrafts, textiles); providing technical assistance to boost the competitiveness of promising activities; and improving the skills of the people displaced from coffee. Programs to address the employment losses also seem necessary in the short run, given the large number of workers that will likely be affected. For example, workfare programs (particularly seasonally-targeted schemes) could be strengthened and expanded to provide alternative employment for those dependent on salaries from (seasonal) labor in the coffee sector. Crop Diversification and Non-Traditional Agricultural Products Guatemala's non-traditional agricultural exports sector has grown impressively over the last few years. Over the past decade, exports of various non-traditional crops have increased dramatically. For example, exports of fruits such as mangoes, papaya, berries and melons, increased from US$14 million in 1990 to more than US$300 million in 1999. Traditional exports, such as coffee and sugar, still dominate, however, increasing from US$629 million to US$1,044 million over that same period. Nonetheless, the weakened outlook for revenues from traditional exports combined with the expansion of the production and export of non-traditional crops highlights the potential role of this sector for future opportunities in rural areas. However, the ENCOVI 2000 reveals that very few farm-households produce non-traditional agricultural products, and most are non-poor. The survey allows for the division of households among those that produce export-related agricultural products (further divided in traditional and non-traditional) and all other products.2' Only 23,000 households produce non-traditional agricultural crops as opposed to 650,000 that produce subsistence crops (Table 6.3).22 Households that produce non-traditional agricultural products have better socioeconomic indicators, suggesting that production of these crops may be limited to those households that have better access to resources. Further research could explore the types of constraints that may impede farmers from engaging in the production of these crops. 57 Table 6.3: Types o Crops Produced, by Type of Household Types of cros produced: a Both Both subsistence and Subsistence subsistence and Non-Traditional Traditional only traditional exports exports exports only Household population (in 000's) 651 147 23 28 Indigenous (% within crop category) 58 35 75 44 Poverty (within crop category) Extreme Poor (%) 20 26 5 14 All Poor (%) 73 79 63 66 Household distribution by land ownership (%) Tenants 33 14 17 10 Landowners 0-1 hectares 45 38 54 72 Landowners 1-2 hectares 10 21 17 13 Landowners 2-5 hectares 6 16 9 5 Landowners 5-15 hectares 3 8 2 0 Landowners >15 hectares 3 3 1 2 Total 100 100 100 100 a Traditional exports crops are coffee, sugar, bananas and cardamom. Non-traditional exports are snow peas, sprouts, broccoli, cauliflower, flowers, mangos, melons, pineapple, papaya, okra and berries. Subsistence crops are the remaining crops (e.g. com). Classification based on those of the Asociaci6n Gremial de Exportadores de Productos No Tradicionales (AGEXPRONT). Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Agricultural Technical Assistance Agricultural extension and technical assistance could serve as potentially levers to promote rural growth and reduce poverty, particularly given the importance of crop diversification and the need to improve farm productivity in Guatemala. Nonetheless, data from the ENCOVI show that only 3% of the farmers in rural Guatemala received such assistance in 2000. While most technical assistance provided by non-public institutions seems to be well targeted to poor households, more than 70% of public technical assistance is received by non-poor households, suggesting inadequate targeting of public interventions. Non-Agricultural Opportunities: a Potential Route Out of Poverty? Non-agricultural opportunities may provide a possible channel for escaping poverty. Much of the recent empirical literature on the relationship between non-agricultural incomes and rural poverty clearly indicates a strong relationship between the two.23 First, by diversifying their income portfolios, rural households can augment their incomes and minimize adverse income shocks from farm activities. Second, as non-agricultural incomes increase the households' cash liquidity, households can access farm inputs easier thus raising farm productivity. Finally, the non-farm sector offers poor landless households (otherwise unable to engage in farm activities) an option for income generation. Indeed, rural households that do not depend on agriculture are less likely to be poor. The ENCOVI reveals a negative correlation between non-agricultural incomes and poverty in Guatemala. Landless households have significantly better socio-econornic indicators than those that depend on agriculture: lower poverty rates (see Table 6.2), higher incomes and consumption levels, lower household sizes and higher levels of education for the household head. In fact, almost half of the rural non-poor households are landless, typically working as self-employed entrepreneurs in commerce or manufacturing, or in non- agricultural salaried jobs (e.g., construction, teaching). Only 26% of the landless households are indigenous, which may be an indication that access to specific non-agricultural occupations may be constrained for indigenous households. The earlier findings on indigenous labor-market discrimination via wages also supports the possibility of exclusion. 58 Given the potential of non-agricultural employment in reducing rural poverty, it is important to understand the constraints that specific groups face in accessing these jobs. Numerous constraints are plausible. For example, as the poor generally live in smaller and isolated communities (based on regional population densities), opportunities for non-agricultural jobs may be scarce. In addition, this spatial isolation implies that these communities will be more likely to lack complementary infrastructure (e.g. roads, electricity, telephones). In fact, a study on the role of basic services and non-farm enterprise profitability using the ENCOVI finds that: (a) the probability of having a micro-enterprise in rural areas is significantly higher among households with coverage of modem utilities; and (b) micro-enterprises without access to services such as electricity and telephone connections in the rural areas have significantly lower profits that those who have access.24 Regression analysis confirms that human capital as well as lack of infrastructure are both important constraints for participation in the non-agricultural sector. In particular, employment-type choice models indicate that low human capital (e.g. education) is an important impediment for participation in higher return occupations. In addition, areas where non-agricultural opportunities are more wide-spread increase the probability of being employed in such jobs, suggesting that the role of local infrastructure and access to services is a necessary condition for employment growth in non-agricultural jobs. These findings imply that while the non-agricultural sector could offer an exit path from poverty, overcoming the constraints that are associated with accessing these jobs is an essential part of a rural poverty reduction policy kit. Still more work could focus on explicitly exploring these constraints to fully understand the dynamics of the non-agricultural sector with the rest of the economy. SUMMARY OF KEY ISSUES AND PRIORrTrES This chapter highlights a number of key issues regarding livelihoods and opportunities for the poor: * The poor, women and the indigenous are constrained in both employment opportunities and earnings, with a high dependence on agriculture and the informal sector and few labor benefits. * The indigenous face significant wage discrimination; women also face wage discrirmination, though the wage gap between men and women is smaller than the one between the indigenous and non- indigenous. * Geographic location is an important factor in determining both poverty and earnings opportunities, with smaller municipalities offering fewer options. * While the poor are highly dependent on agriculture (subsistence farming and agricultural jobs), agriculture is not likely to be a dynamic source of new employment opportunities and will continue to shrink as a share of GDP. As such, agriculture is unlikely to serve as a major vehicle for poverty reduction. * Similarly, although land is an important asset, its ownership is highly inequitable in Guatemala. Moreover, the holdings of the poor tend to be: (a) quite small, often providing below-subsistence incomes; (b) untitled; (c) poorly located (geographically isolated); and of poor quality. In addition, full-fledged land reform is unlikely to serve as a major vehicle for poverty reduction due to the high costs and slow pace of land programs, and low agricultural returns. * Coffee, Guatemala's traditional export crop, has suffered a long-term structural price shock, further emphasizing the need for diversification. This shock is likely to have substantial impacts on the poor, both as producers and as workers (permanent and seasonal). 59 * While non-traditional export crops could offer potential opportunities for growth and employment, their scope has been limited to date, with little involvement of the poor. * Rather, non-agricultural employment could provide an important route out of poverty, though a variety of barriers constrain access to these opportunities, particularly education and geographic location. These findings suggest a number of priorities for reducing poverty and promoting opportunities: * Enhancing opportunities should be at the center of the poverty agenda. A recurring pattem that arises in the analysis is the fact that the poor and other marginal populations such as the indigenous are not able to fully participate in, or benefit from, the overall economic system. Therefore, addressing the improvement of employment opportunities is necessary. Some specific areas for policy intervention that emerge from the analysis are: o Reducing the human capital gaps between the poor and non-poor. Education is essential to expand the opportunities of the poor and allow them to access higher-paying jobs. Education and technical training are particularly important to help them access expected growth opportunities in non-agricultural sectors. Investments need to concentrate on expanding the poor's access to education, as well as improving quality so as to boost the returns to education (see Chapter 7). o Lowering transactions costs in accessing markets. By decreasing the strong spatial disadvantage that many of poor face (especially in the rural areas), marginal populations could see dramatic increase in opportunities via the easier access to product and factor markets, both in agricultural and non-agricultural sectors. Thus emphasis on investments in road infrastructure and basic services coverage is essential (see Chapters 9 and 10). o Creating mechanisms to discourage labor-market discrimination for the indigenous and women. * A rural development strategy is also key for Guatemala's overall poverty reduction strategy. As poverty is highly concentrated in rural areas special attention in rural employment and income generation is important. In particular, the two main areas for consideration include: o Promoting growth of non-agricultural sectors, which are likely to be the main engines of rural growth and employment. Despite the potential of the non-farm sector as a vehicle for reducing poverty, numerous barriers prevent the poor from accessing such opportunities, including disparities in education levels, transport and basic infrastructure, lack of access to rural credit, and geographic disadvantages. Interventions should thus focus on removing such barriers, with targeted investments in education and technical training, policies to promote micro-, small- and medium-scale enterprises (MSMEs) which tend to generate employment, and investments in basic services and transport. o Increasing agricultural productivity and diversification. While agriculture is unlikely to generate enough additional employment opportunities to reduce poverty on a large scale in the medium term, it will continue to be an important source of incomes for the poor (at least in the short run). As such, efforts should be made to increase land and labor productivity and to diversify to non-traditional crops. Coffee production should also take greater advantage of markets for specialty coffees. This requires investments in human and physical capital, as well as access to new technologies, financial institutions, and technical assistance. 60 * Finally, safety nets and risk management are important components of a poverty strategy. Even if a successful poverty reduction program is established, certain groups could remain vulnerable (as discussed in Chapter 11), including: o Seasonal migrants who are already affected by the adverse structural terms-of-trade shock in the coffee industry. While they face the possibility for job loss, the welfare of their families is also affected, as they depend on the remittances that they send. A strengthening and expansion of workfare programs may be needed to help supplement their incomes; o Child-laborers, who face considerable trade-offs in terms of under-investments in education. Exploring mechanisms and incentives to motivate tem to stay in school is crucial since these children represent an important future labor force base; o The indigenous, who seem to have limited access to non-agricultural opportunities and face wage discrimination; and o Geographically isolated households, that may not be able to take advantage of economic growth or employment creation programs due to the high transaction costs and barriers of participating in such programs. ' Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. 2 See Technical Paper 1, Vakis (2002). 3Von Hoegen (2000). 4Von Hoegen (2000). 5 Figures 6.1 and 6.2 use only those employed in the informal sector older than age 15. "Other" includes mining, basic services, construction, transport and financial jobs. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. 6 Ministerio de Salud (1998). 7Estimating the magnitude of migration is a complex matter. The ENCOVI 2000 estimates the number of seasonal migrant workers to around 400,000 (excluding children under 7). Other sources such as Ministerio de Salud (1998) put this figure closer to 800,000. Such differences may arise from the sample design but also from the survey definition of migration. The QPES and other qualitative work reveal that migration is an important coping mechanism in times of crisis as well as a regular source of earnings for the poor. Other motivations for migration include the search for better social services such as education and health). 9 ASIES (2000). 10 World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica -Guatemala. " However, 26% of formal-sector workers also receive less than the minimum wage. '2 The ENCOVI survey does not distinguish between those who are registered in the Govemment's minimum wage system versus those who are not registered. 13 Wages have been adjusted for inflation using the CPI. 14 Indeed, the two sharp changes in monthly wages occurred in 1997 in the basic services and transport sectors, and may be linked to the privatization process of electricity and telecommunications industries that occurred that year (see Chapter 9). 5 Taking into account all labor earnings, whether cash or in-kind, and including gross wages/salaries, the "13h month" salary bonus, the value of tips, etc. Wage data have been adjusted for spatial differences using price indices. See Technical Paper I, Vakis (2002) for details. 'Based on a Oaxaca-Blinder decomposition. 1' The World Bank (1999). la Source: World Development Indicators 2001, The World Bank. 19 It is important to note that, because the ENCOVI is a household survey, plantations owned by entities other than households (e.g., corporations, banks) are not captured in the data collected in the survey. The estimates presented here are household-based; conclusions should not be drawn about the poverty status of all coffee producers (e.g., corporations, banks, etc.). 20 AGEXPRONT (2000). 21 Still, due to the survey design, these findings are only suggestive and are not representative. 22 Traditional exports crops are coffee, sugar, bananas and cardamom. Non-traditional exports are snow peas, sprouts, broccoli, cauliflower, flowers, mangos, melons, pineapple, papaya, okra and berries. Non-tradable crops are the remaining crops (e.g. corn). Classification based on those of the Asociaci6n Gremial de Exportadores de Productos No Tradicionales (AGEXPRONT). 23 For a recent survey see Lanjouw and Lanjouw, 2001. 24 See Technical Paper 7, Foster and Araujo (2002). 61 Chapter 7: Education and Poverty "Poverty is the lack of education in the community... Many are still illiterate, they can't communicate with everyone else. " K'itche Villager, KIl (QPES) Improving education and educational outcomes is central to both the Peace Agenda and the poverty agenda. While previous chapters review the importance of education in determining poverty and inequality, and the historical and institutional situation pertaining to human capital development, this chapter seeks to analyze more deeply the issues pertaining to the current education system, with a view of informing policy and highlighting priorities for poverty reduction. After a brief sectoral overview, the chapter considers the issues of: (a) educational stock (literacy and attainment), coverage and equity issues; (b) internal efficiency; (c) disparities in the quality of education; (d) barriers to increased enrollment and attainment; (e) the returns to education; and (f) public spending and equity. It concludes with a review of progress, key issues and priorities in education for poverty reduction. The chapter draws on several types of research, including (a) the results of the ENCOVI 2000, which allow for an examination of education issues from a poverty perspective; (b) the results of the QPES, which included an extensive module on the perceptions of education in the ten rural study villages; and (c) an updated sector review, which provides institutional and public spending information.' SECTORAL OVERVIEW The education sector has undergone important reforms since the signing of the Peace Accords in 1996. These accords establish education as a means to transmit and develop values and knowledge within a multi- lingual and multi-cultural society, thus the curricula must integrate equally the diverse cultures and languages of the country. They also assert that education and training are key factors to achieve equity, national unity, economic modernization and international competitiveness. As discussed in Chapter 4 and below, various targets were set for increasing public spending, and improving literacy and coverage. The Ministry of Education (MINEDUC), which is responsible for regulating, directing, planning, supervising, and evaluating the sector, has been restructured in order to deconcentrate, decentralize, and simplify education administration to promote efficiency and effectiveness. Financial management was deconcentrated under the SIAF program, and since 1998, all central administrative units and Departmental Directorates receive budgetary allocations and are accountable for their use. Important components of decentralization efforts include support to the expansion of the PRONADE program (see Box 7.1), where legally organized communities receive direct transfers of funds to manage the schools; the creation of School Boards (Juntas Escolares) in most schools, that also receive a direct transfer of funds annually for school maintenance; and the creation of Teacher Selection Committees (Jurados de Oposici65n) at the municipal and departmental levels, though these latter were abolished in 2000. The split between public and private provision varies significantly by level of education. Pre-primary and initial education are not required in Guatemala, though a number of programs have developed over the past decade to expand their coverage.2 Data from the ENCOVI suggest that over three quarters of pre- primary students attend public schools. Primary education is compulsory for 7-12 year olds, although many do not enroll and there is a high percentage of over-age students in the primary system (as discussed below). According to the ENCOVI, 88% of all primary students attend public schools, with 79% provided by MINEDUC and 9% under the PRONADE program (see Box 7.1). Secondary education is split into two cycles: basic secondary and diversified secondary.3 Basic secondary provides three years of education (grades 7-9) to those who have completed the sixth grade of primary school. It is expected to provide the academic and technical skills necessary to join the labor force to those who do not further pursue their studies. Diversified secondary provides between 2-4 years (normally, grades 10-12) of schooling for those who have completed grade 9 of basic secondary and consists of four learning tracks: general education, 62 teacher education, commercial education, and technical education. According to the ENCOVI, only 40% of all secondary students are enrolled in public schools, with the private sector playing a more major role in provision at this level. Public sector schools include Ministry Schools (30% of total secondary enrollment) and Cooperative Schools (10% of total secondary enrollment), which operate primarily in urban areas and involve a tri-partite financing arrangement between MINEDUC, Box 7.1 - The PRONADE Program municipal authorities, and legally Organization, Decentralization, and Participation. PRONADE (the organized parents' associations. Programa Nacional de Autogesti6n para el Desarollo Educativo) is a University education is provided decentralized, community-led program that seeks to increase access and by one public sector university, San improve the quality of primary education, especially in rural, indigenous. Carlos (USAC) and nine private and remote areas. Under PRONADE, rural communities with no access to sector universities. Data from the education services receive direct financing from MINEDUC. To qualify, ENCOVI suggest that just under communities must meet at least four criteria: (a) the community must find a 40% of all university students are site and demonstrate ability and interest in managing the new school; (b) the enrolled at the USAC. community must be located at least 3 km. from the nearest public school; (c) the community must have at least 20 pre-primary and primary aged EDUCATIONAL STOCK, children; and (d) the community must not already have any teachers on the official government's payroll. Communities receive financing directly from COVERAGE AND EQuITy the Government to cover teacher salaries, learning materials, and school snacks. Specialized NGOs are contracted by PRONADE to cover Guatemalan literacy is not only administrative training. Financing is fully contingent on extensive below average in Latin America, community participation in all aspects, ranging from hiring teachers to it is far lower. With an illiteracy setting the local school calendar. Each community is represented by a rate of 31% in 2000, only Nicaragua school committee (COEDUCA), which is elected locally and comprised of and Haiti rank worse. Female parents and community members. illiteracy is particularly high (39%), Outcomes and Impact. Various program evaluations4 suggest impressive especially among indigenous results for PRONADE (compared to other public primary schools), women (62%). Progress in teaching including: longer time spent in the classroom, higher attendance, higher and women to read and write lags about more informed community participation, and higher grade promotion rates 20 years behind male literacy, and student retention. Moreover, as discussed below, the ENCOVI 2000 Illiteracy is higher among the poor shows that the program is extremely well targeted to the poor. (46%) than the non-poor (17%) and in rural areas (42%). Regional and ethnic patterns largely mimic those for poverty (Figures 7.1, 7.2). Despite this poor performance, Guatemala has seen improvements over time, with a slight quickening of the pace since the signing of the Peace Accords in 1996 (Figure 7.3). Although educational attainment is quite low in Guatemala, with significant gender, ethnic and poverty gaps, there has been some progress over time. Guatemala is still a "primary" country, with an average educational stock of 4.3 years (for those aged 14+). Educational attainment is higher among the non-indigenous than the indigenous, with an average "ethnic" gap of three years. It is lower for women than men, with an average "gender" gap of one year. Indigenous women in particular have completed very few years of schooling (1.8). Yet the largest gap is observed between the poor and the non-poor, with an average "education-poverty" gap of four years. Overall, Guatemala has made significant progress over time, with today's 19-25 year olds having completed an average of 5.6 years as compared with those over 40 years old having completed only 2.9 years (Figure 7.4). Moreover, the "gender," "ethnic" and "poverty" gaps do appear to be narrowing across generations (Figure 7.4). 63 Figure 7.1 - Poverty, Illiteracy, and Net Primary Enrollment Deficit by Region (source: ENCOVI Figure 7.2 - Poverty, Illiteracy, and Net Primary 2000) Enrollment Deficit by Ethnicity (source: ENCOVI 100 2000) 80 J 100 - _illiteracy 601 - 40 -U-illiteracy 80 1 20 - j60 - -~net enroll 0 +net enroll 40 - e deficit Figure 7.3 - Improvernents In LIteracy, FIgure 7.4 - Educational Attainment (stock): Primary Enrollment over Time Improvements and Closing of Gender, Poverty, sources: Ganuza et. aL. (1999); ENCOVI 2000 -Ethnic Gaps Over Time 90 . °-6.00 ENCOVI 2000 + All Gender Gap 70 -4--litracy .~ 4.0 ~ -a--povertyGa 60 0 2 .00 1---- u-- -0-Avg total years _ a~~~~~~~~~~~~~~~~~~~~~~zC 40~~~~~~~~~erlmn > 0.00 0 c) O ^ o ~~~~~~~~Age Age Age Age 9~~ 9V 9~~~~ 9~~~ 69~~~ , ~40+ 26-40 19-25 14-18 Guatemala has made a sincere effort to improve educational coverage, particularly since the Peace Accords. Using primary enrollment as a yardstick, Guatemala has taken large strides towards universal access. In the early 1970s, primary schools enrolled just over half5 of the target population. Net enrollment rates increased dramatically in only one generation, to 79% in 2000.6 Importantly, progress has been significantly faster in the years since the signing of the Peace Accords in 1996 (Figure 7.3), reflecting a large increase in public spending on education and ambitious programs to expand coverage.7 About half of the increase in coverage was achieved through the PRONADE program, with a targeted expansion in rural areas (Box 7.1). Secondary enrollment also rose sharply in that period (by 39%), with rural enrollment increasing as a share of total enrollment (from 12%-19% of the total from 1996-2000). There has also been a significant push to increase girls enrollment, particularly that of indigenous girls. With USAID and private- sector support, for example, the Ministry of Education initiated a pilot program Eduque a la Nina in 1993 to promote the enrollment of indigenous girls via scholarships, educational materials and community promoters. An evaluation of the pilot shows that, compared with a control group of students that were not part of the program, the scholarship program was effective in promoting attendance, retention, and completion of female students in school and that female teachers constitute an incentive for active 8~~~~~~~~~~~~~~~ participation of female students in the classroom.n Nonetheless, important gaps in coverage remain, particularly among girls, the indigenous, rural children, and the poor. Pre-primary coverage is low overall, particularly among rural, indigenous, and poor children (Table 7.1). At the primary level, key gaps incla)sgirls overall, particularly indigenous girls, 64 one third of whom are not enrolled; (b) the indigenous, particularly girls --and the Q'eqchi; (c) rural children, a quarter of whom are not enrolled; and (d) poor and extreme poor children (Table 7.1). Ethnic and regional enrollment gaps largely mimic patterns observed for poverty (Figures 7.1 and 7.2). Secondary coverage is quite low overall, with only a quarter of the target population enrolled. Secondary enrollment is even lower among the indigenous, rural youths, and the poor and extreme poor (Table 7.1). Table 7.1 - Net Enrollment Rates, by Level and Group % of target aged children that are enrolled in school Pre-Primary Primar Secondary All Male Female All Male Female All Male Female Total 23% 22% 25% 79% 81% 76% 25% 26% 24% Non-Indigenous 27% 27% 28% 84% 71% 86% 32% 32% 33% Indigenous 18% 16% 20% 75% 82% 67% 14% 18% 11% K'iche 21% 18% 25% 71% 78% 64% 17% 23% 12% Q'eqchi 15% 15% 14% 59% 65% 52% 6% 9% 3% Kaqchiqel 23% 17% 30% 77% 78% 77% 23% 28% 17% Mam 15% 15% 15% 75% 78% 71% 9% 9% 10% Other 16% 16% 16% 71% 75% 67% 13% 17% 10% Urban 35% 32% 38% 85% 88% 82% 46% 48% 44% Rural 17% 17% 18% 75% 78% 72% 12% 14% 10% Non-Poor 39% 35% 43% 90% 90% 89% 44% 44% 45% All Poor 17% 18% 17% 78% 81% 75% 13% 16% 10% Extreme Poor 13% 13% 12% 58% 65% 53% 3% 3% 2% Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadfstica - 2000. INTERNAL EFFICIENCY Even when children do enroll in school, they face a number of hurdles as they attempt to advance through the educational system. Symptoms of these obstacles include: (a) delayed initial enrollment; (b) grade repetition; (c) delayed advancement (late grade-for-age, which reflects both delayed initial entry and repetition); and (d) drop out. Each suggests institutional problems inherent in the education system. First, a significant share of Guatemalan children delay initial enrollment in primary school, particularly poor, rural, and indigenous children. Enrollment is delayed in Guatemala both by policy and in practice. In terms of policy, the official age of entry for primary school is age seven - a full year later than in most countries. This means that officially, 12-year olds are still in primary school and 13-19 year olds would still be in Figure 7.5 - Late Initial Entry In Primary School secondary school. In practice, a substantial % of students aged 7-12 enrolling In first grade at age share of Guatemalan children initially enroll 8+ even later than the official entry age. In fact, ENCOVI 2000 the ENCOVI shows that about a fifth of all 35% - 26% 29% primary-aged students enrolled in first grade 25% 20%0i at least one year later than the official entry 2 * 9% 14% age.9 Late initial enrollment is particularly 15_ 1i common among rural, indigenous, and poor 0% - _E students (Figure 7.5). ,o Late enrollment is costly. First, it can lead to truncated schooling as older children face greater opportunity costs on their time and pressures to work. The ENCOVI shows this to be the case for non-indigenous children, for whom late enrollment translates into less total years of schooling upon withdrawal. In contrast, indigenous children 65 who are enrolled late are also more likely to be withdrawn at a later age. As such, for indigenous children, late enrollment means delayed overall schooling, but not necessarily less schooling. Second, late initial enrollment is also associated with overall lifetime earnings. For example, calculations using the ENCOVI suggest that a high-school graduate with a two-year enrollment delay earns 15% less over his lifetime than one that started at the official entry age. Second, despite progress, a significant share of primary-school children are held back from grade advancement, suggesting deficiencies in the educational system. Data from the ENCOVI suggest that the overall repetition rate for primary school is 12.8%.'° Official statistics on repetition are slightly higher, but do show some progress since the signing of the Peace Accords in 1996.11 Nonetheless, repetition rates in the first two grades are so high (Figure 7.6), that they signal serious deficiencies in the educational system. There are several possible factors that commonly explain this sort of repetition pattern: (a) inherent characteristics - such as Spanish language barriers, malnutrition, and a lack of academic support in the home -- make it difficult or impossible for many students to perform adequately; (b) the curriculum may be unrealistically demanding; (c) school quality and teacher "seriousness" may be judged by "high standards" that they demonstrate by holding children back; and (d) schools are thought to be "too full," so lower grades are used as filters. The fact that repetition rates fall in the later grades should not be understood to mean that the situation eventually improves in the sense that students finally learn and repeat less. Rather, this is best seen as indirect evidence that many more repeaters drop out of school before reaching those grades (discussed below). Repetition rates are significantly higher, for poor children (Figure 7.6) than non-poor children. Almost a quarter of poor students a made to repeat the first grade. The poverty-repetition linkage could reflect inherent characteristics (lack of academic support in the home, malnutrition), school quality issues, or overcrowding in schools. Interestingly, repetition rates do not differ significantly by gender or ethnicity at the primary level. Likewise, repetition patterns at the secondary level signal problems with transitions between levels and cycles. The institutional split between cycles of secondary education (see above) is clearly betrayed in repetition rates across grades, which peak at the beginning of each cycle (in grade 7 and again in grade 10, see Figure 7.7). Likewise, while repetition rates are higher overall for the poor than the non-poor (Figure 7.7), they are even higher for the poor at grade 10 (22% of poor tenth graders compared with only 1% of non-poor tenth graders), signaling that poor children face difficulties in transitioning from the basic cycle to diversified secondary. These discontinuities in repetition rates across grades suggest that students face changing standards, which create bottlenecks at the beginning of each educational cycle. The standards between levels are not harmnonized: what for many students constituted a passing level of effort in grade 6 is judged insufficient in grade 7; what was acceptable in grade 9 is suddenly below par in grade 10. Again, several explanations can account for this pattern. There may be successively fewer schools serving higher grade levels, and those schools that do offer higher grade levels tend to be of better quality.'2 In this scenario, students that pass at smaller ("worse") schools find they are unprepared for the standards of the bigger ("better") schools that offer higher grades. Another possibility is that the bottom grade at each level acts as a filter in an overcrowded system. Under this view, for instance, the top three secondary school grades serve as a filter for students that are university-bound, where lower quality students are intentionally "flushed out" during the basic education cycle. In any case, the observed discontinuity between cycles is a "red flag" for quality differentials and should be explored more fully. Third, as a result of late initial enrollment and repetition, a significant. share of students are "over- aged" for their grades. Both official statistics and ENCOVI data suggest that late "grade-for-age" - which reflect cumulative schooling delays from all sources (late initial entry, grade repetition, and temporary withdrawals) - is a significant problem in Guatemala. The ENCOVI shows that 19% of all 15-year-olds are still in primary school. Official statistics also point to such a problem: some 11-17% of basic secondary education students were apparently over-aged in 2000, though this figure drops to 2.6% for diversified secondary (purportedly due to the "filtering" forces described above). Multi-variate regressions13 analyzing 66 the determinants of grade-for-age suggest important links to poverty. Higher household per capita consumption (an indicator of current well-being) and the child's height-for-age (a proxy for past well-being and malnutrition) are both associated with more rapid progress through the school system. Given all other characteristics, girls tend to advance faster than boys. There is also considerable inter-ethnic variation. The non-indigenous, K'iche, Mam and "other indigenous" children all advance through the grades at similar paces. In contrast, Q'eqchi children are about a third of a year behind similar non-indigenous (taking into account other factors), while the average Kaqchiqel student is about one-tenth of a year ahead. Interestingly, grade repetition has about a one-to-one impact on late grade-for-age (repeating one grade translates into almost exactly a one-year delay), whereas the trade-off with late enrollment is less than one-to-one (children whose initial enrollment was delayed by two years are on average less than two years behind those who enrolled at age 7). Finally, while drop out rates have improved over time, they are still high and a significant share of poor and extreme poor children drop out of the school system before they reach grade 6. Official statistics suggest that drop-out rates have improved somewhat over time, declining from 8.2% in 1996 to 7.0% in 2000 for primary school, and from 4.7% to 3.6% for secondary over the same time period. While it is difficult to calculate drop-out rates directly using the ENCOVI, the data do suggest that they are particularly high for the poor. In the QPES, the main reasons that children drop out were seasonal migration and child labor, domestic responsibilities, and poverty and the costs of attending school, according to parents and teachers in the study villages. Figure7.6 - Primary School Repetition Rates by Figure 7.7 - Secondary School Repetition Grade, Poverty Group, ENCOVI 2000 Rates by Grade, Poverty Group, ENCOVI c _________________________________________2000 2 25 - . o 20 - 14 15 4 e~~~r1 9 °~X ,n n ,11 1, 8 4 3 3, 2,,l ,n ,l b~~~~~~~~~~~~~~~ lb lb NIZ -, Ne 1Os &O o Grade Level, Poverty Group Grade Level, Poverty Group DISPARITIES IN THE QUALY OF EDUCATION The quality of education received by poor children lags that received by the non-poor. There are two ways to assess educational quality: the quality of learning (via standard achievement tests) and the quality of inputs (materials, books, teachers, etc.). Results of student testing under the National Program of Evaluation of School Achievement (PRONERE) reveal significant inequities in the quality of the primary education system.14 While student achievement in reading and math is generally low, it is even lower among rural students, boys, and indigenous students. The evaluations also show that urban schools have more qualified and better trained teachers, better infrastructure and more books. The ENCOVI 2000 and QPES offer several additional signs that the poor lack access to key quality-related inputs: 67 * Lower household spending on books and school supplies among the poor and indigenous suggests lower use of educational materials in their learning process (Table 7.2 below). There is substantial variation in absolute costs by poverty group, ethnicity and geographic area, with much higher spending overall by the non-poor, non-indigenous and urban residents - even in public schools. These differences arise primarily from spending on materials and books. Even in public schools, the non-poor spend close to three times more than the poor on these items. Spending on informal fees (fund-raisers, donations) - which contribute to educational and extra-curricular activities - is also higher among the more privileged groups (even in public schools). * A larger share of the poor and indigenous report not having textbooks, which tend to be highly correlated with other indicators of school quality. While only 5% of all primary students and 10% of all secondary students report not having any books, a higher share of the indigenous report not having them. Moreover, even when they report having books, the poor and indigenous spend significantly less on books and materials than the non-poor and Box 7.2 - Schooling in Guatemala: Children's Perspectives non-indigenous, suggesting that these latter groups have access The QPES interviewed children on their perceptions of their schools. In terms these latter groupshaveaccess of their teachers, children in all but one village (LI) perceive their teacher as to more (and possibly better or an important person in their lives who teaches them, gives them snacks (la more up-to-date) books and refacci6n), and serves as a second parent. In LI, the children voiced opposing materials than the opinions, with some saying the teacher was good because he teaches them, and disadvantaged groups. others saying he was "bad" because he hits them. The children in all ten villages universally identified (a) learning to read and write; (b) learn Spanish * The poor and indigenous (in the indigenous villages); (c) drawing; and (d) singing as their favorite children are more likely to aspects of school. Their least favorite aspects included (a) fighting between classmates; (b) that they aren't given time to play; and (c) that the schools are belong to communities dirty. Physical punishment was identified as common practice in 9 of the 10 reporting an insufficient number schools (the exception being in KA2 where the teacher takes away "points" as of teachers.'5 The QPES the main form of discipline). In terms of language of instruction, the children villages also report high in the 8 Mayan villages expressed a range of views, with some preferring student-teacher ratios in multi- instruction in Spanish, others preferring bilingual teaching, and others favoring teaching in their maternal indigenous languages. Children in 7/10 of the grade classrooms, ranging from villages see e'ducation as offering them a better future. Those who didn't 31-52 students per teacher in the make this linkage belong to the communities of QE2 and KAI (both finca indigenous villages (average of villages), and M2 (also quite poor and highly dependent on seasonal migration 38) and 25 in both Ladino to thefincas for income). communities. Indeed, vacancies and teacher turnover in rural areas are high; this could be associated with the lack of incentives for rural teachers, who face difficult working conditions but do not receive differentiated salaries than their urban counterparts.'6 * Poor and indigenous children are more likely to belong to communities reporting inadequate school facilities (infrastructure, classrooms, desks, etc.).'7 Informants in all ten QPES villages also report problems with the physical infrastructure of their schools, including: poor quality and maintenance of the school itself; lack of classrooms; lack of furniture, materials; and a lack of basic services (e.g., water and latrines). Other problems identified in'the QPES include: problems with school meals and meal preparation (e.g., lack of a proper kitchen to prepare school meals), lack of parental support for education (identified by the teachers), and.irregular pay of.teachers. * The returns to education are fairly low in Guatemala (see below), particularly at the primary level, which could signal deficiencies in quality. 68 o While Danguage barriers may prevent higher enrollment and advancement - particularly in the early years of primary - potential language and cultural losses also need to be taken into account in the national curriculum in order to preserve Guatemala's rich cultural heritage (Box 7.3). Box 73 - Language, Culture and Education One of Guatemala's great assets is its rich cultural and linguistic diversity. Indeed, a cross-cutting theme in the Peace Accords is the preservation and promotion of Guatemala as a multi-cultural, multi-lingual society. Unfortunately, there are signs of erosion of this cultural asset, with formal education playing somewhat of a detrimental role. First, fluency in indigenous languages is being lost over time. As discussed in Chapter 4, the ENCOVI 2000 reveals that a smaller share of subsequent generations of ethnically-indigenous people speak indigenous languages (77% for those aged 7-13 compared with 89% for those aged 40+). Moreover, the loss is particularly strong with inter-ethnic marriage: while 90% of those between aged 7-25 who boast "full Mayan linguistic lineage" (4 Mayan-speaking grandparents and 2 Mayan-speaking parents) speak indigenous languages, only 41% speak the language if only 3 grandparents are Mayan speakers (even if both parents are), and only 18% speak indigenous languages if only one parent does (even if all 4 grandparents do). Indeed, the language abilities of parents and grandparents were significant and positive determinants of mayan-speaking abilities in multi-variate regressions.'8 Second, it appears that education is playing an adverse role in the loss of indigenous languages. Controlling for other factors (including the language abilities of parents and grandparents), there is a 1.8% decline in the probability of native fluency in an indigenous language for each year of schooling completed by the mother and a further 2.4% decline for each year of schooling completed by the child.19 ThIird, some Mayan communities perceive a conflict between traditional oral teachings and modern education. All Mayan communities in the QPES note substantial differences between traditional oral teachings (at home) and modem education (at school). According to elders and parents in the study communities, the main roles of traditional teachings are to teach values, respect (respeto, saludar), culture and customs. Most also note that children learn gender roles and skills at home (with boys getting apprenticeships in productive skills and roles and girls learning reproductive, domestic roles). Elders in all ten study communities perceive that these norms are being lost over time. Regarding formal education, most QPES informants perceive that education plays an important economic role (creating opportunities, overcoming poverty), and many point to the importance of acquiring "knowledge" (conocimiento) and to the role formal education plays in helping them overcome ethnic barriers and exclusion so that they may reach the "ladino world" of opportunities (e.g., by learning Spanish at school). Nonetheless, these very benefits of formal education apparently come with a cost. Five of the eight Mayan communities in the QPES perceive a conflict between formal education and traditional oral teachings. This conflict is usually described as the school children losing respect for traditions and elders, becoming "lazy" at home, or having to do homework instead of traditional (gender-based) duties. The Government has recently undertaken a number of initiatives to preserve and promote Guatemala's cultural heritage in the education system. These include promoting bilingual teaching methodologies under the DIGEBI program, and a new program to prepare a National Cultural Development Plan, deconcentrate the cultural education services of the Ministry of Culture and Sports (MCS), and develop a National Cultural Resources Information System (NCRIS), with the support of the World Bank under the Universalization of Basic Education Project. Nonetheless, there has been an important push to improve education quality in Guatemala, particularly since the Peace Accords. The Ministry of Education, with bilateral and multi-lateral financing, has implemented various education models to address quality issues in primary education, including: (a) providing teacher training in multi-grade teaching methodologies (Nueva Escuela Unitaria, NEU schools) piloted in several departments; (b) promoting bilingual schools and methodologies in departments with large indigenous populations (Directorate General of Bilingual, Intercultural Education, DIGEBI schools); (c) providing training in bilingual multi-grade methodologies (DIGEBI-NEU combinations) piloted in the Department of Alta Verapaz; and (d) developing the "schools for excellence" piloted in various schools throughout the country. Results of various program evaluations for these pilots are generally positive.20 BARRI1ERS TO ENROLLMENT AND ATTAINMENT Many factors contribute to enrollment decisions, including supply-side factors (availability and proximity of schools, quality of schooling), and a range of inter-related demand-side factors, such as the ability to cover direct costs (fees, transport, uniforms, materials, etc.), opportunity costs (earnings from child labor, domestic 69 responsibilities), migration (seasonal or permanent location changes), expected returns, parental education history, household composition, and attitudes and beliefs. This section explores potential barriers to enrollment and attainment, with a view to informing policy as to possible priorities. Primary School: Supply- vs. Demand-Side Constraints Demand-side factors - rather than a lack of schools - appear to be the main obstacles to increased primary enrollment. Numerous indicators reveal that a lack of schools is not the main constraint to increased enrollment in Guatemala. First, household perceptions point to demand-side factors, rather than a lack of school facilities, as the main barriers to enrollment. Only 7% of non-enrollees at the primary level cited supply-side factors - such as distance to schools, inadequate class space, and lack of facilities - as key obstacles (see Figure 7.9).2' Second, late enrollment and drop-out, for example, both suggest that the children have a school to go to, but they - or their parents - are declining to enroll. Third, gender disparities in enrollment rates strongly reveal such choices, given that about half the population is Figure 7.8 - Supply- vs. Demand-Side Barriers to School male and half is female, and that Enrollment, ENCOVI 2000 boys and girls are randomly distributed among families. ... Mt Overall, 24% of all primary- aged girls are not enrolled, ..... compared with only 19% of i same-aged boys (Table 7.1 above). As such, at least a fifth M -i ...U of all girls who are not enrolled 4%_ .t-&. do have a school to go to and _ - could be enrolled. We know ..*%sL43..s.. this because their brothers are 0%. enrolled: Such biases are even more evident among the indigenous: net non-enrollment rates are 33% for indigenous girls as compared with 18% for indigenous boys. As such, about 45% of all primary-aged indigenous girls who are not enrolled do have access to a school and could be enrolled. Likewise, for the poor, about a quarter of all un-enrolled primary-aged girls could be enrolled in the sense of having access to a school.22 Finally, cluster analysis of the ENCOVI strongly suggests that very few children fail to attend due to a lack of facilities. Since households in the ENCOVI were sampled in clusters (within primary sampling units, PSUs), enrollment in school by any child within a cluster signals the existence of a primary school facility. Using such information, non-enrollees can be classified into: (a) those who do not attend because there are no facilities available (pure supply-side gap); (b) those who fail to enroll even when a school is available (pure demand-side constraints); and (c) those who do not enroll because there is no school available, but still would not enroll if school facilities were provided (mixed supply- and demand-side factors).23 This analysis reveals that only 2% of all non-enrollees lack primary school facilities (and this share is similar in both urban and rural areas, see Figure 7.8). This means that the problem of non- enrollment cannot be solved by building more schools. Rather, policies should seek to ease key demand-side constraints, as discussed below. As discussed above, the quality of schooling should also be addressed as low quality leads to low returns, which in turn lead to low demand for the educational investment. Primary School: Main Demand-Side Constraints Indeed, gender, ethnicity, parents' education, and poverty are all important demand-side determinants of enrollment and attainment. Controlling for other factors,24 girls are less likely to enroll at any age than boys. Likewise, being indigenous reduces the likelihood of enrollment at any given age. 70 Furthermore, if a person has both characteristics, it is even less likely that she will enroll.25 Paternal and maternal education and total consumption per capita all have positive effects on the likelihood children will enroll. As such, education not only plays a key role in determining current poverty (as discussed in Chapter 3), but also plays a crucial role in the inter-generational transmission of poverty. Education not only begets higher household incomes, but also contributes directly to the education - and hence future earnings - of today's children. Even after controlling for per capita consumption, household size and birth order also affect the chances that a child will enroll, with older children (lower birth order) and children in larger households being less likely to enroll. Child malnutrition (height-for-age) is also positively correlated with enrollment and attainment, even after other factors are taken into account. Similar results are observed in multi-variate regressions that examine the determinants of dropping out of school, once enrolled.26 In this case, girls and indigenous children are significantly more likely to drop out than their boys or non-indigenous counterparts. Interestingly, late-enrolled indigenous children (interaction variable) are less likely to drop-out, signifying that late enrollment delays schooling but does not necessarily translate into less schooling. Again, parental education and household size are also important determinants of drop out. Guatemalan households point to economic fuctors as fte maen constraints to incireased prinary enrollment. "Lack of money" was the single most common reason given for not enrolling in primary school, accounting for 38% of non-enrollees (Figure 7.9).27 This pattern did not differ significantly by gender, ethnicity, or area, though it was slightly higher for the poor (39%) than the non-poor (32%). This suggests that direct costs of attending school are prohibitively high, particularly for the poor. Indeed, the costs to households of public schooling are quite high (see below). "Lack of interest" is the second main impediment to primary enrollment, according to Guatemalan households, accounting for 16% of absentees.28 This might reflect the relatively low returns to education at the primary level (see below), or poor quality of education (not interesting to the children). Economic factors clearly do not explain the significant gender biases observed in Guatemala (particularly given higher female rates of return, as discussed below). As such, a combination of economic, cultural and information/outreach solutions might be needed in order to improve girls' opportunities to attend school. Secondairy School: Supply- vs. Demand-Side Constraints A combnautAon of sujpRly- and demand-side factors pirevemt ncireased eunrolment et the secondary Devel. Cluster analysis of ENCOVI data suggests that less than half of all non-enrollees fail to enroll in secondary school due to a lack of school facilities (Figure 7.8). Demand-side factors are even more dominant in deterring enrollment in urban areas. In contrast, demand- and supply-side factors both prevent rural youths from attending secondary school.30 Interestingly, in indicating the main reasons for non- enrollment, Guatemalan households tend to downplay the lack of facilities as a deterrent to increased enrollment. In fact, despite the apparent lack of secondary schools in Guatemala, only 3% cited supply-side factors -- such as distance to schools, inadequate class space, and lack of facilities - as the main reasons for not enrolling (Figure 7.10). Instead, most attribute non-enrollment to demand side factors, such as the direct costs of schooling or competing demands for their time (work, domestic responsibilities). Indeed, economic factors were by far the main reasons reported in the ENCOVI for not enrolling in secondary school (accounting for 69% of all non-enrollees, Figure 7. 10).3' Specifically, "had to work" was cited by about 29% as the main reason, followed by the direct costs of schooling (25%). Domestic responsibilities accounted for about 15% of non-enrollees. While the share citing direct costs did not vary significantly by gender, ethnicity, poverty group, or area, there were significant variations in the split between work and domestic responsibilities by gender. Close to half of all truant boys cited "work" as the main reason for not enrolling, compared with 14% of girls. Close to 28% of truant girls cited domestic responsibilities, as compared with only 1% of boys. "Lack of interest" accounted for just under a fifth of all non-enrollees, and this share did not vary significantly across the demographic groups. 71 Figure 7.9 - Reasons for not Enrolling in Primary School % of Children Aged 7-12 not Enrolled, ENCOVI 2000 AP NP - - Ind. 7 ] | Y /2 iS! TSupply-Side - Total Non-Ind. ILack of money Female _ i _ 233 [:1 Work Female T9%'0 1 Domestc duties Male _ _ _ _11 | ..... * Not interested E AII other Rural_ ?11 Z Urban E _ Total .eYO __ _ 0% 10% 20% 30% 40% 50% 60% 700/o 80% 90% 1 00% Figure 7.10 - Reasons for Not Enrolling In Secondary School % of Children Aged 13-19 not Enrolled, ENCOVI 2000 tw fl1 . - i br. - l ;D . *s.c c_ ..e, h8 -l. 0z AP ,a 1 1 bo NP u Ind.=_ §i ! i l W l2in M Supply-Side - Total Non-Ind. _ 1U3o * Lack of money Female _ t ~ _ | _ ;j OWork Female 143~ 1 281:.6 910-V 0 Domestic duties Male _ ENot interested po.~w El All other Rural OnE_ F 29°o6 I t71 111111111111111er RUra Urban 270 I 10 Total _290 1 15e- 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Secondary School: Main Demand-Side Constraints The direct costs of schooling are indeed high. Guatemalan households allocate an average of 5% of total consumption to education, though this share varies significantly between the poor (3%) and the non-poor (6%). Overall, it costs households an average of Q.650 per student to attend primary school, including fees, uniforms, transport, and books and materials. This cost is lower for Ministry schools and significantly lower 72 for students at PRONADE schools,32 as compared with private schools (Table 7.2). Secondary school is considerably more expensive, averaging Q.2,95 1 per student, or 68% of the annual per capita poverty line.33 Even public secondary school provided by the MINEDUC is five times more expensive than public primary (Table 7.2). Cooperative schools are not much cheaper. This presents a significant barrier to the poor for attending school, as highlighted by Guatemalan households in the ENCOVI (see above). Opportunity costs to schooling are also high, as evidence by the large prevalence of child labor in Guatemala. As discussed in Chapter 6, one in every five children aged 7-14 is employed. Older children are more likely to work, with 29% of children aged 10-14 working as compared with 8% of those aged 7-9. Three quarters of child laborers come from poor households, and about 80% reside in rural areas. About two thirds of child laborers are boys. Importantly, child laborers complete about half the number of years of schooling than their non-working counterparts.34 The QPES also reveals that working conflicts with schooling: child labor was cited as the main cause of absenteeism and drop out in most of the study communities. Work-school conflicts are often seasonal, increasing during harvest periods which vary by crop (for example, coffee harvests run from November to December, sugar in October to April but especially from November to February, and cardamom twice a year). Such seasonal variations throughout the country suggest that a possible decentralization and tailoring of the school calendar to local conditions might be warranted to reduce the likelihood of work-school conflicts. Table 7.2 - The Direct Cost of Schooling Annual Costs to Households Per Student, Primary and Secondary Schools, Quetzales/Student Ministry PRONADE-prim. Private All Schools Schools Cooperative-sec. Schools Other Primary - All 650 341 92 3,434 374 Non-Indigenous 905 405 100 4,012 444 Indigenous 262 245 88 1,111 236 Urban 1,352 577 316 3,884 767 Rural 257 222 76 1,747 225 Non-Poor 1,305 554 306 3,913 658 All Poor 206 210 74 731 223 Primary - By Type of Cost (%) Fees 35% 5% 3% 58% 44% Uniforms 9% 13% 13% 6% 7% Books and Materials 33% 53% 44% 17% 28% Transport 8% 3% 2% 11% 5% Other (fundraisers, donations) 15% 26% 39% 7% 17% Secondary - All 2,951 1,705 1,525 4,017 1,758 Non-Indigenous 3,216 1,618 1,658 4,365 2,296 Indigenous 2,095 2,057 1,343 2,610 1,230 Urban 3,377 1,929 1,546 4,336 2,439 Rural 2,034 1,297 1,504 3,014 1,187 Non-Poor 3,493 2,067 1,762 4,404 2,578 All Poor 1,298 956 1,138 1,845 1,110 Secondary- By Type of Cost (%) Fees 41% 9% 26% 51% 30% Uniforms 6% 8% 8% 5% 2% Books and Materials 28% 45% 39% 22% 34% Transport 10% 16% 5% 9% 17% Other (fundraisers, donations) 15% 23% 22% 13% 18% Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. THE RETURNS TO EDUCATION While the rates of return to primary are education are low, they are higher for secondary school and university, particularly for women. Earnings regressions35 by level of education completed show that the rates of return to education increase in a non-linear function (Table 7.3). For example, a male who has completed primary education is expected to receive 11% more than a male with no education (which 73 translates into an hourly earnings increase of about 2% per year of primary education). In contrast, the returns to secondary education are much higher. A male with secondary education receives 27% more than a male with no education (or 6% more per year of secondary schooling).36 This pattern is similar for females,37 but actual rates of return are higher for women, highlighting the importance of girls' education. The overall low returns to primary education suggest inadequate quality of schooling or lack of opportunities for low- skilled workers. They also corroborate the high share of non-enrollees who indicate that a "lack of interest" is their main reason for not enrolling: the returns perhaps do not justify the costs of the education investment decision. A lack of ability to speak Spanish is Table 7.3 - Returns to Education, by Gender correlated with lower rates of return. Male Female Overall Controlling for other factors, men and Primary completed (%) ' 11 17 15 women who speak Spanish earn more Secondary completed (%) 27 54 51 than 30% more than those who do not.38 University completed (%) 74 76 74 This is also true for bilingual speakers, Overall (%) 2 3 6 6 highlighting the importance of language Returs to education is the % increase in hourly wages due to educational level increases 'compared with no education 2 per year of education Income inequalities emerge from both differences in educational attaimnent and the returns to education. As discussed in Chapter 3, differences in education account for over half of total inequality in Guatemala. Further analysis suggests that inter-generational inequality is attributable to a combination of (a) inequality in educational attainment (years of schooling), which has been equalizing over time due to the expansion of educational coverage; and (b) inequalities in the returns to schooling, which could signal disparities in the quality of education.39 Inequalities seem to emerge in particular at the secondary level: a very small share of the poor attend secondary school and the returns to this level far exceed those at the primary level. Additional years of schooling are needed to compensate for the low returns to primary education as a means to escape poverty. An "education-poverty line" can be calculated taking into account household size and differences in the level of returns by gender and ethnicity.40 Such calculations reveal that for a five- person household, a male ladino household head needs 12.5 years education to earn enough to pull his family out of poverty (more than a secondary education), assuming he were the only income-earner in the household. For a similar household, if the male were indigenous, he would need to have completed 17.6 years of schooling (almost a university education) to earn enough to escape poverty. Ladina women would need 15.1 years, as compared with Indigenous women who would have had to complete 23.4 years to pull their households out of poverty. The ethnic and gender differences reflect disparities in the actual levels of returns to schooling (not the rates)." PUBLIC SPENDING AND EQUITY Level and Composition of Public Spending on Education Guatemala has made substantial progress in increasing public spending on education since the Peace Accords. Although still quite low relative to other LAC or lower-middle income countries,42 public spending on education has expanded by an average of about 20% per year in real terms since 1996, rising faster than GDP and total government expenditures (Table 7.4). Primary schooling absorbs the bulk of public expenditures on education, which is consistent with equity and poverty concerns. Primary schooling receives close to half of public spending on education in Guatemala (Table 7.4). Given existing gaps in coverage and the fact that the majority of the poor still do not complete primary school (discussed above), this allocation seems appropriate. Moreover, public spending on primary schooling is slightly progressive, largely due to the impressive targeting of the PRONADE program 74 (see below). Spending on secondary education is quite low (7% of total public education spending in 2000), much lower than spending on tertiary education (USAC), which is highly regressive (see below). The spending category that has grown the most since the Peace Accords is "other," which includes a range of adult education and special programs. Table 7.4 - Public Expenditures on Education 1996 1997 1998 1999 2000 Average Education Expenditures (mn Q.) a 1,505 1,900 2,617 3,285 3,629 n.a. As % of GDP 1.6% 1.8% 2.1% 2.4% 2.5% 2.1% As % of Total Gov't Exp. 15.2% 15.1% 15.7% 17.1% 18.3% 16.1% Education Expenditures (mn Q, base year=2000)b 1,978 2,307 3,062 3,642 4,054 | n.a. Annual % change n.a. 16.6% 32.7% 18.9% 11.3% 19.9% By Level (% of total)b Pre-Primary 8.0% 9.8% 8.7% 6.0% 7.4% 8.0% Primary (non-PRONADE) 49.2% 42.0% 40.0% 44.7% 44.2% 44.0% PRONADE n.a. n.a. 3.3% 3.8% 4.9% 4.0% Basic Secondary 7.8% 4.9% 5.0% 5.3% 5.0% 5.6% Diversified Secondary 4.0% 6.8% 5.3% 3.0% 1.9% 4.2% Tertiary (USAC) 14.0% 11.1% 12.1% 10.0% 9.3% 11.3% Construction' 10.2% 9.3% 9.4%' 9.2% 9.6% 9.5% Otherd 6.8% 16.2% 16.2% 17.9% 17.7% 14.9% By Economic Classification (% of total)b Capital Expenditures 21.0% 22.0% 20.1% 26.6% 19.1% 21.8% Personnel Expenditures 63.0% 61.8% 57.8% 57.7% 64.6% 61.0% Non-Personnel Recurrent 16.0% 16.2% 22.1% 15.7% 16.3% 17.3% Sources: a. SlAF/Ministry of Finance; b. Banguat, Ministry of Finance, and MINEDUC (Planning Unit), as reported in Anderson (October 2001). c. Information not available by level of education, but heavily weighted towards primary schools. d. Includes adult education, literacy training, other training programs, special education programs, and miscellaneous expenditures not classified by level of education. The Distributional Incidence of Public Spending on Education Public spending on pre-primary school is well targeted overall, with an even stronger pro-poor bias for the Figure 7.11 - Incidence of Public Spending on Primary PRONADE program. Over three School: International Comparison quarters of public spending on pre- primary schooling is received by the Guate-PRONADE - l poor (Table 7.5). This reflects the fact lalal that a substantial share of non-poor pre- Panama * _ primary students attend private Jamaica 001 (poorest) institutions. This share received by the *02 poor is even higher for spending on the Colombia _ . IQ3 PRONADE Program. Ecuador -. 004 Ecuado _ Q5 (richest) While public spending on primary Guatemala _ 5 Ic education is distributionally neutral, Nicaragua spending on the PRONADE program - which accounts for a large share of 0o% 20% 40% 60% 80% 100 the recent expansion - is extremely % well targeted. Spending on primary 75 education essentially mirrors the distribution of the population: the poor account for about 56% of the population and receive about 54% of public spending on primary education (Table 7.5). Indeed, public spending on primary school is not as well targeted in Guatemala as in other LAC countries (Figure 7.11), reflecting the significant coverage gaps observed among the poor in Guatemala. In contrast, public spending on PRONADE is extremely well targeted, with only 8% of all benefits going to the non-poor - one of the best targeting records in LAC. The success of PRONADE targeting is important particularly in light of the fact that PRONADE accounted for a substantial share of new net enrollment since the Peace Accords. In contrast, public subsidies for secondary and university schooling are highly regressive. The non- poor receive 67% and 94% of public spending on secondary and university education respectively (Table 7.5). Public spending on Cooperative schools at the secondary level is likewise regressive, with 62% going to the non-poor. In fact, the distsribution of public spending at the secondary and university levels is more regressive than in other countries, reflecting extremely low enrollment by the poor in Guatemala (Figures 7.12 and 7.13). Demand-side interventions - such as school feeding, scholarships and other subsidies - are poorly targeted to the poor. Scholarships in particular disproportionately benefit the non-poor, with close to half of all scholarship benefits accruing to the top quintile of the population (Table 7.5). The top two quintiles are likewise the main beneficiaries of the school transport subsidy, receiving 83% of total subsidy benefits. School feeding and the Bolsa de Utiles (school materials) programs are somewhat better targeted, benefiting primarily the middle quintiles of the population. The targeting of demand-side interventions to the poor is particularly important since such programs represent transfers designed to ease cost-related barriers to enrollment. Overall, public spending on education is neutrally distributed. Reflecting the sum of the above patterns by level, public spending on education in Guatemala essentially follows population patterns, with a slight bias against the poorest quintile (Table 7.5). It is, however, more progressive than the existing distribution of total consumption in the economy (Table 7.5). 76 Table 7.5- Distributional Incidence of Public Spending on Education and Educational Support (Demand-Side) Programs % of total benefits received by each group By Quintile By Poverty Group By Ethnicity By Area Total Ql Q2 Q3 Q4 Q5 XP AP NP Ind Non-Ind Rural Urban Education - Total 100 17 21 21 21 21 13 55 45 37 63 59 41 Pre-Primary 100 39 18 24 14 4 30 78 22 64 36 76 24 o/w Ministry Schools 100 35 19 27 14 5 26 76 24 65 35 74 26 o/w PRONADE 100 59 16 10 15 0 49 85 15 60 40 88 12 Primary 100 21 25 23 21 10 16 65 35 42 58 69 31 o/w Ministry Schools 100 18 24 24 22 11 14 62 38 40 60 67 33 o/w PRONADE 100 44 37 12 6 1 37 92 8 65 35 93 7 Secondary 100 3 12 23 31 32 2 33 67 22 78 36 64 o/w Ministry Schools 100 3 12 22 31 33 2 33 67 21 79 35 65 o/w Cooperatives 100 3 16 27 30 25 2 38 62 42 58 51 49 University 100 0 0 6 11 82 0 6 94 37 88 8 92 Demand-Side Programs School Feeding 100 16 25 27 20 11 12 63 37 43 57 79 21 Scholarships 100 9 4 23 16 48 3 30 70 47 53 28 72 BolsadeUtiles 100 18 24 24 20 13 14 60 40 35 65 69 31 (materials) 100 0 2 15 56 27 0 16 84 8 92 3 97 School Transport Memo for Comparison Share of total population 100 20 20 20 20 20 16 56 44 43 58 61 39 Share of poor population 100 36 36 29 0 0 n.a. n.a. n.a. 58 42 81 19 Share of total 100 5 9 13 20 54 4 24 76 24 76 37 63 consumption I Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Figures reflect differential unit subsidies by level and program as well as the number of students enrolled in public institutions regardless of their age. Quintiles are individual consumption quintiles. Figure 7.12 -IncidenceofPubilc Spending on ScondarySchool: Rgure 7.13 incidence of Pubic Spending on insmrnatlonal Comperwnon Univereity: International Comparison Colombia -- e_,_ Ecuador- Jamaica DO1 (poorest) Al5a.1 Panrarna 0 aoe Nlzcamguarihst p GuateCccp _ T - ~~~~~~~~~~~~~~uamd Guatemala _ -- 0. 1% 0% 60V W% 10% 0% 20% 40% 60% 80% I100% SUMMARY OF KEY ISSUES AND PRIOR1TIES There have been significant improvements in the education sector over time, particularly since the signing of the Peace Accords in 1996. Notably, * The sector has undergone important institutional and structural reforms, including the deconcentration of financial management and the development of the PRONADE program and school boards, which help decentralize key decisions to the local level; * Public spending on education has increased significantly since 1996 and the bulk of such spending goes to the primary level; 77 * Literacy and educational attainment are increasing over time - with important reductions in the disparities between genders, ethnicities, and the poor versus the non-poor; * Coverage has accelerated at all levels since the Peace Accords, particularly the primary level, and about half of the expansion was achieved through the PRONADE program; * The expansion in coverage appears to have been well-targeted, since much was achieved through the PRONADE program which is extremely well targeted to the poor; and. * Official statistics on indicators of internal efficiency (repetition rates, drop out) suggest improvements. Nonetheless, important challenges remain. In particular: * Significant coverage gaps and disparities remain, particularly for the poor, girls, and rural and indigenous children; * Very few poor even make it to the secondary level (few complete primary school) - and inequalities in earnings are largely generated at this level; * Despite progress, indicators of internal efficiency (late enrollment, repetition, drop out) suggest serious structural problems in the educational system; * The low returns to primary education suggest deficiencies in the quality of schooling; * Beyond a physical lack of schools, key barriers include "demand-side" factors such as the direct costs of attending (at both the primary and secondary levels) and opportunity costs (e.g., work and domestic duties which conflict with secondary enrollment) * Public spending on primary school, which absorbs the bulk of public funds, is distributionally neutral, though spending on the PRONADE program is extremely well targeted. Public spending on secondary and university education is highly regressive. * Demand-side programs - such as scholarships and the school transport subsidy - are highly regressive, with most benefits going to the non-poor; other programs - such as school feeding and the bolsa de utiles program - are slightly better targeted, but benefit mainly the middle quintiles of the population. In light of these challenges, a number of policy recommendations seem appropriate. Specifically: * Continued increases in public spending on education are needed to allow for a further expansion of enrollment to cover remaining gaps, particularly at the pre-primary and primary levels; * Expansion of coverage should be promoted particularly for girls (especially indigenous girls) and indigenous children (e.g., by expanding the "Eduque a la ninia" programs); * Expansion should be implemented via the PRONADE program (with quality improvements as necessary) given the benefits of this program in terms of community and parental participation. However, as supply-side gaps are filled, the Government should consider easing eligibility criteria so as to allow poor communities that already have schools to be eligible for the PRONADE-type community-based school-management model. To target this expansion, the poverty map could be used to identify eligible schools and preserve PRONADE's exemplary targeting record; * Efforts should be made to better target public spending and demand-side programs so as to make more equitable and efficient use of existing resources; the poverty map could be used to help target programs and future interventions; * The Government should seriously consider lowering the age structure for primary school, such that the official starting age is 6 rather than 7; this would lower the ending age from 12 to 11, which is 78 important since older children face more competition for their time (opportunity costs via work and domestic duties); o The Government should examine and improve quality, curriculum and performance standards, particularly at grades 1, 7 and 10, so as to improve student retention, transition and grade advancement at those grades; bilingual education (particularly in the early grades) is also important to improve learning, retention, and repetition rates in those transition years; o As part of its efforts to decentralize key decisions to the local and community levels, the Government should consider decentralizing the school calendar so that communities could adjust to local harvest schedules so as to reduce competition between school and work; o Demand-side support programs (such as scholarships) should be promoted and expanded at both the primary and secondary levels, as demand-side factors appear to be key constraints at both levels; however, such programs need to be restructured (consolidated, see Chapter 12) and explicitly targeted to the poor; and o Investments should be made in early childhood development to promote: (a) improved child nutrition at an early age, since nutritional status is a significant factor in determining enrollment and attainment and since nutritional deficiencies emerge at a young age (see Chapter 8); and (b) early educational opportunities, including links between traditional schooling and pre-primary schooling. 'Much of the analysis of the ENCOVI is presented in GUAPA Technical Paper 3, Edwards (2002). Additional analysis of the ENCOVI (returns to education and supply- and demand-side constraints) was conducted by Renos Vakis and is contained in Annex 6 and GUAPA Technical Paper 1. The education module of the QPES was designed and analyzed by Rodriguez (2001). see GUAPA Technical Paper 4. The updated sector review was conducted for the World Bank by Anderson (2001), see GUAPA Technical Paper 2. 2These include (a) the Proyecto de Atenci6n Integral al Nifio (PAN), supported by MINEDUC, which has expanded rapidly since the Peace Accords and provides initial education for 0-5 year olds; (b) monolingual and bilingual pre-primary schooling, which is directed to 4-8 year old children and has likewise expanded rapidly since 1996; and (c) CENACEP, an accelerated pre-school program that provides 35 days of basic skills to children over 6.5 years old before they enter first grade. See Anderson (October 2001) for more details on these programs and their coverage. 3 See Anderson (October 2001) for more details. 4See Anderson (October 2001) for an overview of these various program evaluations. WorldBank WDI 2001. 6 Source: World Bank calculations using ENCOVI 2000 - Instituto Nacional de Estadistca Guatemala. 'Source: Ministry of Education, Education Statistics Yearbook 1996-2000, as cited in Anderson (October 2001). s Anderson (October 2001). 9 Share of children ages 7-12 at the time of the ENCOVI 2000 who ever enrolled in school. Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemnla. lo These ENCOVI 2000 calculations are based on the share of children who are currently in school who report repeating the grade they are currently in. They do not include those who were told they had to repeat, but subsequently dropped out. " The Ministry of Education reports a repetition rate of 14.5% in 2000, down from 15.3% in 1996. See Anderson (October 2001). 12 Anderson (October 2001) provides descriptions of the sector that are consistent with this explanation. 3 See GUAPA Technical Paper 3 (Edwards, 2002) for more details and regression results. '4 The Universidad del Vale was contracted to carry out the testing. Tests were administered to third and sixth graders in 1997, 1999 and 2000. See Anderson (October 2001) for a summary of key findings. I5 Calculations using the ENCOVI 2000 community questionnaire, which is not representative for al communities in Guatemala, but covers 60% of households in the household survey. 16 Anderson (October 2001). " Calculations using the ENCOVI 2000 community questionnaire, which is not representative for al communities in Guatemala, but covers 60% of households in the household survey. Analysis by Edwards (2002). '9 Based on multi-variate regressions of the determinants of indigenous language ability for those ages 7-25 that have at least one indigenous-speaking grandparent. Data from the ENCOVI 2000. Analysis by Edwards (2002). See Anderson (October2001) for details. 21 Other reasons cited - such as age, sickness, nmigration, and "othere (non-coded) - al individually accounted for small shares of the reasons for not enrolling. 22 Calculating this as the difference between net "non-enrollment" rates (poor girls: 25% vs. poor boys: 19%) as a share of total net non-enrollment for grls (6/25 = 24%). World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. See Annex 6 for additional details on the methodology for decomposing supply- and demand-side constraints to increased coverage of education, health and basic utility services using data from the ENCOVL 79 24 This section presents the results of multi-variate "hazard function"regressions which explain the probability that a child will ever enroll as dependent on a variety of individual and household characteristics. For details and results of various specifications, see GUAPA Technical Paper 3 (Edwards, 2002). 25 In other words, the interaction effect is also significant and negative. 26 Results of multi-variate hazard regressions for dropping out of school, once enrolled (ages 6-25). For details and results of various specifications, see GUAPA Technical Paper 3 (Edwards, 2002). 27 Informants in the ENCOVI were asked for the primary reason children did not enroll in school. Percentages are for the target age cohort of non- enrollees (aged 7-12). World Bank calculations using the ENCOVI 2000, Instituto Nacional de EstadLstica - Guatemala. r The share was fairly constant across genders, ethnicities, and areas, though was slightly higher for the non-poor (21%) than the poor (15%). 29 See Annex 6 for details on the methodology for decomposing supply- and demand-side constraints to increased coverage of education, health and basic utility services using data from the ENCOVI. 30 The rural and urban decompositions do not necessarily average into the decomposition for the overall country since they rely on different take-up rates for each area as well as different rates for the gap due to supply reasons. 3' Informants in the ENCOVI were asked for the primary reason children did not enroll in school. Percentages are for the target age cohort of non- enrollees (aged 13-19). World Bank calculations using the ENCOVI 2000, Insttuto Nacional de Estadfstica - Guatemala. 32 Given PRONADE's reliance on community and parental participation, these out-of-pocket costs underestimate the true costs to families to the extent that they do not cover parental contributions in kind and in labor. 33 Poverty line established by INE at Q.4319 per person per year, see Chapter 2. 34 Among 10-14 year olds, those who work complete 1.78 years of schooling as compared with 3.38 for those that don't or 3.35 for those that go to school and work. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadtstica - Guatemala. 35 See GUAPA Technical Paper 3 (Edwards, 2002) for details on the methodology used for these estimations (based on a Heckman selectivity model). 36 Regressions using the years of education were also estimated. The overall rates of return are 3% per year of schooling for males and 6% for females. 37 Specifically, a female who has completed primary school is expected to receive 17% more than one with no education (or almost 3% per year of primary completed). A woman who has completed secondary school receives an average of 54% more than one with no education (or 9% more per year of secondary completed). See GUAPA Technical Paper I (Vakis, 2002) for details. 35GUAPA Technical Paper 3 (Edwards, 2002) for details. 39 See GUAPA Technical Paper 3 (Edwards, 2002) for details. 4° Although the rates of return per additional year of school are higher for girls, the overall level of earnings are lower. 4' See GUAPA Technical Paper 3 (Edwards, 2002) for details on these "education-poverty line" calculations. 42 Public spending on education averaged 4% and 5% of GDP for LAC countries and lower-middle income countries respectively. Source: World Bank World Development Indicators Database. 80 I Chapter 8: Health, Malnutrition and Poverty Poverty results. in "death for not bringing the sick quickly to a health post. " Ladino Villager, LI (QPES) "My wife is sick but I can't bring her to a doctor or health center because I don't have money to pay for the visit or the trip (pasaje) " Mam Villager, M2 (QPES) This chapter seeks to analyze more deeply the issues pertaining to health and health care, with a view of informing policy and highlighting priorities for poverty reduction. The chapter begins with a review of the main health challenges facing Guatemala. Reflecting the gravity of the problem, it then assesses the problem of malnutrition, a key health challenge in Guatemala. It then explores the issues of coverage and access, concluding that both supply- and demand-side factors explain remaining access gaps. The chapter then reviews some of the key supply-side issues, including reforms and progress since the Peace Accords, public spending in the sector, institutional fragmentation and inefficiencies, and insurance coverage. Demand-side factors are then considered, including utilization patterns of various health services, and demand-side obstacles to greater coverage (particularly economic and cultural factors). Finally, the chapter concludes with a review of progress, key issues, and priorities in health for poverty reduction. The chapter mainly draws on analysis of the ENCOVI 2000,' supplemented by institutional information and qualitative data from the QPES. MAIN HEALTH CHALLENGES Unlike many other countries in Latin America, Guatemala is only at the beginning of the demographic and epidemiological transition. The population is very young, with a median age Figure 8.1 - Population Structure by Poverty Group, ENCOVI 2000 of 18, which is more on par with Africa (median age 18) than LAC (24), and far ! . younger than Europe and North America (37 a and 36).2 The poor population is particularly young, signaling a higher dependency ratio (Figure 8.1). It is also growing rapidly. Fertility (five children per woman) is the highest in Latin America.3 Guatemala also exhibits the second lowest average age of Z34 mothers at first birth in LAC, which -0-24 internationally has been shown to have 1014_ deleterious consequences for child health and 04 . survival of childbearing. The population o 30 - a I= growth rate is far higher than the average for LAC or other lower-middle income countries (Table 8.1). In terms of diseases, major causes of death in Guatemala still include treatable and communicable diseases, such as diarrhea, pneumonia, cholera, malnutrition, and tuberculosis. While the incidence of cholera and dengue has decreased rapidly over the past five years, the prevalence of malaria, diarrhea, and acute respiratory infections (ARI) have increased significantly.4 At the end of 2000, 4,000 cases of AIDS were officially recognized with a potential sub-estimation of 50%. The share of women infected by the virus has increased significantly over time, and the rise in infected women in reproductive age implies that there is also a higher probability of vertical transmission.5 Guatemala is among the worst performers in terms of health outcomes in all of Latin America (Table 8.1). Life expectancy at birth is the lowest in Central America, and far lower than the average for LAC countries or lower-middle income countries. Infant mortality (40-45 per thousand)6 is also the highest in Central America, and far higher than LAC average or the average for lower-middle income countries. Only Bolivia and Haiti perform worse for life expectancy or infant mortality in LAC; Guatemala does worse 81 than other low-income HIPC/IDA countries, such as Nicaragua and Honduras. As discussed below, malnutrition rates are also among the highest in the world. Finally, maternal mortality is extremely high at 270 per 100,000; the only other countries in LAC with higher rates of maternal mortality are Haiti and Bolivia.7 Guatemala has made some progress in improving health outcomes (such as infant mortality and life expectancy) over the past 20 years, but it's progress has been slower than the low-income HIPC/IDA countries of Bolivia, Nicaragua and Honduras. Progress in reducing malnutrition has been significantly slower in Guatemala than in other countries; in fact, it was the slowest in the LAC region.8 The patterns of health indicators do suggest worse conditions for the poor, rural, and indigenous populations, but the correlations depend on the indicator. In terms of demographic behavior, fertility rates are definitely higher among the poorer regions (Norte, Nor-Occidente, Peten), rural Figure 8.2 - Health Indicators by Quintile, residents, indigenous women, and those with % of Children < 6, ENCOVI 2000 less education. (which seems to have the 70 - strongest distinguishing effect, with those with 60 - no education having fertility rates 2.3 times 50 ---Malnutrition higher than those with secondary education).9 430 e Diarrhea In terms of morbidity, the prevalence of 20 -ARI diarrhea among children is higher among 10 those in the poorer quintiles (Figure 8.2), rural 0 residents, indigenous populations, and the poorer regions (e.g., Norte). In contrast, no , o9 significant difference is observed in the prevalence of ARI among children by quintile (Figure 8.2) or ethnicity.'0 In terms of outcomes, the prevalence of malnutrition is significantly higher among indigenous and rural children, and among those in the poorer quintiles (Figure 8.2), as discussed below. While infant mortality is higher among the indigenous and those with lower education, there is no distinguishable difference between urban and rural areas. Moreover, regional patterns for infant mortality do not closely follow those for poverty, with higher infant mortality in the Central (57), Sur-Occidente (58), and Nor-Oriente (54) regions (which do not have the highest poverty rates in the country). Interestingly, parts of these regions cover the "finca zone" with a significant share of plantations (coffee, sugar, banana) that serve as catchments for migrant laborers and their families." The 1995 DHS survey, does, however, suggest a correlation between "wealth" and infant mortality, with infant mortality rates for children in the poorest "wealth quintile" that were twice those for their counterparts in the top quintile.'2 Table 8.1 - International Cornparison of Various Health Indicators, 1999 Infant Mortality Life Expectancy Population Growth Dependency Ratio GNI/Capita, PPP Haiti 70 53 2.0 0.8 630 Bolivia 59 62 2.2 0.8 2,300 Giiatenia1a- . ..... .... \ - . 40-45 5 ; - 2.6 0.9 Dominican Republic 39 71 2.0 0.6 5,210 Peru 39 69 2.0 0.6 4,480 Honduras 34 70 3.0 0.8 2,270 Nicaragua 34 69 2.7 0.8 2,060 Brazil 32 67 1.7 0.5 6,840 El Salvador 30 70 1.5 0.7 4,260 Mexico 29 72 1.9 0.6 8,070 Ecuador 28 69 2.3 0.6 2,820 Paraguay 24 70 2.9 0.8 4,380 Colombia 23 70 2.0 0.6 5,580 Panama 20 74 1.9 0.6 5,450 Uruguay 15 74 0.7 0.6 8,750 Costa Rica 12 77 2.4 0.6 7,880 Chile 10 76 1.6 0.6 8,410 LAC Average 30 70 1.8 0.6 6,620 Lower Middle-Income Countries 32 69 1A 0.5 4,250 World 54 66 1.6 0.6 6,870 Source: WDI 2001. Infant mortality per 1000 live births. Population growth for 1980-99. Dependency ratio = dependents as % of working age population. 82 A Focus ON MALNUTRITION: A RED FLAG! Indicators of Malnutrition Malnutrition among Guatemalan children is extremely high - among the worst in the world. Guatemala has among the worst performances in the world in terms of child growth attainment, with an overall stunting rate of 44% of all children under five (height-for-age, HAZ).'3 With these rates, malnutrition affects a total of 756,000 children under 5 nationwide. Not only is the prevalence of chronic malnutrition (stunting, HAZ3 higher in Guatemala than any other country in LAC, it is also twice as high as the second highest rate in the region, observed in Bolivia in 1998 (27%). Guatemala's malnutrition rate is also among the highest in the world. Among those countries for which there is reliable information, only Bangladesh and Yemen have higher stunting rates (55% in 1996/7 and 52% in 1997, respectively). These estimates are consistent across time and across types of surveys (for example data from DHS surveys yield stunting rates (HAZ) for Guatemala of 50% in 1995 and 46% in 1998). Furthermore, they are not the result of some genetic aberration: stunting rates among the population of Southern Mexico - which arguably could have similar genetic hen'tage as much of the Guatemalan population (particularly the Mayan population) - arefar lower than those in Guatemala (29% in 1998).'4 Moreover, malnutrition is declining slower in Guatemala than in other countries. Guatemala has made some progress in reducing malnutrition, Figure 8.3 - Poverty and Malnutrition by from 59% in 1987 to 44% in 2000. However, the Bhnicity, % of kidividuals Below the Full yearly percentage reduction (1.7% p.a.) has been Poverty Line and % of Children < 5 w ho are the slowest in the LAC region. The annual Stunted (HAZ), B1COV I 2000 reduction was nearly twice as fast in Bolivia, 80 which was the second slowest at reducing 60+ :.. malnutrition in the region; the reduction in Brazil 40 i- 11 . ... .manut. was far faster. This lack of adequate progress 20~ H 4poet paradoxically contrasts with significant progress in ', other areas (basic services, health, education), as .4 ', discussed in subsequent chapters. Simulations suggest that, even if malnutrition in Guatemala were to fall at the pace projected by a worldwide panel of elasticities (i.e., faster than its historical pace), it would not improve enough to meet the Millennium Development Goals (MDGs) by the year 2015, as discussed in Chapter 5. Malnutrition is most serious among children under age two. Since malnutrition is a cumulative phenomenon, malnutrition rates increase with child age. The fastest increases in stunting occur in the first 24 months of life, particularly during the weaning period (from 6-24 months).'5 As such, the most effective nutrition programs should target children under the age of two, rather than at school age, which is the focus of most existing interventions (e.g., school feeding). There is a strong correlation between poverty and child malnutrition. Four fifths of malnourished children in Guatemala are poor. Malnutrition is much higher among poor children than non-poor children (64% of extreme poor and 53% of all poor children versus 27% of non-poor children). Children in the poorest quintile are almost four times more likely to be malnourished than their counterparts in the top quintile (62% and 16% respectively). 83 The incidence of malnutrition closely mirrors the geographic and ethnic patterns of poverty. As such, anthropometric measures appear to be good objective indicators of living standards. Like for poverty, malnutrition is much higher among rural children than urban (51% vs 32%). It is also Figure 8.4- Poverty and Malnutrition by much higher among the indigenous than the Region, % of individuals below the full poverty line non-indigenous (58% vs. 33%), and this gap and % of children < 5 who are stunted, ENCOVI holds true even controlling for other factors, as 2000 discussed below. Figures 8.3 and 8.4 show that 100 the patterns of both poverty and malnutrition are 80 60 quite similar across specific ethnic groups and 40- * *El geographic regions. The only exception is 20] - - u n. imalnut, among the Q'eqchi, who have quite high bpvr poverty, but relatively lower chronic 0 malnutrition (stunting). Q'eqchi children do, , sP cs Cb, however, have higher acute malnutrition (wasting) and diarrhea rates (which tend be correlated). The lDeterminants of Malnutrition Malnutrition is the product of the interaction of many factors, including individual and household decisions and behaviors (such as feeding practices), community infrastructure, the cultural and natural environment, national policies, and international economic conditions. Multi-variate regressions and proximate correlate analyses were used to identify the relative importance of many of these factors. Based on this analysis, the main determinants of malnutrition for children under age 5 include: o Education. Parental education is among the most important determinants of children's growth attainment, even after controlling for other factors. Education improves parents' ability to manage nutrition, disease and sanitation. It also influences other socioeconomic characteristics, such as parental age at marriage, number of children and their status in the community. In turn, malnutrition among school-aged children (which largely reflects early child malnutrition) has a significant impact on grade advancement in school, as discussed in Chapter 7. o l[llness. Morbidities, especially diarrhea and respiratory infections, are both causes and consequences of malnutrition. Stunting rates are much higher among children with frequent and early-age exposure to diarrhea or respiratory infections. In tum, malnourished children are more likely to be susceptible to such diseases. Disease prevention and treatment, together with improving the availability and quality of water and sanitation are critical for fighting chronic malnutrition. o Family Planning. Pregnancies at a young age, numerous children, and short intra-birth spacing - all very common in Guatemala (as discussed above) -- are associated with deficient child growth pattems. o Mother's Nutritional Status. Reflecting inherent genetic traits, there is a strong relationship between the mother's height and her child's nutritional status. As expected, the influence of a father's height is significant, but smaller than the mother's height because of the mother's additional effect on children's nutritional status through the womb environment. Including variables for parents' height also controls for genetic factors. o IBreastfeeding. Breastfeeding and appropriate weaning practices are one of the most important household behaviors that can influence nutritional outcomes and that can be influenced by policies and programs (e.g., information campaigns). Exclusive breastfeeding for at least the first six months 84 of life provides an adequate source of nutrients and antibodies and eliminates the risks of illness from contaminated utensils or water. It also helps repress fertility by extending the duration of post- partum amenorrhoea. Indeed, breastfeeding proved to be a significant determinant of malnutrition in regression analysis in Guatemala. * Basic Utility Services. The availability of piped water, sanitation, electricity, and garbage collection systems is significantly correlated with better nutritional outcomes. * Geography, Ethnicity and Language Ability. Malnutrition is also worse in rural areas, among the indigenous, and among children whose mothers do not speak Spanish, even after controlling for other factors. As discussed above, malnutrition is almost twice as high among the indigenous as the non-indigenous. Even after controlling for other characteristics,'6 a significant share of this gap is unexplained and likely arises due to health and behavioral factors (e.g., weaning practices, hygiene, cooking behaviors). * Poverty. Per capita consumption has a very significant and positive effect on children's nutritional status, though the direction of causality is not clear. Poverty boosts malnutrition by restricting individuals' access to basic services and other assets, increasing exposure to disease, and reducing access to food. On the other hand, high rates of malnutrition jeopardize future economic growth by reducing the intellectual and physical potential of the population. Malnutrition Interventions Little funding is devoted to the problem of malnutrition in Guatemala. Nutrition programs received about US$27 million in 2000, most of which was used in school feeding programs for primary-school-aged children. The School Lunch program of MINEDUC received 9.3 million, while almost US$12 million paid for the school breakfast program sponsored by the Office of the Vice President. Food for preschoolers was administered through PAIN and Hogares Comunitarios, both of which covered a total of 42,000 children under five, or 14% of the total number of children under five living in extreme poverty, or 2.8% of the total number of poor children under five. Moreover, existing resources are misdirected and do not attack the main sources of the problem. Virtually all public funds earmarked for "nutrition" are directed to school feeding programs or short-term emergency assistance (e.g., food handouts or food-for-work programs in areas struck by natural disasters). A variety of other grass-roots level schemes exist, but these are small and quite dispersed. While school feeding programs can serve as useful incentive mechanisms to promote increased enrollment and attendance at school, they have little effect on malnutrition because (a) they reach the malnourished or potentially malnourished population too late since they are directed to school-aged children when most stunting or wasting occurs in early phases of life (particularly between 6-24 months of age); and (b) a lack of available food is rarely the main cause of long-term malnutrition. Moreover, while emergency food programs are useful for alleviating short-term acute food security problems, such food handouts are not necessarily the most effective or sustainable solutions to the larger, long-term malnutrition problems. Past efforts in micronutrient fortification have lagged, and anemia among women is quite high. At one time, Guatemala stood out as a pioneer in micronutrient fortification, as one of the first countries to adopt fortification practices. However, interruptions in fortification, weak regulation and poor targeting have made Guatemala one of the worst performers in terms of implementation of these schemes.'7 Without proper fortification or adequate dietary sources of iron, the percentage of women affected by anemia is the second highest in Central America (35% in 1995), which results in adverse consequences for infant mortality and low-birth weight babies. An aggressive, integrated approach to malnutrition needs to be adopted in Guatemala. Targeted and concerted actions are needed in the areas of health (particularly child and maternal health), access to basic 85 utility services (e.g., potable water and sanitation), and education (nutrition information). Moreover, specific nutritional interventions are needed. These include: (a) promotion of exclusive breastfeeding, proper health, hygiene, and complementary feeding practices; (b) growth monitoring and promotion of pregnant women and children under age two; (c) micronutrient supplementation (particularly for iron) and reinstating fortification programs; and (d) deworming treatments and oral rehydration therapy. The main target population for these schemes should be pre-school children (particularly those under 24 months of age) and mothers (including pregnant and lactating women). To improve the effectiveness and reach of these programs (and avoid having them get "set aside"), these interventions should be integrated into the MSPAS basic health care package and provided at the community level through outreach workers selected by the community but contracted either by NGOs or the MSPAS. This would also allow for the institutionalization of nutrition activities within MSPAS, which would eventually contribute to the much needed rationalization of the myriads of independent, often incompatible, nutrition efforts in Guatemala. Nonetheless, although institutionalization can be done within MSPAS, strong collaboration should be established with other key ministries and stakeholders at the highest levels, given the multi-faceted nature of malnutrition. ACCESS TO HEALTH CARE Access to health care services is a key determinant of health outcomes. International evidence suggests that access to - and use of - health care services is highly correlated with health outcomes. Other factors -- access to potable water, education, and household behaviors - also play important roles. While estimates vary, a significant share of Guatemallans apparently lack access to health care services. The World Health Organization (WHO) defines access as living within one hour of traveling time to a health care facility. Data from the ENCOVI suggest that average traveling times are about 45 minutes overall, but that they are much higher in rural areas (just under an hour). According to the ENCOVI, about 59% of households lacked access to a health facility (had to travel more than 60 minutes of traveling time).'8 This lack of access was worse among the poor (63%) than the non-poor (52%). It is also worse in rural areas (64%) than urban (48%), and in the Norte, Central, Suroccidente, Sur-oriente, and Peten regions, most of which have very high rates of infant mortality. As discussed in Chapter 9, adequate road access is crucial for rural populations to access health services: travel times for those without motorable roads are considerably higher (69 minutes) - well beyond the WHO definition of access. Qualitative data also suggest that health services are more lacking than other basic social services. In the QPES, for example, only three of ten rural villages had access to a health clinic or health post, though all ten villages reported access to midwives. In contrast, data from the Ministry of Health (MSPAS) suggest that only 9% of the population lacked access in 1999, coming down from 46% without access at the time of the Peace Accords in 1996, largely due to improved coverage of the SIAS program (including traveling health promoters). Even when health facilities are present, they are often understaffed and lack medicines and equipment.19 According to data from the ENCOVL the poor and indigenous are more likely to live in communities reporting insufficient medicines, medical equipment, beds, ambulances, and medical staff (particularly maternal-child specialists such as gynecologists, obstetricians, and pediatricians). Various qualitative studies also point to these problems.20 The World Bank World Development Indicators (2001) suggest that there are significantly fewer physicians per capita in Guatemala (0.9) than other lower-middle income countries (1.9) or the LAC region as a whole (1.6). The ENCOVI suggests that waiting times - a sign of understaffing - are considerably long, averaging 46 minutes for all types of facilities. They are substantially longer for public and IGSS hospitals (66 and 87 minutes respectively) and shorter at community centers (33 minutes). 86 Figure 8.5 - Reasons for Not Seeking Health Care Treatment When Needed, % of those who did not seek treatment when ill and believed treatment was necessary; ENCOVI 2000 Total - !I Q5 richest M Distance/lack transport * Lack doctors/nurses I4 [ OWaiting time Q4 _ __ I _ ODid not have time * Lack money for service 03 ElI" Ei Lack money for transport Q3 _ w * Don't speak my language lO Don't trust caregiver 02 l Other reasons Q1 poorest 0% 20% 40% 60% 80% 100% Access is constrained by both supply and demand-side barriers. Two types of indicators support these findings. First, the ENCOVI 2000 gathered data on reasons people did not seek treatment when they were 21 ill but judged that treatment was necessary. Demand-side factors - particularly economic constraints - account for the bulk of the coverage gap (Figure 8.5). Nonetheless, supply-side factors - including distance and lack of transport to health facilities, a lack of doctors and nurses, and lengthy waiting times - also appear to present significant constraints to improved access. Overall, these factors accounted for 17% of the coverage gap. These supply-side constraints were higher among those in the poorer quintiles (25% in the poorest quintile compared with 8% in the richest, Figure 8.5). Supply-side constraints were also seven times more frequent in rural than urban areas, reflecting the greater presence of health facilities in Guatemala's cities. Second, data on health treatment patterns within "communities" (primary sampling units, PSUs) in the ENCOVI sample allow for a decomposition of coverage gaps between (a) those who did not seek treatment because there were no health care facilities available (pure supply-side gap); (b) those who did not seek treatment even when facilities were available (pure demand-side constraints); and (c) those who did not seek treatment because no health care facilities were available, but still would still not seek treatment if such facilities were provided (mixed supply- and demand-side factors).22 This analysis reveals that overall, about 13% of households do not have access to any health care facility (supply-side constraints), but about half of these would not seek treatment even if facilities were to be made available (mixed supply- and demand-side constraints, see Table 8.2). Demand-side constraints prevented about 87% of the population from seeking health treatment, even when facilities were present. In terms of public facilities, which are significantly cheaper than private (as discussed below), a much larger share of the population lacked physical access to such facilities, particularly in rural areas where 62% lack public health facilities (though demand-side constraints would also prevent many of these from seeking treatment). In conclusion, a combination of both supply- and demand-side constraints limit the ability of households to seek health care services in Guatemala, with supply-side constraints playing a more dominant role in rural areas than urban. Both demand- and supply-side constraints to improved access are discussed in greater detail below. 87 Table 8.2 - Decomposition of Coverage Deficit - Access to Health Facilities (for households that do not use facilities) Demand side Both supply- and Supply-side Total constraints only demand-side constraints constraints only Publc facility - All 45 44 11 100 Public facility - Urban 55 35 10 100 Public facility - Rural 38 51 11 100 Private - All 69 21 9 100 Private facility - Urban 72 17 11 100 Private facility - Rural 68 25 8 100 Other facility - All 55 35 10 100 Other facility - Urban 41 47 12 100 Other facility - Rural 66 26 8 100 Any facility - All 87 6 7 100 Any facility - Urban 84 8 8 100 Any facility - Rural 89 4 6 100 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala See Annex 6 for details on methodology. SUPPLY-SIDE FACTORS Some Improvements Since the Peace Accords...But Little Impact on Health Outcomes The Peace Accords included significant commitments to improving health and health care services in Guatemala. The main health-related objectives included: (a) increasing public spending on health by 50% from 1995-2000 with at least 50% of the budget devoted to preventative health; (b) reducing infant and maternal mortality by 50%; (c) eradicating measles by 2000 and maintaining polio eradication certification; and (d) decentralizing and deconcentrating health service provision. Some progress has been made in reforming the system. Important steps have been taken on the institutional side, with health being one of the pilot ministries (along with education) to deconcentrate financial management under the SIAF program, as discussed in Chapter 4. The Government also launched the Sistema Integrado de Atenci6n de Salud (SIAS) in response to the Peace Accords in order to expand coverage to the rural and indigenous population and improve health services. The SIAS system is highly decentralized, often working through NGOs, cooperatives, and municipalities. It provides both preventive and curative care with the aim of bringing services closer to communities and promoting community participation. It often sponsors traveling "promotores de salud" in rural areas. The Government also adopted a National Health Plan for the period from 2000-2004 that focuses on (a) decentralization with social and community participation; (b) providing primary health care with an emphasis on promoting information, self- awareness and responsibility; and (c) improving equity, efficiency, quality and sustainability with a focus on the most disadvantaged groups. Finally, available data do suggest a shift in public spending towards preventive care, which is appropriate given Guatemala's epidemiological status and equity concerns. Despite these efforts, spending and health outcomes have not improved significantly. Although public spending on health has increased slightly, it still falls short of targets set by the Peace Accords (see below). Infant mortality still hovers around 40-45 - far from the target for 2000 of 20 per 1,000 live births. Maternal mortality also far exceeds the target of 48.5 per 100,000. Measles vaccinations likewise fell short of their target for 2000 (see Chapter 4). Moreover, malnutrition remains stagnant at an extremely high level, as discussed above. Closely related, Guatemala has seen an increase in the prevalence of diarrhea, malaria, and acute respiratory infections. 88 Public Spending: Increases Insufficient, Not Well Targeted Public spending increases have been insufficient. As a share of GDP, total public spending on health has increased over the period from 1995-2000, but still falls short of the Peace Accords targets (Table 8.3). Over the period from 1990-98, total per capita spending on health care (public and private) averaged US$155 in PPP terms, which was far lower than its Central American neighbors,23 or the averages for LAC (US$452) or lower-middle income countries (US$190). Nonetheless, the composition and execution of public spending have improved somewhat. Available data suggest that the Government has increased its allocations to preventive health care (over tertiary or other levels of health care), which is vitally important to poverty reduction as it focuses on the main health problems faced by the poor (Table 8.3). Moreover, changes in public expenditure management - including the incorporation of health spending into the SIAF system and the deconcentration of financial management - have improved the execution of public health spending somewhat since their introduction in 1997 (Table 8.3). Nonetheless, recurrent spending seems to have edged out capital expenditures over time, thus reducing spending on longer-term investment activities. Table 8.3 - Public Spending on Health, 1995-2000 1996 1997 1998 _ 1999 2000 2000-target Public Spending on Health, mn Q.' 627.2 872.1 1177.8 1594.1 1557.7 n.a. % of GDP 0.7% 0.8% 0.9% 1.2% 1.1% 1.3% % of Total Public Spending 6.3% 6.9% 7.1% 8.3% 7.9% n.a. Capital exp/total public health exp. 30.5 35.3 18.0 23.6 17.6 n.a. Preventive/total public health exp. n.a. 43.0 46.0 49.0 52.0 >50 Executed/Planned Spending (ratio)'a 62.1 91.5 91.6 89.6 94.3 100 Sources: a. SIAF/Ministry of Finance (executed spending). b. World Bank (February 2000). Public spending on health is not very well targeted. Overall public health spending Figure 8.6 - Distributional Incidence of Public Spending on benefits the middle quintiles Health, Intemational Cornparison disproportionately more than their share in the ENCOV12000, IADB 1999, World Bank (2001a) population (Table 8.4). This is not dissirnilar 100% to other countries in LAC, though - 80% Guatemala's poorest quintiles still receive t 60% richest slightly less benefits from public health 40% - -03 o ~~~~~~~~0 spending than their counterparts in other 20% - - - * countries (Figure 8.6). By poverty group, the X3 0% - , ,, ,Q1 poorest poor receive about 53% of total public OR spending on health, close to their share in the overall population (56%). Nonetheless, public ___ spending on health is more progressive than the current distribution of total consumption in the economy (Table 8.4). By type of facility, public spending on hospitals is by far the most regressive, with the top two quintiles receiving 51% of total net public subsidies on hospitals, as compared with only 29% for the lowest two quintiles (Table 8.4). In contrast, public spending on health posts and community centers is very well targeted (Table 8.4), reflecting higher usage of these types of facilities by the poor. 89 Table 8.4- Distributional Incidence of Public Health Spending, by Facility % of net public subsidies received by each group By Quintile By Poverty Group By Ethnicity By Area Total Ql Q2 Q3 Q4 Q5 XP AP NP IND Non-Ind Rural Urban Health - Total 100 17 18 23 25 17 12 53 47 40 60 64 36 Hospital 100 13 16 21 29 22 9 45 55 33 67 58 42 Health Center 100 20 23 28 20 9 16 65 35 51 49 67 33 Health Post 100 40 22 27 6 5 29 84 16 53 47 98 2 Community Center 100 39 20 23 8 10 34 75 25 71 29 87 13 Memo for Comparison Share of total population 100 20 20 20 20 20 16 56 44 43 58 61 39 Share of poor population 100 36 36 29 0 0 n.a. n.a. n.a. 58 42 81 19 Share of total consumption 100 5 9 13 24 54 4 24 76 24 76 37 63 Source: World Bank calculations using the ENCOVI 2000, Instiuto Nacional de Estadistica - Guatemala. Figures reflect differential unit subsidies by provider as well as utilization patterns by provider. Quintiles are individual consumption quintiles. Duplications and Fragmentation in the Health Sector Institutionally, public services are somewhat fragmented between two large providers. The Ministry of Public Health and Social Action (MSPAS) is responsible for providing curative and preventive care for the entire population at practically no charge to users. The largest actor in the sector, MSPAS also has the constitutional mandate for defining sectoral policies and coordinating other actors in the sector. The other large public actor is the Guatemalan Social Security Institute (IGSS), which provides retirement benefits and health services to eligible formal sector workers in their families. The IGSS runs separate facilities from MSPAS. While members of the IGSS can use MSPAS facilities, only affiliated members can use IGSS facilities. As discussed in Chapter 12, coverage of the IGSS is very low, even as a share of formal sector workers. Other providers in the sector include private providers, including traditional health practitioners (see Box 8.1) and modem providers, which have emerged primarily in urban areas in response to rising incomes and dissatisfaction with the quality of public sector care. The private sector also dominates the provision of pharmaceuticals, with very limited state regulation. NGOs (both international- and national- based) are also active in the health sector. In reality, there is little coordination between providers, with significant duplications and gaps in coverage. Use of various facilities by the insured also reveals duplications and inefficiencies in coverage. Interestingly, a substantial share of those with private health insurance - who tend to be in the richest quintiles of the population - still use highly subsidized public facilities (public or IGSS hospitals, health posts and centers, Table 8.5). In contrast, an equal share of those covered by IGSS use IGSS hospitals and private hospitals, perhaps reflecting their dissatisfaction with services at IGSS hospitals (which had the longest waiting times of all facilities). Finally, a substantial share of those with no insurance (who tend to be poorer, as discussed below) still used relatively more expensive private hospitals, reflecting gaps in coverage of the public health system and disparities in the quality of care received in public and private facilities. Table 8.5 - Use of Different Health Facilities by People with Type of Insurance Coverage Type of Public IGSS Private Health Post Community Pharmacy Home Other Insurance Hospital Hospital Hospital or Center Center Private 5.3T. ; .17.0. 53.0 .7.7: 0.3 6.1 _ 8.3 2.3 IGSS .776.07- 38.6 . 37.1 { 4.6 1.2 6.9 3.0 2.7 No Insurance 10.1 1.3 39.5. 26.9 2.2 8.3 5.8 6.1 Source: World Bank calculations using the ENCOVI 2000, Institutio Nacional de Estadistica - Guatemala. 90 Inadequate Insurance Coverage Health insurance coverage of Table 8.6 - Share of Individu Is Covered by Health Insurance any kind is very low in Total Ql Q2 Q3 Q4 Q5 Urban Rural Guatemala. Only 11% have Private insurance 2.2 0.0 0.0 1.0 2.0 9.0 5.5 1.0 any kind ofhealthinsur IGSS 8.3 3.0 3.0 6.0 12.0 18.0 14.5 5.0 any kind of health mnsurance No Insurance 89.0 97.0 97.0 93.0 86.0 73.0 81.0 94.0 (Table 8.6). Most of those who Source: World Bank calculations using the ENCOVI 2000, Institutio Nacional de Esrad(stica - have insurance are affiliated Guatemala. with IGSS, while the rest have private insurance. The great majority of those who have either type of insurance belong to the richest quintiles and live in urban areas (Table 8.6). DEMAND-SIDE FACTORS Underutilization of Health Services The ENCOVI suggests a serious under-utilization of health services among disadvantaged groups. These patterns can be seen for children, individuals and prenatal and delivery care. Specifically: * Children with diarrhea or ARI lack treatment, particularly the poor, indigenous, and rural children. Over half of all Guatemalan children with diarrhea or ARI are treated by relatives or non- medically qualified personnel (Table 8.7). Mirroring this pattern, over half are treated at home. A much higher share of children in the poorest quintiles lack treatment by medical professionals. Only in the highest quintile do children have a higher probability of being seen by a doctor than by any other person. Non-indigenous and urban children are twice as likely to be treated by a doctor than indigenous or rural children. * Overall, a far smaller share of the poor, indigenous and rural residents are treated by medical professionals. Two thirds of those in the poorest quintile do not seek treatment by medical professionals (doctors, nurses or promotores), versus just over one third of those in the top quintile (Table 8.8). Likewise, whereas over half of urban and indigenous residents each seek treatment from medical professionals, only a third of rural and indigenous residents do. * Professional pre-natal care is likewise far less likely among disadvantaged groups. The type and quality of pre-natal care during pregnancy and delivery are very important factors for the health of mothers and their children. While a fifth of all women did not seek any form of pre-natal care, the share is much higher among the poorest quintiles (about one third) than those at the top of the spectrum (less than one tenth). Likewise, a higher share of indigenous and rural women do not undergo pre-natal visits. Of those who do seek care, poor, indigenous and rural women are far more likely to be seen by midwives, whereas the non-poor, non-indigenous and urban women are more likely to be treated by doctors (Table 8.9). * An extremely high share of women from disadvantaged groups give birth at home. Over four fifths of women in the poorest quintile give birth at home, compared with 12% of those in the top quintile (Table 8.9). Similarly, over two thirds of rural women and three quarters of indigenous women deliver at home. Delivery in the home is of particular concern for poor women, since they lack the sanitary conditions needed for safe delivery: as discussed in Chapter 9, half of all households in the poorest quintile lack piped water (and the water is not potable even for those with connections), and over a quarter lack any type of sanitation. The lack of such basic utility services is similar for rural and indigenous households. Only 14% of women in the poorest quintile deliver in the presence of a doctor, nurse or promotor; most (71%) are accompanied by a midwife, and the remainder (15%) are not seen by any medical professional (Table 8.9). Interestingly, women in the poorest quintile are twice as likely to have complications with birth as those in the top quintile, but 91 women in the top quintile are 13 times as likely to have a cesarean sectioned birth as those in the bottom quintile.24 Barriers to Access Demand-side factors - particularly econoniic barriers - limit access to health services. As shown in Figure 8.5 above, 83% of households cite demand-side factors as the primary reasons for not seeking treatment when needed.25 This share is slightly lower among the poorer quintiles (75%), for whom supply- side constraints figure more prominently. Nonetheless, the primary barrier for all quintiles is economic: some 60% do not seek treatment because they lack money to pay for health services. Another 6% do not seek treatment because of the costs of transport to health services. Table 8.7 - Utilization of Health Services by C ildren with Diarrhea or ARI B Quintile By Ethnicity By Area Total QI Q2 Q3 Q4 Q5 Indigenous Non-Indigenous Urban Rural Prevalence of Diarrheaa 31.3 33.8 35.2 30.5 27.8 24.1 35.6 27.5 24.7 34.5 Prevalence of ARI^ 47.9 47.4 50.1 47.6 47.2 46.0 48.8 47.0 41.2 51.1 By Personnel' Non-medically trained 6.4 6.1 5.7 7.5 6.2 5.9 7.9 4.9 6.3 6.4 Nurse/Promotor 17.8 24.1 23.4 17.5 9.4 4.2 20.7 15.2 7.8 22.0 Doctor 25.8 14.1 18.5 22.0 39.4 56.2 18.3 32.7 44.6 18.1 Parent/relative 50.0 55.4 52.4 53.1 45.0 33.8 53.1 47.2 41.3 53.6 By Type of Facility" Public Hospital 2.3 1.4 1.1 2.4 4.7 3.1 1.7 2.9 3.3 1.9 IGSS Hospital 4.1 1.0 3.2 3.1 8.6 8.6 2.9 5.2 11.0 1.3 Private hospital/clinic 10.8 3.5 4.6 8.0 17.2 38.6 6.6 14.8 19.6 7.3 Health post/center 19.3 21.9 23.9 20.7 14.2 8.2 20.5 18.3 15.1 21.1 Community center 3.2 7.1 2.2 1.7 2.2 0.0 4.0 2.4 1.0 4.2 Pharmacy 4.6 3.7 4.4 5.0 6.1 4.0 5.6 3.7 5.1 4.4 At home 52.7 59.3 55.1 56.0 45.4 36.2 55.7 49.8 42.6 56.8 Other 3.0 2.3 5.5 3.1 1.5 1.1 .3.1 2.9 2.8 3.1 Source: World Bank calculations using the ENCOVI 2000, Institutio Nacional de Estadfstica - Guatemala. a. % of children < 6 with diarrhea or ARI in the month preceding the interview. b. % of children with diarrhea or ARI visiting different types of health facilities or personnel. Table 8.8 - Utilization of Health Services by Individuals B Ouintile I Ethnicity BArea b______________ Total Ql Q2 Q 4 Q5 Indigenous Non-Indigenous Urban Rural By Personnelb Non-medically trained 4.8 4.5 4.9 6.6 5.4 3.3 5.9 4.1 3.6 5.7 Nurse/Promotor 9.5 16.5 13.9 11.8 7.8 2.3 13.7 6.6 3.6 13.7 Doctor 36.1 13.6 22.6 32.6 39.0 58.1 24.0 44.4 51.8 25.0 Parent/relative 19.2 29.8 23.0 21.0 17.2 11.0 22.8 16.8 14.7 22.5 Self Medication 19.5 17.2 25.0 18.3 21.5 16.8 20.7 18.8 17.8 20.8 Did nothing 10.9 18.4 10.7 9.7 9.1 8.6 12.9 9.4 8.6 12.5 By Type of Facility' Public Hospital 9.3 8.5 10.8 10.2 11.8 6.8 8.1 9.9 8.8 9.7 IGSS Hospital 6.8 2.2 3.2 5.7 8.2 9.3 3.7 8.4 10.5 3.2 Private hospital/clinic 40.0 12.6 24.8 26.9 39.6 62.5 30.2 45.4 51.5 29.3 Health post/center 22.8 39.9 35.5 34.3 23.2 5.7 29.6 19.1 13.8 31.4 Community center 2.0 5.4 2.1 2.2 1.7 0.9 3.2 1.3 0.8 3.1 Pharmacy 8.0 8.1 9.2 10.3 9.0 5.5 11.5 6.0 5.6 10.2 At home 5.5 8.2 7.1 4.0 2.0 6.9 6.5 4.9 5.2 5.7 Other 5.7 15.1 7.3 6.4 4.4 2.5 7.3 4.8 3.9 7.5 Source: World Bank calculations using the ENCOVI 2000, Institutio Nacional de Estadistica - Guatemala. a. % of individuals with illness or accidents in the month preceding the interview. 92 Table 8.9 - Prenatal Care and Treatment Duri Delivery, Pregnant Women_ _ B Quintile By Ethnicity By Area Total Ql Q Q3 Q4 QS Indigenous Non-Indigenous Urban Rural Prenatal Care No Prenatal Checksa 21.3 34.2 28.0 22.7 10.2 8.8 26.2 17.3 13.3 26.8 By Personnelb Non-medically trained 2.4 4.4 2.6 1.8 1.3 2.0 4.2 1.1 1.4 3.0 Midwife 33.7 59.7 45.4 36.3 20.3 6.1 49.2 22.7 14.4 46.7 Nurse/Promotor 11.0 15.7 15.6 12.0 8.6 2.5 12.7 9.8 5.4 14.8 Doctor 52.4 19.5 36.2 49.3 69.8 89.1 32.9 66.2 78.4 . 34.8 Relative 0.6 0.7 0.2 0.6 0.9 0.2 1.0 0.2 0.4 0.7 By Type of Facility" Public Hospital 12.5 3.9 9.7 15.6 19.2 12.9 6.6 16.6 16.0 10.1 IGSS Hospital 9.3 1.2 4.1 7.4 13.5 20.9 4.3 12.8 18.0 3.5 Private hospital/clinic 17.0 2.0 4.4 8.1 24.0 48.7 9.4 22.3 30.9 7.6 Health post/center 24.3 25.5 31.8 31.6 21.5 9.6 24.2 24.3 19.1 27.8 Pharmacy 0.2 0.0 0.2 0.6 0.0 0.0 0.4 0 0.0 0.3 Midwife's house 13.5 18.7 16.9 17.3 11.0 2.6 17.2 10.8 7.0 17.8 Athome 21.8 45.6 31.3 18.4 9.2 4.8 35.8 11:9 7.8 31.2 Other 1.6 3.1 1.5 1.1 1.6 0.6 2.1 1.3. 1.4 1.8 Delivery: Treatment and Place _ By Personnel' Non-medically trained 2.7 4.1 2.5 3.4 2.2 0.9 3.4 2.2 2.4 2.9 Midwife 47.1 70.5 63.1 49.4 29.8 12.3 63.1 34.8 25.0 60.2 Nurse/Promotor 4.5 3.1 4.8 4.5 6.1 3.9 3.4 5.3 4.3 4.6 Doctor 40.1 11.0 21.5 36.6 61.2 83.0 19.9 55.6 67.6 23.8 Relative 5.6 11.3 8.2 6.1 0.7 0.0 10.2 2.1 0.7 8.5 By Type of Facilityb Public Hospital 25.7 9.8 18.3 30.0 41.2 32.1 13.6 35.1 35.4 20.0 IGSS Hospital 7.6 1.3 3.4 4.7 12.7 19.5 3.5 10.8 15.8 2.8 Private hospital/clinic 8.0 0.1 0.7 2.5 11.7 31.4 3.2 11.6 16.9 2.7 Health post/center 3.3 1.5 3.3 4.9 3.4 3.8 1.8 4.5 4.0 3.0 Midwife's house 3.1 3.4 2.9 4.7 2.9 1.2 3.2 3.1 2.3 3.6 Athome 51.8 83.5 70.5 52.3 28.2 11.9 74.5 34.2 25.2 67.4 Other 0.5 0.4 0.8 0.9 0.0 0.3 0.2 0.7 0.4 0.6 Source: World Bank calculations using the ENCOVI 2000, Institutio Nacional de Estad(stica - Guatemala. a. % of non-pregnant women who have been pregnant. Pregnant women are excluded to avoid censoring in the data. b. % of women who had prenatal checks. nhere are significant Table 8.10- Average Costs of Prenatal an Delivery Care. b ype~ Facilitv variations in the costs of Private health care in Guatemala. Quetzales Public IGSS Hospital Health Post Midwife's Own ____uetzal_s Hospital Hospital or Clinic or Center House Hoi.is Other Guatemalan households Prenatal Care 304 168 635 262 130 89 130 allocate an average of 3.7% Urban 407 183 747 394 252 334 446 of total consumption to Rural 231 87 354 163 102 62 50 health, though this share Giving Birth 274 500 3,014 1,262 224 105 120 varies signUrban 349 500 3,471 1,927 161 126 n/a varies significantly Rural 218 n/a 956 120 250 100 120 between the poor (1.4%) World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala and the non-poor (4.2%). Service costs vary considerably by provider and level of service. For example, out-of-pocket expenditures per visit to public facilities average Q19, with 66% of all households paying nothing for public services. Within the public domain, these expenditures range from Q7 per visit to public health posts (with 74% paying nothing) to Q43 per visit to public hospitals (with 60% paying nothing, including 54% of those in the top quintile). In comparison, out-of-pocket expenditures per visit to private facilities average Q130 (seven times higher than the average cost of public facilities). The average costs of prenatal care and deliveries likewise vary significantly by provider and level of service (Table 8.10), with private providers being much more costly. Although public facilities are quite cheap (and indeed, many pay nothing per visit), costs may still present a significant barrier to access - as identified by households in the ENCOVI survey - if they lack access to public facilities. Hence access is constrained not only by an availability of health services, but also by an availability of affordable (public) services for the poor. 93 Box 8.1 - Health and Ethnicity: The Need for a CulturaDly Sensitive Approach Cultural Barriers to Health Care A qualitative study of ethnicity and health conducted by Solares (1997)' reveals perceptions of discrimination, fears of hospitals, and patient miscommunication among the indigenous population. In some cases, these fears lead to practices such as reducing pregnancy weight gain to avoid having to go to the hospital to give birth (as highlighted by Solares below): *Perceptions of discrimination. "Los ind(genas sienten ... que en el Puesto de Salud y en el Hospital los tratan mal, los reprenden por causas ajenas a ellos como el no poder Ilegar a una hora prefijada o no entender bien el castellano.... Su incomprension dentro del Puesto de Salud no s6lo depende de no poder hablar ni comprender bien una lengua extranjera como es el castellano, sino que se agrava por razon de su analfabetismo tan acentuado en la poblacion ind(gena...... *The cold world of the hospital. "Cuando sus parientes quedan encerrados y ocultos en la instalaci6n hospitalaria, se les impide saber de ellos y s6lo reciben las vagas indicaciones de acudir otro d(a.... Por eso es que las embarazadas hacen lo posible por reducir su peso a fin de que el nino nazca delgado y le evite ast a su madre el tener que acudir al hospital por causa de una "complicacion". El hospital para la parturienta indigena significa el desarraigo absoluto de su vida familiar, de su nucleo domestico y de su casa que es donde ocurre el alumbramiento, con la compailia defamiliares que las apoyan. *Miscomnunication and confusion of the patient. La prdctica medica corriente de callar informacion al paciente, choca profundamente con la manera tradicional del mundo indigena.... Por ejemplo, recetan a la gente pero no les indican el funcionamiento ni que efectos puede causarles. Todo esto produce una gran confusi6n (... agravada en la mayor(a de los casos por la incomunicacion linguistica), la cual explica hechos como el de las concepciones erroneas de interpretar las vacunas como medios para "curar" enfermedades indistintamente; el padecer posteriormente alguna afecci6n o el sufrir reacciones normales en las vacunas, pero desconocidas para la familia del pequenio paciente. Traditional Medicine Another reason why Guatemalans do not seek "formal" health care is the widespread practice of traditional medicine. Although modem medicine is becoming increasingly dominant, traditional beliefs about sickness and health continue to inform rural Guatemalans' health ideas and choices. For example, as in other rural areas of LAC, the indigenous ethno-medical framework revolves partly around one primary set of oppositions: between hot and cold. Illness is believed to be caused by an imbalance of hot and cold, and treatment is believed to be effective only if the prescribed medicines or foods are the opposite temperatures of the disorder so as to restore the hot-cold balance.b Many people use both traditional and modem medicine, which are usually perceived as complementary rather than in conflict with one another.' Recent findings suggest that modem medicine is being used more frequently for the treatment of child illness than traditional remedies." Traditional Health Practitioners In the absence of - or as an altemative to - formal health facilities, the comadrona (midwife), curandero (herbalist) and the ajq'ij (Mayan priest) offer traditional and culturally acceptable services, often together with SIAS promotores de salud (health practitioners)." f The comadrona usually treats pregnant women, serves as a gynecologist, and sometimes as a pediatrician. She is typically a woman in her forties, who shares the same language, values and cultural beliefs as her patients. The curandero deals with the physical and psychic well-being of society, and is believed to have special abilities that allow him direct contact with superior forces. He normally treats diseases such as eye problems, anger and indigestion, usually with medicinal plants, candles and animal parts. The ajq 'ij is not only responsible for the health of the village, but also for upholding customs (e.g., the Mayan calendar and celebrations). He is normally an expert in herbs, including those used in rituals and for medical treatment! Sources: a. Jorge Solares (1997). b. Logan (1973). c. Cosminsky and Scrimshaw (1980). d. Heuveline and Goldman (1998). e. QPES (2002); f. MSPAS, OPS/OMS (2001). Besides costs, cultural factors - such as language, trust, and perceptions of discrimination - may constitute barriers to access to health services, particularly among the indigenous. Indeed, about 3% of indigenous households cite such barriers as the primary reasons for not seeking services when ill, according to the results of the ENCOVI survey. Qualitative research reveals the need for a culturally sensitive approach to health care. For example, a study by Solares (1997) highlights perceptions of discrimination, the "cold world" of the hospital, and miscommunication and confusion of patients (Box 8.1). Indeed, perceptions of discrimination in hospitals were also identified by villagers in the QPES community of M2. Another reason why Guatemalans do not seek "formal" health care is the widespread practice of traditional medicine (Box 8.1). Indeed, the extensive use of midwives for treatment might reflect not only a lack of alternative medical professionals, but also preferences reflecting a better cultural match between midwives and their clients (Box 8.1). 94 SUMMARY OF KEY ISSUES AND PRIORITIES Guatemala has made some progress on health since the Peace Accords, mainly on sectoral reforms. Notably: * The Peace accords made significant commitments to improving health and health care services; * Financial management of public health care was deconcentrated under the SLAF program; * The Government launched the decentralized Sistema Integrado de Atencion de Salud (SIAS) program and a National Health Plan; and * Public spending has shifted towards preventive care, which is crucial for treating the health problems faced by the poor. Nonetheless, significant challenges remain, particularly for improving health outcomes. Notably: * Key health outcomes - malnutrition, infant mortality, maternal mortality, and morbidity (e.g., diarrhea, ARI) - are not improving as fast as they should, and Guatemala remains among the worst health performers in LAC. Health outcomes are significantly worse among the poor, indigenous and rural residents, suggesting a need for better targeted interventions; * The extremely high prevalence of malnutrition - and its resistance to improvements - is particularly worrisome and should be deemed a top priority for public action; * A significant share of the population lacks access to affordable health services, particularly the poor and rural and indigenous residents (especially those who lack access to motorable roads); a combination of supply- and demand-side factors appears to be blocking improved access; * On the supply side, services are fragmented; insurance coverage is minimal; inefficiencies in public funding are generated by use of highly-subsidized public facilities by the few who are insured (virtually exclusively the non-poor); moreover, even when facilities are available, they often lack medicines, doctors or staff; * Public spending on health has not increased sufficiently and public spending is not well targeted to the poor; * On the demand side, economic barriers (direct costs of health care) present the main constraint to improved access. Although public health care is highly subsidized, private health care is relatively expensive. As such, in situations in which only private services are available, disadvantaged groups lack access due to economic constraints; * Cultural barriers further constrain access of the indigenous population to health care; and * Complements to health care - access to potable (not just piped) water and improved sanitation - are also lacking in their coverage, further exacerbating adverse health outcomes (see Chapter 9). In light of these challenges, a number of policy recommendations seem appropriate. Specifically: * Accelerated increases in public spending on health are needed to allow for improved coverage of public health services, with an emphasis on expanding preventive care, reducing and treating 95 infectious and parasitic diseases associated with poverty, and improving the key outcomes of infant and maternal mortality and malnutrition (including information and behavioral change programs); * Improving knowledge and access to effective family planning methods, especially in rural areas, would result in both a reduction of population growth and an improvement of reproductive and child health indicators; * Efforts should be made to better target public spending (and possibly behavior-conditioned demand- side programs) so as to make more equitable and efficient use of existing resources; the poverty map could be used as a tool to help target programs and future interventions; focusing on expanding coverage of, and adequate inputs for, rural health posts, community health centers, and traveling SIAS promoters in remote areas would also result in an improved degree of self-targeting due to the disproportionate use of these services by the poor; o Efforts should be made to promote culturally-sensitive health care practices, including working with traditional community-based health practitioners, recruiting indigenous health promotores, sensitizing hospital staff towards respectful treatment of indigenous patients, and information outreach programs; * A critical review of existing malnutrition interventions (across agencies, both public and private) should be conducted with a view towards (a) identifying programs that have worked (both in Guatemala and internationally); (b) streamlining and restructuring existing programs to better focus on young children, growth monitoring, information and behavioral change via community-based interventions. * Specific interventions should be undertaken to reduce malnutrition as a priority area. These include: (a) promotion of proper health, hygiene, and feeding practices; (b) growth monitoring of pregnant women and children under aged two; (c) micronutrient supplementation (particularly for iron); and (d) deworming treatments and oral rehydration therapy. The target population for these schemes should be pre-school children (particularly those under 24 months of age) and mothers (including pregnant and lactating women). To improve the effectiveness and reach of these programs (and avoid having them get "set aside"), these interventions should be integrated into the MSPAS basic health care package and provided at the community level through outreach workers selected by the community but contracted either by NGOs or the MSPAS. This would also allow for the institutionalization of nutrition activities within MSPAS, which would eventually contribute to the much needed rationalization of the myriads of independent, often incompatible, nutrition efforts in Guatemala. * Demand-side interventions (e.g., transfers conditional on behavioral change, growth monitoring, etc.) should be considered. They could be channeled through self-targeted health posts/community centers. * A full sectoral review should be conducted to examine more closely supply-side issues, including, inter alia: (a) an analysis of the institutional capacity of MSPAS; (b) an evaluation of the SIAS system, with an emphasis on identifying ways in which to improve its impact on health outcomes; (c) an assessment of the provision, gaps and duplication of services by the various health care institutions; (d) a full assessment of access - not only of access in terms of the physical availability of services but also in terms of the availability of affordable (public) services in rural areas; and (e) an analysis of the adequacy of complementary inputs such as staffing, medicines and equipment; * The Government should develop a monitoring system for health outcomes, including better and more regular measurement of maternal and child mortality; an analysis should also be undertaken to 96 further examine the determinants of maternal and child mortality so as to better tailor policy recommendations towards improving these outcomes; * In the medium term, the Government should seek additional steps to improve the efficiency and quality of services offered, such as: (a) introducing a system of referral and counter-referral; (b) implementing a hospital reform program that leads to a more efficient use of existing human and capital resources in public health care facilities (e.g., by introducing incentive schemes such as performance agreements); (c) charging the population covered by IGSS and private insurance for services received in the MSPAS facilities; (d) allowing the non-IGSS population to use IGSS facilities and charging the client rather than MSPAS for their use; and (d) considering contracting out certain services within IGSS and MSPAS to private services (in addition to those contracted to NGOs to expand the coverage of basic health care). ISee GUAPA Technical Paper 5 (Gragnolati and Marini, 2002) for the detailed analysis. 2 Sources: Guatemala: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estad/stica - Guatemala; rest of world: United Nations (2001). 3Source: Guatemala: DHS 1998. The average is 2.6 for LAC and 2.1 for lower-middle income countries. Source: World Bank (2001b). 4See GUAPA Technical Paper 5 (Gragnolati and Marini, 2002). 5 Nufiez (2001). 6 Estimates of infant mortality for Guatemala vary. The 1998/99 Demographic and Health Survey (DHS) puts the average at 45. The World Bank World Development Indicators (2001), based on official health statistics, puts the estimate for that same year at 40. 7World Health Organization. 8 See GUAPA Technical Paper 6 (Marini and Gragnolati, 2002) for details. 9 Source: DHS 1998/99 (survey does not allow for analysis by poverty group). 10 Diarrhea and ARI information from World Bank calculations using ENCOVI 2000, Instituto Nacional de Estad(stica - Guaterala. " Infant mortality indicators from DHS 1998/99. 12 Wealth measures include consumer items (e.g., television, cars), dwelling characteristics (e.g., flooring materials, water, toilet facilities), and other assets. Source: Gwatkin. Rutstein, Johnson, Pande, and Wagstaff (2000). 13 These stunting rates (height-for-age, HAZ) are consistent for various recent surveys, as are other measures of malnutrition (weight-for-age, WAZ, and weight-for-height (WHZ). Specifically, in 1995, the DHS recorded 49.7% for HAZ, 26.6% for WAZ, and 3.3% for WHZ; the 1998 DHS recorded 46.4% for HAZ, 24.2% for WAZ, and 2.5% for WHZ; and the ENCOVI 2000 recorded 44.2% for HAZ, 22:3% for WAZ, and 2.5% for WHZ. 1 Malnutrition indicators for Southem Mexico were as follows: 28.9% stunting (height for age), 1.6% wasting (weight for height), and 11.8% underweight (weight for age) in 1998/99 based on DHS data as reported by the World Health Organization (WHO). In contrast, comparable indicators for Guatemala that same year (DHS data for 1998/99) were: 46.4% stunting, 2.5% wasting, and 24.2% underweight. Estimates for Guatemala from the 2000 ENCOVI were: 44.2% stunting, 2.8% wasting, and 22.3% underweight. Results from the Intemational Biological Programme (Eveleth and Tanner 1976) and from more recent studies (Eveleth and Tanner 1990) reveal that, on a worldwide scale, height differences among children under five years from different countries are relatively small in comparison with large differences between and within countries due to environmental factors. Similarly, Martorell (1985), shows that, while differences due to social class were large, the differences that could be attributed to genetic factors were small (using a sample of height measurements from seven-year old boys from Brazil, Costa Rica, Guatemala, Haiti, Jamaica, Nigeria, India, and Hong Kong). Looking at growth data from well-to-do preschool children of different ethnic groups, Habicht et. al. (1974) reached the same conclusion. '5 The median duration of breastfeeding in the ENCOVI sample is 16 months, with longer periods among lower quintiles. 16 Using the Oaxaca-Blinder decomposition. 17 See GUAPA Technical Paper 6, Marini and Gragnolati (2002) for additional details on these schemes. Is This is based on data gathered for travel time to take children affected by common maladies (diarrhea or ARI). This could underestimate access since it doesn't account for (a) those who didn't seek treatment (perhaps because they lack access); or (b) those who didn't report the illness. '9 Calculations using the ENCOVI 2000 community questionnaire, which is not representative for all communities in Guatemala, but covers 60% of households in the household survey. 20 Annis (1981); Pebley et. al. (1997); QPES (2001). 21 These numbers exclude those who were ill but deemed treatment unimportant (sickness not bad enough to require health care). World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica -Guatemala. 22 See Annex 6 for additional details on the methodology for decomposing supply- and demand-side constraints to increased coverage of education, health and basic utility services using data from the ENCOVI. 23 For that same period, the per capita PPP averages were: Honduras US$210, Nicaragua US$266, El Salvador US$298, Panama US$410, and Costa Rica US$509. Source World Bank WDI 2001. 24 Some 17% of women in the poorest quintile report complications, as compared with 7% of those in the top quintile; 23% of women in the top quintile had cesarean sections as compared with 1.8% of women in the poorest quintile. Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. See GUAPA Technical Paper 5 (Gragnolati and Marini, 2002) for details. 75These numbers exclude those who were ill but deemed treatment unimportant (sickness not bad enough to require health care). World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica -Guatemala. 97 Chapter 9: Basic Utility Services and Poverty "Poverty is not having potable water in all the houses. " Mam Villager, Ml (QPES) This chapter seeks to analyze more deeply the issues pertaining to the coverage of basic services (electricity, water, sanitation, sewerage, telephone), with a view of informing policy and highlighting priorities for poverty reduction. Specifically, the chapter considers the issues of: (a) the benefits of basic services; (b) recent reforms and increases in financing for basic service expansion since the Peace Accords; (c) access and equity in the coverage of basic services; (d) barriers to improved access (supply- vs. demand-side); (e) the time and investments needed to achieve universal coverage; and (f) cost and subsidy issues. It concludes with a review of progress and remaining challenges in the basic services sector for poverty reduction. The chapter draws primarily on an extensive analysis of the ENCOVI 2000.' THE BENEFITS OF BASIC SERVICES Modern utility services offer important economic, productivity, and health benefits. Data from the ENCOVI reveal that households that have access to modem utility services reap important advantages:2 * First, the cost of modern utility services is often considerably lower than traditional alternatives. The clearest example is that of households without electricity who pay implicit prices of more than US$11 per kilowatt-hour (more than 80 times the price of electricity) to illuminate with candles and wick lamps and power appliances with dry cell batteries. * Second, access to modern services can substantially enhance the productivity of households and household-based micro-enterprises. Rural households with access to piped water and liquid propane gas for cooking save around six man-hours per week compared with households who must go out to collect water and fuel wood.3 Furthermore, micro-enterprises with access to water and electricity are twice as profitable than comparable enterprises without access to these services, and the effect of a cellular telephone on micro-enterprise profitability is even larger. * Third, some traditional substitutes for modern utility services are associated with adverse health impacts and may contribute to infant mortality. Although it is difficult to isolate the underlying causality, children from households with access to piped water and adequate sanitation are significantly less likely to suffer from diarrhea and overall physical stunting (malnutrition). These effects would likely be even stronger with improvements in the quality of piped water (e.g., potability). SECTORAL OVERVIEW Reflecting these benefits, the 1996 Peace Accords acknowledged the pivotal importance of modern utility services in the Guatemalan development process and made a commitment to expanding coverage to disadvantaged groups in order to make-up for historic neglect. This commitment was not an empty one and has indeed given rise to very significant and tangible changes in the utilities sectors in Guatemala. On the one hand, Guatemala took major steps to allow private sector participation and promote the development of competition. Specifically: * The national telecommunications operator TELGUA (previously GUATEL) was privatized in 1998, three new cellular licenses were issued, and the local and long distance markets were opened-up to immediate competition as part of one of the most radical liberalization processes in LAC. * The three main electricity distribution companies EEGSA, DEORSA, and DEOCSA were also privatized, and the generation market opened up to competition, although the state-owned enterprise INDE continues to hold have of the generation capacity and to control the transmission network. 98 * Water and sanitation was the only sector where reform measures did not Figure 9.1 - Social Fund Investments in Rural prove possible. As such, service Infrastructure since 1993 (sources: FIS, FONAPAZ, FSDC) continues to be provided by municipal so utilities in urban areas and 45 community-based organizations in 40 rural areas. The metropolitan region is 35- partly served by the state-owned 0 Electdcit enterprise EMPAGUA, but small- E Water a0 .sartation scale private sector operators also play X 15 - and important role. 10 5 At the same time, the volume of resources o channeled towards expanming rural service 1993 1994 1995 1996 1997 19 199 200 provision has increased substantially through a variety of new and existing institutional mechanisms. First, the investments made by the three main social funds (FIS, FONAPAZ, and FSDC) in rural electrification, water and sanitation, climbed from US$17 million in the period prior to the Peace Accords (1993-96) to US$152 million in the period since the accords (1997-2001). It is important to note, however, that this increase reflected an overall increase in social fund expenditures, rather than a shift towards infrastructure sectors. Moreover, there is evidence that water, sanitation and electricity investments by the social funds have begun to tail off (Figure 9.1). Second, in an example of a successful attempt to improve both equity and efficiency via market-based reforms, part of the sales proceeds from the privatization process were earmarked to finance rural service expansions. Thus, the US$110 million raised from the sale of the electricity distribution companies will be used to cover one third of the cost of a new Rural Electrification Program, which aims to connect 2,633 communities to the national grid during the period 2000/05. In addition, US$7.5 million raised from spectrum auctions for mobile telephony services were allocated to a special fund (FONDETEL) designed to support the expansion of public telephones in rural areas. ACCESS AND EQUrIY Current Coverage and Equity A significant share of Guatemalan households lack access to basic services. Overall, about 70% of Guatemalan households have water4 and electricity. Almost 90% have some kind of basic sanitation,5 though fewer than half have sewerage. About 20% subscribe to either a fixed line and/or a cellular telephone service. Around 16% of Guatemalan households do not have access to any kind of modern network utility service. Interestingly, households who only have one utility service (23% in all) are most likely to choose electricity, even when other services (such as piped water) are available in their communities. While overall coverage rates are average for Central America, they lag slightly the average for Latin America and other lower-middle income countries (Table 9.1). Table 9.1 - Internati nal Comparisons C verage of Basic Services (Percentage of ho seholds with access) Electricity Piped water' Basic sanitation* Telephone Guatemala 73 69 87 20 El Salvador 80 52 81 20 Nicaragua 69 61 84 16 Panama 79 86 93 41 LAC Average n.a. 85 78 n.a. Lower-middle income average n.a. 80 54 n.a. Notes: 'Piped water in dwelling or yard. * Includes toilets and latrines. El Salvador and Honduras quintiles based on income aggregate. Sources: El Salvador (Encuesta de Hogares de Propositos Multiples 1997); Guatemala (ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala); Honduras (Encuesta Nacional de Ingresos y Gastos de los Hogares, 1999); Nicaragua (LSMS 1998-99); Panama (LSMS 1997); averages for LAC and lower- middle income countries (World Bank World Development Indicators 2001). 99 Access to modern utility services is highly inequitable. While piped water w Figure 9.2 - Coverage of Basic Services by Quintile, and electricity are almost universal in ENCOVI 2000 urban areas, they reach little more than X 2 100 - half of rural households (Table 9.2). .1 00 Relative to the poorest quintile, the r. 60 richest quintile of the population are , 40 - electricaty twice as likely to have a water or n 20 - eanita electricity connection, and four times as 0 .U-fixed phone likely to have sewerage (Figure 9.2). e oZ ct ie Just over half of all poor households are X< connected to water or electricity. o5 0- Sanitation is more equitable, reaching close to 80% of poor or rural households. Almost no poor, rural, or indigenous households have telephone connections. Households living in the Norte and Peten regions are the most underserved for all services. About one third of rural households lack access to any kind of modem utility service. This figure rises to 40% for households in the - 6 lowest quintile. These inequities are typical in the Central American region. Table 9.2 - Coverage of Basic Services, by Area and Quintile (Percent of households in each group) National By area By quintile Urban Rural 1 2 3 4 5 Electricity 73 95 56 39 64 78 90 95 Water 69 88 54 50 62 63 76 92 Sanitation 87 97 79 73 80 88 95 98 Sewerage 38 76 09 06 18 32 54 81 Fixed telephone 15 31 03 0.3 01 03 14 58 Cellular telephone 10 18 03 0.1 01 03 11 34 Community public telephone 64 89 44 37 53 65 79 83 Lack access to any service 16 2 27 39 21 15 6 2 No service = lack of all network services and latrine. Network services = electricity, piped water in dwelling or field, telephone (fixed or cellular), and toilet connected to sewerage. Source: World Bank calculations using the ENCOVI 2000. Instituto Nacional de Estadfstica - Guatemala. Quintiles are individual consumption quintiles. Significant Progress Since the Peace Accords Guatemala has witnessed significant progress since the Peace Accords, in terms of expanding coverage and reducing inequities. Coverage has accelerated considerably in recent years, reflecting increased levels of investment in the utilities sector (as discussed above). Coverage indices for electricity, water and sanitation increased by about 15 percentage points in the period after the Peace Accords (1997-2000) versus about 10 percentage points for the period preceding the accords (1993-96). Taking into account population growth, the expansion of new connections was in general about 50% higher in the years following the Peace Accords.7 Furthermore, the acceleration of coverage was quite generalized affecting both urban and rural areas, as well as poor and non-poor populations. For telephones, the overall teledensity index rose almost fivefold from the period from 1997-01, largely due to the explosion of cellular telephony. Moreover, disparities in coverage were reduced, with new connections going disproportionately to traditionally disadvantaged groups. This is not surprising, since most other groups were already being served. To detect whether there has been an improvement in the targeting of services towards excluded groups, it is necessary to compare the probability that an unserved household in any particular category would become connected during the period immediately preceding and following the Peace Accords (Table 9.3). At the national level, the probability of an unserved household receiving a connection increased by approximately 80% for electricity, water and sanitation. All types of households experienced a significant 100 increase in the probability of being connected. Importantly, however, disadvantaged groups gained disproportionately, increasing their probability of being connected by well over 100% in most cases (Table 9.3). For example, the probability of being connected to electricity increased by 183% for the extreme poor, 115% for the poor, and 90% for the non-poor. This disproportionate gain has not been enough, however, to compensate for their historic disadvantage. Thus, notwithstanding the large gains, in absolute terms, the probability of being connected to electricity for a poor household (0.28) is still lower than for a non-poor household (0.55). Table 9.3 - Probability that an Unserved Household was Connected (Proporti n of unserved households receiving a connection) Electricity Piped water4 Sanitary services' National 1993-1996 .19... .19... 31... 1997-2000 .36 .34 .55 % change 89% 79% 77% Urban 1993-1996 .38- .31.. .50.. 1997-2000 .70 .53 .82 % change 84% 71% 64% Rural 1993-1996 .13. .14 .22... 1997-2000 .29 .28 .48 % change 123% 100% 118% Extreme poor 1993-1996 .06 .13 .21- 1997-2000 .17 .26 .37 % change 183% 100% 76% All Poor 1993-1996 .13 .15.. .25... 1997-2000 .28 .29 .44 % change 115% 93% 76% Non-poor 1993-1996 .29 .24* .38 1997-2000 .55 .41 .72 % change 90% 71% 89% Indigenous 1993-1996 .16 .18 .30 1997-2000 .30 .32 .52 % change 88% 78% 73% Non-indigenous 1993-1996 .21* .19-- 31-* 1997-2000 .42 .35 .57 % change 100% 84% 84% Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Notes: The null hypothesis of equality of the probability of coverage before and after Peace Accord is rejected at: ^ 99% level. 95% level, 90% level. 'Piped water in dwelling or yard. 'Includes toilets and latrines. For telephones, although the richest quintile accounts for about half of new cellular subscriptions (as second telephones), there is evidence that ceilular telephones are having a wider social impact. In rural areas, cellular telephones are as common as fixed lines and two-thirds represent first telephones for households that lack a fixed-line service. Moreover, there is considerable anecdotal evidence that cell phones are being used to provide an informal public "pay phone" service for rural communities. Rural access to telecommunications has improved substantially, with the number of rural public telephones increasing by 80% since the Peace Accords. Thus, 50% of rural households now have a public telephone in their community, and 80% live within 6 kilometers (about half an hour) of a public telephone.9 101 Remaining Coverage Gaps Notwithstanding this momentous progress, a significant coverage gap remains. Well over half a million households are still without electricity and piped water (Table 9.4). Some 200,000 are without any form of sanitation, while about 1.3 million rely on latrines as opposed to conventional sewerage. The households that remain unserved are predominantly rural and predominantly poor. Table 9.4 - Coverage Gap for Modern Utilities (Number of unserved households) Electricity Piped water' Basic sanitation' Improved sanitation Total no. of households National 585,933 686,893 288,807 1,353,895 2,191,451 By area Urban 45,189 113,235 24,156. 224,291 951,654 Rural 540,744 573,658 264,651 1,129,604 1,239,797 By quintile I 266,931 220,182 116,340 411,318 438,437 2 155,116 163,797 84,249 349,173 427,908 3 98,428 164,199 52,064 304,708 446,068 4 44,513 104,894 25,003 .203,850 442,583 5 20,945 33,821 10,161 84,846 436,455 % unserved - Nat'l 27 31 13 62 n.a. ANale. 'piped water in dwelling or yard; 'includes toilets and latrines. Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Quintiles are individual consumption quintiles. BARRIERS TO IMPROVED ACCESS: SUPPLY VS. DEMAND-SIDE The availability of basic infrastructure is not the only constraint to improved access - demand-side factors play an important role too. Disentangling these constraints is crucial Figure 9.3 - Decomposing the Coverage Deficit: Supply- or for policy decisions. For some nO Demand-Side Constraints? (ENCOVI 2000) households, the service is simply not 80% 21 25 available in the communities in which 0 El % demand side they live. This is essentially a supply-side a 60% %both D&S problem that requires increased e 20% % supply side investment in infrastructure expansion. 0% Other households, however, simply fail to 0 take-up the service even when it is . c' available in the community. This is essentially a demand-side problem that will not be solved by building more infrastructure. The relative importance of demand- and supply-side constraints is revealed by decomposing the coverage deficit into the share of unserved households that: (a) live in communities where the service is available but don't connect (pure demand-side constraints); (b) would connect if the service were made available (pure supply-side constraints); and (c) that live in communities where the service is not available but would still not connect even if it were (mixed supply- and demand-side constraints). According to this decomposition,'0 between 20-40% of the coverage gap is caused purely by demand-side factors and could be resolved without major investments in infrastructure (Figure 9.3). Between 32-59% is caused primarily by supply-side factors. Finally, between 8-44% of the coverage gap, depending on the service, would require both physical expansion and demand-side measures. Economic factors are important demand-side barriers for lower-income households. Controlling for other factors, household consumption is significantly correlated with the take-up of all modem utilities, except electricity." This finding suggests that connection charges for all services may represent a barrier for lower-income households. Indeed, connection charges for all services except electricity in urban areas 102 represent a significant proportion of the poverty line. Furthermore, the cost of connecting to utility services often goes beyond the direct connection charge. There is often a substantial complementary investment that must be made in adapting the dwelling to the new service. For example, internal wiring for electricity can cost around US$100, while internal plumbing for water and sewerage can cost several hundred dollars. In addition, once connected, households face recurring costs of using the service, as discussed further below. Other significant demand side factors include gender (male headed households less likely to connect), years of education, ethnicity (indigenous less likely to connect to sewerage and telephone), the presence of a household business (more likely to be connected to electricity and telephone), and area (urban more likely to connect to all services). ACHEVING UNIVERSAL COVERAGE: TIME AND COSTS It will take more than eight years to reach universal coverage for all services except basic sanitation, given current levels of population growth and actual average annual rates of service expansion (Table 9.5.12 Only a doubling of current rates of expansion, or a stabilization of population, would permit universal coverage to be reached in the medium term; that is between 3 to 12 years depending on the service. Table 9.5 - How Far Away is Universal Coverage? (Anticipate date of universal cover ge) Present effort levels Present effort levels sustained doubled Electricity 2006 2003 Water 2007 2004 Basic sanitation 2003 2002 Improved sanitation 2014 2007 World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Based on typical unit costs for service expansion, the total cost of meeting universal coverage across the electricity, water and sanitation services is estimated at US$1.5 billion.13 The electricity service, owing to its relatively high unit cost, accounts for over 40% of this total expenditure, compared with 24% for piped water. Sanitation would absorb about 32% of these costs, with. public telephones accounting for the remainder (about 4%).4 For water and sewerage services, the costs of universalizing access could be reduced by as much as 40% with a "condominial" approach whereby a single branch from the main network is provided to a whole block ("condominium") of houses, who then make their connections along this common branch instead of providing a separate branch to each household.'5 It is important to note, however, that these estimates assume that supply-side factors account for the full coverage gap. The decomposition of the coverage gap above suggests that this is not the case, and that a significant share of the deficit could be bridged by removing demand-side barriers that prevent households from connecting to existing networks. Overall, it is estimated that this factor could reduce the cost of meeting universal access by as much as 30%, from US$1.5 billion to US$1.0 billion, provided appropriate interventions were instituted. COSTS AND SUBSIDIES FOR BASIC SERVICES Households spend about 10% of their budgets on basic services. Over half of this expenditure goes to energy for cooking and heating, and another quarter covers the costs of lighting and powering appliances. Less than 1% of total consumption is spent on water services. The overall budget share is higher for households in poorer quintiles, though the composition shifts away from cooking fuels towards telecommunications for richer households (Figure 9.4). 103 In Guatemala, there has been a conscious policy decision to use subsidies to keep water Figure 9.4 - Household Spending on Basic Services and electricity tariffs artificially low. To some (ENCOVI 2000) extent, this is understandable, given that providing o 14 access to utility services is only ultimately .2 meaningful if these services are available to poor households to use. However, the evidence 8 10 * coin de suggests that the policies have been mistargeted, 0 8 ightingandappliances failing to reach desired outcomes, and that their a Water disadvantages are substantial. 0 Telecoms In the electricity sector, subsidies via the "social 2 tariff" are not well targeted to the poor. D o1 Introduced following the privatization of 1 2 3 4 5 Total distribution companies, the objective of the "tarifa Consumption quintile social" is to keep domestic tariffs for those consuming up to 300 kilowatt-hours per month capped at a fixed price of US$0.08 per kilowatt- hour.16 An analysis of the ENCOVI, however, reveals that this costly subsidy has been misguided for several reasons. First, the thresholds that have been set for the social tariff are very high in relation to typical residential consumption. The average household consumes 102 kWh per month, with poor households consuming 48 kWh per month on Figure 9.5 - Target Incidence of Bectricity average and non-poor households consuming 128 Subsidies, ENCOVI 2000 kWh per month. As a result, virtually all 80 residential customers - rich and poor alike - 0- qualified for the subsidy.t7 Second, coverage of O60 the poor is relatively low owing to relatively low 40 22 connection rates among poorer households. While 20 2 3 131iiii three quarters of households in the top quintile 04 t benefit from the subsidy, only 38% of those in the P2orest Q2 03 04 Rchest bottom two quintiles benefit. About 65% of the Q1 Q5 beneficiaries are non-poor households. Third, the subsidy is regressive in terms of its distributional incidence. The richest quintile receives two thirds of total subsidies transferred, as compared with the poorest quintile, which only receives 2% (Figure 9.5).18 By poverty group, 90% of the benefits accrue to the non-poor. Simulations using ENCOVI data suggest that lowering the threshold from 300 to 100 kWh per month would improve matters somewhat by lowering the number of non-poor beneficiaries. Nonetheless, even with this lower threshold, an estimated 75% of benefits would still accrue to the non-poor. A more pro-poor policy would be to instead channel these resources towards expanding coverage of electricity to unserved households. As discussed above, households without electricity pay some 80 times more per kilowatt hour for traditional fuel sources. From this perspective, it would appear to make much more sense to channel the US$50 million annual cost of the 'tarifa social' towards increasing connections to unserved households. It is estimated that an additional 50,000 new connections each year could be financed in this way. Moreover, given that over 70% of households without electricity belong to the poorest segments of the population, such a policy would be quite well targeted. In the water and sanitation sector, tariffs are well below true economic costs and international parameters of willingness to pay. Households pay bills of less than US$2 per month in Guatemala City, and less than US$1 per month in other urban areas. The underlying tariffs are barely US$0.10 per cubic 104 meter compared with typical costs of around US$0.40 per cubic meter for the Latin American region. In spite of these low tariffs, as many as 30% of households with piped water reported that they did not pay for the service in the last month, compared with only 8% for electricity. As a result, households spend barely 0.5% of their budgets on water and sanitation services, which is a fraction of the 3%-5% World Health Organization guideline for what households are typically willing to pay. Moreover, many households spend three times as much on bottled water as on piped water. While low water tariffs may seem attractive, they are also associated with poor service quality. There is substantial evidence that the precarious financial position of water utilities is contributing to a very poor quality of service. In fact, piped water in Guatemala is generally not potable. Three quarters of households with piped water feel it necessary to either buy bottled water or perform some kind of self-treatment. It is particularly striking that the practice regular boiling drinking water is equally prevalent among households with and without piped water (some 40% of both groups). Moreover, water service is irregular. Households report that on average they receive only 17 hours of water per day and face 3.6 days per month without water. Such findings were confirmed in the QPES, where informants in most communities complain about non-potable water and irregular service provision. SUMMARY OF KEY ISSUES AND PRIORrrIES There has been significant progress in expanding the equitable provision of basic utility services since the signing of the Peace Accords in 1996. Notably, e Sectoral reforms have improved competition and efficiency; o The volume of resources channeled towards the expansion of rural service provision has increased substantially through a variety of new and existing institutional mechanisms; e Overall coverage of basic services has accelerated considerably since 1996; and * This expansion has been well-targeted, with new connections going disproportionately to traditionally disadvantaged groups. Nonetheless, important challenges remain. In particular, o Despite improvements, significant coverage gaps remain; o Access to modem utility services remains highly inequitable; * A significant share of those without access to basic services live in communities where the services are present but do not connect due to demand-side barriers, such as the direct costs of connecting to and using services; as such, interventions other than simply supplying the basic infrastructure will be needed; * Given existing rates of expansion, it will take more than eight years to reach universal coverage for all services except sanitation, and the total cost of meeting universality is estimated at between US$1-1.5 billion; o Energy subsidies (under the "tarifa social") are poorly targeted, benefiting primarily the non-poor; and o The quality of piped water services is poor (non-potable and irregular). 105 A number of policy recommendations seem appropriate in light of these challenges. Specifically: * To maintain and, if possible, increase the current level of resources channeled towards the expansion of modern utility services so as to reach universal coverage within a 10-year horizon. * To better target service expansion efforts to traditionally disadvantaged groups, in particular, poor, rural and indigenous households (e.g., using the poverty map). * To develop a strategy for removing the barriers that prevent a significant proportion of excluded households from making connections to services even when these are available in their communities. * To find new financial resources for the FONDETEL rural telephony program and to consider using these to subsidize the extension of cellular networks into commercially marginal areas. * To reform the 'tarifa social' policy by at least reducing the eligibility threshold to 100 kilowatt-hours per month, and preferably replacing it with a program to fund new connections. * To allow water tariffs to rise to a level that allows water utilities to become financially sustainable and thereby improve the quality of service that they offer to the public. * To encourage social funds and other providers to consider measures to improve the quality of water when expanding coverage * To complement expansion of water and sanitation programs with measures to improve household hygiene practices so as to reap the full health benefits of the service. * To complement expansion of electricity and telecommunications coverage in rural areas with measures to promote the productive use of these services by micro-enterprises. See GUAPA Technical Paper 7 (Foster and Araujo, 2002) for a detailed presentation of this analysis. 2 See GUAPA Technical Paper 7 (Foster and Araujo, 2002) for a detailed quantification of these benefits. 3Moreover, the ENCOVI demonstrates clear gender specialization in collection activities, with men and boys accounting for 65% of the labor devoted to the collection of fuel wood, and women and girls accounting for 74% of the labor devoted to the collection of water. See GUAPA Technical Paper 7 (Foster and Araujo, 2002) for details. 4Defined as piped water to property (dwelling or yard). Not necessarily potable. 5Broadly defined to include latrines, septic tanks, and sewerage. 6 See GUAPA Technical Paper 7 (Foster and Araujo, 2002). 7See GUAPA Technical Paper 7 (Foster and Araujo, 2002). These differences were statistically significant. 8 This has the effect of normalizing the number of new connections received against the size of the corresponding unserved population in each group. 9 For households in the ENCOVI community questionnaire sample. 1° See Annex 6 and GUAPA Technical Paper 7 (Foster and Araujo, 2002) for details on this decomposition. " This paragraph presents the results of multi-variate probit regressions in which take-up is the dependent variable (for all households that live in communities in which the service is available). See GUAPA Technical Paper 7 (Foster and Araujo, 2002) for details. 12 Present average rates of service expansion were around 115,000 new connections for electricity, water, and sanitation. Population growth rates are around 2.6% p.a. See GUAPA Technical Paper 7 (Foster and Araujo, 2002). These projections differ from those presented in Chapter 5 as they are based on projections of rates of service expansion that were observed since the Peace Accords, where as those in Chapter 5 are based on average intemational rates of service expansion (worldwide panel of growth-service elasticities). 13 See GUAPA Technical Paper 7 (Foster and Araujo, 2002) for details. 14 In the case of sanitation, two levels of universal service are defined. The first level is universal basic sanitation, which basically entails providing latrines to the 288,807 households that currently have no form of sanitation, and would cost less than US$15 million to achieve. The second level is universal improved sanitation. This entails providing sewerage to all households in conurbations with greater than 50,000 population (notably the Metropolitan area, Quetzaltenango, and Escuintla), and upgrading all other households to a flush toilet with a septic tank. This is a very much more expensive proposition, accounting for almost a third of the overall expansion costs. See GUAPA Technical Paper 7 (Foster and Araujo, 2002). I5 Foster, 2001. See GUAPA Technical Paper 7 (Foster and Araujo, 2002). 6 This threshold of eligibility was reduced from 500 kWh to 300 kWh per month in January 2001, leading to an estimated cost savings of US$7.1 million annually. The new law also obliged distributors to finance the subsidy via cross-subsidization across consumer categories (transferring the subsidy from industrial users to households). Prior to these changes, the cost of the subsidy, estimated at US$57 million per year, was met by INDE on the basis of state transfers. 17 An estimated 99% of residential customers qualified for the social tariff under the original threshold of 500 kWh, and 94% still qualify with the recently lowered threshold of 300 kWh. Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Is These incidence estimates differ slightly from those presented in Chapter 12 because they assessed at the household (rather than per capita) level. 106 Chapter 10: Transport, Poverty and Isolation "During the rainy season, the community is affected because the road is blocked by the mud and the landslides... The result is that we can't buy or transport products to the market. " Mam Villager, M2 (QPES) As an "intermediate good," transport' is widely viewed as critical to reducing poverty and promoting growth by providing access to opportunities, markets, and basic services. Transport also empowers people, providing them with physical, social, and political access. Finally, easy access also reduces vulnerability to natural and man-made disasters. This chapter seeks to provide a "poverty-transport" profile that can 2 simultaneously inform strategies to reduce poverty and expand rural access. Specifically, this chapter will examine (a) the correlations between poverty and road access; (b) the issues of road dependability and quality; (c) the correlations between poverty and access to public transport services; and (d) the effects of physical isolation in terms of increased travel times associated with a lack of access to motorable roads. The chapter concludes with a review of key challenges and recommendations for the transport-poverty nexus.3 ACCESS TO ROADS Relative physical isolation is not uncommon in Guatemala due to a limited road network. As discussed in Chapter 4, the country is not physically united and many villages are fairly isolated, with long- inter-village distances, due to an inadequate road network. The classified Box 10.1 - Classification of Roads in the ENCOVI road network is approximately 14,0004 *Type A roads are surfaced (asphalted or metaled) and accessable to kilometers in length (4,000 km of main motorized transport (carretera pavimentada o balustrada). and secondary roads that are mostly *Type B roads are gravel or unsurfaced and accessible to motorized transport (carretera de tierra o terraceria), though their surface paved and 10,000 km of unpaved qultvaisgely ' ~~~~~~~~quality varies greatly. tertiary and rural roads), or 1.2 km per *Type C roads are dirt roads or tracks (caminos de herradura sin 1,000 people. This compares with Costa balastre) and are mainly used for beasts of burden or walking. Rica - a country that is half the size of *Type D consists of paths (veredas) that are mainly for pedestrians. Guatemala - which boasts a road *Motorable roads consist of Types A and B. network of about 35,600 kilometers, or 11.1 kilometers per 1,000 people. Indeed, access to roads is limited, with a non-poor and urban bias. Some 13% of households in the ENCOVI PSU sample5 lack access to a motorable road (surfaced or unsurfaced) (Table 10.1). This share is higher in rural areas than urban. It also rises with poverty with a higher share of those in the lower quintiles lacking access (Figure 10.1). Regionally, a lack of access to motorable roads is highest in the Region Figure 10.1 - Lack of Access to Paved and Nor-Occidente and Norte, which are Motorable (paved+unpaved) Roads, by Quintile, also the regions with the highest rates of ENCOVI 2000 poverty (Figure 10.2). This geographic ' 100 correlation between poverty and road 8 80- _ access can also be seen in more detail in D 60 - -4No Paved 4o 0 No ave Figure 10.3, which overlays the poverty - -4-NO Motorable map with the roads map. The patterns o 20 for access to surfaced roads are similar, a o - but with a larger share lacking access 0 Q1 Q2 Q3 04 Q5 (60% overall, with over 70% of ENCOVI PSU households in the lowest 107 two quintiles lacking access). The apparent correlation between poverty and the road network could be by design. In other words, since roads are not randomnly distributed, it is likely that they were placed according to some sort of strategy that likely favored the development of the coffee sector and the finca zone, as well as urban areas. Nonetheless, regardless of the direction of causality, the fact remains that the poor are relatively more access-constrained, which limits their opportunities and access to basic services, as discussed below. ROAD QUALITY AND DEPENDABILITY Road reliability is also a problem. Road closures are common in Guatemala. Indeed, while significant progress has been made over the period from 1994-2000 in the rehabilitation and maintenance of the surfaced road network (such that 75% is in good or fair condition), only 45% Figure 10.2 - Poverty and Road Access, by of the unsurfaced classified road Region, ENCOVI 2000 network is in maintainable condition.6 Some 28% of households in the 100 N ENCOVI sample report road closures. 80 - 111 No Paved Close to three quarters of these (or 40 -l No Motorable 20% of all ENCOVI households) 20 -Ah Poverty Rates 0 report closures lasting longer than five days. While the differences between urban and rural areas are not c significant, road closures are more common in the Norte (34%), Sur- Oriente (48%) and Central (32%) regions. The main causes of road closures according to households in the ENCOVI PSU sample are: flooding and winter (36%) and landslides (28%). There appears to be a slight correlation between road type and dependability, with a larger share of households (32%) without access to motorable roads reporting road closures. These findings are consistent with those in the QPES: several villages (e.g., M2 and KAI) note that heavy rains make their single dirt access roads impassable, completely isolating the villages during the rainy season. Road closures further constrain access to opportunities and services. The main activities affected by road closures include: school attendance (36% of those reporting closure), work (28%), and market access (11%). The interruptions of access to work and school are reportedly higher in rural areas. Households in the poorest quintile (45%) are more likely to report interruptions in their lack of access to work than those in the richest quintile (12%). Road quality appears to have improved since the Peace Accords, but with improvements favoring the urban and non-poor populations. Overall, 44% of households in the ENCOVI PSU sample perceive improvements in road quality since 1995, with 16% perceiving worsening conditions. Perceptions are generally more favorable in urban areas (53% perceive improvements in quality) than rural (only 40% perceive improvements). They are much more favorable among non-poor than poor (55% among top quintile vs. 34% among poorest quintile perceive improvements). Finally, there are significant differences by region, with more favorable perceptions in the Metropolitan (53% perceiving improvements) and Sur- Occidente (57%) regions and much less favorable in the Norte (only 28%) and Nor-Occidente (22%) regions. Again, this seems to mimic an apparent historical bias towards urban areas and the finca zone, since many of the municipalities in the Sur-Occidente are dominated by plantations. 108 Table 10.1 - Access to Roads by Type of Road (Households for which PSU level information is Available) HHs with Access to Road Type/ Total s (%) Among those with access to Road Type: Share by Characteristic (%) Type A Type B Type C Type D No Motorable Roads* Type A Type B Type C Type D No Motorable Roads* All 39.8 69.9 37.9 62.5 13.3 100 100 100 100 100 Urban 65.3 58.7 22.3 34.9 10.1 51.1 26.1 18.3 17.4 23.9 Rural 28.3 75 44.9 75 14.7 48.9 73.9 81.7 82.6 76 Poverty Levels Non-Poor 52 67.4 32 58.8 8.9 60.4 44.7 39.1 81.1 31 All Poor 29.4 72.1 43 86.1 17 39.6 55.3 60.9 18.9 69 Extremely Poor 20.5 69.4 42.4 86.1 23.4 7.1 13.6 15.3 18.9 24.3 Regions Metropolitan 65.2 57.3 22.3 34.3 3.2 33 16.5 11.8 11 4.7 North 25.8 66.3 15 75.1 17.8 6.5 9.5 4 12.1 13 North-East 18.7 77.2 51.2 34.9 18.3 4.3 10.2 12.5 5.2 13.7 South-East 32.6 73.4 69.9 77.8 14.4 7.9 10.2 17.8 12 10.1 Central 47.4 83.6 30 61.5 9.5 12.4 12.5 8.3 10.3 7.4 South-West 37 76.2 38.7 69.3 16.1 22.7 26.7 25 27.1 31 North-West 29.3 61.7 48.2 91.7 20.5 9.6 11.5 16.6 19.1 19.5 Peten 47.7 70 51.1 67 2.3 3.6 3 4.1 3.2 0.5 Quintiles First 20.7 70.5 43.5 86.3 22.3 9.2 17.8 20.3 24.4 29.7 Second 29.6 71.7 44.1 77.9 16.4 15.1 20.9 23.7 25.4 25.1 Third 37.1 74.4 40.4 61.3 11.3 18.7 21.4 21.4 19.7 17.3 Fourth 48.3 70.8 33.7 52.5 8.9 26.4 22.1 19.4 18.3 14.4 Fifth 60.7 62.2 28.7 38 8.9 30.6 17.8 15.2 12.2 13.5 Ethnic Groups Non-Indigenous 44.5 69.5 41.8 49.7 11.2 62.5 55.7 61.8 44.6 47.2 Indigenous 33.9 70.4 32.9 78.9 15.9 37.5 44.3 38.2 55.4 52.8 Kiche 43.6 71.7 27.7 86.9 4.4 9.8 9.1 6.5 12.4 2.9 Qeqchi 21.3 59.3 13.2 67.6 23 4.5 7.1 2.9 9 13.9 Kaqchikel 43.1 86.6 30.2 69.2 14 9.5 10.8 7 9.7 9.9 Mam 28.9 72.6 41.4 89.7 17.9 6.3 9 9.5 12.5 11.7 OtherIndigenous 32.1 61.9 50.2 80 20.5 7.5 8.2 12.3 11.9 14.4 Type A road is surfaced (Ca rrerera pavimentada o balastrada); Type B = graveVunsurfaced (Carretera de tierra o terraceria);,Type C = dirt roads/tracks (Carretera de herradura sin balastre); Type D = paths (Veredas). HHs can have access to more than one type of road. A motorable road is a Type A or Type B road. Source: World Bank calculations using ENCOVI 2000, nstituto Nacional de Estadistica, Guatemala 109 Figure 10.3 - The Poverty Map and The Road Network Map, by Municipio Guatemala: Percentage of population poor, by municipio, 2000 Roads Y .~~~~~~1 No Asatd i -~~' /C (L - 3 X iL g ~~~~~0- 90 90 - 100 110 PUBLIC TRANSPORT Access to public transport is limited in Guatemala, with biases towards urban areas and the non- poor. Less than half of the households in the ENCOVI PSU sample report access to public transport. Access is Figure 10.4 - Access to Public Transport, % of Households in PSU Sample with Access reportedly higher in urban areas than ENCOVI 2000 rural and among the non-indigenous 70% than the indigenous (Figure 10.4). 680% There also appears to be a strong 400 - _o correlation between economic status and 20% ___l=__ .i__............. access to transport services: 66% of 0% - those in the top quintile have access o , .& dO ' compared with only 29% in the poorest qI ° \9 X5 quintile. This is important because the Z poor also generally lack alternative modes of transport (few own vehicles for example). There are significant differences in access by region, with three quarters of those in the Metropolitan region reporting access versus one third in the Norte, Nor-Oriente, Sur-Occidente, and Nor-: Occidente regions. When available, demand for public transport services is high. Among those who report having access to public transport services, 92% report using it every day. While 41% of households in the ENCOVI PSU sample perceive improvements in public transport since 1995, many also report problems with public transport. The main problems cited are insufficient busses (51%), particularly in rural areas (56%). The second main problem cited is the bad condition of busses, particularly in urban areas. The road network seems to be a constraining factor for expanding public transport services. There is a significant correlation between the motorable road network and availability of public transport. While half of all households in the ENCOVI PSU sample with motorable road access report access to public transport services, only a quarter of those without access to motorable roads have access to public transport services. This effect is particularly strong in rural areas. THE EFFECTS OF ISOLATION: LIMITED ACCESS TO SERVICES AND OPPORTUNITIES Access to motorable roads confers significant benefits in terms of reduced travel times to key services and opportunities. Travel times were compared for those Figure 10.5 - Access to Health Service: Travel households with and without access to Times With and Without Access to Motorable motorable roads using data from the Roads, ENCOVI 2000 ENCOVI PSU sample. The results v 80 t5 8l suggest rather striking benefits of road 5 60 4 3 5 48 5 8 Qomotorable access: Z! 40 3511 351 43i Dn-3 it motorable 20- * Health Services. Travel C 0 times to health services are o° o' ° significantly longer for those 4°c without motorable roads 'E,+ access (Figure 10.5). Physical isolation is particularly serious in rural areas and among the poor: for these populations, a lack of motorable roads puts health services just out of reach according to the WHO definition of access (travel times of less 111 than an hour). These findings are consistent with the QPES: two villages (KAl and M2) note that inadequate access roads constrain their access health services, particularly when the rains make their single dirt access roads impassable. In fact, when discussing vulnerability, the villagers of KAI specifically identify "giving birth" as a risk because laboring mothers cannot access health services due to inadequate road access, particularly in the rainy season. Interestingly, road access has no effect on urban travel times, likely reflecting the much stronger coverage of urban populations by health services (Figure 10.5). * Access to Opportunities (Markets, Commercial Inputs). In rural areas, road access also appears to be a significant determinant of access to Figure 10.6 - Access to Opportunities: Travel Times opportunities, such as markets With and Without Access to Motorable Roads, Rural and other commercial inputs Areas ENCOVI 2000 (Figure 10.6). Travel times to markets and post offices, for C 120 Owith motorable example, are almost doubled 8100 -nomotorable for rural households in the 60 _{] r n ENCOVI PSU sample without 20 __;|___i _r _____, access to motorable roads. o_ l__ Access to inter-urban and urban bus stops is * significantly constrained by a P00 qo' N-3 -.5 lack of motorable roads. Travel times are also longer for urban households without access to motorable roads (compared to those with access), but still shorter than those in rural areas (even with motorable roads), reflecting a much more extensive network of markets and commercial services in Guatemala's cities. Moreover, as discussed in Chapter 6, households located in smaller municipalities (with less infrastructure including roads) have higher chances of being poor and fewer employment opportunities than those living in larger municipalities (with more extensive infrastructure networks). * Access to Institutions. Likewise, road access seems _ to play an important role in Figure 10.7 - Access to lnstitutions:Travel Times With and Without Access to Motorable Roads, Rural determining access to Areas, ENCOVI 2000 institutions in rural areas. n Access to police and I 120- municipal services, for S 80 _ motorable example, is significantly 0 40_, 1F i _4 constrained by limnited road 5 20 - -i F _ nomotorable access (Figure 10.7). Church access is the least dependent pe '° , e" 05 on the road network, o ,. reflecting the extensive cP presence of churches in rural areas. This finding is consistent with the QPES, in which all but one of the ten communities report at least one church, with most having more than one (and one village, KI1, boasting 17 churches). 112 SUMMARY OF KEY ISSUES AND PRIORITIES Although households do report improvements in the road network and public transport, important challenges remain to extend this infrastructure to the poor, particularly in rural areas. In particular: * Rural residents and the poor are relatively more isolated. There is a significant inverse correlation between access to the motorable road network and poverty; * Year-round access is also crucial, with road closures from rains and landslides further cutting off access to opportunities and services; * The poor also lack access to public transportation, which appears to be correlated with a lack of adequate road networks, particularly in rural areas; * Road improvements implemented since the Peace Accords appear to have favored the non-poor and urban residents; and * Inadequate road access significantly constrains the access of the poor and rural residents to health services, opportunities, and institutions, further exacerbating their isolation. In light of these challenges, policy makers should seek to improve the targeting of future transport investments. First, a shift in focus towards rural roads is needed. Second, while decisions on future transport investments (improvements, rehabilitation, and new roads) clearly. must take into account economic calculations (population density, economic returns), equity concerns should be given more weight than has been done in the past. The poverty map, combined with the roads map and other information on economic activity and service access, should be used by policy makers to determine funding allocations and project location decisions so as to better favor the poor. Transport in this means "accessibility-providing infrastructure and services," including roads and transport services. 2 A subsequent phase of the work will attempt to quantify the impact of transport infrastructure on service access and poverty. 3This chapter is largely based on an analysis of the ENCOVI 2000 that was jointly sponsored by the GUAPA Program and the World Bank's transport team. See GUAPA Technical Paper 8 (Puri, 2002) for details. 4 The classified road network covers surfaced, gravel, and unsurfaced roads. It does not include dirt roads, tracks or paths - which are the only form of access for numerous villages in Guatemala. This unclassified road network is estimated at 12,000 km. Source: Direcci6n General de Caminos. 5 Road information is only available for households in the ENCOVI sample for which community-level data were collected. Community level data was collected for 481 Primary Sampling Units (PSUs). Due to difficulties in administering the community questionnaire in certain areas (particularly urban areas, where "community" boundaries are less clear), this represents about two-thirds of all households in the ENCOVI household survey (with relatively higher coverage of rural areas). All ENCOVI information presented in this chapter refers to this sub-sample, which, while large, is not statistically representative of the population. See GUAPA Technical Paper 8 (Puri, 2002) for details. 6 Project Concept Document for the proposed new Rural Roads Project. 113 PART 3: KEY CHALLENGE: REDUCING VULNERABILITY "As for the earthquake (of 1976), it affected the families because the majority of houses were destroyed. Most had to build shelters between thefields and in the road.... Many still haven't repaired their houses." Kaqchiqel villagers in KA I (QPES) The limited assets of the poor (and the near-poor) - discussed in previous chapters - also makes them particularly vulnerable to the impact of adverse shocks as they lack the means to be able to cope with them. With a wave of recent shocks in Guatemala (Hurricane Mitch in 1998, the recent coffee crisis, droughts and recent deaths from extreme acute malnutrition), the issue of vulnerability has taken center stage in policy discussions. This section of the report brings the "vulnerability lens" to the traditionally static, asset-based poverty analysis. It thus shifts the emphasis from a passive or reactive approach (given poverty, what can be done to reduce it?), to a dynamic or proactive approach (given vulnerability, poverty and risks, what can be done to get help the current poor escape from poverty and reduce the likelihood that others will fall into poverty?). It also helps further characterize the nature of poverty in Guatemala: (a) is it more chronic, with a mass of people statically living in poverty and transmitting it across generations? or (b) is it more transient, with many moving in and out of poverty? or (c) is it a specific set of sub-groups that are chronically poor and vulnerable due to specific features or circumstances? The implications of each of these scenarios for policy and targeting are clearly quite different. Chapter 11 provides an operational assessment of vulnerability, while Chapter 12 reviews existing social protection and social risk management mechanisms to assess their adequacy and offer insights into ways in which to strengthen them. Chapter 11: Vulnerability and Vulnerable Groups This chapter presents the findings of an operational assessment of vulnerability. Vulnerability comes from the notion that certain groups in society are more vulnerable to shocks that threaten their livelihood and/or survival. Other groups are so vulnerable that they live in a chronic state of impoverishment where their livelihood remains a constant state of risk due to certain structural features (structurally or chronically vulnerable groups). The concept of vulnerability as used here has two elements: (a) a person or household's resilience to a given shock, which is largely based on the portfolio of assets at their disposition (the higher the resilience, the lower the vulnerability); and (b) the severity of the impact of the shock (the more severe the impact, the higher the vulnerability). The sources of risk may be natural (e.g., hurricanes) or the result of human activity (e.g., job loss or conflicts). By identifying the sources of vulnerability, and - in a broad sense - the nature of poverty and vulnerability in Guatemala, this chapter seeks to deepen our understanding of the dynamic nature of poverty, with a view towards informing policy and highlighting priorities for poverty reduction and prevention. Vulnerability is analyzed using a combination of quantitative and qualitative data, as well as administrative information on shocks in Guatemala (e.g., mappings of natural disasters). The primary quantitative data source is the ENCOVI 2000, which included a risks and shocks module in both its household and community questionnaires. The household module covers shocks that occurred during the past 12 months, examining (a) whether the shocks triggered a loss in consumption, income or wealth; (b) the main coping strategies used to compensate for these losses; (c) whether the households were able to compensate or resolVe the welfare loss; and (d) the estimated time until successful resolution of the situation. Similar information was collected via focus group discussions at the survey cluster level (primary sampling units, PSUs) in the ENCOVI community questionnaire but using a five-year time horizon (rather than 12 months). Although panel data would better facilitate the analysis of the dynamic concept of vulnerability, such data are not available. Instead, these retrospective modules provide a unique opportunity to analyze vulnerability using available cross-section data. Qualitative information primarily comes from the risks 114 and shocks module of the 10 village QPES study. Detailed results and methodological considerations are discussed in GUAPA Technical Paper 9 (Tesliuc, 2002). This chapter is divided into two sections. The first analyzes shocks as a source of vulnerability in Guatemala, painting a portrait of shocks in the year 2000, their occurrence, severity, duration, distribution by poverty group, and estimated impact. The main coping strategies used by Guatemalan households and communities are also examined. Finally, the first section postulates likely future sources of vulnerability and their potential impact, building on the impacts revealed for the year 2000. The second section examines vulnerable groups in Guatemala in two ways. First, it analyzes the chronic or transient nature of poverty and vulnerability in Guatemala. Second, it broadens the concept of vulnerability beyond the income and consumption sphere to look at groups that are structurally vulnerable due to other dimensions (e.g., malnutrition or education) over the life cycle. Finally, the chapter concludes with a review of key issues and priorities. SHOCKS AS A SOURCE OF VuLNERABILITY IN GUATEMALA This section seeks to describe the sources of vulnerability in Guatemala by (a) analyzing the characteristics of shocks that affected the country in recent years (particularly in 2000, the year the ENCOVI data were collected); (b) reviewing the mitigation and coping strategies used by households when faced with these shocks; and (c) estimating the potential impact of these shocks on welfare, poverty and inequality. Shock Characteristics: Frequency and Correlation Structure Guatemala was largely spared any major "macro" shocks in the year 2000. Macroeconomic indicators were stable, with modest growth (3.3%), low inflation (6% p.a.), and negligible unemployment (1.8%). Indeed, a recent regional World Bank study classifies Guatemala as a low mnacroeconomic risk country.' No massive natural disasters hit the country that year, the latest major disasters were Hurricane Mitch in 1998 and earthquakes in previous years (particularly the 1976 earthquake). Nonetheless, Guatemalan households reported a high incidence of shocks in 2000. Respondents in the QPES recalled al types of shocks: natural, health, economic, social, life-cycle related, political and environmental. Over half (53%) of households reported one or more shocks in the ENCOVI 2000.2 A quarter (23%) reported natural shocks, 17% cited man-made (e.g., economic) shocks, and 13% reported having experienced both types.3 The most common types of shocks are agricultural-related: pest infestations and harvest losses, (Figure 11.1). Other shocks, such as a "fall in income," accident of the breadwinner, job loss, drought, worsened terms-of-trade, tempest, criminal offense, or floods are reported by 2-10% of the population. Not all reported shocks resulted in material losses, which can be interpreted as the result of the households' ability to entirely mitigate the impact of the shock or as "false" complaints. Some shocks without serious material losses could have impacted other dimensions of household well- being, such as social or psychological effects, that were not captured in the ENCOVI (but nonetheless reported in the QPES, Table 11.1). Such could be the case, for example, for the death of household members other than the main breadwinner, land or family disputes, or public protests. Most households experienced multiple shocks. The majority of households reporting shocks in 2000 were hit by more than one shock. A quarter reported having experienced two shocks that year, and another quarter reported three or more shocks. The incidence of multiple shocks is similar for the (ex-post) poor and non-poor. The large share of households reporting multiple shocks signals accumulated vulnerabilities as a possible cause of poverty. Consistent with the absence of any major macro shock, the analysis reveals that all reported shocks were localized (idiosyncratic). A variance decomposition test found that location alone explains less than 25% of shocks that were classified a-priori as covariate. The shocks with higher degree of covariance at the local level were harvest losses and income losses. 115 Shocks tend to hit in bunches. There is strong empirical evidence about the effect of shock-bunching on household welfare. The impact of a shock is harder if the affected household was hit by other shocks as well. A factor analysis revealed the following major types of bunched shocks: (a) agricultural (drought, pests, harvest or terms-of-trade losses); (b) idiosyncratic economic shocks (job loss, bankruptcy, accident or death of the breadwinner, lost remittances); (c) social / violence (family or land disputes, criminal offense); (d) covariate economic shocks (enterprise closure; mass lay-offs); and (v) natural (earthquake, floods, tempests, hurricanes, landslides, or forest fires). Figure 11.1 - Incidence of Reported Shocks During the Year 2000 20- 16- <12 4 - H .~~~Q 01 .4 .- P" . .60 C)~~~~~~~~ed0 pA 0~~~~~R I Last consumption Lost income No Loss World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadisticas - Guatemala Shocks are difficult to predict. A comparison of the information on the incidence of reported shocks from the ENCOVI, the QPES, and administrative and secondary sources reveals that there are important changes in the shock profile over time. The evidence suggests sizeable differences in the incidence of shocks when different reporting periods are used. For example, while no one was affected by a hurricane in 2000, 45% report hurricanes over the past five years. To some extent, this reflects the rare nature of certain shocks (such as hurricanes, earthquakes) - although these shocks can have catastrophic impacts. An implication of this finding is that the cross-sectional incidence of shocks in 2000 is a poor predictor of the future distribution of shocks. It does not allow us to predict how many shocks of specific types will hit the country next year. Nonetheless, while the profile of shocks does not allow us to predict which types of shocks will hit in the future, it does shed light on the likely impact of various types of shocks, as discussed below. Natural shocks are an exception. Exposure to some natural shocks does seem to be largely determined by location and geographic factors. Administrative maps of vulnerability to drought, seismic activities, hurricanes, storms or tempests, frosts, and landslides are highly consistent with the results reported in the ENCOVI 2000. For example, both the flood maps prepared by the Ministry of Agriculture and the FAO, and the data from the ENCOVI, classify the Nororiente Region as a high risk area and the Nor-Occidente Region as a low risk area. As such, vulnerability maps are useful instruments for risk management planning. They can be even more powerful in targeting assistance to those less able to cope with such 116 shocks when combined with poverty maps. Guatemala has experience in collecting information on areas with high vulnerability to natural disasters and some institutional mechanisms to implement disaster-relief programs (Box 11.1). However, these mechanisms face implementation problems (e.g., inadequate financing and human resources to respond in a timely manner) and actual interventions have mainly focused on ex-post coping rather than risk mitigation or prevention. The Effects of Shocks: Multi-Dimensionality, Severity, Du ration The effects of shocks are multi-dimensional. Respondents in the QPES identified a wide range of impacts of shocks (Table 11.1). In addition to economic effects, such as wealth and income, reported impacts include: (a) psychological, such as the demoralizing impacts of job loss, the traumatic impacts of violence (fear, post-traumatic stress syndrome or susto); (b) social (destroying trust and social capital Box 11.1 - Disaster Management Programs within villages); (c) damage or destruction of The Coordinadora Nacional para la Reducci6n de community assets (loss of road access, school Desastres (CONRED) is a disaster management program destroyed, water tank damaged); and (e) impacts on (early-warning and early-response) employing about 1,300 health (death, illness) and education (children can't people and providing training to communities exposed to attend school). In the ENCOVI covariant shocks natural risks and special housing arrangements to reduce tend to be reported more often even if they do not vulnerability to earthquakes. The initiative relies primarily on foreign funding (90%). The budget for CONRED is trigger income or wealth losses (perhaps because quite small (about Q130 mn, or 0.09% of GDP). The they resulted in other non-economic impacts). Ministry of Health (MSPAS) also has a Unidad Nacional Idiosyncratic shocks are almost always associated de Prevenci6n de Desastres. with income or wealth losses. Social shocks (e.g., violence, unrest) are less likely to cause income or wealth losses (but clearly have other psychological and social impacts, as found in the ENCOVI). Some shocks, like earthquakes, fire, or hurricanes, mainly affect household wealth and community assets. Other natural shocks primarily affect income. Table 11.1 -Shocks Reported in the QPES: Types of Shocks and impacts Types of Impacts Cited # of Villages ECON ECON PSYCH. COMM. HEALTH EDUC. Reporting Shock WEALTH INCOME SOCIAL ASSETS Types of Shocks (out of 10) _ Natural Eanhquakes (1976, 86) 7 X X X X Tremors 1 X Hurricane (1998, 2000) 5 X X X X Flooding 2 X X X Landslides 1 x Tornado 1 x Freezing/frost 2 X Forest Fire I X Drought 2 X Crop Loss 3 x Man-Made Violenceof 1980s 3 X X X X X Debt/Collective Debt 2 X Land conflict' 1 x X x Conflict in Community' 1 X X Job loss 4 X X Domestic Violence 3 X Crime and Violence 2 X Terms-of-Trade losses 1 x X Abandonment from migration 1 x x Health Cholera, dengue epidemic 2 X Birth (maternal mortality) 3 x Sickness 7 X X Automobile accidents 1 x x Death 5 s X Source: QPES. PSYCH = psychological (flar, susto); COMM = community assets, such as damaged water tank, road blocked, etc. ECON = econornic. EDUC = educational (kids couldn't attend school due to violence). a. Many other communities identified these conflicts as occurring in their villages (land, religious), but. only a few identified them as shocks. 117 The duration of impact varies by type of shock. In the QPES, which did not restrict respondents to a particular reference period, several villages note long-lasting effects of shocks. For example, many villages report that families still live in homes that were badly damaged by the Earthquake of 1976 (some 25 years later). The shock simply shifted these households from an already poor level of living conditions to an even worse state, with no resolution. Hurricane Mitch also seemed to have catastrophic consequences on some villages, completely wiping out the main productive base of the villages (see Box 11.2). The social and psychological impacts of the conflict of the 1980s also clearly have had lasting effects, according to respondents in the QPES. The ENCOVI 2000 suggests a broad diversity by type of shock in terms of the share of households that were able to resolve the shocks by the time of the interview. Most households were, in fact, able to resolve the shocks and restore their (economic) welfare within 12 months. For example, close to 70% of households reporting fire were able to resolve their situation within a year. The duration of impact seems longer among more frequent shocks, however: less than 20% of households were able to restore their welfare within a year when faced with a worsening terms-of-trade, income (earnings) losses, enterprise closure, public protest, criminal offense, or bankruptcy. Only a third of those reporting pest infestation, harvest losses, lost remittances, or job loss were able to overcome the shock within a year. Box 11.2 - Catastrophic Consequences of Hurricane Mitch: the Case of LI (QPES) The Ladino village of LI, located in the Nororiente Region of the country has faced a drastic worsening of living conditions since Hurricane Mitch struck in 1998. Prior to the hurricane, the main source of income was agriculture, with a somewhat diverse range of production: lemons, papaya, tobacco, melons, eggplant, palms for raw material from artisan work, corn for subsistence, and livestock. Hurricane Mitch severely damaged the land, however, rendering it largely infertile and covered with rocks (the flooding basically washed away the productive topsoil and dumped rocks all over the fields). It also destroyed livestock animals and farm implements. Now most have to search for day labor jobs elsewhere, with some 400 (half the village population) migrating to the capital or the US and leaving their famnilies behind. Moreover, despite having certain community assets, such as electricity and water, the hurricane clearly exposed vulnerabilities in the asset base of the village, which lacks proper drainage and contamination, as well as health services. The villagers blame an ensuing dengue epidemic on Mitch, as a result of stagnant waters, which generated an infestation of mosquitos. The effects of this epidemic were exacerbated by the lack of health services in the village. Moreover, not all income or wealth losses result in a reduction in consumption. Most households were able to smooth their consumption even when faced with shocks. In fact, just over a quarter of all shocks that resulted in income or asset losses forced households to cut their consumption as a way of coping with the shock. In the majority of cases, households were able to mitigate the effects of shocks or use coping strategies other than reducing consumption, as discussed in more detail below. The shocks with harsher impacts on household consumption include economic shocks that most often force households to reduce their income streams: falling earnings (32%), job loss (21%), lost remittances (19%), and worsening terms- of-trade (12%). Somewhat surprisingly, public protests are also associated with reduced consumption (27%). In terms of severity, economic shocks had the highest negative impact on household income, consumption and wealth. Using data from the ENCOVI, shocks were classified by the severity of their impact on income or wealth by creating an index equal to the mean rank of three variables:4 (a) the percentage of households that lost income or wealth given the shock; (b) the share of households that reported reduced consumption as a main coping strategy; and (c) the proportion of households that did not resolve the shock by the time of the interview (Table 11.2). According to this index, economic shocks - including job loss, lost remittances, bankruptcy, worsened terms-of-trade, a fall in income, or inflation - seem to have the highest impact. Using this index, it also appears that there is a positive correlation between shock frequency and severity, with frequent shocks, such as harvest or income losses, having a severe negative impact on household income or wealth. This ranking of the severity of shocks, however, is 118 likely to hold only for the current distribution of shocks, as the ENCOVI was conducted during a year in which no large catastrophic disasters (such as hurricanes or earthquakes) occurred. Table 11.2 - Ranking the Severit3 or Shocks Reduced Shock Not Niedian Losi Income - Cons. Overcome Nlean'Rank Rank Low Impact Forest_Fires 33 3 60 4 3 Land-Dispute 31 0 75 7 1.5 Land_Slides 63 4 57 7 6 Drought 57 6 59 8 8 Family-Dispute 39 7 66 9 10 Death Other_HH Member 57 9 50 9 7 Moderate-Low Impact Pest 65 4 68 9 11 Earthquake 61 9 46 9 9 Fire 100 0 31 10 1.5 Tempest 55 6 72 10 8 Floods 52 10 54 10 5 Hurricane 66 11 53 12 13 Moderate-High Impact Criminal_Offense 90 5 80 14 16 Public_Protest 38 27 80 15 18 Death_Breadwinner 97 5 72 15 16 Accident_Breadwinner 94 10 65 15 18 Lost_Harvest 97 8 69 16 14 Abandon_Breadwinner 77 7 86 16 14 Enterprise_Closure 65 7 86 16 13 Mass_Lay_offs 82 7 85 16 15 High Impact Lost_Remittances 92 19 70 18 17 Job-Loss 97 21 69 20 23 Bankruptcy 98 9 82 20 20 Lost_TOT 96 12 86 22 21 Income_Drop 98 32 83 24 25 Inflation 96 46 88 24 26 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadfsticas - Guatemala Shocks for the Rich and Shocks for the Poor? The poor are more exposed to natural shocks. Consistent with international experience, the poor are more likely to be hit by natural shocks (Table 11.3), probably reflecting their dependence on agriculture as a source of living and their geographic location (e.g., more likely to live in marginal areas). The non-poor are more often the victims of economic shocks. This relationship holds in multi-variate regressions controlling for the impact of shocks on welfare. The probability of experiencing a welfare loss following harvest losses, droughts, pest infestation or worsening terms-of-trade drops significantly with higher household wealth, after controlling for other potential determinants. The reverse is true for economic shocks, such as job loss, bankruptcy, and accidents of the main breadwinner. 119 The poor have lower Table 11.3 - Shocks for the Rich and Poor? resilience than the rich. The Wealth Quintiles5 probability of restoring % of households Total Ql poorest Q Q3 Q4 Q5 richest prba t . Reported Shocks incomes to levels prevailing Natural 28.7 35.4 28.2 32.0 26.7 21.2 before the occurrence of Economic 32.8 32.8 31.5 34.7 33.0 31.8 shocks rises with income. Social 5.7 2.2 3.1 6.4 7.1 9.8 Some 88% of the extreme Life-Cycle 12.4 10.8 12.4 11.8 14.2 12.7 poor and86% of he poor Shocks that Reduced Welfare' poor and 86% of the poor Nhocks thatural 18.6 23.0 21.2 22.2 15.9 10.6 suffered losses in response to Economic 31.7 32.1 30.8 33.6 31.7 30.2 shocks, compared with 83% of Social 3.7 1.0 1.8 5.5 4.6 5.8 the non-poor. This is Life-Cycle 11.0 9.8 11.4 11.0 12.0 11.0 particularly notable for natural Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Natural shocks include: earthquake, drought, floods, tempests, hurricanes, pests, land shocks: while half of those in slides, fire, forest fire; Economic shocks include: enterprise closure, mass lay-offs, job loss, income the top quintile that losses, bankruptcy, lost remittances, worsened terms-of-trade, lost harvest; Social natural shocks shocks include: public protests, criminal offense, land disputes, family disputes; Life-cycle shocks experienced natural shocks include: accident of the breadwinner, death of breadwinner or other, abandonment of breadwinner. suffered resulting . welfare a. Income or wealth. losses, two thirds of those in the bottom quintile did (Table 11.3). Main Coping Strategies Faced with shocks, Guatemalan households rely on their own assets as their main coping strategy; few receive public assistance. For the majority of shocks, the ENCOVI reveals that the main coping strategies include reducing Table 11.4 - Main Co ing Strategies, by Wealth Quintiles consumption or self-help % of households We ath Quintiles (supplying more labor, reporting shocks Total Ql poorest Q2 Q3 Q4 Q5 richest selling or mortgaging Self-Help 35.3 39.4 39.2 31.8 33.6 33.1 selling or mortgaging Informnal/Social Capital 7.4 11.0 8.1 7.4 5.8 5.1 assets, drawing down Private Insurance/Credit 12.6 7.7 14.4 13.9 14.7 11.9 savings), as shown in Government Assistance 0.2 0.4 0.0 0.2 0.0 0.3 Table 11.4. Few NGO/lnt'l Assistance 0.5 0.6 0.5 0.7 0.0 0.5 households report Reduced Consumption 44.0 40.9 37.7 46.1 46.0 49.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 receiving Government or Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. NGO/donor assistance (Table 11.4). Informal coping mechanisms ("social capital"), such as borrowing or receiving help from friends, relatives or neighbors, were the primary coping strategy in instances of family disputes, accidents of the bread-winner, or death of other members of the household. Formal insurance (market-based mechanisms such as credit and private insurance) were most common for insurable risks, mitigating the risks of fire, earthquakes, hurricanes, or land slides. Consumption losses were the main "coping strategy" used in the face of falling household incomes (earnings) and public protests. These findings are very consistent with the findings of the QPES, where self-help and collective (community) action were the prime coping strategies, with little evidence of government or formal assistance (Table 11.5). 120 Table 11.5- Main Coping Stra tegies and Formal Assistance in 10 Rural Villages (for main collective shocks, QPES) Main Coping Strategy/Response Any Formal Assistance? Community Shock Description Type Description KAI Earthquake 1976 SELF Tried to rebuild homes, mill Received a few housing COLLECTIVE Using mill at neighboring finca materials from unknown bilateral donors ("los gringos de francia o italia") Massive labor SELF Tried to get union help (pending action) None dismissals by finca Sought temporary employment on (labor dispute) nearby fincas KA2 Earthquake 1976 COLLECTIVE Organized development committee Received housing materials, food, helicopter help for injured Violence of 1980s SELF, Villagers fled to nearby town, now have None during the violence COLLECTIVE organized development committee to Did receive a school, housing protect town project after the conflict KIl Earthquake 1976 SELF (market- Went into debt and then had to migrate UNEPAR provided some based) to find work to pay back debt housing materials Cholera epidemic SELF Went to hospital, received medicine, Health education campaign at 1990 now treat water health center (promoting water treatment) Hurricane Mitch SELF Went into debt, had to migrate to find None 1998 work to pay back debl LI Hurricane Mitch SELF Migration in search of work None Dengue epidemic SELF Illness, sought treatment None L2 Earthquake 1976 SELF Each family rebuilt home None Tornado 1998 SELF Each family rebuilt home None Ml Violence 1980s SELF, Villagers fled, helped each other repair Some unknown external COLLECTIVE homes, provide shelter and food agency provided housing Forest Fire 1990s COLLECTIVE VillaZe tried to fight fire None M2 Earthquake 1976 SELF Family rebuilt home Municipality provided some housing materials and food Rains, road washed COLLECTIVE Village worked together to transport None out products and repair roads QEI Drought 1998 None None None Hurricane Mitch COLLECTIVE Water committee, collective action None 1998 Collective debt COLLECTIVE Collective action but little progress Some advice from NGO/MAGA Land conflict with COLLECTIVE Contacted govemment officials Yes, soldiers withdrawn (see neighboring Box 13.1 in Chapter 13). community QE2 Drought 1998 SELF Temporary migration in search of work None Conflict within None None ("voluntad de Dios") None community The poor are less equipped than the rich to fight shocks. The poorest quintiles were far more likely to rely on themselves (self-help) or use informal, social capital networks to resolve shocks than the better off (Table 11.4). In contrast, the wealthier quintiles were more likely to rely on private insurance or credit to resolve their situations. Indeed, a multinornial model contrasting three coping strategies (market-based and informal coping versus self-help) reveals that for all shocks, the probability of choosing market-based strategies over self-help increases with household wealth and education of the household head. Interestingly, informal coping mechanisms ("social capital" networks) are more often used by female- headed households and the probability of using market-based coping mechanisms is lower for indigenous people (ceteris paribus). For some shocks, such as income or job losses, poorer households are more likely to resort to a reduction in consumption rather than other strategies. In contrast, for pests, worsening terms-of-trade, 121 or harvest losses, the reverse is true, with richer households more likely to cut consumption when faced with such shocks. Other studies shed light on this phenomenon: while shocks can trigger a reduction in consumption among the rich and poor alike, the poor will be forced to cut their consumption of an undiversified basket of basic staples, whereas the rich will reduce their consumption of luxury goods. The interpretation of these findings is that poor households cannot smooth their consumption when faced with income or job shocks. The poor do seem more able to smooth consumptions by using their assets (mainly labor) when faced with agricultural shocks (drought, pests, or harvest losses). Self-help is more often used by the poor. However, while the poor respond to such shocks by seeking job elsewhere (either by augmenting the labor of those already working or of other members of the family, including children), the better off use their physical and financial assets (savings, formal insurance) to cope with the effects of shocks. The Impact of Shocks The cost of shocks is significant. Two different methodologies were used to estimate the impact of shocks: a multivariate regression model6 and propensity-score matching techniques.7 The overall impact of shocks on average income and consumption in 2000 was estimated at 4% and 1% respectively using the multivariate model, and 8% and 6% respectively using propensity-matching techniques. The shocks increased income inequality by 2% under the multivarite model and 16% under the propensity-matching model. Poverty was also worsened by the shocks, by 2% using the multivariate model and 20% using propensity-score matching. While the absolute impact of the shocks was larger for the non-poor, the relative impact was greater among the poor (as a share of their counterfactual consumption or income) using both models. The most severe impact is associated with economic shocks,8 with an average loss of income of 28% for job loss, 19% for accidents of the breadwinner, and 17% each for lower earnings and bankruptcy. Natural, agricultural shocks had an important, but less severe, impact on household income: 11% for harvest loss, 10% for pest infestation, and 9% each for drought and worsened terms-of trade. These estimates confirm the severity ranking discussed above that signaled larger impacts from economic shocks. Current and Future Sources of Vulnerability Key future sources of vulnerability include worsening terms-of-trade, reduced remittances, and natural disasters - all shocks that could hurt the poor. While the composition of shocks in the year 2000 should not be used to predict which shocks might occur in subsequent years, it can help shed light on the potential impact of various types of shocks, especially when combined with structural information on the likelihood of various shocks. Given the small share of the public sector and its under-developed financial institutions, Guatemala is likely to be relatively cushioned from international financial contagion, debt or currency crises.9 Rather, the main "macro" and covariant shocks that are more likely to affect Guatemalans in the future (2001 and beyond) include: (a) worsening terms-of-trade; (b) a reduction of international remittances; and (c) natural disasters. The profile of shocks discussed above suggests that. these shocks could all have significant poverty impacts: * Coffee shocks: severe and lasting impact likely. The recent fall in the prices of coffee (and to a lesser extent, sugar) has been a substantial blow to a sector that a large number of poor workers depend on for seasonal and permanent livelihoods (see Chapter 6). As discussed above, the profile of shocks shows that this type of shock - job loss and worsening terms-of-trade - could indeed have significant adverse impacts in terms of: (a) severity and magnitude of impact: both job loss (for coffee workers) and worsened terms-of-trade (for coffee producers) resulted in large average income and consumption losses (9-28%, see above); and (b) duration of impact: only a small share of households (20-30%) were able to overcome the impact of this type of shocks within a year (see above). 122 o Natural disasters: high and lasting impact, particularly for the poor. Compared with many other countries, Guatemala is very prone to natural disasters, especially earthquakes and hurricanes. The country is located at the confluence of three tetonic plates, with 30 volcanos that pepper its southwestern highlands. Although the latest massive earthquake dates back to 1976, a series of medium-scale earthquakes have hit Guatemala in recent years. Hurricane Mitch in 1998 affected the coastal regions of the country, with lasting effects on soil erosion and destruction of livelihoods (see Box 11.2). Droughts have also been a problem in 2001. The profile of shocks analyzed above suggests that natural disasters affect a disproportionate share of poor households, either because they are pushed to live in marginal areas and/or because of their limited ability to manage these risks. Moreover, the QPES suggests that such shocks have multiple effects, including economic (reducing wealth and income), communal (damaging community assets), and psychological (causing fear and post-traumatic stress syndrome or susto). Finally, the QPES suggests that the effects of catastrophic natural disasters are lasting, shifting the poor to an even lower standard of living for the long-run. VULNERABLE GROUPS MN GUATEMALA As discussed above, the notion of vulnerability encompasses two elements: (a) the severity of the impact of shocks; and (b) people's resilience to shocks. While the above discussion examines shocks (the event) as a source of vulnerability in Guatemala, this section examines the characteristics of particular households and groups (the people) in an attempt to discern a which groups (or characteristics) might be more vulnerable to the impact of shocks (due to structural features or lower resilience). The first sub-section examines household characteristics to form a type of "vulnerability profile" that predicts the probability that households will be poor in the future. The second sub-section examines structural features of different groups across the life-cycle to identify which might be inherently more vulnerable due to special circumstances. Classifying the Vulnerable: The Probabilty of Being Poor in the Future The profile of vulnerability is similar to that of poverty, but there are differences particularly for urban areas. Vulnerability to consumption-poverty was estimated using a stochastic model of consumption and its variance, taking into account household characteristics (including assets and other risk management instruments) as well as the likelihood of experiencing shocks.'0 The results reveal significant overlaps, as well as some differences, between the profiles of poverty and vulnerability in Guatemala (Table 11.6). Overall, while 56% of the population was poor in 2000, 64% had a probability larger than 50% of falling into poverty in future years (the group called "vulnerable" to consumption poverty). The general patterns of poverty and vulnerability are similar - higher poverty and vulnerability for rural than urban residents, for indigenous than non-indigenous, for those with little education, for those dependent on agricultural incomes, and so forth (Table 11.6). Nonetheless, while the overall number of vulnerable is 14% higher than the number of poor, this difference is higher for certain groups. Most notably, higher vulnerability-poverty ratios were observed in the Metropolitan region (2.2 times higher), and urban areas in general (33% higher). In other words, while observed poverty is low in urban areas and the capital, there is. a significant share of the population that is vulnerable to poverty. 123 Table 11.6 - Prorile of Povert) and Vulnerability - Population Share ot Share of Pocerry 'Nlean Vulnerabiliky Vulnerabilii) to share - -.-poor vulnerable Headcount vulnerability- Headcouni po%eriv ratio Total 100.0 100.0 100.0 56.2 0.58 64.1 1.14 Area of Residence Rural 61.4 81.4 78.3 74.5 0.75 81.8 1.10 Urban 38.6 18.6 21.7 27.1 0.27 36.1 1.33 Region Metropolitan 21.7 6.9 13.6 18.0 0.10 40.21 2.24 Norte 8.1 12.1 10.7 84.0 0.79 84.58 1.01 Nororiente 8.2 7.6 7.6 51.8 0.56 59.41 1.15 Suroriente 8.8 10.7 9.2 68.6 0.64 67.25 0.98 Central 10.7 9.8 9.5 51.7 0.55 57.03 1.10 Sur-Occidente 26.5 30.1 28.6 64.0 0.65 69.16 1.08 Nor-Occidente 12.9 18.8 17.3 82.1 0.80 85.88 1.05 Peten 3.3 4.0 3.7 68.0 0.67 71.59 1.05 Ethnicity of the HH Head Ladino 57.5 42.4 46.1 41.4 0.44 51.5 1.24 Indigenous 42.6 57.6 53.9 76.1 0.75 81.2 1.07 Gender of the HH Head Male 85.3 87.5 88.8 57.6 0.60 66.7 1.16 Female 14.7 12.5 11.2 47.8 0.46 48.8 1.02 Age of the HH Head <25 years old 4.7 3.8 4.2 45.9 0.52 57.4 1.25 25-59 years old 81.3 82.8 82.9 57.2 0.59 65.3 1.14 60yearsoldandover 14.0 13.4 13.0 53.7 0.54 59.2 1.10 HH Head ST Migrant : Non Migrant 92.7 90.3 90.8 54.8 0.57 62.8 1.15 Short Term Migrant 7.3 9.7 9.2 74.6 0.75 80.5 1.08 Education Status of the HH Head, . No._School 38.3 52.9 49.9 77.7 0.76 83.5 1.07 Primary 45.0 43.3 45.8 54.1 0.59 65.2 1.20 Secondary 11.4. 3.3 1.9 16.4 0.12 10.4 0.64 Higher 5.3 0.4 2.5 4.2 0.03 30.5 7.32 HH Head Industry of Employment---- -- Agriculture 47.0 66.0 61.6 80.8 0.80 86.6 1.07 Mining 0.4 0.1 0.4 18.3 0.32 67.9 3.71 Manufacturing 10.1 6.2 6.8 35.5 0.38 44.3 1.25 Gas, Electricity, Water 0.5 0.3 0.5 31.7 0.45 65.7 2.07 Construction 8.3 7.2 8.0 50.2 0.54 64.1 1.28 Commerce 15.2 9.7 9.6 36.8 0.35 41.9 1.14 Transport 3.9 2.5 3.5 36.8 0.47 59.3 1.61 Financial 2.6 0.7 1.5 14.9 0.21 38.9 2.61 Community 12.0 7.3 8.0 35.1 0.33 44.4 1.26 Mean vulnerability is the mean probability of being poor for a particular group (the mean vulnerability index for persons in that group). World Bank calculations using data from the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Education is a strong predictor of both poverty and vulnerability. Households headed by individuals with less education are more likely to be poor and vulnerable to poverty than those with more education. The exception to this is secondary education, which records the lowest level of vulnerability (10%). Those with higher education are vulnerable (30%), this vulnerability is all due to high volatility of consumption 124 (HV vulnerability, as discussed below) rather than low expected mean consumption (LM vulnerability, see below). Certain shocks are more likely to hit the vulnerable. The ENCOVI suggests that agriculture-related natural shocks like drought, pest infestation, and harvest losses are mostly reported by households vulnerable to poverty. Other shocks, such as tempests, worsened terms-of-trade, and accidents of the breadwinner, are equally likely to hit vulnerable and non-vulnerable households. Shocks associated with the formal labor market or entrepreneurship - such as job loss or falling earnings - are reported primarily by non-vulnerable households. Strengthening the ability of the poor and vulnerable to reduce, mitigate or cope with agricultural-related shocks is important for reducing vulnerability. Most poverty and vulnerability arises due to chronic rather than transient conditions in Guatemala. The proximate cause of poverty is low consumption, which can either be a relatively chronic condition (e.g., due to a low level of assets and endowments) or a transient situation (e.g., due to a temporary setback from a recent shock). For vulnerability, there are two proximate causes: low expected consumption and high variance of consumption." For policy formulation, it is useful to decompose the pool of vulnerable into two mutually exclusive groups: (a) those who are vulnerable due to high volatility of consumption (labeled the "HV vulnerable"); and (b) those who are vulnerable due to low expected mean consumption (labeled the "LM vulnerable"), as shown in Figure 11.2.12 An analysis of the ENCOVI suggests that low expected consumption rather than volatility is the primary cause of both poverty and vulnerability in Guatemala: * Chronic poverty dominates: While 56% of Guatemala's population was poor in 2000, the majority of these (79%) were chronically poor (44% of the total population), whereas only a fifth were transient poor (12% of the total population), as shown in Figure 11.2. * Chronic vulnerability dominates: Similarly, while 64% of Guatemala's population was estimated to be vulnerable to poverty in the future, the main reason for their vulnerability is low expected mean consumption (LM vulnerability), which accounted for 79% of total vulnerability (or 51% of the total population), whereas only a fifth are vulnerable due to high volatility of consumption (13% of the total population), as shown in Figure 11.2. The chronic nature of poverty and vulnerability has important policy implications. Namely, interventions should emphasize building the assets of the poor, as discussed in previous chapters (Chapters 5-10) and emphasized in the Peace Accords (see Chapter 4) and in the Government's poverty reduction strategy (see Chapter 14). 125 Figure 11.2- Classification of Poverty and Vulnerability: Transient vs. Chronic? Numbers = % of total population Observed Poverty Status Based on Consumption Poor Non-Poor 56.2% 43.8% Chronic Poor Vulnerable to Chronic Poverty Expected consumption Vulnerable (LM vulnerable) (LM vulnerable) < poverty line >50% chance of being 44.4% 6.9% 51.3% r poor in future 64.2% o 52 (LM = 51.4%e) Frequently Poor Vulnerable to frequent poverty Expected consumption g 5 (HVY= 12.8%) (HV vulnerable) (HV vulnerable) > poverty line 4.1% 8.7% = 12.8 (frequently 3 poor) + 35.9 (non- c. Non-Vulnerable Infrequently Poor Non-vulnerable non-poor vulnerable) = < 50% chance 7.7% 28.2% 48.7% 35.8% Poor Chronic poor + transient poor; transient poor = frequently poor + infrequently poor. Vulnerable group = LM vulnerable + HV vulnerable. LM vulnerability group = chronic poor + vulnerable to chronic poverty; HV vulnerability group = frequently poor + vulnerable to frequent poverty. Classification scheme adapted from Bidani and Richter (2001). Estimates from World Bank calculations using data from the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Numbers may not add exactly to 100% due to rounding. Other Types of Vulnerability: Chronically Vulnerable Groups and a Life-Cycle Approach Certain groups are highly vulnerable due to special circumstances. There are other sources of vulnerability in Guatemala besides consumption poverty. Figure 11.4 uses a classification of risks by life- cycles, to expand the vulnerability concept in other areas, such as nutrition, education, health, access to basic services and exposure to natural disasters. For each type of risk, the share of people at risk is presented for four categories: the extreme poor, total poor, vulnerable to poverty, and total population in that group. There is a monotonic increase in the incidence of risks and household welfare, with the highest exposure to these other vulnerabilities among the extreme poor. For each type of risk, the risk matrix lists possible interventions that may contribute to risk reduction, mitigation or coping. The matrix also offers a subjective prioritization of risk groups, based on two criteria: the relative size of the group (particularly the number of poor that also fall into that group), and whether the risks or circumstances are likely to have lasting, even inter-generational effects on the transmission of poverty (e.g., education and malnutrition- related risks). Among the risks that are of particular concern include: malnutrition; low school enrollment, late school entry and grade repetition; child labor; low earnings; low health coverage of the elderly; lack of access to basic services among the poor; and higher exposure to natural disasters (Figure 11.4). Seasonal migrants and their families also appear to have higher poverty and vulnerability rates than those who migrate permanently or the general (non-migrating) population. 126 Friu 11.3 Risks by Main A9e Groeuo Leading Indicaiors ol Ri5ks, Uncovered Poor/Vulnerable. and Suggested Interventions Main Hiska .. Lauding naceators -IndicatorsVaue . . Number of Affected Persons Proposed Interventlons ; xtreme . Years of Seected Risks Poor Total Poor ruble -Total Extreme Poor: TotalPoor Vulnerable Total Priority Risk Preventiors Mitigation Risk Coping Age Groups -1) lii 1 3) 14) lS) (6). ; 1 ( (ia) () () 0 0-5 Malnutrition Stunting 63% 52% 49% 44% 293,988 756,379 794,321 945,974 ECD, growth monitoring Care of malnourished Wasting 4% 3% 3% 3% 16,742 39,570 51,355 60,611 Pre-school Not Covered 98% 97% 96% 94% 456,270 1,416,562 1,545,838 2,052,302 . ECD Programs Care of maenouprshed lncr. coverage primary, demand-side Income support tied to school 6-13 Low HumanDev Not enrolled in school 32% 22% 20% 16% 155,673 357.962 389,047 420,585 I.. interventions attendance Laet entry 41% 33% 31% 27% 200.492 645.399 565,965 700,076 Raise quality Remedial education Improve and enforce child labor Income support tied to school Child Labor Child labor 20% 16% 17% 15% 97,546 294.398 312,749 387,542 legislation attendance Incr. coverage secondary (supoly and 14-17 Low HumanDev Not enrolled In school 75% 67% 63% 54% 124,264 410,297 446,765 577,502 demand-side) income support Targeted schollarships Late enty 29% 21% 19% 16% 47,162 128,470 135,445 172,339 Raise quality Remedial education Improve and enforce child labor Income support tied to school Child Labor Chid labor 53% 56% 63°% 51% 87,222 344.399 380,651 541,197 legislation attendance 18-24 Low HumanDev Not enrolled in school 93% 91% 87Y% 81% 153,344 610,756 676,519 1,104,339 Incr. coverage high-school Income support Targeted schollarshios Employment Youth Unempbyment 2% 1% 2% 2% 2,971 8.871 12,344 27,488 Income support Job search support Underemployment 17% 18% 198% 17% 27.324 124.149 139,972 237,094 Labor Intensive growth Remedial education Income support Job search support Low wages 29% 24% 23% 18% 48.028 160,453 174,993 249,785 Labor intensive growth Remedial education 25-60 Low Income Unemployment 0% 1% 1% 1% 2,046 10,075 17,000 41,031 Workdare Income support Job search support Underemployment 24% 22% 21% 22% 103,380 379.571 419,473 758.109 Labor intensive growth Remedial education N) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Income support Job search support Low wages 31% 25% 22% 18% 134,991 423,077 436,873 627,033 Labor intensive growth Remedial education No IGSS coverage 62% 64% 63% 63% 272,712 1,082,284 1,244,863 2,207,606 Enforce regulations Expand TAM Nutritlon Obesity 0% 0% 1% 2% 345 7.798 22,384 78,121 Provide Health Insurance for Poor over 60 Chronic diseases No Health insurance 87Y% 95% 92% 86% 63.641 288,072 286,683 584,217 Elderty Low incomte No pension coverage 96% 96Y6 94% 88% 63,327 290,292 294,995 597,626 instiute Social Pension Nutrition Obesity 0% 0YO 0 Y/ 1% . 242 7568 4,033 Other Population groups. large Rousing subsidies Land Housing Poor housing Poor housing conditions 10% 9% 7% 5% 183.827 509,251 516,144 582,134 Promote savings and mortgages t_ling No Water 50% 41% 39% 31% 884,601 2,638,151 2,858,295 3,524.266 ~Improve coverage (supply and demand- Basic Services NoWater 50% 41% 39% 31% 884,601 2,638,151 2,858,295 3,524,266 side Interventions) Subsldizo connections for poorest Improve coveorage (supply and demand- Poor basic services No Sewage 86% 68%/ 81% 66% 1,715,914 5,500,787 5,904,865 7,648,734 side interventions) Subsidize connections for poorest improve coverage (suppty and demand- Etminate user subsidy; subsidize No Electricfty 669 44% 40% 29% 1,162,089 2,633,552 2,684,040 3.282,762 side interentions) connections for poor Disaster management. Incentives for Natural Disasters Dameges due to ND 26% 31% 31% 29% 468,110 1,957,760 2,270,000 3,267,620 relocalin Temporary shelter provision Improved housing Supply of foodlmedicine (1) Share of extreme poor individuais in the age-category affected by the risk (e.g., first row share of chiidren aged 0-S years old in extreme povertYaffected bystunling) (2) Share of poor lndividuats In the age-category affected by the risk (e.g. share of children aged 0-5 years old in poverty afSected by stunting) (3) Share of vulnerable individuals In the age-category affected by the risk (e.g. share of chl5dren aged 0-5 years old classiSied as vulnerable affected by stunling) (4) Share o1 individuals affected by the risk In that age category (e.g. share of children aged 0-5 years old affected by stunting) (5) Number o0 extreme poor persons in the age-category affected by the risk (e.g. number of extremeiy poor children aged 0-5 years old affected by stunting) (6) Number o0 poor persons in the age-category affected by the risk (e.g. number of poor children aged 0-5 years old affected by stunting) (7) Number of vulnerable persons in the age-category affected by the risk (e.g. number o0 vulnerable children aged 0-5 years old affected by stunting) (8) - - loP priority; " medium priolty; *= priority. (9) Usts suggested public. private or In1ornal interventions that can be used In Guatemala to reduce the occurrence 0o the risk or mitigate (creds, insurance, hedging, portfolio diversHltcation) the Impact of the risk (10) Lists suggested public, private or lnifofnal interventions that can be used In Guatemala to help households cope with the impact of the risk Underemployment is defined as working less than 40 hours a week, and low pay as recording eaming in the lowest deciie of the eaming distribution (less than 150 Oz per month) SUMMARY OF KEY ISSUES AND PRIoR1TIEs This chapter highlights a number of key findings with respect to vulnerability and the dynamics of poverty. Specifically: * Households in Guatemala experienced a high incidence of shocks in 2000, and most experienced multiple shocks with varying duration of impact. * The effects of shocks are multi-dimensional, affecting not only income, wealth and consumption, but also community assets, the psychological and social well-being of individuals, families and communities, health and education. * The poor are more exposed to natural disasters and agriculture-related shocks. They also have lower resilience to shocks than the non-poor. * The cost of shocks is significant. Economic shocks have larger and more severe impacts than other types of shocks. * Faced with shocks, Guatemalan households tend to rely primarily on their own assets. The main coping strategies include reduced consumption or self-help. Few households report receiving any formal governmental or non-governmental assistance in the face of shocks. The poor are less equipped than the non-poor to fight shocks, and are more likely to reduce consumption (regrettable, of basic staples) or use existing assets (particularly labor). The non-poor are more likely than the poor to use market-based insurance mechanisms. * Key sources of vulnerability in the future include: (a) worsening terms-of-trade and job loss, such as those associated with the crisis in the coffee sector; (b) lost remittances from the global economic slowdown; and (c) natural disasters. All are likely to have lasting and severe impacts on the poor. * Poverty and vulnerability are primarily chronic - rather than transient -- phenomena in Guatemala, reflecting low average consumption (current and expected) due to low levels of overall assets (e.g., education). D Certain sub-groups of the population are inherently or structurally vulnerable due to special circumstances. Specifically, key vulnerable groups include young children, who are vulnerable to malnutrition and lack of development; school-aged children, who are vulnerable due to lack of educational opportunities and child labor; the working poor, particularly those in agriculture, due to low earnings and susceptibility to natural shocks; poor households lacking basic services; seasonal migrants and their families; and poor, rural households living in areas prone to natural disasters. In light of these findings, a number of policy recommendations seem appropriate as options for a strategy to reduce vulnerability. Notably, strategies to reduce vulnerability should emphasize: * Children. A strategic emphasis on children - particularly child-focused interventions to reduce malnutrition (see Chapter 8). and promote early childhood development --is crucial to avoid an inter-generational transmission of poverty and vulnerability; * Building the assets of the poor. Given the chronic nature of poverty and vulnerability implies that interventions should emphasize building the assets of the poor, as discussed in previous chapters 128 and emphasized in the Peace Accords and the Government's poverty reduction strategy. As discussed in Chapter 12, social protection programs can play an important role in building the assets of the poor. Specifically, when designed properly, conditional-transfer programs can be quite effective in helping ease demand-side constraints, which have been shown to constitute important limitations for improved coverage of key assets, such as education (Chapter 7), health (Chapter 8) and basic utility services (Chapter 9). * Disaster management and relief, which should be expanded and improved, given the disproportionate exposure of the poor and vulnerable to natural disasters and agriculture-related shocks. The introduction of catastrophic insurance may also merit consideration. Such interventions should be well-targeted to the poor and delivered in a timely manner. Since exposure to some natural disasters does seem to be largely determined by location and geographic factors, administrative maps of vulnerability to drought, seismic activities, hurricanes, storms, frosts, and landslides could be quite useful instruments for risk management planning. Many such maps have been prepared by the Ministry of Agriculture and FAO, and are available to policy makers. Their ability to target limited funds for disaster relief would be greatly enhanced if used in conjunction with poverty maps (since those who are already poor are less equipped to cope with shocks). Since the impact of natural disasters often includes damage to, or destruction of, community infrastructure (in addition to income and wealth losses at the household level), the social funds may serve as the institutional channel by which such relief and infrastructure rehabilitation is implemented. de Ferranti, Perry, Gill, and Serven (2000). Gaviria (2001) reports that in the first semester of 2000, 36% of urban Guatemalans reported income losses and 26% reported falling consumption. 3These estimates omit inflation, which was mentioned by 68% of households as a shock. Nonetheless, inflation was quite low in 2000 (6% p.a., falling from 11% in 1996) and none of the respondents in the QPES reported inflation as a shock. It is believed that ENCOVI respondents were simply indicating a general complaint about the level of prices (the cost of living), as they did in their perceptions of poverty and welfare responses. 4The ENCOVI did not collect information on the magnitude of income or consumption losses due to shocks. As such, this analysis combines the information on the shares of households that lost income, wealth or consumption and the share that did not resolve the shock within a year, as a way to rank the severity of shock impact. 5 Wealth quintiles are used to circumvent the circularity of being poorer and being affected by a shock, households were ranked from poorest to richest using a wealth index instead of observed consumption. As suggested by Filmer and Pritchett (1998), this factor is a latent variable that captures the long-term socio-economic status of the household. This measure of wealth correlates strongly with (log) household consumption (Pearson coefficient = 0.77), as well as other indicators of well-being such as education and health status. 6 Counterfactual income and consumption was first estimated using an augmented specification of the typical consumption regression including "dummy" variables for the main shocks and interactions between shocks and wealth. The consumption or income in the absence of the shock was then estimated by setting al shock variables to zero, and welfare "with and without" the shocks was then compared. See GUAPA Technical Paper 9 (Tesliuc, 2002) for details. 7See GUAPA Technical Paper 9 (Tesliuc, 2002) for details. 8 Patterns in this paragraph come from regression results from the multivariate model. 9 Although recent problems in Guatemala's own financial sector have damaged overall confidence, as discussed in Chapter 5. 10 Specifically, in addition to the typical household characteristics, the model included the likelihood of experiencing a loss following any of the seven most frequent shocks: drought, pests, job loss, income (earnings) losses, accident of the breadwinner, worsened terms-of-trade, and harvest losses. Separate models were estimated for the Metropolitan Region and for the remaining rural and urban areas. See GUAPA Technical Paper 9 (Tesliuc, 2002) for details. " Chaudhuri and Datt (2001) elaborate on the parallel that exists between the classification in Figure 18 and the more familiar distinction between transient and chronic poverty: "Loosely speaking, households who are HV-vulnerable are in a sense more likely to be only transitorily poor, whereas households who are LM-vulnerable are more likely to be chronically poor... The two taxonomies differ fundamentally because of the different questions they pose. The distinction between the transient poor and the chronic poor is based on the question: how often is the household poor? .... [Whereas] the distinction between HV-vulnerable and LM-vulnerable households is based on the question: why is the household poor?" 12 Specifically, the HV vulnerable have estimated vulnerability in excess of 0.5, but estimated mean consumption above the poverty line. In absence of shocks (or more generally, consumption volatility), these households will not experience poverty. The LM vulnerable have mean consumption below the poverty line and may or may not have high consumption volatility. Even in the absence of shocks, this group will remain in poverty (and their vulnerability may even increase). 129 Chapter 12: Social Protection, Private Transfers and Poverty This chapter reviews the effectiveness and efficiency of Guatemala's numerous and scattered social protection programs with a view to informing policy and highlighting priorities for reducing poverty and vulnerability. An adequate social protection system is an important element of a comprehensive strategy to reduce poverty and vulnerability. Social protection (SP) has been traditionally defined as "a set of public measures aimed at providing income security for individuals."' The final goal of public social protection policies is to increase the welfare of the population, and to that end, these schemes have generally included social assistance (SA) and social insurance (SI) programs. Social assistance programs are generally designed to help individuals or households cope with chronic poverty or transient declines in income that would cause them to live in a situation of poverty or worsening poverty. As such, they help alleviate poverty and reduce vulnerability to poverty. SA programs as a whole make up what is commonly referred to as "the social safety net," and include programs such as transfers (in cash or kind),2 subsidies,3 and workfare.4 Social insurance schemes include publicly-provided or mandated insurance for unemployment, old age (pensions), disability, survivorship, sickness, and so forth, which are designed to help mitigate income risks. Private transfers can complement public social protection interventions. In Guatemala, private transfers - particularly remittances, are a particularly important source of income. As such, this chapter also considers private transfers. The chapter begins with an overview of SP programs. It then assesses the coverage, targeting incidence, adequacy, and overall effectiveness of these schemes. It examines the potential impacts of social protection programs on poverty and inequality, and assesses their cost effectiveness. Finally, the chapter concludes with a review of key issues and priorities. It builds on an inventory of social protection programs conducted by the Universidad Rafael Landivar (2001), an analysis of this inventory by Santiso (2001), and an in-depth analysis of the effectiveness and efficiency of SP programs using the ENCOVI 2000.5 OVERVIEW OF SP PROGRAMS: TYPES, MAGNITUDES AND SPENDING Overall Magnitude and Trends in Public SP Spending Public spending on social protection is low by international standards, reflecting the low level of overall public resources in Guatemala. Numerous social protection programs are managed by many different agencies in Guatemala, as discussed below. As such, accounting for SP spending is complicated and estimates of the total magnitude of such spending vary. Two approaches are generally used: (a) disaggregating official government spending accounts by type of spending (usually by major category or ministry); and (b) building up spending estimates from an inventory of programs. Using the first approach, IMF estimates indicate that total public spending on social protection ("social security and welfare") absorbed 1.0% of GDP and 8.3% of total government spending in 2000 (Table 12.1). This compares with reported spending on education and health of 2.5% and 1.1% of GDP respectively (representing 18.3% and 7.9% of total government spending).6 It also compares with spending, on social investment funds, which accounted for 0.6% in 2000.7 With these estimates, the level of public SP spending has increased from 0.7% of GDP in 1996 to 1.0% in 2000. Using the second approach, an analysis of an inventory of programs conducted by the University of Rafael Landfvar (URL, 2001) yields a slightly higher estimate, with total public spending on social protection absorbing 1.8% of GDP and 12.4% of total government spending in 2000 (Table 12.1).8 Using these estimates, the level of public SP spending has increased from 0.8% of GDP in 1996 to 1.8% in 2000. Although SP spending appears to be rather low by international standards (Table 12.1), it is not so low relative to other social sectors in Guatemala and current levels mainly reflect low overall public finance base in Guatemala (total public revenues represented about 10.5% of GDP in 2000). SP spending is also quite low in relation to the poverty gap. The estimated annual cost of eliminating to total poverty gap is Qzl 1.1 billion, or 8.4% of GDP. Total SP spending pales in comparison to this gap 130 (Table 12.1). As such, poverty reduction via redistribution is unlikely; growth is necessary. The low level of spending on social protection (and the social sectors in general) is one of the reasons Guatemala ranks so poorly on poverty and social indicators (such as life expectancy, infant mortality, nutrition, literacy, and school coverage), as discussed in Chapter 2. Table 12.1 - Public Spending on Social Protection, 2000 Million Qz. % of GDP % of TS % of SS % of SPS EMF estimates Social Security/Welfare 1,524 1.0 8.3 17.5 n.a. MEMO: Total Government Spending 18,317 12.4 100.0 n.a. n.a MEMO: Total Social Spending 8,722 5.9 47.6 100.0 n.a. URL/Santiso estimates. Social Protection - total 2,698 1.8 12.4 29.3 100.0 Social Insurance 1,090 0.7 5.0 11.6 40.4 Social Assistance 1,608 1.1 7.4 17.7 59.6 MEMO: Social Investment Funds 2,418 1.7 11.1 26.3 n.a. Some International Comparisons Argentina - SP n.a. 5.0 n.a. 32.0 100.0 Si n.a. 4.1 n.a. 26.0 82.0 SA n.a. 0.9 n.a. 6.0 18.0 Brazil - SP n.a. 10.8 45.0 n.a. 100.0 Si n.a. 10.3 42.9 n.a. 95.4 SA n.a. 0.5 2.1 n.a. 4.6 Mexico - SP n.a. 4.3 27.0 (fed) 44.0 (fed) 100.0 SI n.a. 3.2 20.1 32.7 74.4 SA n.a. 1.1 6.9 11.3 25.6 Nicaragua- SP n.a. 1.10 3.00 7.30 100.0 SI n.a. 0.01 0.03 0.09 0.9 SA n.a. 1.10 2.90 7.21 99.1 Sources: RMF (April 2001); URL (2001); Santiso (2001); intemational comparisons: Lindert/World Bank (on-going database) Overview of Public SP Programs The public social protection system in Guatemala is fragnmented, reflecting the lack of an overall strategy and a scattering of programs across many agencies. In 2000, there were some 36 different public social protection programs (39 if sub-programs are included, Table 12.2), including two main social insurance programs (accounting for 40% of total SP spending) and 34 social assistance programs (absorbing 60% of total SP spending). Guatemala's social insurance system provides minimal coverage of the population, risks financial crisis, faces allegations of corruption, and is regressive. Social insurance is run by the Instituto Guatemalteco de Seguridad Social (IGSS) and covers workers in the formal private and public sectors across the country.9 Established in 1946, social security includes several main sub-programs: accident coverage; maternity and sickness; disability; old age (pensions) and survival. More recently, a pilot program (TAM) was launched in 1998 to provide social insurance to agricultural migrant workers and their families in the departments of Escuintla and Suchitepequez. While the social security system is said to cover the entire country, not all services are available in all departments. For example, employee contributions in the Department of Guatemala are 4.83%, as compared with only 2.83% in Alta Verapaz. This is due to the fact that all programs are covered in the capital, but only accident and disability, old age and survival programs are available in Alta Verapaz. The social security system is in disarray and at risk of a financial crisis. As of the end of 1997, the Guatemalan State owed an estimated Q148 million to the IGSS due to its failure to pay contributions as an employer. More fundamentally, the inability to match expenditures with social security contributions raises questions about the sustainability of the system as currently designed. In mid 2001, for example, the IGSS accident-maternity-sickness (IVS) program had a deficit of Q166 million, contrasting with its own surplus of Q178 million in 1998. Moreover, the IGSS has been embroiled in recent corruption allegations, with the General Accounting Office reporting irregularities 131 and unreliability of the accounting system of IGSS in 2000. Finally, as shown below, coverage of IGSS programs is minimal and the incidence of IGSS benefits regressive. Guatemala lacks a comprehensive social safety net, with numerous programs scattered across many agencies, shifting institutional responsibility, duplications, gaps, and often regressive benefits. Social assistance is provided by numerous agencies and includes at least 34 different programs (Table 12.2). The main categories of programs include: (a) scholarships; (b) food-for-work programs; (c) various social assistance/service programs; (d) school feeding; (e) PRONADE, a decentralized program that provides cash transfers and school meals along with education services; (f) micro-credit; (g) disaster management; and (h) a variety of subsidies (land, housing, school transport and electricity). Institutional responsibility for these programs is dispersed among many agencies and duplications in types of programs abound. The seven different scholarship programs, for example, are currently being restructured in an attempt to improve their coherence. Possible options for restructuring including transferring responsibility for coordination to the Vice Presidency, with MINEDUC as the main implementing agency. Likewise, responsibility for the three main school feeding programs has been transferred several times in the past year, from MINEDUC to the Vice Presidency and then to the Estado Mayor Presidencial (military). Delays in food delivery and quality problems with school feeding programs have also been highlighted recently in the press, with questions regarding nutritional value and taste. Criteria for targeting the various SA programs differs widely across programs, with some using geographic criteria (though rarely based on the poverty map) and others using broad categorical eligibility (e.g., girls in poor rural areas, victims of human rights violations, orphans, poor elderly, landless peasants, breast-feeding mothers, refugees, etc.). As a result, many social assistance programs are regressive, as shown below. Overall Magnitude of Private Transfers Private transfers - including remittances and private charity - are an important source of income in Guatemala. As analyzed in more detail below, private transfers account for 46% of all transfers received by households in Guatemala (10% for charity and 36% for remittances from family members, as compared with 29% for public SA programs and 26% for public SI benefits).'° Charity is provided by a range of agencies, including international donors, NGOs, churches, private-sector agencies and so forth. In addition, several private-sector industries have established private foundations to provide social assistance, social services, and basic infrastructure to workers and their families in these sectors as part of a recent wave of "social business responsibility" (the largest are FUNDAZUCAR and FUNRURAL, representing the sugar and coffee sectors respectively)." Average Magnitude of Household-Level Benefits There is considerable variation in the average benefit level per household of various SP programs and private transfers in Guatemala (Figures 12.1 and 12.2). For public programs, the ENCOVI 2000 shows that the smallest transfer is the electricity subsidy, which provides beneficiaries with an equivalent cash benefit of Q.132 per year. At the other extreme, beneficiaries of old-age pensions receive an average of Q.14,500 per household per year. These compare with private transfers, which average Qz 5,800 in total, Q.5,700 for remittances and Qz 3000 for charity. The variation between benefit levels within programs is even higher. For instance, the ratio between the pension of the top 95% recipient household (p95) and the bottom 5% (pS) - shown on Figure 12.1 through the vertical bands - is over 100 times for all types of pensions. Most social assistance programs exhibit lower dispersion in benefit level across households, with p95/pS ratios ranging from 10 to3O times. The largest dispersion in benefit levels for each type of transfer is to be found, as expected, for private transfers, with p95/05 ratios between 100 and 300 times. 132 Figure 12.1 - Average Benefit Levels and Dispersion Among Social Protection Programs X 20000 _ -17500- 65 15000 1250000 7500 5C6000 . > - A ( E 0 Z E c . E 8 g CI) ~~~~~co 6a L Note: Vertical bands indicate the dispersion of benefit levels, from 5th to 95th percentile. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadisticas - Guatemala Figure 12.2 - Average Levels and Dispersion Among Private Transfers 20000 0 17500 -15000 ° 12500 8 10000 7500 _____ 5000 _ r 2500 Local Remittances Foreign Remittances Charity Note: Vertical bands indicate the dispersion of benefit levels, from 5h to 95t' percentile. World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadisticas - Guatemala 133 Table 12.2 - Overview of Public Social Protection Programs Types & No. of Cost 2000 Description of Program Number of Agency Covered in the Programs (inci. beneficiaries Responsible ENCOVI sub-programs) I I (official fig) (Category) Social Insurance Social Security 0 1,090 mn Covers formal private and n.a. IGSS Yes - Partial (IVS) 0.74% GDP public sector workers, including Old Age 5.0% TS accident coverage, matemity Survivorship 11.9% SS and sickness, disability, old age, Alimonya 21.3% SPS and survival TAM n.a. New program for agricultural 50,021 IGSS? No migrant workers and their families; pilot in Suchitepdquez and Escuintla Social Assistance Scholarships 013 mn 7 programs covering students in broad range: Being Yes (7 programs) 0.01% GDP primary, secondary (basic and 80 - 31,195 restructured Beca escolar 0.06% TS intermediary) schools; focus is total approx: Various agencies Bolsa utiles 0.14% SS on rural areas; some focused on 35,000 (MINEDUC has escolar 0.25% SPS girls (programa atencion de la largest, nina); some have sub-nat'l focus FONAPAZ) In-kind transfers, 035 mn 4 food-for-work programs with broad range: FONAPAZ/WFP No mainly food-for- 0.02% GDP varying geographic coverage 175-65,000 WFP work 0.16% TS and size total approx: FIS (4 programs) 0.38% SS 11,575 Municipalities 0.69% SPS Various Social 0129 mn Various programs covering broad range. SBS Partially Assistance 0.09% GDP variety of groups: babies, total approx: SOSEP "Other SA Programs 0.59% TS children, orphans, youths, young 31,106 SAS benefits, (9 programs) 1.41% SS delinquents, breast-feeding o/w: MINEDUC and gov't aid" 2.53%SPS mothers, poor rural women, SBS: 7,921 MSPAS single mothers, poor elderly, SOSEP: victims/perpetuators of domestic 23,185 violence, poor marginalized, etc. SAS: n.a. School Feeding 0143 mn School snack (refacci6n) 992,692 Being Yes (3 programs) 0.10% GDP School breakfast 903,177 restructured and galleta escolar 0.66% TS School lunches 7.000 in flux. Mainly desayuno, leche 1.56% SS total: MINEDUC en polvo, vaso de leche, vaso 2.8% SPS 1,089,869 (also PRONADE de atol ____ ___ NGOs)_ _ _ _ _ PRONADE 0357 mn Decentralized, community- 294,041 PRONADE Yes for SA (I program, 0.24% GDP managed education program. parts, as multiple benefits) 1.64% TS Provides primary schooling, captured under 3.88% SS school meals,b cash transfers, school feedig, 6.98% SPS training, TA for construction of bolsa utiles classrooms transfers Micro-credit n.a. Nat'l micro and small enterprise n.a. MINECON No (2 programs) program under MINECON FONAPAZ FONAPAZ PDP Program Disaster 0130 in Programs to help communities n.a. CONRED No Management (2 0.09% GDP vulnerable to natural and MSPAS-unidad programs) 0.60% TS environmental risks and shocks nacional de 1.41 % SS (national coverage) prevenci6n de 2.54% SPS desastres Subsidies 0801 mn Land Fund (Q106 mn (4 progs) land: 12,915 FONTIERRAS Yes. partial 6 (9) programs 0.55% GDP Housing Subsidy (Q295 fnn) housing: FOGUAVI Sch. transport 3.67% TS School Transport Subsidy (Q27 7,623 FONAPAZ Electricity 8.71% SS mn) Sch.transp: Comision de 15.65% SPS Electricity Subsidy (Q372 mn) 76,374 Vivienda Electricity: MINEDUC 10,212,000 [NDE TS = total spending; SS = social spending; SPS = social protection spending a. Alimony pensions are included in the social insurance system as they are a mechanism regulated by social norms and enforced by the state authority, although implemented outside IGSS b. PRONADE school meals program costs covered under PRONADE category rather than school feeding 134 COVERAGE OF SP PROGRAMS AND PRIVATE TRANSFERS This section analyzes the coverage of SP programs using data from the ENCOVI (the distribution or "incidence" of benefits is discussed in the next section). The ENCOVI collected data at both the household and the individual level for a variety of SP programs. Unfortunately, these do not cover the full range of SP programs provided in Guatemala (see comparison in Table 12.2 above).'2 As such, unless otherwise noted, the remainder of the paper adopts categories of programs collected in the ENCOVI. We estimate that these programs cover between two thirds and three quarters of total social protection spending in Guatemala. Coverage of Public SP Programs Coverage of social insurance programs is extremely limited, with non-poor and urban biases. Overall, 7% of the population lives in households13 that receive social insurance, with only 3% receiving old-age pensions (Table 12.3). Coverage of social insurance is eight times higher Figure 12.3 - Coverage of SA Programs by Poverty Group for the top quintile as the bottom quintile, (ENCOVI 2000) and higher for urban residents and the non-indigenous. o s0% , 80% O 70%- Coverage of social assistance programs o t 6 0%/ is much more extensive and more 0 evenly distributed. Close to four-fifths of 50% the population receives some form of 2 40%- _ S300% - social assistance, and this share is fairly * uniform across quintiles, ethnicities and . 20% 10% areas (Table 12.3). Nonetheless, coverage , 10% differs substantially by specific program, 0i both in terms of overall coverage and by o _ q7, poverty group (Figure 12.3). The " ' electricity subsidy, for example, covers almost all households with electricity, and thus covers a higher share of the non-poor ENon Poor EAll Poor 0 Extreme Poor than the poor. The majority of the population also lives in households benefiting from school feeding programs, and these have a slight coverage bias in favor of the poor. Scholarship and transport subsidies have minimal coverage that favors the non-poor. 135 Table 12.3- Coverage of Social Protection Programs and Private Transfers Beneficiaries as % of population in each group By Quintile By Poverty Group By Ethnicity By Area Total Ql Q2 Q3 Q4 Q5 XP AP NP Ind. Non-Ind Rural Urban All Public SP Programs 80 75 80 83 83 79 74 79 81 79 81 78 84 All Social lnsurance (SI) 7 2 3 5 8 15 2 4 11 3 9 4 11 Pensions 3 1 1 2 3 9 1 1 6 1 5 1 6 Survivorship I 1 0 0 2 3 1 1 2 1 2 1 2 Alimony 3 1 2 3 3 5 0 2 4 1 4 1 5 All Social Assistance (SA) 79 74 79 83 83 76 73 78 80 79 79 77 82 School Feeding 52 60 63 60 51 25 59 61 40 57 48 58 42 Snack 38 39 43 46 42 21 38 42 33 42 36 38 39 Breakfast 28 36 40 34 21 7 34 37 16 32 24 41 6 Powdered Milk 2 2 2 2 2 1 1 2 1 2 1 2 1 Glass of Milk 5 5 7 5 4 2 5 6 3 6 4 6 3 Glass of Atol 35 41 40 42 34 18 41 41 28 39 33 38 31 Scholarships 3 3 2 4 3 3 3 3 3 3 3 2 4 School Materials 23 22 26 24 21 10 33 28 16 26 20 26 18 School Transport Subsidy 2 0 1 2 5 3 0 1 4 1 3 0 5 Electricity Subsidy 53 28 45 58 65 67 24 43 65 49 56 44 66 Other SA 7 9 8 7 7 5 9 8 6 10 5 7 7 All Private Transfers 28 17 24 28 34 36 16 23 35 23 31 24 34 Remittances 21 12 18 19 24 31 11 16 27 17 23 19 24 Local 13 8 12 12 13 19 7 10 16 10 15 11 16 Foreign 9 4 7 8 11 13 3 6 12 7 10 8 9 Charity 13 8 12 14 16 14 7 10 16 10 15 10 17 Public + Private 84 78 83 87 88 87 77 82 87 82 86 82 89 Memo for Comparison Share of total population 100 20 20 20 20 20 16 56 44 43 58 61 39 Shareofpoorpopulation 100 36 36 29 0 0 n.a. n.a. n.a. 58 42 81 19 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Categories are not additive as people can receive more than one type of transfer. Quintiles are individual consumption quintiles. Coverage of Private Transfers Box 12.1 - Access to Market-based Risk Mitigation Mechanisms Private transfers accrue more to non- Only 2% of the population is covered by at least one form of private poor, urban and non-indigenous insurance. The entire group of insured is non-poor. residents (Table 12.3). Overall, about Financial deposits are also used exclusively by the non-poor. Overall 28% of the population lives in households 18% of Guatemalan households have financial deposits, most with that receive private transfers, including private banks (85%). By poverty status, access to financial deposits varies from 30% for the non-poor to 3% for the poor. The main cause remittances (local and foreign) and private of not having financial deposits is a lack of savings (92% of cases). donations (charity). The non-poor are The use of credit services is more evenly spread across the population. twice as likely to receive private transfers Overall, 13% of households applied for and received loans in 2000. as the extreme poor. Urban and non- Although this share is fairly constant across the poverty groups, loan indigenous residents are also more likely sizes differ significantly, with the non-poor receiving an average of to receive such transfers. Remittances are Q1 1,091 as compared with Q2,129 for the poor. even more biased towards advantaged Source: World Bank calculations using the ENCOVI 2000 groups, particularly those from relatives living abroad. Other market-based risk mitigation mechanisms are also biased towards the non-poor (Box 12.1). 136 Figure 12.4 - Duplications and Gaps in Social Risk Management Arrangements Venn Diagram Venn Diagram N = 7276 N = 7276 Private Transfers Social Assistanice Social_Protect (28 0/%T Social Insuran (7 5 176) 1306 1714 (18 %) (24%) 8 Feb 2002 % d trtal F it tt dta (7 Feb 2) 8 Feb 2002 % d total F b: tt_h dta ( 7 Feb 2O) Venn Diagram Venn Diagram N = 7276 Chari t N=7276 h%oc_&eriLfaissslwce / \ Olher-SociaLPAsistance 111 64 9 10% \ 145> 401 J 1235 13 Local_Remittances/e F(r8ei, _Rehttnce 413 = /Sch larsh ps (15~~~~5219 (45176%4) 521 4164 8Feb2002 % d tcMaI FI(E: n_hhdia ( 7Feb2aX ) rFb0 %dKI RI,* UJ0a,.0zo Ove rl ap b etwee n S ocdial R is k M ana ge ment Ar ran g eme nts! 137 Duplications and Gaps in Coverage Duplications and gaps in coverage of social protection and private transfers abound (Figure 12.4). Specifically: * Duplications in Public and Private Transfers. The degree of overlap between public social protection and private transfers is large (upper left-hand panel of Figure 12.4). Specifically, 23% of all households receive support from both the Government and private sources. Private transfers reach only 6% of households not covered by public social protection schemes. * Overlaps between Social Insurance and Social Assistance. Major duplications exist between SI and SA at the household level (upper right-hand panel of Figure 12.4). Three quarters of households that receive pensions also receive other social assistance benefits (such as the energy subsidy). Besides the energy subsidy, one quarter of households that receive pensions also receive other social assistance benefits (mainly programs for children). These duplications are difficult to detect by the administration as most programs target individuals not households. * Duplications between Social Assistance Programs. Given the extensive coverage of school feeding programs, overlaps between SA programs also abound (lower right-hand panel of Figure 12.4). Most children who receive scholarships, school materials, or other social assistance programs also receive school feeding or have siblings that do. Besides school feeding, there is little overlap between the rest of the programs. * Duplications of Private Transfers. The degree of overlapping of private transfers is smaller (lower left-hand panel of Figure 12.4). Households that receive local remittances are generally not the same as those that receive foreign remittances. About half of those receiving private donations do not receive remittances. * Gaps in Coverage. Some 23% of the extreme poor and 18% of all poor were not covered by any type of public or private transfers, as compared with only 13% of the non-poor. The public social protection system fails to reach a quarter of the extreme poor and a fifth of the poor. Virtually all of the poor are excluded from the social insurance system. DISTRIBUTIONAL INCIDENCE (TARGETING OUTCOMES) OF SP PROGRAMS While the last section examined the coverage of programs (in terms of "people" or beneficiaries), this section examines the distribution of benefits (the target "incidence") of social protection programs using data from the ENCOVI 2000. Two concepts are used: (a) absolute target incidence, which measures average benefits received by any particular group as a share of total benefits (or the targeting outcomes of a program); and (b) relative incidence, which measures the average benefits received by any particular group as a share of average total consumption for that group (i.e., the relative "importance" of a program). Incidence of Public SP Programs Overall, public social transfers in Guatemala are regressive in absolute terms, but progressive in relative terms. Specifically, the top quintile receives close to half (46%) of all public social protection spending, as compared with the bottom quintile, which receives only 8% (absolute incidence, Table 12.4). In other words, the richest receive significantly larger absolute public transfers than the poorest. Nonetheless, these transfers are relatively more important to the poor than the non-poor (relative incidence, Table 12.5). Specifically, public transfers represent 9% of total consumption for the poorest 138 quintile, as compared with 5% for the top quintile. Looked at another way, public transfers were less regressive than the existing distribution of total consumption and income. Table 12.4- Absolute Target Incidence of Social Protection Programs and Private Transfers Average transfers received by each group as % of total transfers received by entire population By Quintile By Poverty Group By Ethnicity By Area Total Ql Q2 Q3 Q4 Q5 XP AP NP Ind. Non-Ind Rural Urban All Public SP Programs 100 8 13 15 18 46 6 33 67 25 75 45 55 Al Social Insurance(SI) 100 1 3 5 15 76 1 9 91 9 91 20 80 Pensions 100 1 2 4 12 81 1 6 94 9 91 17 83 Survivorship 100 4 4 4 13 75 2 11 89 9 91 16 84 Alimony 100 1 6 10 24 60 0 16 84 10 90 30 70 All Social Assistance (SA) 100 14 21 24 21 20 10 54 46 39 61 66 34 School Feeding 100 16 25 27 20 11 12 63 37 43 57 79 21 Snack 100 13 21 26 26 14 10 55 45 39 61 59 41 Breakfast 100 17 28 29 17 9 12 68 32 43 57 92 8 Powdered Milk 100 30 26 14 16 14 7 65 35 62 38 56 44 Glass of Milk 100 16 29 25 19 12 11 65 35 49 51 75 25 Glass of Atol 100 17 22 25 23 14 13 57 43 42 58 64 36 Scholarships 100 9 4 23 16 48 3 30 70 47 53 28 72 School Materials 100 18 24 24 20 13 14 60 40 35 65 69 31 School Transport Subsidy 100 0 2 15 56 27 0 16 84 8 92 3 97 Electricity Subsidy 100 2 3 9 22 65 1 12 88 16 84 18 82 Other SA 100 13 20 16 17 34 11 48 52 46 54 53 47 All Private Transfers 100 4 8 14 21 54 2 22 78 24 76 41 59 Remittances 100 4 8 14 20 55 2 23 77 24 76 43 57 Local 100 4 8 11 16 63 3 20 80 18 82 31 69 Foreign 100 4 7 16 23 49 2 25 75 29 71 52 48 Charity 100 2 9 13 24 51 2 20 80 22 78 32 68 Public + Private 100 6 10 14 19 50 4 28 72 24 76 42 58 Memo for Conmarison Share of total population 100 20 20 20 20 20 16 56 44 43 58 61 39 Share of poor population 100 36 36 29 0 0 n.a. n.a. n.a. 58 42 81 19 Share of total consumption 100 5 9 13 20 54 4 24 76 24 76 37 63 Share of total income 100 2 7 11 18 62 4 24 76 23 77 35 65 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadtstica - Guatemala. Categories are not additive as people can receive more than one type of transfer. Quintiles are individual consumption quintiles. In contrast, social insurance is highly regressive in both absolute and relative terms. Three quarters of all social insurance benefits go to those in the richest quintile of the population; the poorest quintile receives only 1% (Table 12.4). Moreover, social insurance benefits represent 4% of total consumption for the richest quintile, as compared with less than 1% for the poorest (Table 12.5). Social assistance programs are generally better targeted, though target outcomes vary significantly by program. Overall, the top quintile receives a larger share of social assistance transfers than the poorest quintile (Table 12.4). Nonetheless, SA programs are relatively more important to the poor (contributing 8.4% of total consumption of the poorest quintile) than the rich (accounting for just 1.1% of total consumption of the top quintile, Table 12.5). Target outcomes vary significantly by program. Scholarships, for example, are very poorly targeted, with the top quintile capturing close to half of all scholarship benefits (Table 12.4). Similarly, the top two quintiles receive 83% of all subsidies to school transport. Likewise, 65% of all electricity subsidies accrue to the top quintile. These programs are quite small however, in terms of their contribution to average consumption in any quintile (Table 12.5). In contrast, school feeding and the school materials assistance program (bolsa de utiles escolares) benefit primarily the middle quintiles (Table 12.4), but are still relatively more important to the poorest (Table 12.5). 139 Incidence of Private Transfers Private transfers are regressive in both absolute and relative terms. The top quintile receives over half of all private transfers (remittances and donations) in Guatemala (Table 12.4). Private transfers are also relatively more important for the rich, contributing 5% of total consumption for the top quintile as compared with 3.5% for the bottom quintile (Table 12.5). Interestingly, local remittances are more regressive than foreign remittances or charity in absolute terms. Table 12.5 - Relative Incidence of Social Protection Programs and Private Transfers (the "importance" of the transfers) Avera ze transfers received by each group as %.of average total consumption for each group By Quintile By Poverty Group By Ethnicity By Area Total Ql Q2 Q3 Q4 Q5 XP AP NP Ind. Non- Rural Urban Ind All Public SP 5.8 9.1 8.5 6.8 5.3 5.0 8.9 8.0 2.5 5.9 5.8 7.0 5.1 All SI 2.8 0.8 1.0 1.1 2.1 3.9 0.7 1.0 0.3 1.1 3.3 1.5 3.5 Pensions 1.8 0.4 0.4 0.5 1.1 2.7 0.5 0.4 0.1 3.3 2.2 0.9 2.3 Survivorship 0.4 0.2 0.2 0.1 0.2 0.5 0.2 0.2 0.1 0.7 0.4 0.2 0.5 Alimony 0.6 0.1 0.4 0.5 0.7 0.7 0.0 0.4 0.1 0.1 0.7 0.5 0.7 All SA 3.0 8.4 7.5 5.7 3.2 1.1 8.2 7.0 2.1 4.8 2.5 5.5 1.6 School Feeding 1.9 6.2 5.7 4.2 2.0 0.4 6.0 5.2 1.6 3.4 1.5 4.2 0.6 Snack 0.4 1.0 0.9 0.8 0.5 0.1 1.0 0.9 0.3 0.6 0.3 0.6 0.2 Breakfast 1.1 3.6 3.5 2.4 0.9 0.2 3.5 3.1 1.0 1.9 0.8 2.7 0.1 Powdered Milk 0.0 0.1 0.1 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 Glass of Milk 0.1 0.2 0.2 0.1 0.1 0.0 0.2 0.2 0.0 0.1 0.0 0.1 0.0 Glass of Atol 0.4 1.3 1.0 0.8 0.5 0.1 1.4 1.0 0.3 0.7 0.3 0.7 0.2 Scholarships 0.1 0.2 0.1 0.3 0.1 0.1 0.1 0.2 0.1 0.3 0.1 0.1 0.2 School Materials 0.4 1.3 1.1 0.7 0.4 0.1 1.4 0.9 0.3 0.5 0.3 0.7 0.2 Sch.Transport 0.1 0.0 0.0 0.1 0.2 0.0 0.0 0.1 0.0 0.0 0.1 0.0 0.1 Subsidy Electricity Subsidy 0.3 0.1 0.1 0.2 0.3 0.4 0.1 0.1 0.0 0.2 0.3 0.1 0.4 Other SA 0.2 0.6 0.5 0.3 0.2 0.1 0.6 0.4 0.1 0.4 0.2 0.3 0.2 All Private Transfers 5.0 3.5 4.7 5.3 5.2 5.0 3.0 4.7 5.3 4.8 5.0 5.5 4.6 Remnittances 3.9 3.1 3.5 4.2 3.9 3.9 2.5 3.8 4.0 3.9 3.9 4.6 3.5 Local 1.6 1.3 1.3 1.4 1.3 1.9 1.3 1.4 1.8 1.2 1.8 1.4 1.8 Foreign 2.3 1.8 2.3 2.8 2.6 2.0 1.2 2.4 2.6 2.6 2.1 3.2 1.7 Charity 1.1 0.4 1.2 1.1 1.3 1.0 0.5 0.9 1.2 1.0 1.1 0.9 1.1 Public + Private 10.8 12.6 13.2 12.1 10.5 10.0 11.9 12.7 3.9 10.7 10.8 12.5 9.8 Memo: Avg Total Cons. 6161 1580 2639 3884 6076 16632 1460 2580 10754 3531 8108 3668 10122 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. Categories are not additive as people can receive more than one type of transfer. Quintiles are individual consumption quintiles. OVERALL EFFECTIVENESS OF SP PROGRAMS The indicators of coverage, absolute target incidence, and relative target incidence (importance/adequacy) all reveal important information about the effectiveness of SP programs. This section seeks to combine those multiple indicators for a more comprehensive review of these programs, in particular with respect to their effectiveness in reducing poverty. Figures 12.5 and 12.6 plot in a single graphs the three related concepts of coverage, absolute target incidence, and adequacy for various social protection programs and private transfers based on a simulated model that classifies the poor based on a counterfactual of consumption without the transfers. The x-axis presents the coverage of the "ex ante" poor (the poor without the transfers, which represent 61% of the population). The share of total benefits received by the "ex ante" poor is plotted on the y-axis (absolute target incidence). Adequacy (relative incidence) is captured by the size of the "bubbles" in the graphs. A "perfectly-targeted program" would be located on the upper right-hand side of these graphs, with a large bubble (equal to the size of the poverty gap before the transfer). 140 While none of the programs are close to "perfect" in terms of targeting or coverage, some are better than others (Figure 12.5). In particular, only social assistance provides relatively high coverage with relatively large transfers. Social insurance and private transfers all cover a smaller fraction of the poor. In terms of adequacy (relative importance), remittances and social insurance are the most important for their poor recipients (as a share of ex ante consumption). None of the programs are very well targeted in terms of the absolute share of funds going to the ex ante poor. Since the ex ante poor represent 61% of the population, even social assistance is at best a neutral benefit in terms of absolute target incidence (with 66% of all benefits going to the ex ante poor). Figure 12.5 - Effectiveness of Social Protection and Private Transfers in Reducing Poverty 11 0 90%e- ---- ---- I ---- Lr--l----J- ---T--- I ---- ILF---- l 0 % 1, J - - - - o ADYe- u 1 X l J____1___ ~~Social\l .C 63Ye- wla_ | f _ T--- zAssistancey n ~~~Remittancesi /l .. 497ez r~~~~~~ > -rI---- L ----- I---- r---------l - o - - I I I F C % A F F FFSocial I I oc I 3ma1nsuranc* - - ' - - - - ,' - - - -L- - - - ' - - - -I' - - - -L- - - - 30tances anF .= 20 7e ---- ----a----r----l-- In - - --- L- - -- - --l n 07 -----J---- I----_L_---- _J ---- I----L---- - - J ---- l Chnt F F F F -10% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%l Coverage of the poor: % of ex-ante poor recehAng the benefitl Note: The size of the bubbles is proportional to the relativel importance of the transfer in the consumption of the recipient World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadt'sticas - Guatenmala Withirn social assistan1ce, the effectivenless of programs varies slignDDfierntly (Figure 12.6). Three types of programs are observed. First, the school feeding programs (combined) and the school materials program have the best target incidence and relatively high coverage. Second, the two subsidy programs, for electricity and school transport, have extremely low targeting incidence and adequacy. From a poverty-alleviation perspective, these programs should be redesigned or eliminated. Third, scholarships have extremely low coverage of the poor, target incidence, and adequacy. Again, these programs should be redesigned to better target the poor. 141 Figure 12.6 - Effectiveness of Social Assistance in Reducing Poverty , 100, I I~~~~~ w - - E/o- e ShI - er r-8------ O lv ------- -SchoolFeedI ing-- C - - o --l- - materials - - ' her S assistance i I I CO) ----D/ - - --- --------------ShoFedn- - - - - -, - - - - c 0o - - ---- - IL ---- --- ---- - I I Elcr~t -- - .--0 ubsdy .I Other SA , . - - - - I - a)~~~~~ I , I I I , I a 'Scholarships I _o ----4,09A HkI", H--k---H----k---H--- ..I I I .f, Note Th siz of th bble is prprinl toterltv V --/- School importantransport' trasfr i fi t rci World Bank calculation us2subsidyol d i - G I I. , I l l I I , ,U r Siultin sugs ltl imac of soil prtcto prgrm onpvryolnqaiy aafo -2 % -19% 0 /a 1q% 2q% 3q% 4q% 5q% 6q% 7q% 8q% 9q% 1 0 Coverage of the poor: % of ex-ante poor receityng the benefit f Note: The size of the bubbles is proportional to the relative importance of the transfer in the consumption of the recipient World Bank calculations using the ENCOVI 2000, Instituxto Nacional de Estadtsticas - Guatemala IMPACT ON POVERTY AND INEQUALITY Simulations suggest little impact of social protection programs on poverty or inequality. Data from the ENCOVI 2000 were used to simulate poverty and inequality in an absence of social protection programs or transfers.'4 Total transfers (public and private) succeed in reducing poverty from 61% without the transfers to 56%, and the poverty gap index from 0.29 to 0.23 and the poverty severity index from 0.19 to 0.12. Nonetheless, these reductions were only significant for the poverty gap (depth) and severity indices, but not the headcount index. In other words, total transfers will not lift many people out of poverty, but will reduce the depth and severity of their destitution. While social insurance has no significant impact on poverty, social assistance does have a significant impact on the depth and severity of extreme poverty. Similarly, while an elimination of all transfers (public and private) would increase inequality from a Gini of 48 to a Gini of 51, some transfers (e.g., social insurance and the electricity subsidy) actually contribute to an increase in inequality. COST-BENEFIT ANALYSIS Comparing the costs's and benefits of social protection programs reveals that social assistance programs are the most efficient in reducing poverty. Social assistance programs cost between Q1.4 to Q2 to reduce the poverty gap by Ql. School feeding programs, for example, costs Q1.5 on average for each Q1 reduction in the poverty gap. Some social assistance programs, however, are quite cost inefficient: school transport subsidies and scholarships cost Q6 and Q3 to reduce the poverty gap by Ql. 142 The energy subsidy is even less efficient, costing Q8 for every Ql reduction in the poverty gap. In contrast, social insurance programs are extremely inefficient for poverty reduction, costing between Q5 to Q9 for a reduction of Ql in the poverty gap. SUMMARY OF KEY ISSUES AND PRioRi[Es The above analysis reveals a number of key messages and recommendations: * Strategic priorities for social protection in Guatemala should (a) seek to maintain the current focus on children, given their inherent vulnerabilities and prospects for long-term transmission of poverty; and (b) seek to build the assets of the poor, given the dominance of chronic, rather than transient, poverty as discussed in Chapter 11. Public social protection programs can play an important role in building the assets of the poor. Specifically, when designed properly, conditional-transfer programs can be quite effective in helping ease demand-side constraints, which have been shown to constitute important limitations for improved coverage of education (Chapter 7), health (Chapter 8) and basic utility services (Chapter 9). Some transfers (social assistance) could also be desirable to alleviate the poverty and suffering of the extreme poor. * Although low by international standards, public spending on social protection has increased since the Peace Accords. Much could be done within the existing budget envelope to improve the effectiveness and efficiency of public social protection spending as an instrument to reduce poverty and vulnerability, including: (a) eliminating programs that are poorly targeted, inefficient and ineffective; and (b) consolidating, rationalizing, improving the targeting and expanding coverage of remaining programs. * Better targeting could be achieved using various tools, including geographic targeting via the poverty map (e.g., selecting eligible municipalities with high concentrations of poor or extreme poor people, as has recently been done for scholarships), categorical targeting (e.g., no connection to electricity, which is a very good proxy for the extreme poor), piggy-backed targeting via self-targeted services (e.g., channeling benefits through health posts or community health centers which are well targeted to the poor), or some form of means testing (or a combination of targeting tools). o Certain social assistance programs should be eliminated due to extremely poor targeting, ineffectiveness, and cost inefficiencies. The electricity subsidy is a candidate ripe for elimination - the funds would be better used to further expand electricity coverage for the rural poor (as discussed in Chapter 9). If a sudden single elimination seems politically unfeasible, the eligibility threshold could be substantially reduced (e.g., to 100 KwH or less) as an intermediate step on the way to a phased elimination. The school transport subsidy is another candidate for elimination or redeployment to rural areas (where a lack of public transport to school is an issue) since, as currently designed, it offers virtually no benefits to the poor and is virtually entirely urban focused. * The seven miniscule and poorly-targeted scholarship programs should be consolidated, streamlined, and explicitly targeted to the poor. A well-designed scholarship program has the potential to be an effective conditional-transfer that could ease demand-side constraints to enrollment and attendance for the poor (as discussed in Chapter 7). To do so, the programs should be effectively targeted to the poor (individuals or entire schools in very poor communities) and benefits should be tied to well- monitored daily attendance (preferably paid on a more frequent basis according to attendance registries). At present the dispersed set of programs do not satisfy either condition. * School feeding programs can also provide strong incentives for regular school attendance in poor communities. Indeed, children in the QPES study cited the "refacci6n" (school snack) as one of the main reasons the like going to school. School feeding programs should not, however, be considered 143 as nutrition interventions, since (a) malnutrition occurs at a much younger age; and (b) Guatemala's nutrition problems do not arise primarily from a lack of food (see Chapter 8). Guatemala's school feeding programs could be strengthened by: (a) improving the targeting of school eligibility based on the poverty map combined with educational attendance and enrollment information; (b) consolidating and rationalizing the implementation of the various programs to improve efficiency and institutional responsibility (and avoiding the recent state of flux that has commonly disrupted program execution); and (c) decentralizing food procurement and preparation to the communities by providing eligible and certified communities (or hub communities serving several eligible satellite communities) block grants to purchase and prepare food locally, which would have the benefits of improving community participation, stimulating local economies, and reducing the costs of shipping and storage of food (and possibly improving quality by allowing for discretion and local tastes). The programs should also investigate the use of powdered milk, which could actually be dangerous for children if combined with unpotable water. * The school materials program (bolsa de utiles) could be improved with explicit targeting (see above). * Regarding the creation of new social assistance programs, Guatemala may want to consider conditional transfer programs channeled through self-targeted health services (health posts and community health centers) that require certain health interventions for children to be eligible (e.g., growth monitoring, vaccinations). These have the dual advantage of alleviating short-term poverty (by providing relief via a transfer payment) and reducing long-term poverty (by providing incentives to build incentives). Well-targeted conditional transfers have proven to be quite effective in other countries (e.g., Progresa in Mexico). An expansion of well-targeted workfare programs may also be considered to help provide seasonal employment alternatives, particularly for those dependent on migration to the coffee fincas, in light of recent terms-of-trade shocks. * Disaster management and relief should be expanded and improved, since Chapter 11 shows that the poor in Guatemala are disproportionately more exposed to natural disasters and agriculture-related shocks: Such interventions should be well-targeted to the poor and delivered in a timely manner. Since exposure to some natural disasters does seem to be largely determined by location and geographic factors, administrative maps of vulnerability to drought, seismic activities, hurricanes, storms, frosts, and landslides could be quite useful instruments for risk management planning. Many such maps have been prepared by the Ministry of Agriculture and FAO, and are available to policy makers. Their ability to target limited funds for disaster relief would be greatly enhanced if used in conjunction with poverty maps (since those who are already poor are less equipped to cope with shocks, as shown in Chapter 11). Since the impact of natural disasters often includes damage to, or destruction of, community infrastructure (in addition to income and wealth losses at the household level, see Chapter 11), the social funds may serve as the institutional channel by which such relief and infrastructure rehabilitation is implemented. * When budgets and administrative capabilities permit, the Government should seek ways to improve the public social insurance coverage of the poor. Recent attempts to cover specific vulnerable groups, such as agricultural migrant workers, seem promising and should be considered for expansion. * Private transfers do constitute an important source of income, accounting for almost half of all transfers between households in Guatemala. Nonetheless, they should not be viewed as a substitute for government assistance of the poor, since they are highly regressive. Nonetheless, attention should be paid to the effects of global economic recessions on such transfers, since they do contribute to the incomes of the poor. 144 X(Holzmann and Jorgensen 2000). 2 Transfer schemes include non-contributory payments in cash (e.g., family or child allowances), near cash (e.g., food stamps, vouchers), and in- kind (e.g., food supplementation, school feeding). 3Subsidies, such as those on food, energy and housing, artificially lower the price of certain goods or services either for the entire population (generalized subsidies) or for certain sub-groups (e.g., self-targeted food subsidies). Subsidies are explicit if their cost is paid for by the govemrnent and implicit if it is bome by the producers of the goods or services. Workfare programs are transfer schemes that require the beneficiaries to work for their benefits. Workfare benefits are paid either in cash or kind (e.g., food-for-work schemes) or some combination. 5 See GUAPA Technical Paper 10 (Tesliuc and Lindert, 2002) for details. 6 SlAF/Ministry of Finance (communication of 2-12-02). 7IMF estimates. IMF (April 2001). Other estimates put social funds spending at 1.64% of GDP in 2000. Santiso (2001). 8 Inventory conducted by URL (2001); analysis conducted by Santiso (2001). 9 Much of this paragraph is based on Santiso (2001), which draws largely on the inventory conducted by URL (2001). ° World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadtsticas - Guatemala. 11 Fuentes & Asociados (April 2001). 12 Programs with low beneficiary coverage, such as land or housing subsidies, were only scarcely captured in the ENCOVI and will be omitted for lack of estimability. 13This analysis uses the household, not the individual as a reference unit. Nonetheless, to ensure comparability with poverty statistics, coverage figures are reported as on a per capita basis (weighted by household size). There are several reasons for this presentational fomiat. First, although most SP benefits provide individual benefits, some target households (e.g., energy subsidies and some form of govemment aid). Moreover, private transfers are recorded only at the household level. Thus the household is the smallest common denominator where all transfers can be presented and compared. Second, benefits provided for one individual will have spillover effects for the other members in the households, which will adequately be captured at this level of aggregation. Coverage is thus presented as the share of people residing in recipient households in a particular population group. 14 See GUAPA Technical Paper 10 (Tesliuc and Lindert, 2002) for details on methodology and results. I5 Due to data limitations, the costs used for this analysis include only the value of the benefits provided by the programs as reported by the beneficiaries in the ENCOVI. They do not include administrative costs or potential incentive effects. See GUAPA Technical Paper 10 (Tesliuc and Lindert, 2002) for details. 145 PART 4: KEY CHALLENGE: EMPOWERMENT "The services at the municipality are deficient. Sometimes there are lots of people and you have to wait two days [to obtain documentation]... When there's a political campaign, they treat you well, after that, no... They always tell us 'wait over there' and when a Ladino enters, they always say 'come on in' (pasen adelante)." Q'eqchi villagers, QEI (QPES) Chapter 13: Building Institutions and Empowering Communities As discussed in Chapter 4, one of the key remaining challenges for the Peace Agenda is the modernization of the state and a strengthening of community and social participation.' This chapter contends that these institutional forces are also crucial for the "Poverty Agenda," influencing the menu of options available to the Government in future efforts to reduce poverty and the way in which these options are carried out. Indeed, poverty is not only the result of economic processes, but also of interacting economic, social and political forces. In particular, it is driven by the accountability and responsiveness of state institutions. Social institutions - kinship systems, community organizations, and informal networks - also greatly affect poverty outcomes, helping communities manage public goods, cope with risks and shocks, and leverage external assistance. In light of the importance of these factors for both poverty and the Peace Agenda, this chapter reviews key institutional challenges in the areas of (a) public sector management; (b) governance; and (c) community participation and social capital. The chapter also considers the role of other important actors in development, namely the private sector, NGOs, and religious organizations. PUBLIC SECTOR MANAGEMENT Guatemala's democratic transition faces a number of challenges regarding effective public sector management. An effective government is a vital necessity for development and poverty reduction. At a minimum, governments should effectively seek to create an environment that is conducive to economic growth, provide basic public goods - such as defense, law and order, protection of property rights, public health and stable macroeconomic management -- and protect the poor via anti-poverty programs and disaster relief. Other important functions include the management of externalities (basic education, pollution), the regulation of monopoly power, and the provision of social insurance, financial regulation and consumer protection to overcome imperfect information and improve equity.2 Although Guatemala has made serious efforts since the Peace Accords to improve living conditions and promote a more inclusive, democratic society (as discussed in Chapter 4), these efforts have been hampered by a weak public sector. Such weaknesses curb the Govemment's ability to deliver services, create an environment conducive to growth, and reduce poverty. Key challenges in this area include: (i) a fragile and inadequate fiscal base; (ii) weak public expenditure management; (iii) a weak civil service; and (iv) an overly centralized government. These issues are treated in turn below. Inadequate Tax Base Despite some progress, Guatemala's inadequate tax base remains a fundamental challenge to improving the government's ability to deliver effectively and to fulfilling a key target established by the Peace Accords. Successive post-war governments have made some progress in increasing total govemment revenues, which rose from 9.2% of GDP in 1996 to an estimated 11.1% in 2001.3 Such increases were aided by the adoption of a variety of tax measures, including increasing the VAT rate from 7% to 10% in 1996, and again from 10% to 12% in August 2001 amnid strong popular opposition, as well as efforts to improve the efficiency of the tax collection agency and clamp down on tax evasion. Nonetheless, the current level of revenues fails to meet the targets set by the Peace Accords (see Chapter 4) and remains relatively low. In comparison, current revenue averaged 14.2% for lower-middle income countries and 20.1% for all of LAC for 1998.4 Moreover, the outlook for public finances in coming years is blurred by 146 the weak state of the domestic economy, the likely impact of the US and global recession, and the cost of consolidating and modernizing the country's financial system. Without adequate resources, the Government's hands are tied in its ability to deliver the services and investments needed to reduce poverty. The inadequate revenue base thus constitutes one of the key challenges for both the Peace Agenda and the "Poverty Agenda." Public Spending and Public Expenditure Management The Government has made considerable progress in increasing public spending and improving public expenditure management. Public spending has increased since the signing of the Peace Accords, with notable gains in sectors that are crucial for poverty reduction (Table 13.1). Moreover, progress has been made in improving the management of public expenditures with the introduction and implementation of the Integrated Financial Management System (SIAF) since 1998. Accomplishments include inter alia: extending SIAF coverage to some 82% of public spending for budget formulation, execution, cash management and internal audit; eliminating arrears to suppliers; improving transparency with full, real-time open access to account information; implementing reforms in public procurement processes; and improving public investment planning and better coordination with SEGEPLAN. Furthermore, the gap between planned and executed spending has improved since the introduction of SLkF (Table 13.1). Table 13.1 - Public Spending 1996-2001 1996 1997 1998 1999 2000 2001 Total Spending, mn Q.' 9,914.8 12,618.2 16,637.0 19,239.2 19,801.2 22,182.2 % of Total Spending: 100% 100% 100% 100% 100% 100% Health 6.3 6.9 7.1 8.3 7.9 7.9 Water, sanitation, environment 1.8 1.9 1.2 1.8 2.2 2.2 Education 15.2 15.1 15.7 17.1 18.3 20.0 Housing 0.0 0.5 4.0 2.1 0.1 0.6 Justice 3.0 3.6 3.3 3.5 3.9 4.2 Urban and Rural Development 7.0 7.6 7.3 6.8 6.3 8.2 Spending/GDP (%):' Total 10.4 11.7 13.4 14.2 13.4 13.8 Health 0.7 0.8 0.9 1.2 1.1 1.1 Water, sanitation, environment 0.2 0.2 0.2 0.2 0.3 0.3 Education 1.6 1.8 2.1 2.4 2.5 2.8 Housing 0.0 0.1 0.5 0.3 0.0 0.1 Justice 0.3 0.4 0.4 0.5 0.5 0.6 Urban and Rural Development 0.7 0.9 1.0 1.0 0.8 1.1 Executed/Planned Total 78 83 100 91 93 100 Health 62 92 92 90 94 134 Water, sanitation, environment 51 69 97 74 87 134 Education 112 93 95 92 98 96 Housing 7 96 95 98 83 53 Justice 98 100 100 93 91 95 Urban and Rural Development 92 91 83 97 78 90 Sources: SIAF/Ministry of Finance (Feb. 12, 2002). a. Spending amounts refer to executed amounts (devengado). Note that the executed/planned ratio may be underestimated for 2001 since some executed amounts may not have been recorded at the time the figures were provided. Nonetheless, public spending remains low and public expenditure management still suffers from a number of weaknesses. Despite increases, the overall level of public spending remains relatively low as a share of GDP (Table 13.1), hampered largely by the inadequate revenue base (see above). In comparison to Guatemala, the averages for lower-middle income countries and the LAC region overall were 18.8% and 147 21.0% of GDP respectively in 1998.5 In addition, despite widespread coverage of the SIAF, the system needs to be extended to cover the remaining 18% of government spending, most notably to bring in the municipalities and social funds. Spending allocations are also fairly inflexible, with a substantial share of the budget earmarked for specific uses by legal mandates (24% of the total budget).6 Moreover, the current budget planning process is weak, involving an overly detailed focus on incremental changes in budget line items or agency spending ceilings rather than a more comprehensive understanding and review of national priorities and government spending tradeoffs by all stakeholders (as would occur under a medium-term expenditure framework). Annual budgeting also hinders proper planning for multi-year projects and makes them more susceptible to political maneuvering and electoral cycles. In addition, public spending is poorly targeted, with a limited share of public resources actually reaching the poor. As shown in Table 13.2, the poor receive a slightly smaller share of public spending on education and heath than their relative share in the population, and disproportionately less public spending on social protection. Investments by the social funds are slightly better targeted, though they still only transfer a slightly disproportionate share to the poor (62%) as compared with their share in the population (56%), as shown in Figure 13.1.7 Nonetheless, social spending is relatively more progressive than the current distribution of income and consumption (Table 13.2), and would thus be inequality reducing. Table 13.2 - Distributional Incidence of Public Spending, by Sector and Socio-Economic Group % of total benefits received by each group By Quintile By Poverty Group By Ethnicity By Area Total QI Q2 Q3 Q4 Q5 XP AP NP Ind. Non-Ind Rural Urban Education - Total 100 17 21 21 21 21 13 55 45 37 63 59 41 Health - Total 100 17 18 23 25 17 12 53 47 40 60 64 36 Social Protection - Total 100 8 13 15 18 46 6 33 67 25 75 45 55 Social Insurance - Total 100 1 3 5 15 76 1 9 91 9 91 20 80 Social Assistance-Total 100 15 23 25 21 16 11 58 42 41 59 70 30 Memo for Comparison Share of total population 100 20 20 20 20 20 16 56 44 43 58 61 39 Share of poor population 100 36 36 29 0 0 n.a. n.a. n.a. 58 42 81 19 Share of total consumption 100 5 9 13 20 54 4 24 76 24 76 37 63 Shareoftotalincome 100 2 7 11 18 62 4 24 76 23 77 35 65 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estadtstica - Guatenmala. See Chapters 7, 8, 9, and 12 for details on the breakdown of the target incidence by sector. Quintiles are individual consumption quintiles. A Weak Civil Service Figure 13.1 - Estimated Share of Social Fund Spending Received by Poor; Sources: UNDP (2001b) Combined The ability of the Government with Data from the ENCOVI 2000 to deliver key public services 100%- and programs is hampered by 80% 65/ 58%0 62% 56% 60% low administrative capabilities 40%0-_ and a weak civil service. Weak 20% hiring practices, a lack of 0% continuity, and uncertainty a LL o regarding policies and budgets 6 I oo hinder the effectiveness of a. Guatemala's civil service: * Relatively small civil service. Reflecting the relatively low share of public sector spending in GDP, Guatemala's civil service is quite small by international standards.8 The ENCOVI suggests that just over 200,000 people were employed by the public sector, representing just 5% of total employment and 3% of the labor force. The largest public employers were education (62%), the Ministry of Interior and National Security (Gobernaci6n, 148 16%), and health (14%).9 The number of civil service has been growing at a rate of 1.5% per year since 1996, or by about 10,000 workers in total for the five years since the Peace Accords were signed. The main areas of growth have been in education, which grew by 3.8% p.a., and Ministry of Interior and National Security (Gobernaci6n), which almost doubled in the five years since the Peace Accords (growing at an annual rate of 16.7%). o Weak incentives, lack of continuity. Because the official civil service salary structure is lower than for comparable professions in the private sector,'0 the Government has resorted to hiring temporary workers and "consultants" as a widespread practice in order to offer better incentives for hiring qualified people. As a result, the ENCOVI reveals that only 15% of public-sector workers have permanent contracts, while 62% have temporary contracts and 23% report having no contract at all." Such temporary status has obvious adverse impacts for efficient public sector management, including an undermining of institutional memory, continuity, technical capacity, and staff motivation. o Arbitrary hiring and evaluation practices. Moreover, hiring practices appear to be fairly arbitrary. According to a survey of current and former civil servants, carried out by CIEN (August 2001), although 64% of survey respondents indicate having been hired based on merit criteria, some 77% also respond that other factors, such as political connections, factored in hiring decisions. Two-thirds indicate that changes in government yield substantial changes in the policies they are implementing. Some 88% complain of uncertainty in their budgets, with divergences between planned and actual spending allocations yielding significant changes in their work programs. In addition, almost half (47%) indicate that their work is not evaluated according to clear performance standards. Finally, there is little representation of the indigenous in the government structure or the ethnic make-up of the civil service. Decentralization Finally, the government is overly centralized, particularly in light of Guatemala's heterogeneity. The Central Government accounts for 68% of total public spending.'2 The budget and financial management are also highly centralized under the Ministry of Finance, though a recent pilot with the Ministries and Health and Education gave them increased control over their own budgets. Government offices and services are disproportionately concentrated in the capital. Given the heterogeneity of Guatemala's population, economy, and topography, decentralization of many public functions could bring the government closer to the client and improve the delivery of public services. Guatemala has made some initial progress on the long path towards enabling decentralization, for example with the introduction of general revenue transfers to municipalities in the 1980s,'3 some deconcentration of sectoral management to department levels, the channeling of resources via newly created and elected Regional and Departmental Councils, and the creation of the Presidential Commission for the Modernization and Decentralization of the State. Since the early 1990s, several mechanisms have emerged with decentralized management of specific programs and activities, such as the Sistema Integral de Atenci6n de Salud (SIAS) and the Programa Nacional de Autogesti6n para el Desarollo Educativo (PRONADE). Social funds, which work directly with communities to expedite the delivery of infrastructure and services, have also increased their share of government spending (though they remain centralized agencies). As for decentralization to municipalities per se, Congress recently passed three laws to broadly define areas of responsibility (the Ley de los Consejos de Desarrollo Urbano y Rural, the C6digo Municipal, and the Ley General de Descentralizaci6n), although leaving the final definition of responsibilities to laws pertaining to specific sectors and central government decisions (reflecting the differing sectoral needs and characteristics). Such laws now need to be put into effect. 149 GOVERNANCE Related to its weaknesses in public sector management, Guatemala scores fairly poorly on most governance indicators. There is strong empirical evidence of a causal relationship between good governance and better development outcomes, including higher per capita incomes, lower infant mortality, and higher literacy. 14 World Bank researchers recently compiled a massive cross-country database of some 300 governance indicators, yielding six key clusters of composite measures: voice and accountability, political instability, government efficiency, regulatory burden, rule of law, and corruption.'5 Guatemala scores poorly for most of these (Table 13.3). Nonetheless, a summary indicator of country policy and institutional quality (CPIA) suggests that Guatemala has made some progress over time, particularly since the Peace Accords (Figure 13.2). Key challenges for the development and poverty reduction agenda include fighting corruption, improving the rule of law and the justice system, and reducing political instability, as discussed in more detail below. Table 13.3 - Composite Governance Indicators, International Comparisons, 2001 World Rankings (lower is better), Numbers in parentheses refer to Guatemala's ranking out of LAC countries Voice and Political Government Regulatory Rule of Law Corruption Accountability Instability Effectiveness Burden Guatemala 105 (21/26) 130 (23/26) 112 (21/26) 83 (15/27) 147 (25/27) 116 (21/26) El Salvador 68 49 87 23 119 86 Honduras 84 70 108 104 153 111 Nicaragua 86 65 116 106 134 122 Panamd 49 50 81 25 80 94 Costa Rica 16 24 37 26 50 32 CA Median 76 58 98 55 127 99 Number of 173 161 159 169 170 161 Countries (N=) - Source: Kaufmann, Kraay, and Zoido-Lobat6n (January 2002). Database and definitions can be accessed at the following website: http://www.worldbank.org/wbi/govemance/govdata2001.htm Figure 13.2- Country Policy and Institutional Assessment (CPIA), 1977-98: LAC and Guatemala CPIA 1977-98 3.50 3.00 - -. - LAC 2.50 - Smoothed 2.00 - - Median 1.50 - -Guate 1.00 - ,,,,,,,,,, ,,,,,,, Smoothed IZP ;' R Ibc lil (bbqlbMedian Source: Collier and Dollar (20XX). Corruption Guatemala's corruption problem is serious. Indeed, Guatemala ranks among the worst 28% of countries in the world on an international composite measure of corruption (Table 13.3). The current Government has been plagued by a stream of corruption scandals, including allegations of misuse of public funds, mis- procurement, embezzlement, and cover-ups. 16 A recent public opinion survey found that 92% of Guatemalans perceive that corruption is high.17 A recent report by Transparency International alleges that 150 24% of the annual national budget in 2001 would be lost to corruption and financial mismanagement'8 - a particularly serious problem given Guatemala's already low public finance base. Guatemala was also recently placed on the OECD-backed Financial Action Task Force (FATF) "money-laundering blacklist," though the recent passage of the "anti-money laundering" law in November 2001 should contribute to improving the country's standing in this respect.19 Although a special public prosecutor's office was set up in 2000 to deal with corruption cases, to date only 13 cases have been taken to court, with just two 20 convictions. Corruption occurs not only at government levels, but even at the village level; in the QPES, numerous villages reported problems with misuse of funds by community development committees (or specialized committees such as the water committee). On-going extension and implementation of the SLAF system should help improve transparency and reduce discretional spending and misuse of public funds. In addition, the Government has recently launched an important initiative to combat corruption. This initiative-formalized in a Letter of Intent from the Government-involves the creation of a national program to promote global transparency in Guatemala. With support from the World Bank and the international community and the participation of civil society, it will seek to further diagnose and analyze the issue of corruption in Guatemala and develop a national plan to combat corruption. This initiative represents an important step in improving governance and promoting transparency in Guatemala. Corruption affects the poor in several ways. Corruption affects the poor indirectly via it's adverse impact on economic growth, which is critical for poverty reduction.21 The poor are also affected directly, since corruption (e.g., bribes) levies a type of "regressive tax," (though such "taxes" are worse than income taxes because their amounts are uncertain and the revenues are not plowed back into the provision of public services). In fact, households in the ENCOVI ranked corruption/bad government as the second main cause of poverty in Guatemala (tied with "high prices" and following a lack of adequate employment opportunities, which was ranked as the top cause). There are also clear links between Guatemala's political instability, protests and corruption - both recently and historically with corruption at the roots of the original uprisings in the civil war (see above). The Government's credibility for raising taxes has also been intimately tied with corruption and misuse of funds in popular opinion and media commentary. Simply put, unless the Government can clean up the use of public funds, continued resistance to increased funding of the public sector is likely, further jeopardizing the Government's ability to deliver services and reduce poverty. The Rule of Law, Justice, Crime and Violence Guatemala is also plagued by a widespread lack of the "rule of law." Empirical evidence clearly links an efficient justice system and an upholding of the rule of law with better economic performance.22 The justice system is crucial for enforcing property rights, convicting crirninals, and protecting citizens' rights. Such guarantees are particularly crucial for the poor, who cannot afford to protect themselves or seek legal representation. Guatemala ranks among the worst 14% of countries in the world for a composite indicator of "rule of law," which measures the extent to which people have confidence in and abide by the rules of society, including the incidence of crime, the effectiveness and predictability of the judiciary, and the enforceability of contracts (Table 13.3).23 Within LAC, only Haiti and Honduras rank worse than Guatemala for effective rule of law. Some progress is evident, however, with improvements since the 1980s and in the period since signing of the Peace Accords in 1996.24 Indeed, Guatemala has a very weak administration of justice, indicated by lengthy case delays, limited and unequal access to justice, a lack of transparency and predictability in court decisions, extremely formal application of written law, and a lack of training of judges. In a 1998 user perception and public opinion survey of justice administration,25 89% of those interviewed indicated that there is a lack of adequate justice, characterizing the system as "corrupt," "inefficient" and "overly-centralized." The main obstacles for effective justice were identified as corruption, impunity, and violence. Over three quarters of those interviewed perceive the justice system as expensive, not only because of high legal fees, but also because of difficult access to justice services (including insufficient number of courts, geographic isolation, and 151 language barriers). Frustrations towards the justice system are also evident because the general public does not have a unified understanding of the authorities responsible for justice administration. The ENCOVI 2000 reveals that most Guatemalans perceive justice to be in the hands of a variety of centralized authorities, including the Supreme Court (77%), Justice Ministry (72%), judges (86%), the President (55%), as well as the National Civil Police (PNC, 56%), and the Army (45%). The latter attribution is particularly surprising since the responsibility of the Army is to defend citizens, and not administer justice. This likely reflects the experiences of the civil war when the Army arbitrarily assumed the role of the justice system. In addition, a third of the population perceives communities as responsible for providing justice, with a higher share of the poor (41 %) perceiving community-responsibility for justice than the non- poor (26%), perhaps reflecting relatively less access of the poor to the formal justice system than the non- poor. Indigenous groups were also more likely to perceive community-responsibility for justice, probably reflecting the historical exclusion and isolation of these groups. Such findings are echoed in the QPES, in which villagers report resolving conflicts either by going directly to top Govemment officials or locally within the community with little of decentralized courts (see Box 13.1). Closely related to the weak justice system are the extra-judicial killings, which occur in Guatemala with alarming frequency. Indeed, lynchings have become a well-known practice in Guatemala, as citizens decide to enforce the law themselves due to perceived failures in the formal justice system. Between 1996 and 2000, some 337 lynchings were reported by MINUGUA, resulting in some 635 victims.26 One third of these occurred in 1999 alone. Over half were executed in the westem regions of the country (Nor-Occidente and Sur-Occidente), which also claim relatively high poverty rates and were heavily involved in the civil war (as discussed in Chapter 4). The 1998 justice survey revealed that 69% of those interviewed approved of lynchings as a form of justice (76% among the indigenous).27 Likewise, crime and violence - particularly in urban areas - have increased in the wake of the armed conflict. Indeed, the ENCOVI suggests that "violence, alcoholism, and family problems" was ranked by households as the third most common "community problem" in Guatemala (8%). This figure is particularly high in the Metropolitana Region (26%), where such problems were ranked second only to public services. Those in the richest quintile perceive such problems to be relatively more important than those in lower quiittiles. Some 15% of all households report the occurrence of a violent event in the past 12 months, with theft (9%) being the most common type of crime reported, followed by assault (5%). Such events are reportedly more likely in urban areas (with 24% of urban households reporting them), in the Metropolitan Region (29%), and among the non-poor (21%), even when other factors are taken into account.28 A qualitative study of urban violence also suggests that crime and violence is on the rise in the post-conflict period, inflicting significant intra-family, social, economic, and human capital costs.29 Alcohol consumption was perceived as the most critical cause of violence, in all communities studied. Rebuilding trust in the judicial and law enforcement system was seen as crucial for reducing the occurrence of crime and violence. Nonetheless, for communities in which the reformed PNC were present, a higher degree of trust was reported. In rural communities, the QPES suggests that domestic violence and conflicts with and between villages (over land, religious difference in particular) are fairly widespread problems, but with little or no involvement of the formal justice system in their resolution (Box 13.1). 152 Box 13.1 - Internal and External Conflicts: the Case of QEI (QPES) Despite being small and ethnically homogeneous (Q'eqchi), the village of QEI has been plagued by intemal and extemal conflicts, relating primarily to differences over land and religion. Located in the Renion Norte, the village was founded about 30 years ago when the founding families migrated from other regions in search of land. Distribution of the land was managed by the founding families (particularly a prominent village leader), who led the initiative to gain access to the land and distribute the plots (primarily to males) in accordance with the procedures of INTA (the National Institute for Agrarian Transformation). Unequal land distribution has generated conflicts within the community. Although some residents are grateful to the leader for his initiative in managing the land, others complain that his family secured access to bigger and better pieces of land. Not all families have plots. The village of QEI has also experienced a serious land conflict with a neighboring community. Apparently the other community hired soldiers to force the villagers of QEI to abandon their land by burning their houses, damaging their plantations, and whipping their community leader. In response, the community sent a letter to the President of the Republic and visited the Ministry of the Interior (Gobernaci6n). As a result, the soldiers were withdrawn and numerous residents from the other community were arrested. It is interesting to note, however, that the village did not turn to any local judicial authorities for help in resolving the conflict - the only way they were able to get action was through centralized authorities in Guatemala City. Religious affiliation also generates conflicts inside the community that started due to confrontation between Evangelical (majority) and Catholic religious leaders. The tone of these confrontations has heightened and villagers now indicate that they attend services in other villages to avoid problems. Political Instability Political instability continues to reign in Guatemala. Although the country is currently experiencing its longest effort at sustaining a democratic system, which was bolstered by the Peace Accords, political instability remains a pervasive feature. Indeed, Guatemala still ranks among the top 20% of most politically unstable countries in the world (Table 13.3). Within LAC, only Colombia and Haiti rank worse than Guatemala for this indicator. Recent confrontations between the Government and private sector, divisions within the ruling party, a series of corruption scandals, various allegations of unconstitutional modifications of laws, weak management of public finances, protests and strikes in response to increases in the VAT, and perceived inadequate response to economic shocks and natural disasters have all contributed to a current climate of heightened instability.30 Indeed, Guatemala maintains a score of "D" on the Economist Intelligence Unit's political risk ratings (with "E" being the highest risk), though the EIU does not forecast the risk of a military coup as likely.3' Such instability further worsens Guatemala's prospects for growth and poverty reduction at a time when economic crisis calls for strong leadership and effective government. COMMUNITY PARTICIPATION AND SOCLAL CAPrTAL32 Communities have an important role to play in promoting their own development and ultimately in reducing poverty. Government is not the only actor in development and poverty reduction. In fact" in the absence of effective Government provision of public goods and services, communities are often forced to rally together to make decisions, manage public goods and collective resources, cope with shocks, and leverage external assistance. Moreover, the empowerment of communities and the promotion of participation are central themes in the Peace Accords (as discussed in Chapter 4). Social capital is an important asset that can reduce vulnerability, increase opportunities, and empower local communities. Social capital is typically defined as norms, trust, and reciprocity networks that facilitate mutually beneficial cooperation in a community. Social capital generally develops in social interaction within, between and beyond communities, and can therefore be defined in three dimensions: o First, bonding social capital develops within communities and constitutes the strong ties connecting family members neighbors, close friends and business associates. These ties connect people who share 153 similar demographic characteristics. Some examples of bonding social capital are religious groups and neighborhood committees. * Second, bridging social capital, created between communities, is defined as horizontal connections to people with broadly comparable economic status and political power. School committees, professional associations, groups that manage community-level public goods, and social groups are examples of bridging social capital. * Finally, linking social capital describes interaction beyond communities and consists of the vertical ties between individuals and people or formal institutions in positions of influence. Connections to politicians and formal development organizations are examples of linking social capital. International evidence suggests that better off individuals and communities tend to have "more" social capital, particularly more extensive networks (e.g., bridging, linking).33 Impact of Civil War on Social Capital Despite decades of civil war, the overall level of participation in organizations and collective action is comparable to that of other countries. In terms of participation in organizations, the ENCOVI 2000 shows proportions of individuals participating in any type of formal group sum to 23%. This compares with Argentina, where participation in any group is close to 20%.34 On average, Guatemalans participate in 1.09 organizations and households participate in I.11.35 In the United States and Tanzania36, the mean value of membership in organizations per person is 1.8 and 1.5 respectively. The average number of associations each household is member in Indonesia and Bolivia37 is 5.5 and 1.4 respectively. Participation rates for collective action activities are significantly higher than membership rates in organizations, with 58% of the population participating in bridging activities and 23% participating in linking activities.38 Unlike membership in organizations, community activities do necessarily not entail long-term commitments and time and monetary investments, which have been identified as barriers to participation (see below). Social capital appears to have been both undermined and strengthened by the prolonged civil war. In some cases, the war encouraged solidarity and community cohesion: confronting the war further Box 13.2 - Strong Village Bonds: the Case of Ml (QPES) Horizontal connections (bonding social capital) within the Mam village of Ml are quite strong, despite the fact that the village was invaded by both the army and the guerrillas during the civil war. Numerous formal organizations are active in the community, including a development committee that boasts representation of all families; sub-committees for education, water, and irrigation; an agricultural organization; and. two women's groups. The community also operates a fairly successful revolving credit fund, which started with seed money from a Canadian organization but is now run entirely by the community, and a micro-enterprise committee for women. The village also appoints two rangers (guardabosques) to manage the common woods and their use (charging fees for people to gather wood). Villagers cite examples of mutual assistance: helping an orphan family, helping those facing a death in the family (with men providing firewood and women providing food), and financial transfers being provided by the church to poor families. The villagers also reportedly helped each other when they were ransacked during the conflict of the 1980s (and their homes were burned). Nonetheless, despite their participation in specific groups, the women of Ml acknowledge that they have little influence on community decisions. The community also seems to have significant ties to extemal organizations (linking social capital), receiving aid and projects from FODIGUA, DECOPAZ, PRONADE, UNICEF, the E.U. and NGOs. In contrast, other than remittances from migrant members in the US and Mexico, the community does not seem to have much in the way of "bridging" networks. This relative lack of influence over others perhaps explains their almost paradoxical complaints that they "lack organization, leaders" unlike "the ladinos who do help each other and collaborate with each other." They also indicate that they feel excluded from broader Guatemalan society due to the fact that most do not speak Spanish and to their physical isolation and distance from other communities. They likewise perceive discrimination in being treated or receiving medicines from the nearest hospital. 154 enhanced villagers' resolve to work as a unit and protect themselves from other threats. In the Kaqchiqel village of KA2 (QPES), for example, the war had devastating consequences, including a severe massacre in which the village's leadership was exterminated and the burning of crops and homes, which forced villagers to flee for two years. Nonetheless, the community has since organized several committees (e.g., development comrnittee, women's groups, a support group for widow victims of the war, a school committee). The development committee now acts as the "maximum authority," not only to solve problems, leverage assistance, and organize community activities, but also as a protective, security body (for example, protecting the village against a recent onslaught of gangs). The village of MI provides another example of strong social capital despite the war (Box 13.2). On the other hand, the violence of the 1980s tested community ties and trust in the village of KII (see Box 4.3 in Chapter 4). Moreover, the ENCOVI also shows that participation rates in any type of organization are far lower for residents of the Region Norte, which is one of the epicenters of the later phase of the war. Such results suggest that the war may have prompted fear of involvement in organizations (or actively repressed them), and hence undermining social capital. Stronger Within Village Bonds, Weaker Bridges Social capital in Guatemala is mainly concentrated in strong horizontal, within-village connections. This pattern is reflected in both the QPES, where most connections are within villages and sometimes with external development agencies (see for example the case of MI in Box 13.2) and in the ENCOVI 2000. Overall, participation rates in bonding organizations (18%) are over three times higher than those for bridging organizations (5%), with many communities having become somewhat closed in upon themselves as a result of war and physical isolation. Membership rates in bonding organizations is mainly driven by participation in religious groups (18% of the population), which constitutes the most common form of social capital in Guatemala. Recreation (3%) and groups that supervise public goods (2%) are the next most common forms of participation in organizations. Participation in income-generation groups, school committees, community service associations, and social and special interest groups is below one percent (each). Likewise, participation in collective activities reflects a rather limited scope of networks. The most common collective action activities are community construction, participation in labor exchange agreements, the provision of voluntary labor, and the collection of monetary or in-kind donations (all bridging activities); linking activities are the least common (e.g., contacting local politicians, contacting government officials, participating in information or electoral campaigns, or notifying judicial authorities). Concentration of Social Capital Among Privileged Groups importantly, social capital seems to be concentrated among the privileged groups in Guatemalan society. As such, programs considering community-based targeting or community-driven development (CDD); such schemes should seek to ensure participation of excluded groups (e.g., women, the poor, the illiterate) so as to avoid a continuation of these patterns. These patterns are confirmed in both the QPES and the ENCOVI, and robust to specifications in multi-variate logit regressions39: o Regional variations are strong and significant. The Metropolitan Region reports significantly higher membership rates for formal organizations, whereas participation rates in the Norte Region, which had been one of the epicenters of the civil war, are significantly lower. By type of organization, participation in recreation groups was far stronger in the Metropolitan Region, whereas participation in religious groups was far lower in the Norte Region (than all other regions). Interestingly, the opposite is true for collective action, which is significantly weaker in the Metropolitana Region - perhaps because metropolitan residents can rely more on government- provided services to solve their problems for them. Indeed, Metropolitan residents were far less likely to participate in collective activities such as community construction, labor agreements, or community workshops, than those in other regions. 155 * Similar patterns are noted by urban and rural areas. Urban residents participate more than their rural counterparts in bridging organizations, particularly recreation groups. In contrast, rural residents participate more in organizations that supervise public goods and in collective action activities (particularly labor agreements and community construction) - perhaps reflecting a relative lack of the state's rural outreach in these activities. * Guatemalan men tend to participate more than women (Figure 13.3). This is true for both membership in organizations and participation in collective action. The only exception is for participation in bonding organizations, which are dominated by religious organizations and in which women tend to participate more than men. Villagers in. the QPES communities confirmed this finding: in virtually all villages, men and women agreed that women do not participate in community decision-making. Some women indicate that when the do try to participate, they are "mocked" by-the men (due to machista attitudes; see for example the case of K12 in Annex 5). * The non-poor participate more than the poor in formal organizations (Figure 13.3). Such differences even stronger for bridging organizations. The dominance of formal organizations by elite or better off families is confirmed in the QPES villages. In the Ladino village of LI, for example, the distribution of water is managed by the local authority ("el comite"), which has been presided over by the same person for the past 20 years. Participation in this conitd is not democratic and excludes those who cannot afford the operations expenses of being members. Study participants also repeatedly complained that the comitd manages resources arbitrarily and has no accountability. Those who feel discontent, however, seem to have little power to make any changes and this produces divisions within the community. In the K'iche village of K12, the poor are self-excluded from community activities. They explain that they have "too much work" and don't have time to participate. Informants in other QPES villages also note that time and financial costs of participation present barriers for participation by the poor in organizations. The ENCOVI suggests that the poor participate slightly more in collective action activities, but the differences are not significant. * Education is also highly correlated with all 'types of participation (collective action and organizations), as is Spanish-speaking ability among the indigenous, and age. Education not only provides skills at the individual level, but also contributes to the community by preparing individuals to assume leadership roles, leverage external assistance, and represent the community in local and national institutions. Informants in several QPES villages do indicate that illiterate members are excluded from community decision making (see Box 4.3 in Chapter 4). * Although there are some differences by ethnicity, for example with much lower participation rates among the Q'eqchi, these are not systematically significant when other factors, such as region, education, or language ability, are taken into account. Benefits of Social Capital Nonetheless, the benefits of social capital to community welfare are significant and diverse. First, the QPES offers a rich array of examples of how communities organize to manage public goods and collective resources, and make community decisions. In the QPES village of Kll, for example, the community organized to solve problems with drinking water and to acquire improved stoves to protect women's health while cooking (though there are allegations of misuse of funds by the water committee, see Box 4.3 in Chapter 4). In MI, the community has established rangers (guardabosques) to protect communal woods (see Box 13.2). Second, social capital is often called upon in the face of shocks. This is evident in both the ENCOVI, in which community-based actions are second only to self-help for dealing with shocks, and in the QPES, in which communities report various forms of mutual assistance when hit with shocks (see for 156 example Box 13.2). Third, higher social capital seems to be strongly associated with the ability of communities to leverage external assistance. Communities in the ENCOVI sample with higher than average participation in organizations also report more help from formal institutions and more success in leveraging funds (see Table 1.3.4 below). Figure 13.3 - Social Capital Participation by Figure 13.4 - Participation in Organizations, Gender (ENCOVI 2000) by Poverty Group (ENCOVI 2000) 80 40 70°3 35- 0650 8-c30 -27 Types ofOgnztin,CiecieAto 0 Mate 13 Non-Poor cg 40 - 25 20SAlPo O. 30 2422 242 E Female 20mllPo 16T1720 242 2 - 17 17 and Poor 43 6oP oC o C C17 9~6 0 > ~~~~~~~~~~~~~Any Bonding Bridging Types of Organizations, Collective Action Tpso raiain Table 13.4 - Social Capital and External Assistance (Percentage of Total Communities in ENCOVI sample) Number of Organizations in the Community Below Median Above Median Help from any Institution a 61.8 94.2 Help from Government 23.5 42.6 Help from Social Funds b 29.5 60.4 Help from Institutions c 9.7 35.2 Help from NGO 21.2 32.8 Success Leveraging Funds 84.1 90.1 Success in Project Application 59.0 65.6 Success Obtaining Support from Institutions 64.6 77.4 Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. a. The Institutions include: Government, Politicians, Church, School, NGOs and Social Funds. b. Social Funds include: FIS, FONAPAZ, FSDC and other Social Funds. c. Institutions include: Politicians, Church and School. Recent Legal Reforms to Empower Communities The recent passage of three laws40 pertaining to citizen participation and decentralization constitutes an important step in empowering local communities in Guatemala. Specifically, the passage of these laws represents a significant step towards creating the framework and legal structures for bringing the Government closer to the community, developing certain responsibilities to the local level, and empowering communities and their councils (consejos comunitiarios) to participate in decision-making. Building on the findings of the ENCOVI and QPES (above), efforts should be made to ensure the participation of women, the poor and other traditionally excluded groups in the implementation of this new framework. 157 ROLE OF OTHER ACTORS In addition to communities and governments, other actors - such as the private sector, NGOs, and religious organizations - have an important role to play in reducing poverty and promoting development. While it is beyond the scope of this study to do a full stakeholder or institutional analysis, it is important to acknowledge the role of these other actors, particularly in Guatemala where the limited size and effectiveness of the state makes their participation all the more crucial. Historically, the private sector, has had somewhat of a mixed relationship with the poor. On the one hand, certain segments of the private sector (particularly large-scale, formal enterprises) had vested interests that supported the exclusionary policies that afforded them land and cheap labor inputs over the past centuries. On the other hand, today's private sector has a vested interest in progress on development, which relies on the success of the peace process, which in turn relies on a reversal of the exclusionary forces that sparked the later phases of the war in the first place. Perhaps reflecting acknowledgement of this basic reality, the private sector has explicitly developed organizations, or "fundaciones" to promote social business responsibility in many industries. As of 1995, nine such foundations from various industries had united under the Consejo de Fundaciones Privadas de Guatemala, the largest of which are the Sugar Foundation and the Foundation for Rural Development (FUNR'URAL, representing ANACAFE and the coffee sector).4' Projects supported by the foundations commonly support activities in education, training, housing, municipal development, health, environment, and social protection. Such projects are commonly co-financed with joint participation of the public sector or donors. There has also been a recognition among several private-sector business groups that bold reforms are needed to enhance productivity and promote development and security in the country. Moreover, the private sector is not homogeneous, and a significant share of private sector activity is conducted by micro- enterprises in Guatemala, with self-employment accounting for one third of all employment in 2000.42 Indeed the strong entrepreneurial spirit that characterizes much of Guatemala seems to be one of the country's main sources of growth and employment. * Hundreds of NGOs are also active in Guatemala, providing services in education, health, agriculture, environment, culture, human rights, and so forth.43 Geographically, there are distinct patterns in the distribution of NGOs between development-oriented organizations and those focusing on human rights. There has recently been a conversion of several NGOs that are shifting from providing humanitarian assistance towards playing the role of effective development brokers, often partnering for the delivery of public services (e.g., under the PRONADE and SIAS programs). Development-focused NGOs are concentrated largely in the Metropolitana Region (Department of Guatemala) and the center and alti-plano of the country. Human rights organizations are concentrated largely in the Departments of Quichd, Huehuetenango, Alta Verapaz, Solola, San Marcos and Chimaltenango (many of which were largely affected by the war) as well as Guatemala. There is a significant under-coverage of NGOs in departments in the Eastern regions of the country. * As discussed above, membership in religious organizations is the most common form of social capital in Guatemala (based on results of the ENCOVI 2000). Besides providing moral and spiritual support, QPES communities suggest many examples of churches providing faith-based assistance to the needy (e.g., to orphan families in Ml or helping villages mitigate the effects of shocks like the earthquake of 1976, Hurricane Mitch, and the war). The villagers also clearly associate spiritual relations with welfare (see Chapter 2). To promote the links between poverty reduction, development and religion, an Inter-Faith Development Dialogue was recently established in Guatemala, with representation from the Mayan, Catholic, Protestant, and Jewish faiths. Nonetheless, conflicts between religions can be quite divisive in Guatemala's villages. 158 Indeed, conflicts between religious leaders or groups were also a common theme raised by informants in the QPES study villages. SUMMARY OF KEY ISSUES AND PioRrriEs The above analysis reveals a number of key messages and policy implications: e The effectiveness of Guatemala's Government affects the menu of options for reducing poverty and the ways in which these options are carried out. Despite some progress, Guatemala's efforts to improve living conditions and promote a more inclusive and democratic society have been hampered by a weak public sector. Key challenges in this area include: o Improving the tax base to generate more revenues for increased public spending, particularly in the social sectors, basic utility services and rural development; o Improving public expenditure management, with stronger links to policy, planning and priorities; o Improving the targeting of public spending to the poor; o Making the public sector more accountable and responsive; o Strengthening administrative capabilities and the civil service; and o Bringing the government closer to the client and improving service delivery, particularly given the heterogeneity of Guatemala's population and communities. o Good governance is also important for growth and poverty reduction. Guatemala scores poorly on most governance indicators, despite some progress. Key challenges for the development and poverty reduction agenda include: (a) fighting corruption; (b) improving the rule of law and the justice system; and (c) reducing political instability, which further hurts the climate for growth and investment. While it is beyond the scope of this study to propose a governance strategy, some actions that could be effective include: report cards for key ministries and services to track perceptions of service quality and corruption, client satisfaction surveys, and formal adoption and implementation of an anti-corruption charter. Recent initiatives to fight corruption should be expanded and implemented. o Improving govemance and public sector management are prerequisites to expanding tax revenues and public spending. Without sincere improvements in public sector effectiveness, accountability and transparency, continued opposition to increased funding of the state is likely. The recent passage of three laws" pertaining to citizen participation and decentralization constitutes an important step in promoting greater social accountability in Guatemala. o Communities have an important role to play in promoting their own development, particularly in light of Guatemala's weak public sector. Social capital can offer significant benefits to community welfare, including managing public goods, coping with shocks, and leveraging external assistance. Nonetheless, social capital in Guatemala is mainly concentrated in strong horizontal, within-village connections, reflecting the physical isolation of many communities and decades of exclusion and war. Moreover, social capital appears to be concentrated among the more privileged groups in Guatemalan society, to the exclusion of women, the poor, and those with less education. An important policy implication of these findings is the importance of explicitly promoting the participation of these traditionally excluded groups in programs that rely on community-driven development (CDD) or community-based targeting. The recent passage of three laws45 pertaining to citizen participation and decentralization constitutes an important step in empowering local communities in Guatemala. * Other actors - particularly the private sector, NGOs, and religious groups - are active players in the fight against poverty. Given the limited size and scope of the Government, partnerships should be sought with these actors to help advance the poverty-reduction agenda. 159 lMINUGUA (June 2001). 'World Bank (1997). 3World Bank macroeconomic database for Guatemala. 4Source: World Bank (2001b). 5 Source: World Bank WDI 2001. 6CIEN (November 2000). 7The target incidence of public spending by social funds should be treated as a rough estimate. It was calculated using regional spending figures by the social funds (source: UNDP 2001b), combined with poverty rates (source: ENCOVI 2000) for each region (hence assuming that within each region, spending is allocated neutrally according to population shares). 8 Estimates of the size of the civil service vary, ranging from 150,000-205,OOOCIEN (August 2001) and Perlman (1995). These estimates translate into a ratio of approximately 132-180 civil service workers per 10,000 population. Comparable figures were 867 for Uruguay, 648 for Venezuela, 641 for Argentina, 340 for Peru, 330 for Ecuador, and 252 for Chile (all in 1990-91). World Bank (1996). 9 CIEN (August 2001). IS CIEN (August 2001). Indeed, the ENCOVI shows that hourly earnings for those with permanent contracts are lower than those with temporary contracts (Q15.3 per hour as compared with Q19.1 per hour on average). " It is important to note, however, that although the ENCOVI is representative for households in Guatemala, it is not representative for public civil servants. 12 Some of this "central govemment" spending, however, is transferred to decentralized programs such as PRONADE or SIAS. World Bank draft Project Appraisal Document for a proposed Third Integrated Financial Management Reform Loan. 3~ Rojas (February 1999).-- 4 Kaufman, Kraay, and Zoido-Lobat6n (October 1999). '5 Database compiled by Kaufmann, Kraay, and Zoido-Lobat6n (October 1999). Database and documentation can be found at the following website: http://www.worldbank.org/wbi/govemance/wp-governance.htm 16 EIU (November 2001); numerous articles in La Prensa, Siglo-XXI, and other daily newspapers. '7 Survey carried out by Vox Latina - Acci6n Ciudadana (April 2001), as cited by CIEN (May 2001). 1S Transparency Intemational (September 2001) and EIU (November 2001). 19 Problems included poor secrecy rules, failure to classify money-laundering (other than that related to drug trafficking) as a criminal offense, laws preventing the authorities from sharing information with other countries in investigations, and the lack of a specific body to investigate such transactions. EIU (November2001). 20 EIU (November 2001). 21 There is ample cross-country evidence that higher levels of corruption are associated with lower growth. World Bank (2002). 22 World Bank (2002). 23 Kaufmann, Kraay, and Zoido-Lobat6n (October 1999). 24 For example, Guatemala's ranking on the Intemational Country Risk Guide's Rule of Law index roe from 1.1 in the 1980s to 2.3 in the period from 1990-96 to 2.8 in the period from 1996-98 (on a scale of 1-7 where I is the most risky). 25 Aragon and Associates (June 1998). 26 MINUGUA (2000). 21 Aragon and Associates (June 1998). 2 Logit regression results provide the significance of these factors, controlling for other characteristics; figures in parentheses are cross-tabulations. Source: World Bank calculations using data from the ENCOVI - Instituto Nacional de Estadfstica - Guatemala. 29 Moser and Mcllwaine (2001). 30 EIU (January 2002a). 31 EIU (January 2002b) and EIU (November 2001). These risk ratings range from "A" (lowest risk) to "E' (highest risk). In Central America, Honduras and Nicaragua also scored a "D" on political risk, while Panama and El Salvador ranked as "C," and Costa Rica ranked as "B." 32 All numbers presented in this. section come from World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala unless otherwise specified. Village examples come from the QPES. For additional details, see GUAPA Technical Paper 12 (IbAfiez, Lindert and Woolcock, 2002). 33 World Bank (200 le). 34 Ledermann (2001). 35 The average number of organizations at the household and individual level has a downward bias because the survey asked respondents to identify only the three main groups in which the individual is member. This bias however might not be significant since from 29,414 who answered this portion of the questionnaire, 8,316 participate in one group, 732 participate in a two groups and 105 participate in three groups. Glaeser et al (1999) and Narayan and Pritchett (1999) 37 Grootaert (XXX) and Grotaert and Narayan (2001). 38 Collective action activities include the following. For bridging: collecting funds, community workshops, labor agreements, donations in cash or kind, community childcare, construction of community infrastructure. For linking: contacting govemment officials, information campaigns, electoral campaigns, contacting local politicians, notifying judicial authorities. 39 See GUAPA Technical Paper 12 (Ibahez, Undert and Woolcock, 2002) for details. 40 Specifically, the Ley de Consejos de Desarrollo Urbano y Rural (Decreto 11-2002), the Codigo Municipal (Decreto 12-2002), and the Ley General de Descentralizaci6n (Decreto 14-2002). 41 Fuentes & Asociados (April 2001). 42 Source: World Bank calculations using the ENCOVI 2000, Instituto Nacional de Estad(stica - Guatemala. 43 Browninget. al. (1998). S pecificaly, the Ley de Consejos de Desarrollo Urbano y Rural (Decreto 11-2002), the Codigo Municipal (Decreto 12-2002), and the Ley General de Descentralizaci6n (Decreto 14-2002). 45 Specifically, the Ley de Consejos de Desarrollo Urbano y Rural (Decreto 11-2002), the Codigo Municipal (Decreto 12-2002), and the Ley General de Descentralizacion (Decreto 14-2002). 160 PART 5: CONCLUSIONS AND RECOMMENDATIONS Chapter 14: Priority Actions to Reduce Poverty This chapter builds on the empirical findings in the rest of the report to build an agenda for poverty reduction in Guatemala. Broadly speaking, a concerted strategy should be adopted to reduce poverty by building opportunities and assets, reducing vulnerability, improving institutions and empowering communities. This broad agenda for poverty reduction largely coincides with the central tenets of both the Peace Accords and social policy in Guatemala. Progress on advancing this agenda has occurred, and is possible. Nonetheless, significant challenges remain, particularly those involving improvements in key outcomes and deeper institutional reforms. In light of these main-issues, this chapter proposes a set of priority actions within the broad agenda for poverty reduction. While these actions are needed in general, they can be made more effective if targeted to specific priority groups. As such, this chapter identifies a set of priority target groups and offers suggestions on ways explicit targeting criteria can be built into poverty reduction efforts so as to improve their effectiveness. Finally, the chapter offers some suggestions on monitoring and evaluation. Three caveats are important to emphasize at the outset: * First, while there is no single "blueprint" for poverty reduction, there are some key levers that take central stage for national efforts to reduce poverty, and these are the emphasis of this chapter. Nonetheless, efforts should be made to tailor this broad agenda to local conditions, particularly in a country as heterogeneous as Guatemala. To this end, efforts are underway in Guatemala to develop not only a national poverty reduction strategy, but also localized strategies at the department and municipal levels.' * Second, the policy discussion is aimed primarily at the perspective of policy makers and the role of the public sector; hence it emphasizes interventions that both (a) would have a substantial impact on poverty; and (b) merit the use of public resources in a market-oriented economy. Nonetheless, other actors - such as the private sector and other facets of civil society (e.g., communities, NGOs, religious organizations), also have an important role to play in reducing poverty. The private sector, in particular, will provide the central arena for economic growth and productive activities, which are crucial for poverty reduction. Other facets of civil society (e.g., communities, NGOs, religious organizations) are clear partners in this poverty reduction agenda, and will play crucial roles in prioritization and implementation of public sector actions, as well as the provision of other services and interventions that are beyond the scope of the public sector. * Third, poverty reduction is a multi-dimensional and long-term process. There is no single magic bullet to reduce poverty. Rather, efforts should be made to attack the poverty problem from a multitude of angles, including those to foster opportunity, build assets, reduce vulnerability, and improve institutions and empower communities. Moreover, poverty reduction does not occur over night. Implementation of key actions to reduce poverty takes time, and often the impact of such actions occurs over an even longer time frame (for example, into subsequent political cycles - or even subsequent generations). That said, the sooner actions are undertaken, the more quickly the inter- generational cycle of poverty can be broken. 161 A BROAD AGENDA FOR POVERTY REDUCTION iN GUATEMALA Broadly speaking, a concerted strategy should be adopted to reduce poverty in Guatemala by building opportunities and assets, reducing vulnerability, and improving institutions and empowering communities. A broad agenda of actions in these areas is outlined in Table 14.2 (at the end of the chapter). o Building opportunities. Economic growth is necessary for poverty reduction, particularly given the relatively small size and capabilities of Guatemala's public sector. Moreover, a recurring theme that arises in the analysis is the fact that the poor, particularly the rural poor, women and the indigenous, are not able to fully participate in, or benefit from, the overall economic system. Therefore, improving employment and earnings opportunities is essential, and this depends largely on the actions of the private sector. The pattern of growth needs to be made more "pro-poor," with an emphasis on building opportunities for the rural poor, women and the indigenous. This will depend on two other key areas: building the assets of the poor, as well as improving institutions and the investment climate. Specific actions are outlined in Table 14.2 below and detailed in Chapters 5 and 6.2 e Building the assets of the poor. This is arguably the most important area for poverty reduction in Guatemala, given the chronic nature of poverty in Guatemala, existing disparities, and linkages to the other key areas - including promoting growth, reducing vulnerability, and empowering the excluded. Key assets include: education, health, basic utility services (particularly water and sanitation), land and physical capital, and rural roads. Specific actions for each of these assets are outlined in Table 14.2, and detailed in Chapters 7-10. e Reducing vulnerability. Again, the central path for reducing vulnerability is to build the assets of the poor, since most vulnerability in Guatemala is associated with low expected earnings (due to weak assets) rather than high volatility of consumption. Nonetheless, disaster management is important, given the poor's exposure to natural and agriculture-related shocks. Moreover, much could be done to improve the efficiency and effectiveness of existing public social protection programs. Many of these - e.g., such as scholarships, school feeding - could also play a role in building the assets of the poor by easing demand-side constraints to improved coverage. Some transfers (social assistance) could also be used to alleviate the poverty and suffering of the extreme poor, particularly when tied to participation in health and education activities. Specific actions in this area are outlined in Table 14.2 and further detailed in Chapters 11-12. o Improving institutions and empowering communities. Weaknesses in the public sector and poor governance strongly shape the menu of feasible options and effectiveness of poverty reduction efforts. They also influence the overall climate for investment and economic growth. As such, improvements in this area are deemed to be of high priority, consistent with the strategic emphasis on "modernization of the state" in the Peace Accords, as discussed in Chapters 4 and 13. The role of communities in promoting their own development is also important, as acknowledged in the Peace Accords, and poverty reduction efforts should seek to partner with communities in determining priorities. Nonetheless, explicit efforts should be made to reach out to groups typically excluded from community decision-making (namely, the poor, women, and the uneducated), as discussed in Chapter 13. Partnerships should likewise be sought with private-sector and NGOs to extend and improve service delivery. This broad agenda for poverty reduction largely coincides with the central tenets of the Peace Accords. Indeed, reducing poverty and improving living conditions is central to lasting peace in Guatemala. As discussed in Chapter 4, the Peace Accords went well beyond formalizing the end to Guatemala's civil war, outlining a broad policy agenda that signaled a significant shift towards a more inclusive development path. Key areas of emphasis related to economic development and poverty reduction include: a focus on human development, goals for productive and sustainable development, a program for the modernization of the 162 democratic state, and strengthening and promoting participation. The rights of the indigenous and women were also highlighted as cross-cutting themes throughout the accords, in an attempt to reverse the historical exclusion of these groups. Moreover, this broad agenda for poverty reduction is consistent with the current thrust of social policy in Guatemala. Given the importance of improving living conditions to lasting peace, poverty reduction has taken center stage on the current social policy agenda. In particular, the Government recently outlined its poverty reduction strategy in an important policy document "Estrategia de Reducci6n de la Pobreza" (ERP),3 presented at the Consultative Group meetings in February 2002. General principles emphasized in the ERP include: a rural focus, using the poverty map for targeting (see Box 14.1 below); efficient and transparent public spending; decentralization; and participation. Key action areas ("ejes vitales") include: (a) promoting growth with equity; (b) investing in human capital (emphasizing health, education and food security); and (c) investing in physical capital (particularly water and sanitation, rural roads, electricity, and rural development). Cross-cutting issues ("temas transversales") in the ERP include multiculturalism and interculturalism, gender equity, and vulnerability. SOME PROGRESS ... AND KEY ISSUES Progress has occurred - and is possible. In the six years since the signing of the Peace Accords, Guatemala has taken important steps on this new, more inclusive development path. In particular, progress has occurred in the areas of building assets and improving institutions: * Public sector management, particularly public financial management with the introduction and implementation of the Integrated Financial Management System (SIAF) since 1998, as discussed in Chapters 4 and 13; * Public revenues and spending, with increases in revenues and public spending, particularly for the education and basic utility services sectors, as discussed in Chapters 4, 7, 9, and 13; * Education coverage, with notable increases in improving coverage and narrowing disparities between genders, ethnicities and poverty groups, particularly since the signing of the Peace Accords in 1996, as discussed in Chapter 7; and * Basic utility services, with an expansion of coverage of water, sanitation and electricity services and a reduction in disparities in access to these services, particularly since the signing of the Peace Accords in 1996, as discussed in Chapter 9. Progress has also occurred in other areas. For example, for land, numerous entities have been created and initiatives launched, though their reach has been limited to only a few thousand households (see Chapter 6). Importantly, these steps signal that progress is possible; despite the magnitude of the challenge of changing the course of the country's history. Nonetheless, significant challenges remain to build opportunities and assets, reduce vulnerability, improve institutions and promote empowerment. Within this broad agenda, several key issues should be considered as top priority: * Growth has slowed, particularly in rural areas. As discussed in Chapter 5, economic growth has slowed in recent years, with a sharp decline in coffee prices and the global economic slowdown, as well as concerns relating to the overall investment climate and weaknesses in the banking sector. Taking into account population growth rates of 2.7% p.a., GDP per capita actually fell and poverty is projected to have risen in 2001 and 2002. Rural growth rates in particular have declined over time. Given the high concentration of the poor in rural areas, economic growth rates - particularly those in 163 rural areas - need to improve if Guatemala is to make significant progress in reducing poverty and meeting the goals set by the Peace Accords, the MDGs, and the ERP. Related to the meager performance of the overall economy, households do not perceive significant improvements in living conditions. While communities in the ENCOVI do indicate they perceive progress - and attribute it to improvements in basic services - households are decidedly more pessimistic about changes in their welfare since the Peace Accords, as discussed in Chapter 2. They attribute these perceptions to economic factors, such as a lack of increases in incomes and opportunities (factors that directly affect "their wallets"). o Coverage of education, health and basic services remains insufficient and biased against the poor. Despite progress, significant gaps and disparities remain, particularly for the poor, girls (for education and health), and rural and indigenous residents. Demand-side barriers, not just a lack of physical infrastructure, largely account for inadequate access to these services. Quality of education, health and basic services also appears to be deficient, as discussed in Chapters 7, 8 and 9. O Health outcomes - particularly malnutrition and infantlmaternal mortality - have not improved in line with targets set by the Peace Accords or the MDGs, as discussed in Chapter 8. o Weaknesses in public sector management and governance continue to hamper Guatemala's quest for a more inclusive and prosperous society, as discussed in Chapter 13. Additional reforms are needed to make the public sector more accountable and responsive, strengthen administrative capabilities and the civil service, strengthen public financial management, and bring the government closer to the client. Key governance challenges for promoting economic development and reducing poverty include: (a) fighting corruption; (b) improving the rule of law and the justice system; and (c) reducing political instability, which further hurts the climate for growth and investment. o Public revenues and spending are still inadequate, and could be better targeted. First, existing levels of revenues remain insufficient for Guatemala to make a significant dent in poverty reduction by expanding assets and service delivery, as discussed in Chapters 4 and 13. However, improving governance and public sector management is a prerequisite to expanding tax revenues and public spending. Without sincere improvements in public sector effectiveness, accountability and transparency, continued opposition to increased funding of the state is likely. Second, public spending is poorly targeted, with a significant share going to the non-poor, as discussed in Chapter 13. PRIORITY ACTIONS FOR PoVERTY REDUCTiON IN GUATEMALA In light of these issues, certain actions stand out as top prioritv. As discussed above, a broad, multi- dimensional strategy should be adopted to reduce poverty in Guatemala by building opportunities and assets, reducing vulnerability, and improving institutions and empowering communities. A broad agenda of actions in these areas is outlined in Table 14.2 below. Within this broad agenda, actions should be further prioritized using the following criteria: (a) likely poverty impact; (b) political, institutional, and administrative feasibility; (c) economic feasibility and costs; and (d) their need and justification for public sector resources. Such prioritization will likely require further dialogue and analysis (e.g., institutional assessments, costing of actions, public expenditure analysis). As a first cut, certain actions should be considered as top priority, based on a cursory review of such criteria: (1) Promoting economic growth and productive opportunities, particularly in rural areas. Guatemala must raise its rate of economic growth if it is to make significant progress in achieving key development and peace targets (as discussed in Chapter 5). This is true internationally, but particularly relevant for Guatemala, given the limited scope for public sector action and redistribution. In this context, the main engine of growth is likely to come from the private sector, with the public sector playing a supporting role affecting growth mainly insofar as it stimulates private-sector investment and productive activities. Yet the actions of the public sector in this supporting role are crucial. In particular, priority actions include: 164 * Maintaining macroeconomic stability; * Enforcing a tight fiscal position, with a careful plan for strengthening tax collection and redirecting public spending towards the social sectors so as to build assets that are crucial to both growth and poverty reduction; * Fostering a climate that is conducive to private investment and growth, including improvements in governance and public sector management (see Chapter 13); * Promoting growth with special emphasis on sectors that are likely to generate substantial employment for the poor. Additional analytical work is needed to define a more comprehensive pro-growth strategy (see below). Nonetheless, while a thorough sectoral analysis of growth is beyond the scope of this study, available data do suggest certain levers that would have stronger impacts on poverty reduction than others for urban and rural areas: o In urban areas, this requires policies to support labor-intensive sectors, particularly micro-, small- and medium-enterprises (MSMEs), as well as education and technical training. o In rural areas, this means developing non-agricultural activities that are better remunerated and have better long-term prospects than traditional agriculture. As discussed in Chapter 6, key interventions to support growth in non-farm activities include: (a) increasing and improving the targeting of investments in education and technical training; (b) increasing investments in transport and basic infrastructure, which are crucial for the diversification, growth and inclusion of the poor in the rural economy and with facilitating the adjustment to the coffee crisis; and (c) policies that promote micro-, small- and medium-enterprises (MSMEs), a segment of the private sector that tends to generate a lot of employment. While agriculture is unlikely to generate enough additional employment opportunities to reduce poverty on a large scale in the medium term, it will continue to be an important source of incomes for the poor (at least in the short run). In this context, diversification efforts should focus on non-traditional products with better demand and price prospects than traditional export crops (as discussed in Chapter 6). Policies should also continue to facilitate productivity improvements (such as technical assistance), so as to boost the earnings of those who remain in agriculture. Investments in infrastructure (e.g., rural roads to improve marketing opportunities and education to improve farm-management practices) will likewise be important. (2) Investing in education, with priority actions to improve quality and access to pre-primary and primary education. Both theory and empirical analysis of the ENCOVI demonstrate the crucial role of education in promoting economic growth;4 reducing poverty and malnutrition; reducing vulnerability by making the labor force more agile and able to adjust to shocks; and reducing inequality, social disparities and exclusion. Since Guatemala is still a "primary" country on average (with average attainment of 4.3 years) and since the poor in particular fail to complete primary school, investments should still focus on expanding and improving primary education, as discussed in Chapter 7. Investments in early childhood education (pre-primary) are also crucial as they (a) increase the likelihood of success at the primary level and (b) reach children at a critical phase of their physical, cognitive,' and social development. As such, priority actions should focus on: * Increasing access to primary education, largely through demand-side interventions, since supply-side constraints are no longer binding for most of the population. This expansion should be implemented via the PRONADE program given the benefits of this program in terms of community and parental participation (see Chapter 7). However, as supply-side gaps are filled, the Government should consider easing eligibility criteria so as to allow poor communities that already have schools to be eligible for the PRONADE-type community-based school-management model. To target this expansion, the poverty map could be used to identify eligible schools and preserve PRONADE's exemplary targeting record; * Improving the quality of education, curriculum and performance standards so as to improve internal efficiency and the returns to education, particularly at the primary level; and * Investing in early childhood development to promote: (a) improved child nutrition at an early age, since nutritional status is a significant factor in determining enrollment and attainment and since 165 nutritional deficiencies emerge at a young age (see Chapter 8); and (b) early educational opportunities, including links between traditional schooling and pre-primary schooling (see Chapter 7). (3) Investing in health, with an emphasis on expanding access and usage using both supply- and demand-side interventions. Again, both theory and empirical analysis using the ENCOVI point to important linkages between health and productivity (economic growth), vulnerability (health shocks), and poverty. Guatemala's health outcomes have lagged significantly behind those in other countries as well as the targets set by the Peace Accords and the MDGs, as discussed in Chapters 4 and 5. A significant share of the population still lacks access to health facilities - or fails to use them when available - due to a mix of supply- and demand-side constraints, as discussed in Chapter 8. As such, priority actions should seek to improve health outcomes by: * Expanding access to affordable health care using both supply- and demand-side interventions (see Chapter 8). Such interventions should be targeted to the poor and priority groups (for example, using the poverty map, as discussed below); * Emphasizing preventative care, infectious and parasitic diseases, reproductive health, and key outcomes (mortality, malnutrition); and • Expanding access to potable (not just piped) water and improved sanitation to complement the basic health care package. (4) Integrating actions to reduce malnutrition into the basic health-care package. The high and stagnant rates of malnutrition in Guatemala are unacceptable, as discussed in Chapter 8. Their lasting effects also result in inter-generational transmission of poverty. Reducing malnutrition should be designated as a top priority. Malnutrition interventions should be integrated into the MSPAS basic health care package and provided at the community level through outreach workers, so as to improve their effectiveness and reach and foster the integration of malnutrition as a key concern into the health system. The target population for these schemes should be pre-school children (particularly those under 24 months of age) and mothers (including pregnant and lactating women). Priority actions include: * Promotion of proper health, hygiene, and feeding practices; o Growth monitoring of pregnant women and children under aged two; o Micronutrient supplementation (particularly for iron); and * Deworming treatments and oral rehydration therapy. (5) Reducing isolation and improving communications by investing in rural transport and roads. Many communities in Guatemala are still relatively isolated due to a lack of road access, as discussed in Chapter 10. Empirical analysis using the ENCOVI has demonstrated the effects of isolation on opportunities, productivity, vulnerability (shocks), and access to services (as discussed in Chapters 6, 10 and 11). Expanded rural transport helps build the assets of the poor, promote economic growth and opportunity, reduce vulnerability, and empower communities. Priority actions in this area include focusing on improving and expanding the network of motorable roads in rural areas, particularly those with untapped economic potential and a high concentration of poor people. (6) Improving governance and the effectiveness of the public sector. Actions are needed to reduce corruption, improve transparency, improve public expenditure management, and better target existing resources to the poor, as discussed in Chapter 13. Such actions will have multiple benefits, including: (a) making the most of existing scarce resources and improving service delivery, which is crucial under any scenario, but even more important in the event that growth were to slow; (b) fostering a climate that is more conducive to economic growth; (c) assuring that public resources reach the poor (needed for impact); and (d) improving the credibility of government and its ability to increase revenues in the future (without such improvements, the Government will face continued resistance to tax increases). Priority actions include: * Improving the tax base and increasing public spending, which will depend on improvements in governance and the effectiveness and credibility of the public sector; 166 * Improving the targeting of public spending, particularly for investments in education, health, basic utility services, and transfers (e.g., using the poverty map as an explicit criteria for allocations, as discussed in more detail below); * Improving public expenditure management, with stronger links to policy, planning and priorities; * Expanding and building on recent initiatives to fight corruption (e.g., adopting an anti-corruption charter); * Improving incentives for better service delivery (e.g., decentralization, local "control social," and service "report cards" and client satisfaction surveys); and * Improving the rule of law and the justice system. While a range of short-, medium- and long-term actions are outlined in Table 14.2, some can be undertaken immediately, including: * Promoting economic growth by maintaining macroeconomic stability and fiscal balances, while redirecting public spending towards the social sectors and rural areas (for 2002 and 2003 budgets), as discussed above and in Chapter 5. * Undertaking extensive study of economic growth with a view towards formulating a pro-poor development strategy (see below); * Using the poverty map prepared by SEGEPLAN-INE-URL as a tool for improving the targeting of public spending and poverty reduction interventions in key sectors (education, health, basic services, as discussed below). This should be applied immediately for the execution of the 2002 budget. It should also be systematized as a criteria for allocating spending during the planning of the 2003 budget. * Reviewing quality, curriculum and performance standards in education, particularly for grades 1, 7, and 10 (transition years associated with high levels of repetition and drop-out), as discussed in Chapter 7. * Consolidating and improving scholarship, school feeding, and other demand-side programs in education for the next school year, as discussed in Chapter 7. * Acknowledging the high priority of reducing malnutrition and conducting a critical review of existing malnutrition interventions (across agencies, both public and private) with a view towards (a) identifying programs that have worked (both in Guatemala and internationally); (b) streamlining and restructuring existing programs to better focus on young children, growth monitoring, information and behavioral change via community-based interventions as part of the basic health-care package, as discussed in Chapter 8. * Encouraging the social funds and other providers to adopt measures to improve the quality of water and complement water and sanitation programs with measures to improve household hygiene and water treatment practices, as discussed in Chapter 9. * Promoting community-based development for better local accountability (and "social control"), while taking steps to ensure the participation of groups traditionally excluded from community decision- making (e.g., women, indigenous, uneducated), as discussed in Chapter 13. * Drafting and adopting an anti-corruption charter, as discussed in Chapter 13. Areas for further work and analysis include: * Conducting an in-depth analysis of the (potential and existing) sources of growth in Guatemala with a view to formulating a pro-poor development strategy. The study should not only adopt a "macro" perspective, but should also look at more "micro" issues and a sectoral perspective (e.g., specific economic activities in agricultural and non-agricultural spheres, and productive activities relying on credit, technical assistance, and microenterprises). While this type of analysis is beyond the scope of the present report, it is a top priority for further research (see Chapters 5 and 6). The planned up- coming Country Economic Memorandum (CEM) should make some headway on such analysis; * Conducting additional analysis of public expenditures in the social sectors with the objective of finding ways in which their efficiency and effectiveness could be improved (see Chapters 7, 8, and 13); 167 o Reviewing quality, curriculum and performance standards in education (see Chapter 7); * Conducting critical review of malnutrition interventions (see Chapter 8); and o Reviewing supply-side issues in the health sector, including an evaluation of the institutional capacity of MSPAS and of the effectiveness and impact of the SIAS system, in particular with respect to its ability to improve health outcomes (see Chapter 8 for specific issues). PRIORITY TARGET GROUPS FOR POVERTY REDUCTION IN GUATEMALA The broad agenda for poverty reduction can become even more effective by focusing efforts on key priority groups. For example, while economic growth is needed in general, growth that provides opportunities for the rural poor will be even more effective in reducing poverty. While building assets of the poor in general is essential, priority is needed to tackle the issues of malnutrition and the relative disparities against poor women and indigenous residents. As such, the Government should prioritize among poverty groups, according to the prevalence of poverty, specific risks, and demographic circumstances. Specifically, the analysis reveals several priority groups that should be emphasized in poverty reduction efforts: (a) poor and malnourished children; (b) poor women and girls; (c) poor indigenous households; (d) the rural poor; and (e) specific geographic areas (Table 14.1). Clearly, these groups can have considerable overlaps. For example, a poor or malnourished indigenous girl living in rural areas in the North or North-Western parts of the country would probably qualify for just about any anti-poverty intervention. o Poor and malnourished children. The developmental status of children renders them extremely vulnerable to the risks of living in an impoverished environment. Youth (particularly early childhood) is the point in the life cycle when physical, cognitive, and psycho-social development occurs at its most accelerated pace and is most susceptible to abnormal development from poverty conditions. As such, childhood poverty also increases the likelihood of inter-generational transmission of poverty. About two-thirds of all Guatemalan children live in poverty. Close to half (44%) are stunted, putting Guatemala among the worst performers in the world for malnutrition, as discussed in Chapter 8. Four- fifths of malnourished children are poor. Pre-school children are particularly vulnerable to malnutrition (especially those between 6-24 months). Infant mortality is also alarmingly high. A significant share of poor pre-school and primary-aged children also fail to enroll in school, as discussed in Chapter 7. Finally, child labor is common, particularly among poor children, further compromising their chances of attending school, as discussed in Chapters 6 and 7. In this context, poverty reduction efforts should confer top priority to poor and malnourished children as a key target group. o Poor girls and women. Girls and women face cumulative disadvantages in Guatemala, reflecting historically exclusionary policies (e.g., in land and education, see Chapter 4) and a general culture of machismo. They face limited access to education (with fewer girls attending school even when schools are available, see Chapter 7), constrained employment opportunities, explicit wage discrimination (even after taking into account differences in endowments, see Chapter 6), and traditional exclusion from land ownership. Women are also at risk for health shocks, with Guatemala recording extremely high levels of maternal mortality, as discussed in Chapter 8. Furthermore, women participate significantly less in community decision-making (limited social capital networks), as discussed in Chapter 13. Yet women's roles are crucial in promoting long-term development, with a strong influence, for example, on the nutritional status of children. o Poor indigenous households. The indigenous likewise suffer cumulative disadvantages, reflecting the historical pattern of exclusion and decades of conflict. Poverty is higher among the indigenous. Indigenous children also suffer higher rates of malnutrition and less access to education, which affect their earnings ability in the future, as discussed in Chapters 7 and 8. The indigenous also have less 168 access to health and basic utility services (see Chapters 8 and 9). They are further constrained in employment opportunities (particularly those who don't speak Spanish), and face considerable wage discrimination (even after taking into account disparities in endowments), as analyzed in Chapter 6. Finally, they also report perceptions of discrimination by public officials and service providers (see Chapters 4 and 13). * Rural poor. Poverty is higher in rural areas, and even higher among specific rural sub-groups, including small land-holders, agricultural day laborers, and seasonal migrant agricultural workers (see Chapter 6). The rural poor (particularly these sub-groups) have relatively limited access to services and infrastructure (education, health, utilities, transport, markets), as discussed in Chapter 9. They also have limited employment and earnings opportunities, particularly those living in more geographically isolated areas and smaller municipalities, as analyzed in Chapter 6. They also face lower returns to their labor and are rarely covered by formal labor and IGSS benefits. Finally, they are quite susceptible to shocks, particularly natural disasters, agricultural-related shocks, and recent economic shocks (such as the coffee crisis which has worsened the terms-of-trade for producers and caused job loss for day laborers), as discussed in Chapter 11. The emphasis of the ERP on rural areas is thus correct and should be maintained. * Specific geographic areas. While poverty is clearly a national problem in Guatemala, poverty is significantly higher in the "poverty belt" in the Northern and North-Western regions as well as the departments of San Marcos. The poverty map helps further pinpoint specific municipalities with higher incidence of poverty (see Chapter 2 and Annex 4). The ERP's inclusion of the poverty map as a key tool is thus appropriate. 169 TABLE 14.1- PRIORITY TARGET GROUPS PRIORITY TARGET GROUPS KEY CONSTRAINTS/CHALLENGES POSSIBLE TARGETING TOOLS Poor and malnourished * Poverty * Poverty map combined with information children, especially pre-school * Malnutrition (stunting) on malnutrition and educational (age 0-6) and primary-aged o Not enrolled in school enrollment children (7-13) * Child labor * Self-targeting via health posts and * Vulnerable phase of life cycle community health centers (e.g., a growth . Inter-generational transmission of poverty monitoring program channeled through these facilities) * Community-based targeting * Proxy means testing Poor women and girls * Historical pattem of exclusion * Poverty map * Less access to education * Gender-based targeting (e.g., programs * Constrained in employment and eamings that restrict eligibility to girls, such as opportunities scholarships) * Face wage discrimination * Community-based targeting * Face discriminatory attitudes (culture of * Proxy means testing machismo) * Excluded from participating in community decision making (social capital) Poor indigenous * Historical pattem of exclusion * Poverty map combined with language * Higher poverty and malnutrition map * Less access to education, health services * Proxy means testing * Less coverage by basic utility services * Constrained in employment and eamings opportunities * Face wage discrimination * Face discrinmination in treatment by public officials and other service providers The rural poor, particularly * Higher poverty and malnutrition * Poverty map small land-holders, agricultural * Less access to education, health services * Vulnerability maps (e.g., natural day laborers, seasonal migrant * Less coverage by basic utility services disasters) agricultural workers * Geographic constraints (isolation, roads, small * Proxy means testing, with certain proxies municipalities) emphasized (e.g., land holdings, * Constrained employment and earnings electricity connections, etc.) opportunities * Migration maps (that could be developed * Low returns, limited coverage of labor and from census data) showing municipalities IGSS benefits with significant concentrations of * Susceptible to shocks seasonal migrants Specific geographic areas, * Higher poverty rates, malnutrition * Poverty map, combined with other asset- especially in the "poverty * Lower access to basic services specific maps/info. (e.g., gaps in coverage belt' (Norte, Nor-Occidente, * Geographic isolation, limited road network of roads, education, health services, San Marcos) utilities, etc.) Explicit efforts to target resources and interventions could greatly improve their impact. As discussed in Chapter 13 (and throughout the report), public spending on the social sectors is not well targeted to the poor. Specifically, the poorest quintiles receive disproportionately less public spending on health, education and social protection, than their share in the population. Yet targeting is crucial for many of the actions outlined above if they are to have a real impact on reducing poverty and improving social indicators. The rationale for targeting is that, given budget constraints, stronger impacts are achieved when resources are concentrated on those who need them the most. In Guatemala, this means that explicit efforts are needed to reverse historically contrary tendencies whereby most resources have traditionally been captured by elites as a matter of policy (as discussed in Chapter 4). These efforts should involve explicit criteria for, and monitoring of, budget allocations, eligibility, and project site selection. These mechanisms should be built into existing institutional processes, such as the public expenditure management system. Given budget constraints, certain activities should be actively targeted to the poor. As a "rule of thumb," incremental increases in public spending on areas such as education, health, basic utility services, core 170 communication links, social assistance transfers, or employment schemes should be explicitly targeted to the poor in order to better integrate them into the economy and improve social indicators. In contrast, decisions regarding the allocation of investments in other services, such as more intensive infrastructure, institutional support, or banking services, should generally follow indicators of economic potential (e.g., opportunities for intensification of agricultural or non-agricultural activities), which could also be combined with targeting criteria (e.g., the poverty map). Ideally, a strategy to promote pro-poor growth and reduce poverty would focus on areas that have both a large concentration of poor people, but also a strong potential for future economic activity. The Government can use a variety of tools to better target programs to priority groups. Improved use of limited public resources is crucial for poverty reduction efforts. Ensuring such resources are channeled to key poverty groups (Table 14.1) is a first step in improving the effectiveness of public spending and poverty reduction efforts. The Government has at its disposition several potentially potent tools for targeting its poverty reduction efforts to these priority groups. First, the poverty map, recently constructed by SEGEPLAN-IE-URL (Box 14.1), can be extremely useful (alone or with other targeting tools) in ensuring that resources get channeled to municipalities with high concentrations of the poor. Second, considerable efforts have been made to develop other geographic-based maps and databases, such as an extensive road network inventory/map, vulnerability maps (showing areas prone to specific natural disasters), conflict maps, municipal-level databases on education and health services, etc. The upcoming census will help update many of these maps. A unified geographic information system could combine the poverty map with these other maps and databases to better target specific interventions to the poor. Third, certain services - such as health posts and community health centers - are self-targeted to the poor. Other programs could be channeled through these facilities to take advantage of this inherent self-targeting (and perhaps even promote use of these facilities, which would have positive spill-over effects). Fourth, community-based targeting could be used (perhaps after broader program allocations are made using the poverty map) to select specific individuals eligible for programs (such as poor or malnourished children, girls or women). Given the traditional exclusion of certain groups from community decision-making (see Chapter 13), however, care should be taken to ensure that these patterns are not repeated with such mechanisms. Fifth, a unified proxy-means database (such as those in Costa Rica or Colombia) could be developed for programs targeted to individuals, though this could require significant administrative capabilities. 171 Box 14.1 - Poverty Map Poverty maps can be extremely useful tools for Government policy, in terms of improving the targeting of programs and public spending (either alone or in conjunction with other information and maps), promoting transparency in resource allocation and helping agencies resist political pressures (because allocations are based on objective poverty criteria), monitoring public spending, and tracking poverty trends. New methodologies have greatly improved the construction of poverty maps. Traditionally, poverty maps are constructed using census data and a composite index of basic needs. The disadvantage of this approach is that weights must be assigned in order to weight the relative importance of the various components (or "needs"). These weights are subjective (or arbitrary). The World Bank5 has recently developed-a new methodology for constructing poverty maps which gets around this problem by estimating the empirical relationship between these "basic-need" type variables with monetary measures of poverty (usually consumption). The weights used are thus empirical rather than subjective. The methodology combines census data (which provides the needed disaggregated sample) and household survey data (which provide monetary measures of welfare and poverty) in order to construct a map based that predicts poverty down to the municipal level (or lower). The multi-agency technical team of SEGEPLAN-INE-URL recently developed a poverty map using this new methodology with technical assistance from the World Bank under the GUAPA Program. The map was first developed using data from the Census and the ENIGFAM 1998-99, and will be updated using data from the upcoming census and the ENCOVI 2000. The poverty map was widely disseminated, both in a technical publication and in a "popular" publication prepared by SEGEPLAN and FLACSO. The map is already serving important uses. Specifically, the map was included as an official part of the Government's poverty strategy (ERP). It was also used to design criteria for allocation decisions for the 2002 public investment budget - the first time such decisions were based on objective poverty criteria. It is currently being used in conjunction with roads maps to develop the rural roads strategy and to determine eligibility for. a proposed new rural roads project. Finally, it has been used to verify geographic allocations for a re-targeting and expansion of the scholarships program. MONITORING POVERTY 1REDUCTION EFFORTS Monitoring of both poverty and poverty reduction interventions is necessary, and adequate resources should be made available for this task. First, the MECOVI program seeks to develop an integrated system of household surveys to track living conditions and provide data for the evaluation of the impact of interventions. The system will build on the ENCOVI 2000, and should execute similar surveys every 3-5 years. In addition, INE is currently developing an employment and incomes survey that would be executed on a more regular basis, to fill crucial gaps in Guatemala's information base. Finally, the upcoming Population Census will provide additional information for the monitoring of poverty, including an opportunity to update the poverty map (combined with data from the ENCOVI), as well as infrastructure maps. Second, SEGEPLAN is also developing tools to monitor actions to reduce poverty under the ERP, including: further elaborating the ERP (fleshing out details for specific sectors and developing poverty reduction strategies at the department- and municipal-levels), and developing a system of monitoring indicators for the targets set by the ERP. Ideally, a goal-based poverty reduction strategy would involve a system that relates actions and external conditions to progress in reaching the goals, incorporating evaluation mechanisms and feedback loops. The development of this type of system should clearly be coordinated with: (a) efforts to gather data (e.g., with the MECOVI program); (b) efforts to monitor the targets set by the Peace Accords and the MDGs, (c) the SLAF, which is developing performance monitoring indicators for public expenditure management; and (d) the various executing agencies (e.g., sectoral ministries). Adequate financial and technical resources should be made available to the concerned agencies for the purposes of strengthening these two facets of the monitoring system. 172 Table 14.2 - Menu of tions and Key Actions for Poverty Reduction MAIN CONSTRAINTS MAIN RECOMMENDATIONS' Key Issues I Priority I Key Actions & Time Period for Actions and Impact BUILDING OPPORTUNITIES AND LIVELIHOODS: Priority overall, especially in rural areas * Growth has slowed and isn't very "pro- .. * Maintaining macroeconomic stability, with a careful plan for allocating public poor." Economic growth is crucial for expenditures and strengthening tax collection; ACT: on-going, IMP: ST, MT reducing poverty and building opportunities, ... * Improving the climate for growth, including govemance and public sector particularly given the relatively small size management; ACT: ST, MT, IMP: MT, LT and limited capabilities of Guatemala's . * Improving regulation and supervision of financial sector; ST public sector. . * Promoting growth with emphasis in sectors that are likely to generate * Households do not perceive employment, such as non-agricultural sectors, via education and training, improvements, largely due to constrained transport, basic infrastructure, and support to MSMEs.; ACT: ST, MT; IMP: opportunities and limited eamings LT * Limited opportunities and earnings for * Reducing transactions costs in accessing markets (e.g., with road access, basic the poor, particularly the rural poor, women, services); ACT: ST, MT; IMP: MT, LT and the indigenous: * Creating mechanisms to discourage labor-market discrimination for women o Discrimination for women, indigenous and the indigenous; ACT: MT; IMP: LT o Low profitability in agriculture * Expanding land titling and land markets programs; establishing financial o Constrained entry for non-farn institutions in rural areas; ACT: MT, IMP: MT opportunities * Expanding seasonal employment creation programs (such as existing food-for- work programs) to provide opportunities for the rural poor; ACT: MT; IMP: MT BUILDING THE ASSETS OF THE POOR - EDUCATION: Priority for poor overall, especially for girls, indigenous, rural * Disparities, gaps in access: .. * Continuing increases in public spending on education, particularly at primary o Pre-primary: all poor, esp. rural and pre-primary levels; ACT: ST, MT; IMP: MT, LT o Primary: poor, esp. girls, indigenous * Expanding coverage, especially for girls and indigenous. Expansion should be o Secondary: all poor implemented via decentralized PRONADE program using poverty map to * Demand-side constraints (both primary replace supply-side restrictions as targeting mechanism; ACT: ST, MT; IMP: and secondary) . MT, LT * Supply-side constraints (mainly at * Lowering official age of entry for primary school from 7 to 6; ACT: ST; IMP: secondary) MT * Internal efficiency, quality * Reviewing and improving quality, curriculum and performance standards, * Weak targeting of pubic spending, particularly at grades 1, 7, and 10 (transition years); ACT: ST, MT; IMP: MT, education programs LT * Promoting, expanding, consolidating and improving demand-side programs, with emphasis on girls and indigenous children (e.g., scholarships, school feeding, bolsa de utiles); ACT: ST, MT; IMP: MT, LT 0 Increasing investments in early childhood development; ACT: ST, MT; IMP: MT; LT * Using poverty map and other mechanisms, to better target public spending and demand-side programs (e.g., scholarships, school feeding, bolsa de utiles); ACT: ST; IMP: MT BUILDING THE ASSETS OF THE POOR - HE LTH: Priority for poor overall, especially for girls, indigenous, rural * Health outcomes - malnutrition, infant and ... * Increasing public spending and expanding access to health care combined with maternal mortality, and morbidity - are better targeting (via poverty maps and health posts/community centers); ACT: inadequate and not improving fast enough ST, MT; IMP: MT * Public spending inadequate and not well * Emphasizing preventative care, infectious and parasitic diseases, reproductive targeted health, key outcomes (mortality, malnutrition); ACT: ST, MT; IMP: MT . Significant share of population lacks * Conducting a critical review of existing malnutrition interventions; ACT: ST; access to affordable health care, IMP: ST particularly the rural poor and indigenous * Implementing specific interventions for malnutrition as a top priority: * Supply-side constraints, including community-based information and behavioral change programs; growth fragmented services, minimal insurance monitoring for pregnant women and children under age two; micro-nutrient coverage, waste in public spending, lack of supplements. ACT: ST, MT; IMP: MT, LT medicines, doctors, staff * Focusing on demand-side interventions (e.g., conditional transfers) that could * Denand-side constraints, including cost be channeled through self-targeted health posts/community centers; ACT: ST, and cultural barriers .. MT; IMP: MT * Promoting cultumlly-sensitive health care practices; ACT: ST, MT; IMP: MT * Conducting full review of supply-side issues; ACT: ST; IMP: MT * Developing monitoring system for health outcomes, including better and more regular measurement of infant and matemal mortality; ACT: ST; I4P: MT * Adopting measures to improve efficiency and quality of services delivered (see Chapter 8); ACT: MT, LT; IMP: LT * Facilitating increased awareness of family planning options so as to reduce Guatemala's high population growth rates, which constrain per capita income .______________________________ .___ _ _ _growth; ACT: ST, MT; I4P: LT ... = top priority; *- = medium priority; * = priority; ST = one year period; MT = 1-3 years; LT = more than 3 years; ACT = period for implementation of actions; IMP = period needed for impact on poverty 173 Table 14.2, Cont'd - Menu of Options and Key Actions for Poverty Reduction MAIN CONSTRAINTS I MAIN RECOMMENDATIONS Key Issues I Priorit Key Actions & Time Period for Actions and Impact BUILDING THE ASSETS OF TEE POOR - BASIC SERVICES: Priority for poor overall, especialy for rural, indigenous * Significant coverage gaps and disparities, *. * Maintaining and, if possible, increasing resources for expansion of services; ACT: especially among rural poor and indigenous ST, MT; IMP: MT * Demand-side factors (connections costs) * Targeting service expansion to poor (particularly rural) using poverty map combined * Supply-side constraints (not available) with geographic information on coverage gaps; ACT: ST; IMP: MT 4 Energy subsidies poorly targeted * Developing strategy for demand-side constraints; ACT: ST; IMP: MT * Quality of water is poor (not potable, * Eliminating "tarifa social" energy subsidy and using resources to fund new irregular) connections instead; ACT: ST, but gradually; IMP: MT * Allowing water tariffs to rise to a level that allows water utilities to become finan cially sustainable and improve the quality of service offered; ACT: ST but gradually; IMP: ST-MT * Encouraging social funds and other providers to consider measures to improve quality of water; ACT: ST; IMP: ST * Complementing water and sanitation programs with measures to improve household hygiene and water treatment practices; ACT: ST; IMP: ST BUILDING THE ASSETS OF TEE POOR - TRANSPORT: Priority for rural poor * Geographic isolation for rural poor, due to * Focusing public spending on transport on rural areas; ACT: ST; IMP: ST, MT limited road network and public transport * Expanding and improving motorable road network in rural areas, particularly by services improving existing roads (including dirt roads); ACT: ST, MT; IMP: ST, MT * Road quality and closures limit year-round * Targeting expansion and rehabilitation using combination of poverty map with road access maps; ACT: ST, MT; IMP: ST, MT * Road improvements have favored non-poor, urban areas * Inadequate road access significantly constrains access of rural poor to health services, opportunities, institutions REDUCING VULNERABILITY: Priority for U poor/vulnerable, particularly rural and specific vulnerable groups * Lack of assets makes poor vulnerable to ... * Building assets of poor and key vulnerable groups (see Table 14.1); ACT: ST, MT; shocks, particularly natural disasters and IMP: MT, LT agriculture-related shocks *- * Expanding and improving disaster management relief; ACT: ST, MT; IMP: MT * Key sources of future vulnerability: (a) * Introducing catastrophic insurance schemes; ACT: MT; IMP: MT coffee crisis; (b) lost remittances from global * Improving targeting of social protection programs; ACT: ST, MT; IMP: MT slowdown; (c) natural disasters *- * Eliminating energy subsidy and school transport subsidy; ACT: ST; IMP: ST * Certain sub-groups are particularly * * Consolidating and improving scholarships and school feeding programs; ACT: ST, vulnerable due to special circumstances MT; IMP: ST, MT * Faced with shocks, households rely on own * Improving targeting of bolsa de utiles program; ACT: ST, MT; IMP: ST, MT assets with little formal assistance • Existing social protection programs are poorly targeted and ineffident IMPROVING INSTIT UTIONS AND EMPOWERING COMMUNITIES: Prionit for all poor * Weak public sector hampers poverty * Improving tax base and tax collection; ACT: ST, MT; IMP: ST, MT reduction efforts * Increasing and improving targeting of public spending; ACT: ST, MT; IMP: ST, o Weak tax base, limited public spending MT o Public exp. management needs *.. * Improving public expenditure management, with stronger links to policy, planning, strengthening and priorities; ACT: ST, MT; IMP: ST, MT o Public spending poorly targeted *. * Strengthening the civil service; ACT: MT; IMP: MT o Weak civil service e * Improving incentives for better service delivery (e.g., implementing recently-passed o Overly centralized laws on decentralization, local "control social," and service "report cards" and client * Governance weak, constrains growth and satisfaction surveys); ACT:-MT, LT; IMP: MT, LT poverty reduction efforts: corruption, lack of * Expanding and building on recent initiatives to fight corruption (e.g., an anti- rule of law, inadequate justice system, political ** corruption charter) and making it a top priority. ACT: ST, MT; IMP: ST, MT instability. * Improving rule of law, justice system; ACT: ST, MT; IMP: ST, MT * Social capital lmited, concentrated among * Promoting community-based development but with explicit outreach programs to privileged ensure participation of excluded groups (women, poor, uneducated) in community- o ULmited networks outside villages decision making; ACT: ST, MT; IMP: ST, MT o Community participation lmited for * * Partnering with private sector, NGOs to extend services; ACT: ST, MT; IMP: ST, women, poor, uneducated MT 00= top priority; ;. = medium priority; * = priority; ST = one year period; MT = 1-3 years; LT = more than 3 years; ACT = period for implementation of actions; IMP = period needed for impact on poverty 'These efforts are being led by SEGEPLAN under the ERP initiative. 2 The "priority" column in Table 14.2 reflects not only a rating of relative priorities for poverty reduction (based on the relative importance of these factors for poverty and growth), but also judgments about the feasibility (administrative and political) of certain interventions. 3SEGEPLAN (November 2001). 4Indeed, Loening (2002) demonstrates the empirical impact of education on economic growth in Guatemala. 5 For technical details, see Hentschel, Jesko, Lanjouw, et al. (2000). 174 REFERENCES FOR MAIN REPORT Anderson, Maria Elena (October 2001): "Guatemala: the Education Sector." Draft prepared for the World Bank. Arcia, Gustavo (February 2002). "Educaci6n y Reducci6n de Pobreza - Estrategia para el perfodo 2002- 2005." Draft. Arragon and Associates (June 1998). "User Perceptions of Justice Administration in Guatemala." Draft. Asociaci6n de Investigaci6n y Estudios Sociales (ASIES, 2000). "Caracteristicas y Funcionamiento del Mercado Laboral en Guatemala." Asociaci6n Gremial de Exportadores de Productos No Tradicionales (AGEXPRONT, 2000): "Exportaciones La Experiencia de Una Decada." Bachrach Ehlers, Tracy. Silent Looms: Women and Production in a Guatemalan Town. University of Texas Press. Barrios, Lina Eugenia (2001). El Papel de la Alcaldia Indfgena en el Desarrollo. Britnell, G.E. (November 1951). Problems of Economic and Social Change in Guatemala. The Canadian Journal of Economics and Political Science. Browning, Felix Alvarado, Maribel Carrera Guerra and Abel Gir6n (1998). Perfil de las Organizaciones No Gubernamentales en Guatemala. CEPAL (1991). Magnitud de la Pobreza en America Latina en los Afios Ochenta. Centro de Investigaciones Econ6micas Nacionales (CIEN) (November 2000). Proyecto del Presupuesto General de Ingresos y Egresos del Estado del Ejercicio Fiscal 2001. Informe al Congreso No. 45. Centro de Investigaciones Econ6micas Nacionales (CIEN) (May 2001). "Del Autorismo a la Democracia: Democracia, Transparencia e Instituciones." Centro de Investigaciones Econ6micas Nacionales (CIEN) (May 2001). "Del Autorismo a la Democracia: Democracia, Transparencia e Instituciones." Centro de Investigaciones Econ6micas Nacionales (CLEN) (Noviembre 2000). Informe al Congreso: Proyecto del Presupuesto General de Ingresos y Egresos del Estado del Ejercicio Fiscal 2001. No. 45. Centro de Investigaciones Econ6micas Nacionales (CGEN) (Junio 2001). Informe al Congreso: Ejecucion del Presupuesto de Ingresos y Egresos del Ejercicio Fiscal 2000. No. 46. Centro de Investigaciones Econ6micas Nacionales (CGEN) (August 2001). Carta Econ6mica: En Busca de una Bureaucracia Eficiente. Carta No. 224. by Vilma de Liu. Centro de Investigaciones Econ6micas Nacionales (CGEN) (Noviembre 2000). Carta Econ6mica: El Problema de la Corrupci6n. Carta No. 227. by Maria del Carmen Acenia de Fuentes and Vilma de Lid. 175 Clert, Carine and Ana-Marfa Ibafiez (2002). Exclusion and Poverty in Guatemala's Rural Villages: The Challenge of Tackling Cumulative Barriers. Guatemala Poverty Assessment Report. The World Bank. Comisi6n para el Esclarecimiento Hist6rico (CEH, Truth Commission) (1999). Guatemala: Memoria del Silencio. Comisi6n de Acompafiamiento del Cumplimiento de los Acuerdos de Paz (December 2000). Cronograma de Implementaci6n, Cumplimiento, y Verificaci6n de los Acuerdos de Paz: 2000-2004. Cosminsky, S. and M. Scrimshaw. Medical Pluralism on a Guatemalan Plantation. Social Science and Medicine. Vol. 148: 267-278. Davis, Shelton H. (1988). "Sowing the Seeds of Violence." In Harvest of Violence: The Maya Indians and the Guatemalan Crisis. Robert M. Carmack (Ed). University of Oklahoma Press: Norman and London. Davis, Shelton H. (2001). El Potencial de los Recursos Culturales Para el Desarrollo. de Ferranti, David, Guillermo Perry, Indermit Gill and Luis Serven (2000). Securing our Future in a Global Economy. The World Bank. Edwards, John (2002j. Education and Poverty in Guatemala. Guatemala Poverty Assessment Report. The World Bank. EIm-Economist Intelligence Unit (November 2001). Guatemala Country Report. EIU-Economist Intelligence Unit (January 2002a). Guatemala Country Report. EIU-Economist Intelligence Unit (January 2002b). Guatemala Country Risk Service Report. Eveleth, P.B. and J.M. Tanier (1976). Worldwide Variation in Human Growth. Cambridge: Cambridge University Press. Eveleth, P.B. and J.M. Taiier (1990). Worldwide Variation in Human Growth. Second Edition. Cambridge: Cambridge University Press. Ferreira, F.H.G. and Litchfield, J.A. (1997). "Poverty and Income Distribution in a High-Growth Economy: Chile: 1987-1995." Capitulo 2 "Income Distribution and Poverty: A Statistical Overview." FLACSO (November 2001). Pueblos Indfgenas y Pobreza. Foro de Coordinaciones de ONG de Guatemala. Perfil de las Organizaciones No Gubemamentales en Guatemala. Gacitua Mario, E and C Sojo, with S. Davis (2000). Social Exclusion and Poverty Reduction in Latin America. World Bank and FLACSO, Washington, DC. Foster, Vivien and Araujo, Caridad (2002). Poverty and Modem Utility Services. Guatemala Poverty Assessment Report. The World Bank. Fuentes & Associados (April 2001). "Intervenciones de Protecci6n Social Financiados con Fondos Empresariales Privados." 176 Gragnolati, Michele and Marini, Alessandra (2002). Health and Poverty in Guatemala. Guatemala Poverty Assessment Report. The World Bank. Grandin, Greg (2000). The Blood of Guatemala: A History of Race and Nation. Duke University Press. Habicht, J.-P., R. Martorell, C. Yarbrough, R.M. Malina, and R.E. Klein. 1974. Height and Weight Standards for Preschool Children. How Relevant Are Ethnic Differences in Growth Potential. Lancet 6:61- 614. Hansen, John R. (August 1978). Guatemala: Economic and Social Position and Prospects. A World Bank Country Study. Hentschel, Jesko, Jean Olson Lanjouw, Peter Lanjouw, and Javier Poggi. "Combining Census and Survey Data to Trace the Spatial Dimensions of Poverty: A Case Study of Ecuador." World Bank Economic Review, Vol 14, No. 1, 2000. Heuveline, P. and N. Goldman (1998). A Description of Child Illness and Treatment Behavior in Guatemala. Mimeo (Office of Population Research, Princeton NJ). Ibanez, Ana-Maria, Kathy Lindert, and Michael Woolcock (2002). Social Capital in Guatemala: A Mixed Methods Analysis. Guatemala Poverty Assessment Report. The World Bank. Instituto Nacional de Estadistica - Guatemala and FNUAP (1989). Perfil de la Pobreza: Encuesta Nacional Sociodemografica 1989. Instituto Nacional de Estadistica - Guatemala (1999). Encuesta Nacional de Ingresos y Gastos Familiares 1998-1999. Instituto Nacional de Estadistica - Guatemala (October 2001). Encuesta Nacional sobre Condiciones de Vida ENCOVI 2000: La Pobreza en Guatemala, Principales Resultados. Instituto Nacional de Estadistica - Guatemala (SE), Secretaria General de Planificaci6n y Programaci6n de la Presidencia (SEGEPLAN), Universidad Rafael Landivar (URL), y Programa de Naciones Unidas para el Desarrollo (PNUD), (May 2002). "Encuesta Nacional de Condiciones de Vida 2002: Perfil de la Pobreza en Guatemala." International Bank for Reconstruction and Development (1951). The Economic Development of Guatemala. Report of a World Bank Mission. International Monetary Fund (April 2001). Guatemala: Recent Economic Developments. International Monetary Fund (April 2001). Guatemala: Staff Report for the 2001 Article 2001 IV Consultation. Jonas, Susanne (2000). Of Centaurs and Doves: Guatemala's Peace Process. Westview Press. Kaufman, Daniel, Aart Kraay, and Pablo Zoido-Lobat6n (October 1998). "Governance Matters." World Bank Policy Research Working Paper No. 2196. 177 Loening, Ludger J (June 2002). "The Impact of Education on Economic Growth in Guatemala." Discussion Paper No. 87, lbero-America Institute for Economic Research, Georg-August-Universitait Gottingen. Logan, M.H. (1973). Human Medicine in Guatemala and Peasant Acceptance of Modem Medicine. Human Organization. Vol 32(4): 385-395. Lopez, Humberto (October 2000). "The Cost of Armed Conflict in Centro-America." draft paper. Lopez Rivera, Oscar Augusto. Guatemala: Intimidades de la Pobreza. Universidad Rafael Landivar, Instituto de Investigaciones Econ6micas y Sociales. Marini, Alessandra and Gragnolati, Michele (2002). Malnutrition and Poverty in Guatemala. Guatemala Poverty Assessment Report. The World Bank. McCreery (1976). "Coffee and Class: The Structure of Development in Liberal Guatemala." Hispanic American Historical Review, vol. 56. McCreery (1983). "Debt Servitude in Rural Guatemala, 1876-1930." Hispanic American Historical Review, vol. 64. McCreery (1994). Rural Guatemala: 1760-1940. Stanford University Press. MINUGUA (2000). Informe de Verificaci6n: Los Lincharnientos: un Flagelo Contra la Dignidad Humana." MINUGUA (June 2001). Sexto Informe del Secretario General de las Naciones Unidas Sobre la Verificaci6n de los Acuerdos de Paz de Guatemala. MINUGUA (September 2001). Duodecimo Informe Sobre Derechos Humanos de la Misi6n de Verificaci6n de las Naciones Unidas en Guatemala. Moser, Caroline and McIlwaine, Cathy (2001). Violence in a Post-Conflict Context: Urban Poor Perceptions from Guatemala. The World Bank. MSPAS, OPS/OMS (2001). Propuesta: Establecimiento de las Redes Sociales Locales, su Epidemiologia Cultural, Recursos Etno-terapeuticos, su Interacci6n con el Medio Social y las Redes del Sistema Oficial de Salud. Guatemala. Narayan, D. and L. Pritchett (1999) "Cents and Sociability: Household Income and Social Capital in Rural Tanzania," Economic Development and Cultural Change 47(4): 871-97. Nugent, Jeffrey B. and James A. Robinson (August 2000). "Are Endowments Fate?" Unpublished draft paper. Nuiiez, Cesar Antonio (2001). "Exclusi6n social y VIH-SIDA en Guatemala." United Nations. Rojas, Ferdnando (February 1999). A Strategic View of Decentralization in Guatemala. Draft paper. ODHAG, Guatemala (1998). Guatemala Nunca Mds. Volume IV: Victims of Conflict. Pebley, A.R., E. Hurtado and N. Goldman (1997). "Beliefs about Children's Illness Among Rural Guatemalan Women." In Journal of Biosocial Sciences. 178 Plant, Roger (1995). Rebuilding Civil Society: Rural Workers' Organizations in Guatemala. International Labor Office: Issues in Development Discussion Paper No. 5. Psacharopoulos, George and Ying Chu Ng (1992). "Earnings and Education in Latin America: Assessing Priorities for Schooling Investments." World Bank Policy Research Working Papers Series no. 1067. Psacharopoulos, George (1993). "Returns to Investment in Education: A Global Update." World Bank Policy Research Working Papers Series No. 1067. Psacharopoulos, George and Harry Anthony Patrinos, Eds. (1994). Indigenous People and Poverty in Latin America: An Empirical Analysis. World Bank Regional and Sectoral Studies. Puri, Jyotsna (2002). Transport and Poverty in Guatemala: a Profile Using Data from the ENCOVI 2000. Guatemala Poverty Assessment Report. The World Bank. Rodriguez Santana, Martha (2001). Percepciones Sobre la Educaci6n: Un Estudio Cualitativo y Multi- Etnico En Guatemala. Guatemala Poverty Assessment Report. The World Bank. Sabino, Carlos A. Guatemala: Dos Paradojas y Una Incognita. Atlas Economic Research Foundation. Santiso, Carlos (July 8, 2001). Guatemala: Analysis of Social Protection Interventions in the Public Sector in Guatemala in 2000. The World Bank. Secretaria de Planificaci6n y Programaci6n, SEGEPLAN (February 2001). "El Drama de la Pobreza en Guatemala: Sus Rasgos y Efectos Sobre la Sociedad." Secretaria de Planificaci6n y Programaci6n, SEGEPLAN (November 2001). Estrategia de Reducci6n de la Pobreza: El Camino de la Paz. Gobierno de Guatemala. Secretaria de Planificaci6n y Programaci6n, SEGEPLAN (June 2001). Mapas de Pobreza: Informe Final. Secretaria de Planificaci6n y Programaci6n, SEGEPLAN (November 2001b). Mapas de Pobreza de Guatemala: Un Auxiliar para Entender el Flagelo de la Pobreza en el Pais. Solares, J, 1992, 'Guatemala: Etnicidad y Democracia en Tierra Arrasasada', en Problemas de la democracia, Cuidad de Guatemala. FLACSO. Solares, Jorge (1997). Etnicidad, Salud y Riesgo en Guatemala. OPS/OMS-BID. Tax, Sol (1953). Penny Capitalism: A Guatemalan Indian Economy. Srnithsonian Institution, Institute of Social Anthropology, No. 16. Tesliuc, Emil and Lindert, Kathy (2002). Vulnerability: a Quantitative and Qualitative Assessment. Guatemala Poverty Assessment Report. The World Bank. Tesliuc, Emil and Lindert, Kathy (2002). Social Protection, Private Transfers and Poverty: a Quantitative and Qualitative Assessment. Guatemala Poverty Assessment Report. The World Bank. Tierney, Nancy Leigh (1997). Robbed of Humanity: Lives of Guatemalan Street Children. 179 Tovar G6mez, Marcela (May 1998). Perfil de los Pueblos Indigenas de Guatemala. The World Bank. Transparency International (September 2001). Global Corruption Report. Universidad Rafael Landfvar, Instituto de Investigaciones Econ6micas y Sociales (2001). Inventario de Intervenciones de Protecci6n Social con Fondos Pdblicos. United Nations (2001). World Population Prospects: The 2000 Revision Highlights. United Nations Development Programme (2001). Human Development Report. United Nations Development Programme (2001b). Informe de Desarrollo Humano 2001: Guatemala: El Finaciamiento del Desarrollo Humano. United Nations Development Programme (1999). Guatemala: el rostro rural del desarrollo humano. United Nations Development Programme (2000). Guatemala: la fuerza incluyente del desarrollo humano. Vakis, Renos (2002). Guatemala: Livelihoods, Labor Markets and Rural Poverty. Guatemala Poverty Assessment Report. The World Bank. Villamar Contreras, Marco Antonio (January 1993). Significado de la Decada 1944-1954 Conocido Como la Revoluci6n Guatemalteca de Octubre. von Hoegen, Miguel (2000). Tres Retos para la Inclusi6n en Guatemala: La Inversion Publica, la Educaci6n Formal, y el Empleo en el Sector Formal. Universidad Rafael Landfvar/Instituto de Investigaciones Econ6micas y Sociales. von Hoegen, M. (2001). Las Macro Tendencias del Empleo Formal en la D&cada de 1990. Universidad Rafael Landivar, Guatemala. von Hoegen, M. and Palma (1999). Los Pobres Explican la Pobreza. Universidad Rafael Landivar, Guatemala. World Bank (1995). Guatemala: An Assessment of Poverty. Report No. 12313-GU. World Bank (2001). Rural Poverty reduction in Brazil: Towards an Integrated Strategy. Report No. 21790-BR. World Bank (1996). Guatemala: Building Peace with Rapid and Equitable Growth; Country Economic Memorandum, Report No. 15352-GU. World Bank (1997). World Development Report 1997: The State in a Changing World. World Bank (2000a). Panama Poverty Assessment: Priorities and Strategies for Poverty Reduction. A World Bank Country Study. World Bank (February 2000). Guatemala: Expenditure Reform in a Post-Conflict Country. Report No. 19617-GU. 180 World Bank (2001 a). Nicaragua Poverty Assessment: Challenges and Opportunities for Poverty Reduction. Report No. 20488 NI. World Bank (2001b). World Development Indicators 2001. World Bank (2001c). Honduras Poverty Assessment. World Bank (2001d). The Little Green Data Book 2001. World Bank (2001e). World Development Report 2000/2001: Attacking Poverty. World Bank (2002). World Development Report 2002: Building Institutions for Markets. World Bank and the Carter Center (July 1997). From Civil War to Civil Society: The Transition from War to Peace in Guatemala and Liberia. World Faiths Development Dialogue (November 1999). Poverty & Development: An Inter-faith Perspective. Wodon, Quentin (July 31, 2000). Guatemala: Preliminary Poverty Diagnostic. The World Bank. Wodon, Quentin (2001). Poverty and Policy in Latin America and the Caribbean. World Bank Regional Study Number 467. Woolcock, Michael and Deepa Narayan (2000) "Social Capital: Implications for Development Theory, Research, and Policy." World Bank Research Observer 15(2): 225-49 181 Annex 1. Page 1 ANNEX 1- MEASURING CONSUMPITION USING THE ENCOVI 2000 * MEASURING WELFARE: TOTAL CONSUMPTION 1. Assessing poverty relies on some measure of welfare. Since well-being, or utility, cannot be measured directly, consumption is used as an indirect measure of welfare. Consumption is used instead of income for several reasons. First, consumption is considered a better indicator of standards of living since it fluctuates less than income during a month or year. When incomes change (e.g., in different seasons), individuals tend to use their savings (in cash and kind) to smooth consumption throughout the year. Second, consumption data tend to be more accurate than information on individuals' incomes. International experience has shown that respondents tend to provide more accurate information on consumption than income. The latter is often underestimated or difficult to measure due to informal or in-kind income. Finally, using consumption as a measure of welfare has the advantage that poverty lines can be derived from the same data and not from other information sources. 2. Consumption also has several advantages over other welfare measures, such as indicators of basic needs (as access to water, electricity, and schooling; malnutrition; etc.). While consumption is an objective measure of welfare, indicators of basic needs are based on various subjective definitions, including the level at which such needs would be "satisfied" and the respective weights assigned to their components. Moreover, indicators of basic needs are not responsive to short-term changes, since they mainly reflect public investments. As such, they are less useful for monitoring changes in economic conditions. Although the Poverty Assessment uses consumption as the basis to measure welfare and poverty, the vast array of data available from the ENCOVI 2000 allow for the use of basic social indicators (as malnutrition access and use of basic services) to complement this quantitative measure of poverty. COMPONENTS OF TOTAL CONSUMPTION Overview 3. The ENCOVI 2000 includes the data necessary to construct a measure of total consumption. This measure includes the annual consumption of food (both purchased and non-purchased, including own-production), housing (using an imputed value for owned housing), durable goods, spending on consumer goods and services, basic services (water, gas, electricity), and outlays on health and education. These components are described in detail below. The prices used to value the consumption of these components come mainly from the household and community surveys. A price index was established to adjust for geographical cost differentials (see below). Finally, information on household members was used to convert household consumption (collected in the survey) into a measure of the individual (per capita) welfare, taking into account household size. | . .Box 1 - Components of Total Consumpton - - Consumption of purchased food - . . Consumption of non-purchased food town-production, gifts. donations) . . Transpqnr.andconmurucauon . . ! - . - : Spending on consumer goods - . - . * . Household serVices and legal cosLs . - - - - . . X - - ~~Basic serv ices l water, elecUncty, gas).. - . - - Annual use value of housing - - - - Annual use %alue of durable goods * - -Education . : Health Annex 1. Page 2 Food Consumption 4. Purchased Food. The main data source for purchased foods is Section 12.A of the ENCOVI 2000 household questionnaire ("Spending and Consumption of Food, Drinks, and Tobacco"). Question 3 (variable P12A03) indicates if household members purchased each item during the last 12 months. Using this section, the number of months (question 4, variable P12A04) in which each food item was consumed was multiplied by the average monthly value (question 5, variable P12A05) to obtain the annual value of consumption. 5. To calculate annual spending on foods purchased in supermarkets', Section 12.B of the ENCOVI 2000 ("Place and Frequency Purchases" of food) was used, multiplying the total value of purchases (question 12, variable P12B 12) by the annual frequency of purchases (question 11, variable Pl2B 112). In addition, the annual value of food consumed outside the household was calculated by multiplying weekly expenditures in food and drinks consumed outside the home (variable GHOGAR for variable ITEM = 105) in Section 12.C (spending the last 7 days) by 52 (weeks per year). 6. Adding the annual expenses of all purchased foods, food purchased in supermarkets, and food consumed outside the home yields the total annual spending on purchased foods. 7. Non-Purchased Food. Even though the consumption of these items does not involve a monetary outlay, household welfare increases in the same way as with purchased food. The main data source for the consumption of non-purchased foods is Section 12.A in the ENCOVI 2000 household questionnaire ("Spending and Consumption of Food, Drinks, and Tobacco"). Question 7 (variable P12A07) indicates whether the item was obtained by, own production or through other means (donations, partial reimbursement, or from a business) during the last 12 months. To obtain the annual quantity, the number of months in which each food was consumed (question 8, variable P12A08) was multiplied by the average monthly amount (question 9, variable P12A09A3) and by the corresponding price. 8. To obtain the annual value of non-purchased food consumption, the annual amount was multiplied by a price. In the case of non-purchased food, however, prices and values were not reported (since such quantities were never purchased or sold). Therefore, prices were imputed as follows. First, if the household also purchased the item (in addition to being consumed from non-purchase acquisitions), the price paid for the purchased quantities was used. To impute this paid price, the total value of purchases during the last 15 days (question 6d, variable P12AO6D) was divided by the amount purchased during the last 15 days (question 6a, variable P12AO6A) 4. Second, if this price was unknown (because the good was never purchased), its value was estimated using the prices paid by nearby households (geographically), since they would presumably have access to similar markets. Information from the price questionnaire was used when no price information was available for an specific item in the nearby households. 9. The consumption of food from social programs was also included (from Section 4.c, "Participation and Benefits from Social Programs"). To obtain the annual value of free food consumed outside the home5, the social programs "School Cookie", "School Breakfast", "Powder Milk", "Glass of Milk", and "Glass of atoll" were selected (question 1, variable P04CO1 with values of 1, 2, 3, 4 or 5) and ' If no individual information for food purchases was provided 2 For daily purchases, the frequency was 365, for weekly purchases it was 52, for monthly purchases it was 12 and for annual purchases the frequency was one. 3 After taking into consideration the product unit from question GA109B. 4 Unit weight information was collected from several local markets. This information was used to transform the units reported into pounds or units 5 Not all the food was 100% free. Annex 1. Page 3 the reported value from questions 4, 8 and 12 (variables P04C04, P04C08 and P04C 12) was multiply by 12 for the "powder milk" program, and by 10 for the other programs (month of services related to the school calendar). 10. The total annual value of non-purchased food consumption is obtained by adding the imputed annual expenses of all non-purchased foods consumed at home (internal consumption, gifts, donations) and food from social programs. Spending on Consumer Goods and Services 11. The main data source for outlays on goods and services that are generally consumed in one year or less (such as matches, soap, detergent, newspaper, deodorants, books.' school or non- work related transportation expenses, shoes, clothing, etc.) are include in section 12.c of the ENCOVI 2000. All expenses were reported in question 4 (variable GHOGAR). For expenses during the last 7 days, the value reported was multiplied by 52 weeks to obtain the annual value, for expenses during the last month, the value was multiplied by 12 months for the annual value, and annual expenses were included directly. Question 3 from the same section (variable ITEM) include all expenditures from the last seven days (codes 101-111), from last month (codes 201- 23 1), and from last year (codes 301-326). Transport and communication included the annual value from questions 101, 102, 103, 104, 107, and 314. Spending on consumer goods included the annual value from questions 106, 108 through 217, 221, 222, 223, 225, 301, 302, 303,305, 308 through 313, 315 and 325. It also includes information provided in Section 4.c, questions 4, 8 and 12 (variables P04C04, P04C08 and P04C12) for the non food social programs benefits (variable P04CO1 with values between 6 and 116). Household services and legal included the annual value from questions 218, 220, 224, 231, 319, 322 and 326. Household Services: Energy, Water, Telephone 12. Data on household water, sanitation and communication services expenses come from Section l.a of the ENCOVI 2000 ("Housing conditions"). To obtain the annual value of household water consumption, monthly values from "piped water" (question 17, variable POlA17), and "non piped water" (question 24, variable POlA24) were multiplied by 12 months. For garbage recollection, the monthly expenditure from question 33 (variable POA22) was multiplied by 12. To obtain annual spending on regular telephone, cellular telephone, beeper, intemet and cable connections, the monthly consumption (question 29, variables POIA29A, B, C D and E) were multiplied by 12 months. Data on monthly spending on household energy sources consumption (candles, kerosene, gas, coal, batteries, electricity, firewood and others) come from Section l.b ("Sources of Energy"). Yearly values were derived by multiplying monthly consumption (question 7, variable POIB07) by 12 months. Total annual spending on household services equals the sum of annual spending for each of the household services. Annual Use Value of Housing 13.i The annual use value of the housing must be included in total consumption for each household. Data on housing come from Section l.c of the ENCOVI 2000 household questionnaire ("Housing ownership"). 14. Rented housing. Rent is considered to be a good estimate of the use value of housing for those households that pay for the use of their house, apartment, or other type of home. As such, for rented 6 School transport subsidy, scholarships, school materials, health program, female children program and others. Annex 1. Page 4 housing, the annual rent value was calculated by multiplying monthly rent (question 8, variable P01C08) by 12 months and included in the consumption aggregate. 15. Owned Hiousing (not rented). The annual use value of owned housing was imputed as follows: (i) in most cases, the value estimated by the owners was used; or (ii) for households that did not provide an estimated value, the use value of housing was estimated by a regression (as discussed below). 16. (i) Value estimated by owners. The use value estimated by owners was used for most cases of owned housing. Fortunately, the ENCOVI 2000 asked households that did not rent: "If you had to pay rent for this housing, how much would you pay on a monthly basis?" (question 7, variable POlC07). The answer to this question was used as an estimate of the rental value of the housing and therefore as an estimate of its use value (the estimated value was multiplied by 12 months to obtain the annual value). 17. (ii) Value not estimated by owners. In 0.36% of the 7,276 households, the respondents did not provide an estimate of the rental value for owned housing; consequently, their value was estimated from the average value of nearby households. Value of the Annual Use of Durable Goods 18. Many goods are only partially consumed during the study period, such as cars, refrigerators, stoves, etc. Even if a television set has been purchased during the time period of the survey, it is expected to be used (and hence consumed) during many years to come. To reflect the current welfare that these goods provide to the household, the "value of one year of use" (annual use value) must be estimated and incorporated (rather than the actual purchase cost of these goods), whether the item was purchased in the current year or in previous years. 19. Data on the consumption value of household durables come from Section 14 of the ENCOVI 2000 (Household durable goods"). Since these goods are generally not entirely consumed during one year, the value of their use during the past year had to be estimated. For example, if someone bought a television set this year for Q/.3,000.00, the annual consumption value of this television set is not Q/.3,000.00, since the individual can also use the television during the following year, i.e., the Q/.3,000.00 will be consumed during a time period of more than one year. Food and other consumer goods do not have this characteristic, because if someone buys one liter of milk, this milk will be consumed in less than one year. 20. Three data points are needed to estimate the consumption value of the household durables (i) the age of the durable good (question 3, variable P14A03); (ii) the remaining use life of the durable good; and (iii) the current value of the durable good (question 5, variable P14A05). 21. To obtain the remaining use life of durable goods, we need to know the average lifetime of each good7 or, as commonly referred to, its use life or expected lifetime. If the use life of the durable good is known, we will only need to subtract its age to obtain the remaining lifetime. Fortunately, ENCOVI 2000 data allow for an estimate of the expected lifetime of each durable good.' Assuming that in one year a similar percentage of the population buys a durable good (say a television), it is likely that some individuals will have a new television, some will have televisions that are one-year old, others two-years old, etc. As such, calculating the average age of all televisions sets (average of P14A03) yields the mean life or average age of all televisions. By multiplying the mean life by two, the result would be the expected lifetime of a television set in years. If the reported age (variable P14A03) is subtracted from the expected lifetime of a television set, the remaining use life of each television set is obtained. Finally, 7Each item is identify in question 1, variable ITEM. Annex 1, Page S dividing the current value of a television set (variable P14A05) by the remaining use life yields the annual use value of the television set. 22. Applying this procedure for all durable good and adding the values of each item yields the annual value of the consumption of household durable goods. Education 23. Data on household expenses on education (such as registration and enrollment fees, uniforms, books or material, travel) come from Section VII of the ENCOVI 2000. The ENCOVI 2000 asked households for annual pre-primary school expenses for children under 6 years (questions 3,4 and 5, variables P07A03, P07A04 and P07A05 in Section 7.a) and for studentsage-d 6 and over (questions 12, 13 and 17; variables P07B 12, P07B13 and P07B17 in Section 7.b). Households were also asked for the monthly expenses for children under 6 (questions 6-9, variables P07A06- P07A09 in Section 7.a) and students aged 6 and over (questions 19 through 22; variables P07B19 through P07B22 in Section 7.b). To obtain the annual value of the monthly expenses, they were multiplied by 10 months. 24. Total annual education consumption is obtained by adding the educational expenses and scholarships for all household members. Health 25. The data source for health expenses for the past month is Section 6. Health spending for children 5 years and younger for diarrhea and respiratory problems are in Section 6.c, question 9 (variables P06CO9A-PO6CO9F and P06CO9T). Section 6.d has the information for all the household members expenditures during last month8 on doctor fees (question 11, variable P06D 11), medicines (question 12, variable P06D12), X-rays and tests (question 13, variable P06D13), transport to medical facilities (question 14, variable P06D14), orthopedic equipment (question 15, variable P06D15), glasses, hearing aid, dentures, etc (question 16, variable P06D16), hospitalization (question 18, variable P06D18), and health insurance premiums (question 19, variable P06D19). Expenditures related to pregnancy were reported in an yearly basis in Section 11 (questions 14 and 23, variables P1 IA14 and P1 IA23). 26. Section 12.c has the information for health expenditures for the last 12 months (excluding the previous month and all pregnancy expenses) in question 20 variable GHOGAR (for ITEM = 320). Monthly expenses on accident and death insurance reported in questions 28, 29, and 30 of Section 12.c (variable GHOGAR for ITEM = 228, 229 and 230) were also included. 27. Monthly expenses on accident and death insurance reported in questions 28, 29, and 30 of Section 12.c (variable GHOGAR for ITEM = 228, 229 and 230) were also included (multiplied by 12 months to obtain the annual value). 28. Total annual health spending is obtained by adding all expenditures reported in these questions. Other 29. Total consumption did not include donations (Section 13.b) because it was difficult to avoid double counting. Out of 7,276 households 517 (7.1%) household reported receiving food consumed outside the house or goods (question 2, variables P13BO2B or P13B02C = 1). Since the type of donation received accepted multiple answers, it is not always possible to differentiate from valid from invalid 8 Excluding expenditures reported in section 6.c and pregnancy related health expenditures. Annex 1. Page 6 9 types9. Only 328 households reported valid types of donation and at the same time did not report invalid types. Finally, it was not possible to differentiate from this donations and information provided (and already included) from the social programs. Due to this problems, no information was included from this section. Total Consumption 30. Finally, by adding all consumption values for each component (by household), we obtain the total consumption variable. Thirty three households were excluded from the original figure of 7,309 households (yielding a total of 7,276) because a large share of the consumption aggregate had to be estimated or imputed due to missing values.'° WEIGHTING TOTAL CONSUMPTION BY TEE REGIONAL PRICE XNDEX 31. The cost of living is not uniform throughout the country; as such, the value of total consumption was adjusted to account for regional variation in prices. Price indices were constructed for each Department (22), Area (Urban/Rural) combination (44 indices) using the information collected in the price questionnaire and the household questionnaire (Section 12.a) in the following manner. 32. Using consumption data from Section 12.a, "national average consumption in pounds" was calculated for each food article. This was achieved by dividing the national average value of annual purchased and non purchased food1 ' (without the Social Programs component) by the national average price derived from question 6 (estimated dividing variables P12A06D by P12AO6A).'2 The result is a file with one variable (pounds) for each of the 99 food articles (99 entries). 33. Next, prices for each article were estimated for each Department/Area combination. The average prices were estimated using the Household questionnaire (estimated dividing variables P12A06D by P12A06A). If such information was not available in the ENCOVI 2000 household questionnaire, prices reported in the price questionnaire were used'3. With these prices, the purchase cost of the "national average consumption in pounds" in each Department/Area was estimated. The cost of the national average consumption in pound in each Department/Area was divided by the cost in Guatemala City to produce the Geographical Food Index. 34. To obtain a similar Index for the non-food items, the Price questionnaire, Section B was used. A weighted average for each Department/Area was computed. Items also present in the Guatemalan Consumer Price Index (CPI) were selected (questions 1-12 and 14-22, Variable PRECIO for ITEM = 41- 52 and 54-62). The same weight values used in the CPI were applied. 35. Similar to the Geographical Food Index, all the non-food Department/Area values were divided by the Guatemala City value to produce the Geographical Non-Food Index. 9 Valid types: food consumed outside the house or goods. Invalid types: food consumed in the house (already included in non-purchased food), or cash (people do not consume cash) 10 Also, only households with complete interviews were selected. 1" Described previously under "Food Consumption". 12 This national average price per item was used only for the estimation of the average national consumption quantities. Later, another "national average price" is estimated using a methodology based in the price questionnaire information. 13 A minimum of 15 prices per article for each department/area combination were required. If neither the ENCOVI 2000 household questionnaire, nor the Price questionnaire had enough data points, the average for the Region(8)/Area combination was used (with a minimum of 25 data points), or the Departmental average (with a minimum of 35 data points) or the Regional average (with a minimum of 45 data point). Annex 1. Page 7 36. Finally, to obtain the overall Geographical Index for each Department/Area, the weighted average of the Geographical Food and Non Food Indexes was computed. The weight used is the same proportions between food and non-food observed in the consumption aggregate: 40.5% for the food component, and 59.5% for the non food component. The resulting variable allows for standardization of any expense at the Guatemala City level (to be used as a divisor). 37. Using the Guatemala City average as a basis (Guatemala City = 1), the FACT.GEO variable was found to vary between 0.99 and 1.07. VALUE OF TOTAL CONSUMPTION PER CAPITA 38. For the final step to rank the population by welfare level (consumption) from the lowest to the highest, a share of the total consumption must be allocated to each household member. Per capita consumption is used in the Poverty Assessment, i.e., the total value of consumption of the household divided by the number of household members. There are several other ways of allocating household consumption to the different members, taking into account different requirements, economies of scale, and the presence of public services in the household. Per capita consumption was used due to its transparency, but other methods were used for sensitivity tests of the consumption aggregate. LEVELS OF TOTAL CONSUMPTION PER CAPITA: GUATEMALA 2000 39. The population was ranked from the lowest to the highest level according to total per capita annual consumption (welfare). Per capita consumption varies considerably in Panama (see Fig. Al. 1). On average, annual per-capita consumption is Q/.6,180. The richest ten percent of the population has an average consumption level of Q/.23,543 while the poorest ten percent has an annual average per-capita consumption of Q/. 1,287. Annex 1. Page 8 FIG. All: LEVELS OF CONSUMPTION: REPUBLIC OF GUATEMALA, 2000 % of Level of average annual population per-capita consumption (Q/.) 100 23,543 90 9,862 80 6,940 70 5,243 60 4,240 50 3,537 40 2,916 30 2,369 20 1,876 10 1,287 0 Lowest level of consumption Source: ENCOVI 2000 2000 Annex 2. Page I ANNEX 2 - MEASURING INCOME USING THE ENCOVI 2000 INTRODUCTION This note summarizes the components and the methodology used to construct the income aggregate using the ENCOVI 2000, Guatemala. The income aggregate measures the income obtained by a household in a year. The total household income can be divided between (i) income earned from labor activities; and (ii) income not related to labor activities. In addition, labor income can be further divided between wage and self-employed income, agricultural and non-agricultural, formal and informal. Non-labor income consists of income such as interest earned from savings, pensions and remittances. The above income components are calculated for each household member and transformed in annual income. The total household income is the sum of each of these components for each household member. All variables are previously "cleaned" and reviewed to remove "outliers", missing data and other data problems. All values are also adjusted to account for regional price differences. This notes continues as follows: the next section summarizes the components of the income aggregate. Then the labor income components are discussed in specific detail, followed by a discussion of the non-labor income components (directly using the information in the ENCOVI 2000 data). A list of all the variables used is then provided at the end. INCOME AGGREGATE COMPONENTS Household income can be divided between labor and non-labor activities. The components for each category are summarized as: Labor income * Wages from formal non-agricultural activities * Wages from informal non-agricultural activities * Wages from formal agricultural activities * Wages from informal agricultural activities * Self-employed income from formal non-agricultural activities * Self-employed income from informal non-agricultural activities * Self-employed income from agricultural activities Non-labor income * Rental of equipment and property * Interest and dividends * Remittances * Public assistance and donations * Private assistance and donations * Pensions and compensations * Other income (inheritance, scholarships, lottery winnings) Annex 2, Page 2 LABOR INCOME Labor income classifications Labor income is constructed using chapters X. (economic activities) and XVI (agricultural activities) of the ENCOVI 2000. The first issue to address is the classification of labor activities. Labor activities are divided between wages and self-employed income. Income is defined as wages if a person is an employee in the government or a private company (PlOB 14B < 5 or >6)1. Otherwise, labor income is classified as self-employed. Income is derived from agricultural activities if the person works in any of the agricultural occupational classifications (PIOB02 = 1 or 2 or 5) and as non-agricultural otherwise. Finally, income is classified as formal if a person works for the government, or in the private sector in a company with more than 5 workers, or in a farm that employs more than 5 people (PIOB14 = 1, PIOB12 > 2 and PIOB14 = 2 or 3 or PIOB14 > 4). Otherwise, it is classified as informal. Wages. For each person, the ENCOVI 2000 reports up to three jobs. The following components of annual wages are constructed for each of the three jobs: * Wages and salaries times months worked during the year; (PIOB04 * P1OB22, PlOCOS * PlOC16, PIOD05 * PIOD10); e Bono 14 (PIOB20B, P1OC14B, P1OD08B); * Payment in aguinaldo (P1OB27B); less o Payment to sociai security - IGSS (PlOB13 * PlOB04); * Income tax.2 The level of annual wage income is constructed by first summing the above components for each of the three jobs and then aggregating them at the household level. The income is divided between agricultural and non-agricultural as well as formal and informal based on the definitions above to give the following four wage income categories: o Wages and salaries from formal non-agricultural activities; o Wages and salaries from informal non-agricultural activities; * Wages and salaries from formal agricultural activities; * Wages and salaries from informal agricultural activities; Self-employed non-agricultural income. This refers to independent entrepreneurs that work in non-agricultural activities. For them, their income is the reported net income multiplied by the number of months that they received it for each of the three possible jobs (PIOB lSA * PlOB 15C, PlOBlSA * PIOB15C, PlOB1SA * PIOBl5C). The total income from this activity is constructed by summing income from same type jobs for every person in the household. Self-employed income from non-agricultural activities is divided between formal and informal: * Self-employed income from formal non-agricultural activities In parenthesis, the corresponding variable names and appropriate formulae to construct each variable is provided. 2 Income taxes in Guatemala are based on disposable income (DI). This is calculated as: Net income- Q 36,000 (exempt income)- less social security contributions. Given that, the tax rates are: (i) Dl less than Q 65,000: 15%; (ii) DI between Q 65,000 - Q 180,000: Q 9,750 + 20% of amount in excess of Q 65,000; and (iii) Dl between Q 180,000 - Q 295,000: Q 61,500 + 31 % of amount in excess of Q 295,000. Annex 2. Page 3 Income from independent agricultural activities is the sum of all income minus the costs related to that. In this context, this income represents the net income from agricultural activities. In particular, household self-employed agricultural income is derived from the following: * Revenue from the rent of owned land to others (PI 6A1 1) * Revenue from the sale of crops (P16B06) * Revenue from the sale of processed crop products (Pl6HO3CA through pl6HO3CG) * Revenue from the sale of forest products (P16104A through P16I04E) * Revenue from animal sales (P16JO8B and P16JI IB) * Revenue from the sale of animal products (P16L03CA through P16LO3CJ) * Consumption of own produced. output (constructed in the consumption aggregate) less * Cost for renting land from others (P16A22 if P16A21> I and P16A23A * Period (using P16A23B) if P16A21 = 1) * Cost for agricultural inputs. These are further divided in: / Crop and forest production related inputs (P16C02A through P16CO2E, P16DO2A through P16D02I) - Labor inputs (P16EO2A * P16E02B, P16EO3A * P16EO3B, P16E05, P16E06C) 1 Cost for technical assistance (P16F05) / Cost for inputs related to livestock activities (Pl6MO2A through P16MO2E) Depreciation of agricultural capital equipment.3 NON-LABOR INCOME Income from non-labor income is divided in the following categories: * Rental of equipment and property. This includes the income received from rental of properties, construction, equipment and goods4 (P1 3A02A) and the estimated rental value of owned housing (POlOC7 and constructed in the consumption aggregate). * Interest and dividends from savings accounts and stock holdings (P13A02B) * Remittances. This is support in cash by friends and family (PIOE09, P13BO3E if in cash) * Donations. These include donations and help in cash or in kind gifts received by the government, church, private organizations and friends and family (PlOE06, P13BO3A through P13B03E) and public program assistance (P04C04, P04C08, P04C12, P13A02H- 1). The support in cash by friends and family is reported as remittances. This income is divided between private and public. * Pensions and compensations. This includes child care allowances, orphan and widow pensions, retirement benefits, compensations for work or contract termination, life insurance. (PIOEOlB*12, P1OE02B*12, PIOE03B*12, P13AO2C, P13A02D, P13AO2E, P13AO2G). * Other income. This income includes inheritance, scholarships and lottery winnings. (P13A02F, P13AO2H through P13A02L). 3This represents the value of the annual use of durable goods used in agricultural activities. To estimate this value, the average age of each equipment is calculated (average of P16G04 by equipment type if P16GOI = 1). This is multiplied by 2 to obtain the expected lifetime of the equipment. Then, the reported age is subtracted from the expected lifetime to get the remaining lifetime of the equipment for each household3. Finally, the current value of the equipment (P16G07) is divided by the remaining lifetime to obtain the annual use of the equipment. 4Not related to agricultural activities. Annex 3. Page 1 ANNEX 3 - MEASURING POVERTY USING THE ENCOVI 2000 This annex uses the following method to classify individuals as extreme poor, poor, or non-poor: (i) individuals are ranked according to their level of welfare, as measured by total consumption (Annex 1); (ii) the value of the full poverty line and extreme poverty line is calculated; and (iii) individuals whose consumption levels fall below these lines are classified accordingly. (I) RANKING INDIVIDUALS Defining Welfare. Since welfare, or well-being, cannot be measured directly consumption was used as an -indirect measure of welfare. Consumption is used because it is not subject to the underestimation and biases of an income measure, and because it avoids the subjectivity associated with measures of basic needs and indicators of human development. Annex 1 provides details on the construction of total consumption as a measure of welfare. Individuals were ranked from the lowest to the highest level of annual per-capita consumptiorn (welfare). Figure A3.1 shows major differences in the current per-capita consumption in Guatemala. On average, annual per-capital consumption is Q.6,180. Consumption ranges from an average of Q/.23,543 for the richest ten percent of the population to an average of Q.1,287 for the poorest ten percent. FIG. A3.1: LEVELS OF CONSUMPTION: REPUBLIC OF GUATEMALA, 2000 % of Level of average annual population per-capita consumption (Q/.) 100 23,543 90 9,862 80 6,940 70 5,243 60 4,240 50 3,537 40 2,916 30 2,369 20 1,876 *10 1,287 0 Lowest level of consumption Source: ENCOVI 2000 2000 Annex 3. Page 2 (II) CONSTRUCTING POVERTY LINES Two poverty lines were constructed for this study: an extreme poverty line and a full poverty line. The Extreme Poverty Line. The extreme poverty line represents the yearly cost of the minimum daily caloric requirement recommended for Guatemala (2,172 on average, see Table A3.1), using the observed consumption basket of the entire- population. When the consumption level of any individual is below such value, he/she is unable to consume the minimum recommended calorie level. That is, even if the individual spends all his/her resources on food, he/she would still not be able to acquire the minimum level of recommended calories. The extreme poverty line was calculated as follows: a) Using the ranking based on total annual per-capita consumption, households with the lowest and highest consumption level were dropped (those in the lowest 2% and highest 2% of the population were not included). b) On the basis of the food consumption patterns of the households in the 3% - 98% range, the amount of calories supplied by each type of food' and the percentage of these calories in the total was calculated. For example, for this group of households, corn provides more calories than any other type of food (36.5 percent of the calories consumed). Next in importance are bread (sweet) and sugar, which supply 10.2 and 9.4 percent respectively of the total calories consumed (see Table A3.2 for consumption patterns of all products). c) The mninimum average calorie requirements of a Guatemalan were calculated using data from INCAP: 2,172 kcal/day (see Table A2.1). d) The amount of food required to satisfy the minimum calorie requirements were calculated, using the shares (consumption patterns) for each type of food for households within the 3-98% of consumption. The absolute amounts consumed by this group are adjusted to meet the amounts required to achieve the recommended calorie level (2,172) using their consumption shares. e) On the basis of these amounts, the cost of food required to satisfy the minimum calorie requirements was determined. f) An adjustment to account for "wasted food" was apply. According to nutritional experts in Guatemala only 90% of acquired food reach the mouth of the consumer.2 The cost of food to get 2,172 Kcal./day was divided by 0.9 This is the cost of the minimum calorie requirements, in other words, the value of the extreme poverty line. For Guatemala in 2000, the extreme poverty line was calculated as Q/.1,869 per- capita per year. Figure A3.2 below shows the method used for calculating the extreme poverty line. Using food calorie composition data from "Valor Nutritivo de los alimentos de Centroamerica." Instituto de Nutrici6n de Centro America y PanamA (INCAP) y la Organizaci6n Panamericana de la Salud (OPS). Ciudad de Guatemala, Guatemala, 1998. The rest is left in the plate, thrown away, dropped, etc. Ma. Teresa Mench6, Central America Nutritional Institute (INCAP). Annex 3. Pag:e 3 Population kge in Years Persons % Daily Kcal requirement Daily Kcal contribution 1year * I breast feeding 316,928 3 0 0 1 not breast feeding 69,570 1 738 4.5 1-2 743,855 7 1,200 78.4 3-4 714,964 6 1,400 87.9 5-6 687,707 6 1,675 101.2 Male 79 g496,022 4 2,000 87.1 10 11 309,554 3 2,200 59.8 1 2- 1 3 293,825 3 2,350 60.6 14-15 281,364 2 2,650 65.5 16- 17 268,573 2 3,000 70.8 18-64 2,604,682 23 3,100 709.2 65 & + 193,786 2 2,200 37.4 Fe.mnle j7-9 475,556 4 1,700 71.0 10-11 296,833 3 1,900 49.5 12-13 281,798 2 2,000 49.5 114-15 228,955 2 2,100 42.2 l 16-17 135,127 1 2,150 25.5 l 1 8-49 1,665,876 15 2,100 307.3 l50-64 368,327 3 2,100 67.9 65 & + 209,958 2 1,850 34.1 Preenant 2 14-15 8,137 0 2,385 4.9 l 16-17 42,176 1 2,435 15.0 1 8-49 374,835 3 2,385 69.5 Breast feedine 0 l 14-1 5 6,047 0 2,600 4.0 16-17 31,350 0 2,650 12.2 18-49 279,531 2 2,600 56.5 TOTAdL 11,385,336 100 2172 Minimum Calorie Requirements: Weighted Average 2172 Source: Instituto de Nutrici6n de Centro Am&rica y Panamd (INCAP) CEPALJCELADE -Divisi6n de Poblaci6n, Boletin Demografico No. 66, Julio 2000 Annex 3. Page 4 FIG. A3.2: CALCULATING THE EXTREME POVERTY LINE A. Ranking Individuals B. Calculating the Value of Minimum Caloric Requirements Highest level of total annual per-capita consumption Actual level of average calories for households in the lowest 3-98%. Total level in this group = 2,025/person. For example: 100% 36.5% of calories comes from corn 98% 10.2% of calories comes from bread (sweet) Actual food 9.4% of calories comes from sugar 70 consumption patterns 7.1 % of calories comes from tortillas 60 See patterns for all products in Table A2.2 50 The number of calories corresponds to a physical amount of food (in pounds) 40 Average recommended calorie level = 2,172 30 The quantity of each food item is adjusted to obtain a 20 basket that provides the 2,172 calories maintaining the consumption patterns of households in the 3-98% per- 10% capita consumption range. 3% C. Value of the Extreme Poverty Line Lowest level of total Actual The value is calculated by adding the quantity of items annual per-capita prices estimated in the last step using the prices actually faced consumption by households in the 3-98% group. Using THE ENCOVI 2000 data and allowing for a 10% "waste", the annual value of the extreme poverty line is Q/.1,682 per capita. Zpe = Q\ 1,682 The Full Poverty Line. Total consumption, even among the poorest, almost always includes the consumption of non-food goods and services. As such, the general poverty line includes an additional amount for the percentage of the non-food consumption. The share of non-food consumption is based on the observed consumption patterns of individuals whose food consumption is close to the extreme poverty line. The full poverty line equals the extreme poverty line plus an allowance for non-food consumption, as follows: a) Individuals with food consumption (CA) levels close to (+/-5%) the extreme poverty line (CA = Zp, = 1,869) were selected. These individuals barely meet their minimum calorie requirements. Annex 3, Page 5 b) Consumption coefficients were calculated for this group: that is, the share of total consumption allocated to food (in this case, 44.2%) and non-food products (55.8%). c) To obtain the full poverty line, the value of the extreme poverty line was divided by this share of food consumption (44.2%). Figure A3.3 below shows the method used to calculate the general poverty line. Annex 3. Page 6 uetzals r Consumed daily Kcal Kcal to Yearl cost of 2,172 get 2,172/ 1 ,000 Lb./year/ Household day/ without with 10% rTEM Kca/Lb. Lb. Kcal HH (5 persons) Person % person waste waste Corn (grain) 1,639 0.94 0.57 1,125.2 5,052.4 1,010.5 36.5% 791.8 166.06 184.51 read (sweet) 1,707 1.78 1.04 303.3 1,418.4 283.7 10.2% 222.3 84.39 93.77 ugar 1,743 1.81 1.04 272.9 1,303.2 260.6 9.4% 204.2 77.22 85.80 Tortillas 1,018 1.34 1.32 353.7 987.0 197.4 7.1% 154.7 74.31 82.56 Beans (frijoles) 1,530 2.59 1.69 156.6 656.5 131.3 4.7% 102.9 63.61 70.67 Bread (French) 1,353 2.12 1.57 165.2 612.3 122.5 4.4% 96.0 55.01 61.12 Oil 4,013 4.43 1.10 51.3 563.9 112.8 4.1% 88.4 35.57 39.53 ice 1,634 2.35 1.44 92.4 413.9 82.8 3.0% 64.9 34.00 37.78 Pasta 1,684 3.44 2.04 52.1 240.6 48.1 1.7% 37.7 28.13 31.25 Eggs 591 3.75 6.34 102.8 166.5 33.3 1.2% 26.1 60.42 67.14 Chicken or Hens 573 7.15 12.48 87.0 136.6 27.3 1.0% 21.4 97.49 108.33 Corn tamales 499 2.49 4.98 98.1 134.2 26.8 1.0% 21.0 38.25 42.50 eef Meat 970 13.78 14.21 49.4 131.3 26.3 0.9% 20.6 106.72 118.57 Potato 307 1.15 3.73 135.1 113.6 22.7 0.8% 17.8 24.26 26.96 Oatmeal 1,716 4.31 2.51 22.6 106.4 21.3 0.8% 16.7 15.29 16.99 Bread(slice) 1,212 1.34 1.11 29.9 99.4 19.9 0.7% 15.6 6.29 6.99 Powder Milk 2,252 13.56 6.02 14.4 89.0 17.8 0.6% 13.9 30.66 34.07 Flour (Incaparina) 1,689 3.87 2.29 19.2 88.9 17.8 0.6% 13.9 11.65 12.95 Milk Cream 1,235 6.48 5.25 24.1 81.4 16.3 0.6% 12.8 24.46 27.17 Margarine 3,264 5.61 1.72 8.8 78.9 15.8 0.6% 12.4 7.76 8.62 Cheese 1,230 9.74 7.91 22.7 76.6 15.3 0.6% 12.0 34.68 38.54 UilkCream 238 1.78 7.47 112.7 73.3 14.7 0.5% 11.5 31.36 34.84 andy 1 2,336 11.46 4.91 11.1 70.9 14.2 0.5% 11.1 19.91 22.13 Beef Meat with bone 515 7.34 14.26 49.2 69.4 13.9 0.5% 10.9 56.58 62.87 Flour (Com) 1,657 2.48 1.50 15.3 69.3 13.9 0.5% 10.9 5.93 6.59 lantains . 382 0.97 2.53 50.9 53.3 10.7 0.4% 8.3 7.72 8.58 Flour (Wheat) 1,653 2.35 1.42 11.1 50.2 10.0 0.4% 7.9 4.09 4.54 Panela o rapadura 1,616 2.36 1.46 10.5 46.6 9.3 0.3% 7.3 3.89 4.32 ananas 269 0.80 2.99 58.9 43.3 8.7 0.3% 6.8 7.40 8.22 Cakes 1,766 10.53 5.96 7.9 38.1 7.6 0.3% 6.0 13.00 14.44 Dry Seeds 2,626 5.56 2.12 5.1 36.8 7.4 0.3% 5.8 4.47 4.96 Corn Flakes 1,766 19.58 11.08 7.5 36.1 7.2 0.3% 5.7 22.86 25.40 Chocolate * 2,075 5.50 2.65 5.4 30.7 6.1 0.2% 4.8 4.66 5.17 Cookies 2,111 15.54 7.36 5.3 30.7 6.1 0.2% 4.8 12.93 14.36 Sausaes Pork) 2,996 19.87 6.63 3.7 30.7 6.1 0.2% 4.8 11.65 12.94 Sausages 1,004 12.01 11.96 10.9 29.9 6.0 0.2% 4.7 20.49 22.76 at (Pork) . 3,991 5.28 1.32 2.7 29.2 5.8 0.2% 4.6 2.21 2.46 omatoes 93 1.87 20.04 113.9 29.2 5.8 0.2% 4.6 33.42 37.14 Soups( 1,292 14.97 11.58 8.2 29.1 5.8 0.2% 4.6 19.31 21.45 Powder Milk (children) 2,302 17.75 7.71 4.2 26.6 5.3 0.2% 4.2 11.74 13.05 Pork Meat 978 11.73 12.00 9.6 25.8 5.2 0.2% 4.0 17.70 19.67 Oisquil 108 1.17 10.78 86.6 25.7 5.1 0.2% 4.0 15.87 17.63 Carrots 168 1.34 7.98 54.4 25.1 5.0 0.2% 3.9 11.47 12.75 Tostadas 1,874 9.32 4.97 4.6 23.7 4.7 0.2% 3.7 6.75 7.50 Cabbage 100 0.54 5.36 81.0 22.3 4.5 0.2% 3.5 6.83 7.59 Annex 3. Page 7 _ ! MIRP..my. . * - r - - Quetzals per Consumed daily Kcal Kcal to Yearly cost of 2,172 get 2,172/ 1,000 Lb./year/ Household day/ without & 10% TEM KcalVLb. Lb Kal HH (5 persons) Person % person waste waste Candy2 1,630 9.83 6.03 4.6 20.4 4.1 0.1% 3.2 7.04 7.82 Chicken or Hens inners 763 4.44 5.83 9.6 20.1 4.0 0.1% 3.2 6.71 7.46 Avocado 378 2.26 5.99 19.0 19.6 3.9 0.1% 3.1 6.73 7.48 Oranges and Tangerines 110 0.80 7.34 59.6 17.9 3.6 0.1% 2.8 7.51 8.34 Onions 143 2.38 16.65 43.5 17.0 3.4 0.1% 2.7 16.22 18.02 tce cream 549 4.91 8.95 10.9 16.3 3.3 0.1% 2.6 8.37 9.30 ther Atoll 275 2.49 9.06 20.7 15.6 3.1 0.1% 2.4 8.09 8.99 Pork Mean with bone 808 8.53 10.56 6.9 15.4 3.1 0.1% 2.4 9.29 10.33 Soda pop 202 3.27 16.21 27.1 15.0 3.0 0.1% 2.4 13.91 15.46 Ayote, chilacayote 70 0.40 5.68 71.9 13.8 2.8 0.1% 2.2 4.48 4.98 Paches 499 3.19 6.39 9.6 13.1 2.6 0.1% 2.1 4.80 5.34 om Dough (fresh) 779 2.56 3.28 5.9 12.6 2.5 0.1% 2.0 2.37 2.64 Spices 146 2.98 20.49 29.7 11.9 2.4 0.1% 1.9 13.91 15.45 Fish (fresh) 296 8.30 28.02 14.1 11.5 2.3 0.1% 1.8 18.37 20.41 Cattle inners 554 8.36 15.10 6.8 10.3 2.1 0.1% 1.6 8.88 9.87 om Atoll 173 2.33 13.51 20.7 9.8 2.0 0.1% 1.5 7.58 8.42 Honeyandmolasses 1,280 8.53 6.66 2.5 8.9 1.8 0.1% 1.4 3.41 3.78 Ketchup & tomato paste 427 8.38 19.64 6.5 7.6 1.5 0.1% 1.2 8.59 9.54 Fat (Vegetable) 3,954 4.87 1.23 0.7 7.3 1.5 0.1% 1.1 0.51 0.57 assava 374 2.18 5.84 6.9 7.1 1.4 0.1% 1.1 2.37 2.63 eer 186 7.21 38.74 13.9 7.1 1.4 0.1% 1.1 15.66 17.40 Garlic 572 6.24 10.91 4.4 7.0 1.4 0.1% 1.1 4.34 4.83 imes 67 1.39 20.67 36.4 6.7 1.3 0.0% 1.0 7.92 8.80 Alcohol 1,049 17.56 16.74 2.3 6.5 1.3 0.0% 1.0 6.20 6.88 eets 126 1.42 11.27 18.5 6.4 1.3 0.0% 1.0 4.11 4.57 ucumber 52 1.30 24.79 37.9 5.4 1.1 0.0% 0.9 7.72 8.58 unaandSardines 909 15.01 16.51 2.1 5.1 1.0 0.0% 0.8 4.83 5.37 Butter 2,742 11.73 4.28 0.6 4.8 1.0 0.0% 0.8 1.17 1.30 Mangos 121 2.86 23.63 14.1 4.7 0.9 0.0% 0.7 6.33 7.03 PineApple 139 1.09 7.84 11.8 4.5 0.9 0.0% 0.7 2.01 2.23 Papaya 109 1.15 10.52 14.9 4.4 0.9 0.0% 0.7 2.67 2.97 ell peppers 162 6.28 38.77 10.0 4.4 0.9 0.0% 0.7 9.84 10.93 DryFruit 1,103 6.92 6.27 1.1 3.2 0.6 0.0% 0.5 1.15 1.28 [uice (pre packed) 215 3.09 14.37 5.3 3.1 0.6 0.0% 0.5 2.54 2.82 armalade 1,117 14.78 13.24 1.0 2.9 0.6 0.0% 0.5 2.22 2.46 Lettuce 68 2.10 30.86 15.6 2.9 0.6 0.0% 0.5 5.14 5.71 Watermelon 49 0.90 18.49 15.2 2.0 0.4 0.0% 0.3 2.15 2.39 Yogurt 327 6.67 20.41 2.2 2.0 0.4 0.0% 0.3 2.28 2.54 Arveja 368 4.11 11.17 1.6 1.6 0.3 0.0% 0.3 1.04 1.16 Melons 81 1.04 12.83 7.1 1.6 0.3 0.0% 0.2 1.16 1.28 Condensed & Evaporated Milk 1,033 8.61 8.34 0.5 1.3 0.3 0.0% 0.2 0.64 0.71 Celery 73 4.87 67.04 1.4 0.3 0.1 0.0% 0.0 1.05 1.17 Mushrooms 54 9.98 183.27 0.5 0.1 0.0 0.0% 0.0 0.73 0.82 Salt 0.65 n/a 15.4 - - 0.0% - 1.56 1.73 Sum . 4,528 2,771 2,172 1,682 1,869 Source: ENCOVI 2000, Guatemala Annex 3. Page 8 FIG. A3.3: CALCULATING TIHE FULL POVERTY LINE Ranking Individuals The full poverty line includes the cost of the extreme poverty line, Highest level of total Zpe, plus an additional amount for non-food consumption. per-capita consumption (Quetzales) 23,543 CT = CA + CNA CT = Total consumption CA = Food consumption CNA= Non-food consumption What share of total consumption is allocated to non-food? To calculate the poverty line, actual consumption coefficients of the group 4,240 of individuals with afood consumption near (+/-5%) the extreme poverty line (Zpe) were used. Shares offood In this case, this group allocates 55.8% to non-food consumption and 3,537 and non-food 44.2% to food consumption. consumption These consumption shares were used to calculate the full poverty line: 2,916 / CA = (1-.5584)Zpg 2/369 ....................Zpg = 519/.4416 ....... . . . . CA = Zpe = 1,869 Zpe = Extreme Poverty Line = 1,869 1,876 Zpg = General Poverty Line = 4,233 1,287 Lowest level of total consumption Annex 3, Page 9 (RII) POVERTY MEASURES The poverty indices used in this study are three special cases of additively separable measures developed by Foster, Greer and Thorbecke (FGT, 1984). The general poverty measure is: na n_ [Eq. 1] where: yj = estimated consumption of the ith person in a population of size n Z = the poverty line q = number of persons whose y1 is below poverty line Z and; a = is a non-negative parameter that reflects the measure's aversion to poverty Head Count Index. The first case is that where a = 0. This is the Head Count measure (H) and, as can be seen from Eq. 1, it is simply q/n or the proportion of the population below the poverty line. In short, the Head Count Index provides information on the incidence of poverty. It says nothing about the depth or severity of poverty and treats as equal any two populations where the proportion of the population living in poverty is the same. Poverty Gap. To determine the depth of poverty, a second version of the FGT poverty measure, called the Poverty Gap index (PG), is used. This index is the case where a = 1 (in Eq. 1). The index is the aggregate poverty deficit of the poor relative to the poverty line. FGT P2 (Severity). The third case of the poverty measure is that where a = 2. This measure, often called the Foster-Greer-Thorbecke P2 measure (FGT P2), identifies the severity of poverty and demonstrates the relative inequalities among the poor. It is distributionally sensitive and, essentially, weights the average poverty gaps by the population at each level. Ravaillion (19923) presents a good example to illustrate the differences between the FGT P2 index and the previous two. For example, it is possible for two populations to have the same head count and poverty gap indices but have very different distributions of levels of poverty. Ravaillion (1992) presents the example of two populations A and B where A is made up of four individuals with consumption levels 1, 2, 3, 4 and B is made up of four individuals with consumption levels 2, 2, 2, 4. If the poverty line equals three, the head count for both populations is 75 percent, the poverty gap measure is 25 percent. But the FGT P2 measure is 14 in population A and 8 in population B, thus demonstrating that the poorest person in Population A has half the expenditures of the poorest person in Population B. 3 Ravallion, Martin. "Poverty Comparisons: A Guide to Concepts and Methods," World Bank LSMS Working Paper No. 88, 1992. Annex 3, Page 10 (IV) SENSITIIY ANALYSIS In order to assess the robustness and sensitivity of the poverty lines calculated above, poverty rates are recalculated by varying the poverty line so as to evaluate how sensitive the results are to the specific choice of poverty line. Tables A3.4 and A3.5 present the results of the exercise for general and extreme poverty lines, respectively. For example, for general poverty, a 5 percent increase raises the poverty line from Q\ 4,319 (the baseline) to Q\4,534 (table A3.4). Based on the latter, the new national poverty rate would be 58.9% (compared with the baseline 56.1%). This represents a 2.8 percentage point difference with the baseline poverty rate, or a 5 percent difference. In addition, there is an implied elasticity of poverty line changes to changes in the poverty rate of 1. In fact the last column listing these implied elasticities for different levels of poverty line increases or decreases, shows that there are no extreme changes in poverty rates due to changes in poverty lines. The above also holds for the extreme poverty calculations. Table A3.4 - Poverty line sensitivity: general poverty Une New Poverty % Change in % points Change as % of original poverty rate Change _a. Urban Rural Indig. Non In Nation Urban Rural Indig. Non In Nafion Urban Rural Indig. Non In Nation Nation Original 4,319 27.1% 74.5% 76.0% 41.4% 56.1% 27.1% 74.5% 76.0% 41.4% 56.1% 27.1% 74.5% 76.0% 41.4% 56.1% eatct -10% 3,887 20.9% 68.5% 70.9% 34.6% 50.0% -6.2% -6.0% -5.1% -6.8% -6.1% -22.9% -8.1% -6.7% -16.4% -10.9% 1.09 -9% 3,930 22.6% 68.9% 71.2% 36.0% 50.9% -4.5% -5.6% -4.8% -5.4% -5.2Y% -16.6% -7.5% -6.3% -13.0% -9.3% 1.03 -8% 3,973 22.9% 69.7% 71.7% 36.8% 51.6% -4.2% -4.8% -4.3% -4.6% -4.5% -15.5% -6.4% -5.7% -1 1.1% -8.0% 1.00 -7% 4,016 23.6% 70.3% 72.3% 37.4% 52.2% -3.5% -4.2% -3.7% -4.0% -3.9% -12.9% -5.6% -4.9% -9.7% -7.0% 0.99 -6% 4,059 24.0% 70.9% 72.9% 37.8% 52.7% -3.1% -3.6% -3.1% -3.6% -3.4% -11.4% -4.8% -4.1% -8.7% -6.1% 1.01 -5% 4,103 24.3% 71.3% 73.3% 38.1% 53.1% -2.8% -3.2% -2.7% -3.3% -3.0% -10.3% -4.3% -3.6% -8.0% -5.3Yo 1.07 -4% 4,146 25.0% 72.0% 74.2% 38.7% 53.8% -2.1% -2.5% -1.8% -2.7% -2.3% -7.7% -3.4% -2.4% -6.5% -4.1% 1.02 -3% 4,189 25.7% 72.6% 74.7% 39.4% 54.4% -1.4% -1.9% -1.3% -2.0% -1.7% -5.2% -2.6% -1.7% -4.8% -3.0% 1.01 -2% 4,232 26.0% 73.1% 75.0% 39.9% 54.8% -1.1% -1.4% -1.0% -1.5% -1.3% -4.1% -1.9% -1.3% -3.6% -2.3Yo 1.16 -1% 4,275 26.6% 74.0% 75.4% 41.0% 55.6% -0.5% -0.5% -0.6% -0.4% -0.5% -1.8% -0.7% -0.8% -1.0% -0.9Yo 0.89 0% 4,319 27.1% 74.5% 76.0% 41.4% 56.1%1 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 1% 4,362 27.3% 75.3% 76.4% 42.1% 56.6% 0.2% 0.8% 0.4% 0.7% 0.5% 0.7% 1.1% 0.5% 1.7% 0.9% 0.89 2% 4,405 28.1% 75.9% 76.9% 42.9% 57.3% 1.0% 1.4% 0.9% 1.5% 1.2% 3.7% 1.9% 1.2% 3.6% 2.1% 1.07 3% 4,448 28.3% 76.5% 77.3% 43.5% 57.8% 1.2% 2.0% 1.3% 2.1% 1.7% 4.4% 2.7% 1.7% 5.1% 3.0Yo 1.01 4% 4,491 29.1% 76.8% 77.9% 43.8% 58.3% 2.0% 2.3% 1.9% 2.4% 2.2% 7.4% 3.1% 2.5% 5.8% 3.9% 0.98 5% 4,534 29.3% 77.6% 78.6% 44.3% 58.9% 2.2% 3.1% 2.6% 2.9% 2.8% 8.1% 4.2Yo 3.4% 7.0% 5.0% 1.00 6% 4,578 29.6% 77.9% 78.9% 44.6% 59.2% 2.5% 3.4% 2.9% 3.2% 3.1% 9.2% 4.6% 3.8% 7.7% 5.5% 0.92 7% 4,621 29.8% 78.3% 79.5% 44.8% 59.5% 2.7% 3.8% 3.5% 3.4% 3.4% 10.0% 5.1% 4.6% 8.2% 6.1% 0.87 8% 4,664 30.1% 79.0% 79.9% 45.4% 60.1% 3.0% 4.5% 3.9% 4.0% 4.0% 11.1% 6.0% 5.1% 9.7% 7.1% 0.89 9% 4,707 30.6% 79.4% 80.1% 45.9% 60.5% 3.5% 4.9Y% 4.1% 4.5% 4.4% 12.9% 6.6% 5.4% 10.9% 7.8% 0.87 10%/ 4,750 31.1% 79.9% 80.9% 46.2% 61.00% 4.0% 5.4% 4.9% 4.8% 4.9% 14.8% 7.2% 6.4% 11.6% 8.7% 0.87 Annex 3. Page 11 Table A3.5 - Poverty line sensitivity: extreme poverty Line New Poverty % Change in; % points Change as % of original poverty rate Change Q\. Urban Rural I Indig. Non In Nation Urban Rural Indig. [Non In Nation Urban l ,Rural Indig. I Non In Nation Nation Oga 1,912] 28% 23.8%° 26.4% 7.7%/61 15.7% 2.8%1 23.8%1 26.4/1 7.7/1 15.7%1 2.8%1 23.8%1 26.4%1 7.7% 15.7% elasticity -10% 1,721 1.8% 18.2% 20.6% 5.4% 11.8| -1.0 -5.6% -5.8% -2.3% -3.9% -35.7% -23.5% -22.0% -29.9% -24.8% 2.48 -9% 1,740 1.9% 18.9% 21.4% 5.7% 12.3% -0.9% -4.9% -5.0% -2.0/ -3.4% -32.1% -20.6% -18.9% -26.0% -21.7% 2.41 -8% 1,759 2.1% 19.6%/ 21.9%/ 6.1% 12.8% -0.7% -4.2% -4.5/ -1.6/ -2.9% -25.0% -17.6% -17.0% -20.8% -18.5% 2.31 -7% 1,778 2.1% 19.9% 22.1% 6.3°h 13.0% -0.7% -3.9% -4.3% -1.4% -2.7% -25.0% -16.4% -16.3% -18.2% -17.2% 2.46 -6% 1,797 2.2% 20.4% 22.7% 6.5% 13.4% -0.6% -3.4% -3.7% '-1:2% -2.3% -21.4% -14.3% -14.0% -15.6% -14.6% 2.44 -5% 1,816 2.3% 20.9% 23.3% 6.6% 13.7% 4.5% -2.9°b -3.1% -1.1% -2.0% -17.9% -12.2% -11.7% -14.3% -12.7% 2.55 -4% 1,835 2.4% 21.7% 24.3% 6.8% 14. -0.4% -2.1% -2.1/ 4.9% -1.5% -14.3% -8.8% -8.0% -11.7%b -9.6% 2.39 -3% 1,854 2.6% 22.4% 25.3% 7.0% 14. | .2°b -1.4% 11% 07°b -1.0% -7. 1% 5.9% 4.2% -9.1% -6.4% 2.12 -2% 1,873 2.7%b 22.9% 25.5% 7.3% 15.0% -0.1% -0.9% -0.9% -0.4% -0.7% -3.6°b -3.8% -3.4% -5.2% -4.5% 2.23 -1% 1,893 2.8% 23.3°b 26.1% 7.4%/ 15.3% 0.0% 4.5% -0.3% -0.3%/ 04% 0.00 -2.1% -1.1% -3.9% -2.5% 2.55 0% 1,912 2.8% 23.8% 26.4% 7.7% 15.7%/ 0.0% 0.0% 0.0% 0.0% , 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 1% 1,931 2.9% 24.4% 27.1% 7.9% 16.1% 0.1% 0.6% 0.7%o 0.2% 0.4% 3.6% 2.5% 2.7% 2.6% 2.5% 2.55 2% 1,950 3.0%b 25.0% 27.7% 8.20/6 16.5% 0.2°b 1.2% 1.3% 0.5% 0.8% 7.1% 5.0% 4.9% 6.5% 5.1% 2.55 3% 1,969 3.2% 25.6% 28.6% 8.3% 16.9% 0.4% 1 ;8°b 2.2Y% 0.6% 1.2% 14.3% 7.6% 8.3% 7.8%b 7.6% 2.55 4% 1,988 3.3% 26.0% 29.1% 8.4% 17.2% 0.5% 2.2%b 2.7% 0.7% 1.5% 17.9% 9.2% 10.2% 9.1% 9.6% 2.39 5% 2,007 3.5% 26.5% 29.7% 8.6% 17.6% 0.7% 2.7%/ 3.3% 0.9% 1.9% 25.0% 11.3% 12.5% 11.7% 12.1% 2.42 6% 2,026 3.6% 27.3% 30.4% 9.0% 18.1% 0.8% 3.5% 4.0% 1.3% 2.4% 28.6% 14.7% 15.2% 16.9% 15.3% 2.55 7% 2,046 3.6% 27.6% 30.8% 9.1% 18.3% :0.8% 3.8% 4.4% 1.4% 2.6% 28.6% 16.0% 16.7% 18.2% 16.6% 2.37 8% 2,065 3.7% 28.4% 31.6% 9.4% 18.8% 0.9% 4.6% 5.2% 1.7% 3.1% 32.1% 19.3% 19.7% 22.1% 19.7% 2.47 9% 2,084 4.0% 29.0% 32.4% 9.6% 19.3% 1.2% 5.2% 6.0% 1.9% 3.6% 42.9% 21.8%1 22.7% 24.7% 22.9% 2.55 10% 2,103 4.1% 29.3% 32.8°b 9.7% 19.5% 1.3% 5.5% 6.4% 2.0% ;.3.8% 46.4% 23.1 24.2% 26.0% 24.2% 2.42 Finally, Tables A3.6-A3.8 evaluate the impact of the different ways to compute the consumption aggregate. Specifically, they compare how individuals' classification would change by using per capita consumption as opposed to per adult equivalent consumption. The adult equivalent base used in this exercise was the average for Guatemala (2,172 Kcal/day). Since the extreme poverty line is the cost to buy 2,172 Kcallday, the same line was used for the annual consumption per adult equivalent and the per capita one. In summary, using consumption per capita or per adult equivalent as the unit to classify welfare does not seem to significantly change the results. For example, in classifying individuals in poverty groups (extreme poor, non-extreme poor, non-poor), 91% of the population is the same irrespective of the unit of analysis (as shown by the diagonal cells). Similarly, for quintile classification 83% of the population is the same using either unit of analysis. Table A3.6 - Poverty sensitivity to er capita/per adult equivalent, by poverty Poverty (per capita[ Extreme Non extreme Non Poor Poverty (adult equivalent) Extreme 13.0% 0.9% QQO% Non Extreme 2.7% . 36.90% 2.40,o Non Poor 0.0%1 2.7%.. 41.4°%o No change 91.3% change in one -8.7% changeintwo 0.0% Annex 3. Page 12 Table A3.7 - Poverty sensitivit to per capita/per adult equivalent, by qu ntiles O Quintile (p r capita) Quintile (adult equivalent) 1 2 3 4 5 1 i8.1,° 1.9% 0.0% 0.0% 0.0% 2 1.9% ).-555% 2.6% 0.0% 0.0% 3 0.0% 2.6% ,15.1°% 2.3% 0.0%° 4 0.0% 0.0% 2.3%° 16.1%°8/ 1.6%/ 5 0.0% 0.0%° 0.0% 1.5% 18.5% No change 83.3% change in one 16.7% change in two or more 0.0% Table A3.8 - Pov rty sensitivity to per capita/per adult equivalent, by deciles Deciles (per capita Deciles (adult equivalent) 1 2 3 4 5 6 7 8 9 10 1 ' K 8.9.0.?/iol 1.10% 0.00% 0% 0% 0% 0% o0% 0% 0% 2 1.10% .2z:0' 1.80% 0.10% 0% 0% 0% 0% 0% 0% 3 0% 1.80% .6!10.k% 2.00% 0.10% 0% 0% 0% 0% 0% 4 0% 0.00% 2.10% 5-40% 2.30% 0.20%1 0% 0%° 0% 0% 5 0% 0% 0.00% 2.40% 5.20% 2.10% 0.10% 0% 0% 0% 6 0% 0% 0.00% 0.10% 2.20%°k5.5,6 2.10°% 0.00% 0% 0%1 7 0% 0% 0% 0.00% 0.10% 2.20%.! 6.00°c 1.70% 0.00% 0% 8 0% 0% 0% 0% 0% 0.00% 1.80% .6.70i/% 1.50% 0% 9 0% 0% 0% 0% 0% 0% 0% 1.50% VT70J 0.70% 10 0% 0% 0%1 0% 0%° 0%1 0% 0.00% 0.70% [9430% No change 67.8% change in one 31.1% change in two 0.7% change in three or more 0.0% Annex 4. Page 1 ANNEX 4 - STATISTICAL APPENDIX List of Tables A4.1 - Select Poverty and Social Indicators at a Glance A4.2 - Poverty Indicators using Consumption as Welfare Indicator, for Nation A4.3 - Poverty Indicators using Consumption as Welfare Indicator, for Urban A4.4 - Poverty Indicators using Consumption as Welfare Indicator, for Rural A4.5 - Poverty Indicators using Consumption as Welfare Indicator, for Indigenous A4.6 - Poverty Indicators using Consumption as Welfare Indicator, for Non-Indigenous A4.7 - Poverty Indicators by Welfare Measure, Central America Comparisons A4.8 - Income Inequality and Distribution A4.9 - Consumption Inequality and Distribution A4.1O - Consumption Quintiles: levels and distribution A4.1 1 - Income Quintiles: levels and distribution A4.12 - Income Sources by Income Quintiles A4.13 - Income Sources by Consumption Quintiles A4.14 - Income Sources by Poverty Classification A4.15 - Consumption Patterns: National A4.16 - Consumption Patterns: Urban A4.17 - Consumption Patterns: Rural A4.18 - Consumption Patterns: Indigenous A4.19 - Consumption Patterns: Non-Indigenous A4.20 - Perceptions of Welfare Changes During Last 5 years: Household Questionnaire A4.21 - Perceptions of Welfare Changes During Last 5 years: Community Questionnaire A4.22 - Perceptions of Causes of Welfare Changes, Poverty and Community Problems: Household Questionnaire A4.23 - Perceptions of Causes of Welfare Changes: Community Questionnaire A4.24 - Perceptions of Causes of Welfare Changes: Household Questionnaire, by Region A4.25 - Perceptions of Causes of Poverty Changes: Household Questionnaire, by Region A4.26 - Perceptions of Causes of Community Problems: Household Questionnaire, by Region A4.27 - Perceptions of Causes of Welfare Changes, Poverty and Community Problems: Household Questionnaire, by Rural and Urban Areas A4.28 - Perceptions of Causes of Welfare Changes, Poverty and Community Problems: Household Questionnaire, by Gender of Household Head A4.29 - Perceptions of Causes of Welfare Changes, Poverty and Community Problems: Household Questionnaire, by Poverty Level A4.30 - Perceptions of Causes of Welfare Changes: Household Questionnaire, by Consumption Quintile A4.31 - Perceptions of Causes of Poverty: Household Questionnaire, by Consumption Quintile A4.32 - Perceptions of Community Problems: Household Questionnaire, by Consumption Quintile A4.33 - Perceptions of Causes of Welfare Changes: Household Questionnaire, by Ethnicity A4.34 - Perceptions of Causes of Poverty: Household Questionnaire, by Ethnicity A4.35 - Perceptions of Community Problems: Household Questionnaire, by Ethnicity A4.36 - Perceptions of Causes of Confrontation in the Community: Household Questionnaire A4.37 - Perceptions of Causes of Confrontation in the Community: Community Questionnaire A4.38 - Perceptions of Exclusion A4.39 - Perceptions of Justice Performance Annex 4, Page 2 Table A4.1 - Guatemala: Select Poverty and Social Indicators at a Glance 1989 1994 199t 2000 seerrcoan Population (million) 8.5 9.7 10.8 11.40WDI Population growth rate (%) 2.5% 2.7 2.7 2.7 WDo Infant Mortality (per 1000 lIve bIrth) 72.7 51.1 45.1 38.8 DHS (wreretkno sue 40 fo r 99);: ca 20OW V Under 5 Mortality (per 1000 live births) na 68 59 49.4 DHS: ar..tr 20c0 . WDI Urban Poputaton (% of the total) 38 38.5 392 38.8 WDa: 2erc sW= Er4:Vr Mov Ufe expectancy at birth (yeara) 59.7 83.8 842 65.2 Male 57.3 60.8 61.4 624WDI l 971r101d r94 19t7 Female 62.2 66.5 672 68.2 NO) Total fertility rate (bhirst per woman) 5.4 5.1 5 4.6 WDI ff9965 tged of Iff4: 1997 rfte d 0199) _1989 1994 1998 2000 Adufit literacy rate (%) 39.7 35.8 32.8 31.1 WD91: ePr M20M * ENODVf 2990 Primary school enrollment. gross (%) NA 85.8 101.9 99 WDtn W.W 2000. EN)Vf 2990 Male NA 91.5 107.5 103 wot .ow. tNCoo t s2too Female NA 79.9 98 95 WDt 7C000 . * ENCVI 290o Primary scrhool enroilment. net (%) NA NA 82.7 79 Wolt wpf 2o90. -NCuOvt 20 Male NA NA 85.1 81 WD0: W09912000 99ENXOVf 290 Female NA NA 802 76 wDI: ewow 2owo . ENcoov 21Ow Illteracy rate, young aduit total (% of people 15-24) 27.3 24 21.8 20.7 W0I Rato of Illiterate females to males (% ages 15-24) NA NA NA 1.768WDt: aral 2COw.eNsovttor o Public spending on educaion (% of GDP) 1.8 1.5 2.1 2.5 WDI: emel I. 2Mst. 1n1y f Rb-esLOF 1987 1995 1998/99 20D0 Child ManutAritron Height for age (stunting/chronIc maenutribon) 57.8 49.7 48.4 44.2 DHS aW,d EN05I0 2o0o Weight for age (underweight) 33.2 26.8 24.2 22.3 DHS ew ENCOVi 2cow Weight for height (wasting/acute malnutrition) 1.3 3.3 2.5 2.8 DHS ftwd EN0ov1 290t Immunization, OPT (% of children under 12 monftf) 18 77 78 NA WD1 ImmunIzation, measles (% of children under 12 months) 24 83 81 NA wa0 Puti c health expendaure (% of GDP) NA 0.7 0.9 1.1 sAFtAnusry 01 Fhrun, 1989 1995 1998t99 2000 Living on less than $1 per day (PPP) (% of people) 20 NA NA 16 198 r Enas Sode Ecoronr n 20Do. ENr0Vot UNDPa Urban ((leng on less than $1 day (PPP)) 9 NA NA 5 1aa. Enre S.odE u oEotSCO.ENCOVt UriNDP Rural (ling on less than $1 day (PPP)) 26 NA NA 22 r509Encestsodseocoroe ow.200ENcovt:UNOP. Extreme Poverty (% of population below national extmme poverty line) NA NA NA 15.7 200 = EN0O9IA:09 wEnW (l199 -mnxurssn a5 WrrWo Urban (% below urban extreme poverty Une) NA NA NA 2.8 2000. ENO9VI0 wa anl (at luIng erurpflnn ae Indbo Rura (% below nmral extreme poverty line) NA NA NA 23.8 20o0. ENCOA: We EnIc (uuing nernron en hdhuo Poverty (% of populadon below national poverty line) 62 NA NA 68.2 Urban (% below urban poverty line) NA NA NA 27.1 150091tlann .9 E oo Eref l Rural (% below rura poverty llne) NA NA NA 74. 5NOVI: WB - yb (-hg rFonw m a to GINI Index Using Cortsumption NA NA NA 48 ENtOVI 2090 Using Income NA NA NA 57 ENCOVI 2000 Percentage share of inoome: higfwst 20% NA NA NA 62 ENOOVI 2000 Percentage shame of Income: lowest 20% NA NA NA 3 ENCOVI sow Percentage share of consumption: highest 20% NA NA NA 5 ENmOVI 209D Percentage share of consumption: lowest 20% NA NA NA 53 ENCOovi 2t09 Public expenditure on social secudty and welfam (% of GDP) NA NA NA 1.8% W0 eSurwwes: P enkre r 1 1989 1994 1998 2D00 Sudace area (sq. km) 108890 108890 108890 1088900W90 Population density (people per sq. km of and area) 78.8 89.8 99.6 105 WDI GDP per unit of energy use (PPP $ per kg of oil equivatent) 5.8 6.5 6.5 NA WDi C02 emissons (klt 4234.5 6863.4 9889.3 NA WDI Access to an improved water source (% of population) NA NA NA 69 ENCOVI 200D (GUAPA) Urban (% of urban populaton) NA NA NA 88 EN0V0 2ow (GUIAPA) Rura (% of nire population) NA NA NA 540EN0OV 2001 (GUAPA) Access to Improved sanitation facilites (% of population) NA NA NA 87 EN00V1 2009 I(GULAPA Urban (% of urban population) NA NA NA 97 ENCOVi 20o (GUAPAi Rural (% of rural population) NA NA NA 79 ENCOvi soo (GUAPA) Access to electilty, urban (% of urban poputation) NA NA NA 95 E400Vi 20Clt (GUAPA) Access to electrcity, rural (% of rumi populaton) NA NA NA 56 ENCOVi 20ow (GUAPAi Telephone mainlines (per 1,000 people) 18.1 25.1 47.8 67 ONCOvI 2so (GUAPA) Mobile phones (per 1,0DO people) 0 1.1 10.3 612 09i Penona coamputers (per 1,000 people) NA 2 8.3 11.A WDI Annex 4. Page 3 Table A4.2 - Population, contribution to poverty (contr%), percentage of poor (PO), poverty depth index (P1) & poverty severity index (P2), and cost to elimninate poverty (Valor de la brecha): National - # de personas y su contribucion P 0, P 1, & P 2 Valor do la brecha de pobreza Todos Pobre extremo Todos pobres No Pobre Pob. General Pob. Extrema P. Gene P. Extrema Personas % Personas Conlr% Personas Cow Personas Contr% P O P P2 P0 PI P2 01 000,P00 % 0\. 000,000 % NACIONAL _ _ TOTAL 11,385,441 100.0 1,786,682.0 100.0 6,397,903 100.0 4,987,538 100.0 56.2 22.6 11.7 15.7 3.7 1.3 111,22.0 100.0%l 807.5 100.0Ye POBREZA- _ - _ __ _ Extrema 1,786,682 15.7 1,786,682.0 100.0 1,786,682 27.9 100.0 66.2 44.3 100.0 23.6 8.1 5.107.1 45.9% 806.8 99.9Yo Pobre (todos) 6,397,903 56.2 1,786,682.0 100.0 6,397,903 100.0 . . 100.0 40.3 20.8 27.9 6.6 2.3 11,121.0 100.0% 807.2 100.0% No pobre 4,987.538 43.8 .- - 4,987,538 100.0 . QUINTIL - __I i |- 1 2,277,561 20.0 1,786,682.0 100.0 2,277,561 35.6 - 100.0 63.4 40.9 78.5 18.5 6.4 6,237.9 56.1% 806.8 99.9% 2 2,277,056 20.0 . . 2,277,056 35.6 . 100.0 38.9 15.7 , - , 3,825.3 344% - 0 .0Y 3 2,276,907 20.0 . . 1,843,286 28.8 433,621 8.7 81.0 10.8 1.9 , - * 1,057.0 9.5%- 0.0%. 4 2,276,239 20.0 . . . 2,276,239 45.6 . . . . 0.0% - 0.0Yo 5 2.277,678 20.0 . . . 2,277,678 45.7 _.- - - 0. - 0.0Y AREA Urbana 4,397,854 38.6 123,583.0 6.9 1,192,551 18.6 3,205,303 64.3 27.1 7.8 3.3 2.8 0.6 0.2 1,471.9 13.2% 50.4 6.2Yo Rural 6,987,587 61.4 1,663,099.0 93.1 5,205,352 81.4 1,782,235 35.7 74.5 32.0 17.0 23.8 5.7 2.0 9,647.4 86.7%| 756.1 93.6%/ REGION Metropolitana 2,465,957 21.7 15,524.0 0.9 443,704 6.9 2,022,253 40.6 18.0 3.4 1.1 0.6 0.2 0.1 366.3 3.3% 9.0 1.1% Norte 919,834 8.1 359,308.0 20.1 772,610 12.1 147,224 3.0 84.0 42.3 24.6 39.1 10.1 3.8 1,678.3 15.1% 177.6 22.0Yo Nororiente 932,583 8.2 83,313.0 4.7 483,087 7.6 449,496 9.0 51.8 18.1 8.2 8.9 1.3 0.3 730.2 6.6%| 22.3 2.8% Surorente 998,505 8.8 200,942.0 11.3 684,509 10.7 313,996 6.3 68.6 27.6 14.2 20.1 4.6 1.4 1,191.4 10.7% 67.4 1 0.8%o Central 1,216,330 10.7 106,338.0 6.0 629,328 9.8 587,002 11.8 51.7 17.9 8.2 8.7 1.5 0.4 940.3 8.5% 34.9 4.3Yo Suroccidente 3,013,789 26.5 511,249.0 28.6 1,927,904 30.1 1,085,885 21.8 64.0 25.7 13.2 17.0 4.1 1,4 3,341.0 30.0% 236.2 29.3% Noroccidente 1,466,733 12.9 462,130.0 25.9 1,204,094 18.8 262,639 5.3 82.1 38.5 21.3 31.5 7.7 2.6 2,436.1 21.9% 215.1 26.6% Peten 371,710 3.3 47,878.0 2.7 252,667 4.0 119.043 2.4 68.0 27.2 13.3 12.9 3.4 1.3 1 436.5 3.9% 23.9 3.0Y ETNIA (2gr.) lndigena 4,844,032 42.6 1.281,674.0 71.7 3,687,600 57.6 1.156,432 23.2 76.1 34.2 18.6 26.5 6.6 2.4 7,154.4 64.3% 614.0 76.0%/. No indfgena 6,541,409 57.5 505,008.0 28.3 2,710,303 42.4 3,831,106 76.8 41.4 14.0 6.5 7.7 1.5 0.5 3,966.2 35.7% 192.6 23.8% ETNIA (6 gr.) K'iche 1,073,324 9.4 204,493.0 11.5 691,009 10.8 382,315 7.7 64.4 26.2 13.6 19.1 4.3 1.5 1,214.9 10.9% 88.0 10.9% Q'eqchi 736,163 6.5 279,377.0 15.6 614,315 9.6 121,848 2.4 83.5 42.0 24.5 38.0 10.3 4.1 1,334.9 12.0% 144.8 17.9% Kaqchiquel 1,011,802 8.9 137,603.0 7.7 633,523 9.9 378,279 7.6 62.6 24.2 11.9 13.6 2.9 0.9 1,0570 9.5% 55.5 6.9% Mam 940,865 8.3 321,545.0 18.0 844,308 13.2 96,557 1.9 89.7 43.0 24.2 34.2 9.7 3.6 1,745.5S 15.7% 174.1 21.6% Otros indig 1,081,878 9.5 338,656.0 19.0 904,445 14.1 177,433 3.6 83.6 38.6 21.1 31.3 7.3 2.4 1,802.1 16.2% 151.6 18.8% No indigena 6,541,409 57.5 505,008.0 28.3 2,710,303 42.4 3,831,106 76.8 41.4 14.0 6.5 7.7 1 . 0.5 O 3,966.2 35.7% 192.6 1 23.8%/ SEXO JEFE Masculino 9,716,582 85.3 1,622,618.0 90.8 5,599,652 87.5 4,116,930 82.5 57.6 23.5 12.3 16.7 3.9 1.4 9,873.6 88.8% 730.0 90.4% Femenino 1,668,859 14.7 164,064.0 9.2 798,251 12.5 870,608 17.5 47.8 17.3 8.4 9.81 2.4 0.8| 1,246.1 t11.2 76.2 9.4% Annex 4. Page 4 Table A4.3 - Population, contribution to poverty (con tr%), percentage of poor (PO), poverty depth index (PI) & poverty severity index (P2), and cost to elimiinate poverty (Valor de la brecha) :Urban _____________# de personas y su contribuci6n jPO0, P 1, & P 2 Valor ($) do la brecha do pobrzez Todo I Pobre extremo Todos pobres No Pobre Pob. General I Pob. Extrema P. General P. Extrema TOTAL P~~~~er-sona-s Personas Contr%/ Personas Contr% PFe-rsnas Conr%-i PI P2 PO 0 \ 1 P2 OOO % O\. 000.000 Urbana____ = I_ --i ___ TOTAL ~4,397,854 100.0 123,583.0 1000! 1,192,551 100.0 3.205,303- 100.0 2711 7.8 331 2.81 0.61 0.2 1,471.9 1000% 50.4 100.0% POBREZA Extrema 123,583 2.8 123,583.0 100.0 123.583 10.4 -100.0 65.2 43.0 100.0 21,4 7.1 347.9 23.6% 50.4 100.0%/ Pobre (todos) 1,192,551 27.1 123,583.0 100.0 1,192,551 100.0 - - 100.0 28.6 12.0 10.4 2.2 0.7 1,472.4 100.0%/ 50.4 99.9%/ No pobre 3.205,303 _72.9 - 3.205,303 1000 -______ OUINTIL 1 188,259 4.3 123,583.0 100.0 188.259 15.8 - 100.0 61.1 38.0 65.7 14.0 46496.6 33.7%/ 50.4 100.0% 2 386,320 8.8 388,320 32.4 - 100.0 38.2 15.1 - - 637.8 43.3%/ 0.0%/ 3 745,276 17.0 - - 617,972 51.8 127,304 4.0 82.9 10.5 1.8 --338.3 23.0% -0.0% 4 1,237,370 28.1 - - 1,237,370 38.6 -- - -0.0% 0.0%/ 5 1,840,629 1 41.9 - - - 1,840,629 57.4 - - -- 0.0% - 0.0% AREA Urbana 4,397,854 100.0 123,583.0 100.0 1,192,551 100.0 3.205,303 100.0 27.1 7.8 3.3 2.8 0.6 0.2 1,471.9 1 00. 0%/ 50.4 100.0% Rural - - . - -- 0.0%/ O.0%I REGION Metropolitana 2.078.474 47.3 7,154.0 5.8 289.418 24.3 1,789.056 55.8 13.9 2.3 0.6 0.3 0.1 0.0 203.8 13.8%/ 3.2 6.3% Norte 146,174 3.3 11,406.0 9.2 69,747 5.9 76,427 2.4 47.7 17.0 8.1 7.8 2.1 0.8 107.6 7.30/o 5.7 11.4% Noronente 240,983 5.5 10,296.0 8.3 60,760 5.1 180,223 5.6 25.2 8.1 3.6 4.3 0.9 0.3 84.1 5.7%/ 4.1 8.1% Suronente 230,079 5.2 5,34.0 4.4 97,518 8.2 132,561 4.1 42.4 1 1.8 4.5 2.3 0.5 0.1 117.5 8.0% 2.3 4.6%/ Central 536,293 12.2 26,214.0 21.2 204,577 17.2 331,716 10.4 38.2 1 1.4 4.8 4.9 0.7 0.2 264.7 18.0% 6.9 13.6% Suroccidente 817,755 18.6 29,812.0 24.1 319,395 26.8 498,360 15.6 39.1 12.7 5.5 3.7 0.6 0.2 448.2 30.4% 8.9 17.7%/ Noroccidente 245,867 5.6 29,548.0 23.9 114,362 9.6 131,505 4.1 46.5 18.5 9.6 12.0 3.6 1.3 195.9 13.3% 16.8 33.3% Peten 102,229 1 2.3 3,6. 3.0 36,774 3.1 1 65,455 2.0 36.0 1 11.4 5.0 I 3.7 1.2 10.4 50.5 1 3..4%/ 2.3 4.6% ETNIA (2gr.) Indfgena 1,231,941 28.0 99,619.0 80.6 621,997 52.2 609,944 19.0 50.5 17.3 8.0 8.1 1.8 0.6 917.7 62.4% 42.2 83.6% No indfgena 3 6 1 72.0 23,964.0 19.4 570,554 47. 2.595.359 81.0 18.0 4.1 1.4 0.8 0.1 0.0 555.1 37.7% 8.5 16.8%/ ETNIA (6 gr.) K'Ichie 338,438 7.7 22,342.0 18.1 134,345 11.3 204,093 6.4 39.7 12.6 5.8 6.6 1.3 0.5 184.2 12.5% 8.7 17.2% Q'eqchi 128,729 2.9 16,073.0 13.0 57,384 4.8 69,345 2.2 45.3 18.6 9.6 12.7 2.8 0.9 102.0 6.9%/ 6.7 13.2%/ Kaqchiquel 394,595 9.0 17,084.0 13.8 171.849 14.4 222,946 7.0 43.5 13.4 5.8 4.3 1.0 0.3 227.8 15.5% 7.5 15.0%/ Main 124,695 2.8 12,853.0 10.4 102,876 8.6 21,819 0.7 82.5 29.9 12.8 10.3 0.7 0.1 161.0 10.9% 1.7 3.4% Otros ind!g 247,484 5.8 31,267.0 25.3 155,743 13.1 91,741 2.9 62.9 22.7 11.4 12.6 3.7 1.4 242.8 16.5% 17.6 33.0% No indfgena 3,165,913 72.0 23,964.0 19.4 570,554 47.8 2,595,359 81.0 18.0 4.1 1.4 0.8 0.1 I0.01 555.1 37.7%/ 8.5 1 16.8% -SEXO JEFE - - - -___ Masculino 3,562,698 81.10 99,681.0 80.7 971,630 81.5 2,591,068 80.-8 27.3 7.8 3.3 2.8 0.6 0.2 1,195.5 81.2%/ 42.2 83.7%/ Femenino 835,156 19.0 2,20 19.3 220,921 18.5 614,235 19.2 26.5 7.7 3.1 2.9 0.5 0.1 276.6 18.8%] 8.1 16.1% Annex 4. Page 5 Table A4.4 - Population, contribution to poverty (contr%), percentage of poor (PO), poverty depth index (P1) & poverty severity index (P2), and cost to eliminate poverty (Valor de la brecha) Rural [ de personas y su contribuci6n P 0, P 1, & P 2 | Valor ($) de Ia brecha de pobreza Tod Pobre erMO Todos pobres I No Pobre Pob. General Pob. Extrema P General P. Extrema Personas % Personas Cot sonas Personas nCow% PO 1 P2 P0 P1 P2 M 000,000 % f 000,000| Rural _ _ _ 1 TOTAL 6,987,587 100.0 1,663,099.0 100.0 5,205,352 100.0 1,782,235 100.0 74.5 32.0 17.0 23.8 5.7 2.0 9,647.4 100.0Yo% 756.1 100.0% POBREZA Extrema 1,663,099 23.8 1,663,099.0 100.0 1,663,099 32.0 . . 100.0 66.3 44.4 100.0 23.8 8.2 4,758.9 49.3%1 756.4 100.0% Pobre (todos) 5,205,352 74.5 1,663,099.0 100.0 5,205,352 100.0 . 100.0 42.9 22.8 32.0 7.6 2.6 9,648.3 100.0%o6 756.3 100.0% No pobre 1,782,235 25.5 . . - 1,782,235 100.0 . . . . QUINTIL 1 2,089,302 29.9 1,663,099.0 100.0 2,089,302 40.1 . 100.0 63.6 41.2 79.6 18.9 6.5 5,741.2 59.5%| 756.1 100.0% 2 1,890,736 27.1 . 1,890,736 36.3 . - 100.0 39.0 15.8 - 3,187.7 33.0% * 0. 3 1,531,631 21.9 . . 1,225,314 23.5 306.317 17.2 80.0 10.9 1.9 . . 719.0 7.5% ' 0.0% 4 1,038,869 14.9 . . . 1,038,869 58.3 .- 0.0% . 0.0% 5 437,049 6.3 . . . 437,049 24.5 . . . - 0.0/* 0.0% AREA Urbana - - - 10.0% T 0.0% Rural 6.987,587 100.0 1,663,099.0 100.0 5,205,352 100.0 1,782,235 100.0 74.5 32.0 17.0 23.8 5.7 2.0 9,647.4 100.0% 756.1 100.0%/ REGION Metropolitana 387,483 5.6 8,370.0 0.5 154,286 3.0 233,197 13.1 39.8 9.7 3.7 2.2 0.8 0.3 162.1 1.7%/6 5.6 0.7% Norte 773,660 11.1 347,902.0 20.9 702,863 13.5 70,797 4.0 90.9 47.0 27.7 45.0 11.6 4.4 1,570.6 16.3%| 171.9 22.7% Nororiente 691,600 9.9 73,017.0 4.4 422,327 8.1 269,273 15.1 61.1 21.6 9.8 10.6 1.4 0.3 646.0 6.7%/ 18.2 2.4% Suroriente 768,426 11.0 195,548.0 11.8 586,991 11.3 181,435 10.2 76.4 32.4 17.1 25.5 5.8 1.8 1,073.9 11.1% 85.2 11.3% Central 680,037 9.7 80,124.0 4.8 424,751 8.2 255,286 14.3 62.5 23.0 10.9 11.8 2.2 0.6 675.5 7.0%5 28.0 3.7% Suroccidente 2,196,034 31.4 481,437.0 29.0 1,608,509 30.9 587,525 33.0 73.3 30.5 16.1 21.9 5.4 1.9 2,893.5 30.09/. 227.5 30.1% Noroccidente 1,220,866 17.5 432,584.0 26.0 1,089,732 20.9 131,134 7.4 89.3 42.5 23.6 35.4 8.5 2.9 2,240.2 23.2% 198.4 26.2% Peten 269,481 3.9 44,117.0 2.7 215,893 4.2 53,588 3.0 80.1 33.2 16.5 16.4 4.2 1.6 385.9 4.0O/o 21.61 2.9% ETNIA (2gr.) Indfgena 3,612,091 51.7 1,182,055.0 71.1 3,065,603 58.9 546,488 30.7 84.9 40.0 22.3 32.7 8.3 3.0 6,238.1 64.7% 571.8 75.6% No indigena 3,375,496 48.3 481,044.0 28.9 2,139,749 41.1 1,235,747 69.3 63.4 23.4 11.4 14.3 2.9 0.9 3,409.6 35.3%j 184.6 24.4% ETNIA (6 gr.) K iche 734,886 10.5 182,151.0 11.0 556,664 10.7 178,222 10.0 75.8 32.5 17.2 24.8 5.7 2.0 1,031.1 10.7%/O 79.4 10.5% Qeqchi 609,434 8.7 263,304.0 15.8 556,931 10.7 52,503 3.0 91.4 46.9 27.6 43.2 11.9 4.8 1 ,233.0 12.8% 138.2 18.3%/ Kaqchiquel 617,207 8.8 120,519.0 7.3 461,874 8.9 155,333 8.7 74.8 31.1 15.9 19.5 4.1 1.2 829.5 8.6Yo| 47.9 6.3% Mam 816,170 11.7 308,692.0 18.6 741,432 14.2 74,738 4.2 90.8 45.0 25.9 37.8 11.1 4.2 1,584.7 16.4% 172.4 22.8% Otros indfg 834,394 11.9 307,389.0 18.5 748,702 14.4 85,692 4.8 89.7 43.3 24.0 36.8 8.4 2.7 1,559.2 16.2% 134.0 17.7% No indigena 3,375,496 48.3 481,044.0 28.9 2,139,749 41.1 1,235,747 69.3 63.4 23.4 11.4 14.3 2.9 0.9 | 3,409.6 35.3%1 184.61 24.4 SEXO JEFE Masculino 6,153,884 88.1 1,522,937.0 91.6 4,628,022 88.9 1,525,862 85.6 75.2 32.7 17.5 24.8 5.9 2.0 8,677.0 89.9%o| 688.2 91.0% Femenino 833,703 11.9 140,162.0 8.4 577,330 11.1 256,373 14.4 69.3 26.9 13.6 16.8 4.3 1.5 969.6 10.1% 68.1 9.0°/ Annex 4. Page 6 Table A4.5 - Population, contribution to poverty (contr%), percentage of poor (PO), poverty depth index (P1) & poverty severity index (P2), and cost to eliminate poverty (Valor de la brecha): Indigenous #de personas y su corntribuci6n P 0, P 1, & P 2 Valor (S) de la brecha de pobreza Todos Pobre extremo I Todos pobres I No Pobre Pob. General Pob. Extrema P. General P. Extrema Personas % Personas Contr% Personas | Contr% Personas Contr P0 P1 | P2 PO P1 I P2 QM 000,000 I % OQ 000,000 % Indigena _-- - - 1 1 - - 1 _I TOTAL 4,844,032 100.0 1,281,674.0 100.0 3,687,600 100.0 1,156,432 100.0 76.1 34.2 18.6 26.5 6.6 2.4 7,154.4 100.0% 614.0 100.0% POBREZA Extrema 1,281,674 26.5 1,281,674.0 100.0 1,281,674 34.6 8 100.0 66.8 45.2 100.0 25.1 9.0 3,699.0 51.7% 614.0 100.0% Pobre (todos) 3,687,600 76.1 1,261,674.0 100.0 3,667,600 100.0 . - 100.0 44.9 24.5 34.8 6.7 3.1 7,155.2 100.0% 614.0 100.0% No pobre 1,156,432 23.9 1 . . 1,156,432 100.0 . . . . . QUINTIL 1 1,621,160 33.5 1,261,674.0 100.0 1,621,160 44.0 . . 100.0 64.0 41.7 79.1 19.8 7.1 4,481.4 62.6% 613.9 100.0% 2 1,311,576 27.1 . - 1,311,576 35.6 . 100.0 39.1 15.8 . - . 2,215.8 31-0%| 0.0% 3 933,321 19.3 . 754,864 20.5 178,457 15.4 80.9 11.4 2.1 . . 457.5 6.4% 0.0% 4 672,133 13.9 . . . 672,133 58.1 .- 0.0% .-| 0.0% 5 305,642 6.3 . 305,842 26.5 . . . | - 0.0/ - 0.0% AREA Urbana 1,231,941 25.4 99,619.0 7.6 621,997 16.9 609,944 52.7 50.5 17.3 8.0 8.1 1.8 0.6 917.7 12.8% 42.2 6.9 Rural 3,612,091 74.6 1,182,055.0 92.2 3.065,603 83.1 546,488 47.3 84.9 40.0 22.3 32.7 8.3 3.0 6,238.1 87.2% 571.8 93.1%6 REGION Metropolitana 394,537 8.1 13,292.0 1.0 148,318 4.0 246,219 21.3 37.6 9.7 4.1 3.4 1.1 0.4 165.6 2.3%1 .6 1.4% Norte 749,127 15.5 340,054.0 26.5 665,569 18.1 83.558 7.2 88.9 46.5 27.6 45.4 11.8 4.5 1,504.0 21.0% 169.0 27.5% Nororiente 168,804 3.5 50,374.0 3.9 134,770 3.7 34,034 2.9 79.8 35.8 18.7 29.8 4.6 1.1 261.2 3.7% 14.8 2.4% Suroriente 53,385 1.1 5,252.0 0.4 43,438 1.2 9,947 0.9 81.4 29.8 13.2 9.8 3.3 1.3 68.7 1.0% 3.4 0.5% Central 558,674 11.5 77,822.0 6.1 356.928 9.7 201,746 17.5 63.9 24.9 12.1 13.9 2.4 0.7 599.8 8.4% 25.1 4.1% Suroccidente 1,674,345 34.6 374,507.0 29.2 1,247,389 33.8 426,956 36.9 74.5 32.2 17.1 22.4 5.9 2.1 2,324.7 32.5% 187.2 30.5% Noroccidente 1,147,353 23.7 408,908.0 31.9 1,019,638 27.7 127,715 11.0 88.9 42.7 23.9 35.6 9.1 3.2 2,113.3 29.5% 196.7 32.4% Peten 97,807 2.0 11,465.0 0.9 71S550 1.9 26,257 2.3 73.2 27.9 13.3 11.7 3.7 1.6 117.8 1.6% 6.8 1.1% ETNIA (2gr.) - . - _ - - Indigena 4,844,032 100.0 1,281,674.0 100.0 3,687,600 100.0 1,156,432 100.0 76.1 34.2 18.6 26.5 6.6 2.4 7,154.4 100.0%M 614.0 100.0% No indlaena - . ._._._._._.- . . . . . 0.0I % 0.0% ETNIA (6 gr.) K'iche 1,073,324 22.2 204,493.0 16.0 691,009 18.7 382,315 33.1 64.4 26.2 13.6 19.1 4.3 1.5 1,214.9 17.0% 88.0 14.3% ('eqchi 736,163 15.2 279,377.0 21.8 614,315 16.7 121,848 10.5 83.5 42.0 24.5 38.0 10.3 4.1 1,334.9 | 16%./| 144.8 23.6% Kaqchiquel 1,011,802 20.9 137,6020 10.7 633,523 17.2 378,279 32.7 62.6 24.2 11.9 13.6 2.9 0.9 1,057.0 14.8% 555. 9.0% Mam 940,865 19.4 321,545.0 25.1 844,308 22.9 96,557 8.4 89.7 43.0 24.2 34.2 9.7 3.6 1,745.5 24.4% 174.1 28.4% Otros indig 1,081,878 22.3 338,656.0 26.4 904,445 24.5 177,433 15.3 83.6 38.6 21.1 31.3 7.3 2.4 1,602.1 25.2%| 151.6 24.7% No indigena . . . . - . . 00% 0.0% SEXO JEFE- - Masculipo 4,227,160 87.3 1,171,387.0 91.4 3,261,747 88.5 965,413 83.5 77.2 35.1 19.2 27.7 6.9 2.5 6,400.3 8695% 560.0 91.2% Femenino 616,672 12.7 110,287.0 8.6 425,853 11.6 191,019 16.5 69.0 28.4 14.5 17.9 4.6 1.7 755.5 10.6% 53.9 8.86 Annex 4, Page 7 Table A4.6 - Population, contribution to poverty (contr%), percentage of poor (PO), poverty depth index (P1) & poverty severity index (P2), and cost to eliminate poverty (Valor de la brecha): Non-indigenous # de personas y su contribuci6n P 0, P 1, & P 2 Valor (S) de la brecha de pobreza Todos I Pobre extremo [ Todos pobres I No Pobre Pob. General I Pob. Extrema P. General P. Extrema Personas % I Personas cons% P Coeto/n Personas I ContP/ P 0 P 1 P 2 P 0 P 1 P 2 Q 00000 % Q\- 000°000 % No indigena 7777 TOTAL 6,541,409 100.0 505,008.0 100.0 2,710,303 100.0 3,831,1061 100.0 41.4 14.0 6.5 7-7 1.5 0.5 3,966.2 100.0% 192.6 100.0% POBREZA Extrema 505,008 7.7 505,008.0 100.0 505,008 18.6 - 100.0 64.6 42.1 100.0 20.0 5.9 1,408.2 35.5% 192.7 100.1% Pobre (todos) 2,710,303 41.4 505,008.0 100.0 2,710,303 100.0 . 100.0 33.9 15.8 18.6 3.7 1.1 3,965.5 100.0% 192.7 100.1% No pobre 3,831,106 58.6 - . . 3,831,106 100.0 . . . . . QUINTIL 1 656,401 10.0 505,008.0 100.0 656,401 24.2 - 100.0 62.0 38.9 76.9 15.4 4.5 1,756.1 44.3% 192.7 100.1% 2 965,480 14.8 . . 965,480 35.6 . 100.0 38.6 15.4 . - 1,609.4 40.6% - 0.0% 3 1,343,586 20.5 . 1,088,422 40.2 255,164 6.7 81.0 10.3 1.7 . . 599.4 15.1% - 0.0% 4 1,604,106 24.5 . . . . 1,604,106 41.9 . . . 0.0% 0.0% 5 1,971,836 30.1 . . . 1,971,836 51.5 . . . . . . . 0.0% ' 0.0% AREA Urbana 3,165,913 48.4 23,964.0 4.8 570,554 21.1 2,595,359 67.7 18.0 4.1 1.4 0.8 0.1 0.0 555.1 14.0% 8.5 4.4% Rural 3,375,496 51.6 481,044.0 95.3 2,139,749 79.0 1,235,747 32.3 63.4 23.4 11.4 14.3 2.9 0.9 3,409.6 86.0% 184.6 95.8% REGION Metropolitana 2,071,420 31.7 2,232.0 0.4 295,386 10.9 1.776,034 46.4 14.3 2.2 0.5 0.1 0.0 - 200.4 5.1% 0.4 0.2% Norte 170,707 2.6 19,254.0 3.8 107,041 4.0 63,666 1.7 62.7 23.6 11.3 11.3 2.6 0.8 174.1 4.4% 8.5 4.4% Nororiente 763,779 11.7 32,939.0 6.5 348,317 12.9 415,462 10.8 45.6 14.2 5.9 4.3 0.5 0.1 469.0 11.8% 7.4 3.9% Suroriente 945,120 14.5 195,690.0 38.8 641,071 23.7 304,049 7.9 67.8 27.5 14.3 20.7 4.7 1.4 1,122.8 28.3% 84.2 43.7% Central 657,656 10.1 28,516.0 5.7 272,400 10.1 385,256 10.1 41.4 12.0 4.9 4.3 0.8 0.2 340.2 8.6% 9.7 5.0% Suroccidente 1,339,444 20.5 136,742.0 27.1 680,515 25.1 658,929 17.2 50.8 17.6 8.3 10.2 1.9 0.6 1,016.9 25.6% 49.2 25.5% Noroccidente 319,380 4.9 53,222.0 10.5 184,456 6.8 134,924 3.5 57.8 23.4 11.8 16.7 2.7 0.6 322.9 8.1% 16.3 8.5% Peten 273,903 4.2 36,413.0 7.2 181,117 6.7 92,786 2.4 66.1 26.9 13.4 13.3 3.3 1.2 318.7 8.0% 17.2 8.9% ETNIA (2gr.) Indigena . . . . . . - - 0.0% - 0.0% No indfgena 6,541,409 100.0 505,008.0 100.0 2,710,303 100.0 3,831,106 100.0 41.4 14.0 6.5 7.7 1.5 0.5 3,966.2 100.0% 192.6 100.0% ETNIA (6 gr.) K'iche .- 0.0% 0.0% Q'eqchi . . . . . . . 0.0% - 0.0% Kaqchiquel . . . . . . . . . 0.0% - 0.0% Mam . .0 . . . . . . 0.0% - 0.0% Otros indfg .. . . . . . . . 0.0% 0.0% No ind(oena 6,541,409 100.0 505,008.0 100.0 2,710 303 1 00.0 3,831,106 1 00.0 41.4 1 4.0 6.5 7.7 1.5 0.5 3,966.2 100.0% 192.61 100.0° SEXO JEFE , 0 0.0% Masculino 5,489,422 83.9 451,231.0 89.4 2,337,905 86.3 3,151,517 82.3 42.6 14.7 6.9 8.2 1.6 0.5 3,475.4 87.6%. 170.0 88.3% Femenino 1,051,987 16.1 53,777.0 10.7 372,398 13.7 679,589 17.7 35.4 10.8 4.8 5.1 1.1 0.3 490.7 12.4%1 22.31 11.6%° Annex 4. Page 8 Table A4.7 - Poverty Indicators by Welfare Measure, Central America Comparisons All Poor (below FPL) I Extreme Poor (below XPL) GNI Per % Poor' I Depth' I Severity' I % Poora I De2th SeverityC Capita, PPP Using Consumption as Welfare Measure Guatemala (2000) 56.2% 22.6 11.7 15.7% 1 3.7 1.3 $3,630 Nicaragua (1998) 47.9% 18.3 9.3 17.3% 4.8 2.0 $2,060 Panama (1997) 37.3% 16.4 9.7 18.8% 7.7 4.2 $5,450 Using Income as Welfare Measure Guatemala (2000) 65.6% 35.1 25.9 31.9% 15.1 22.2 $3,630 Nicaragua (1998) 55.1% 26.2 16.0 29.9% 12.2 6.6 $2,060 Panama (1997) 42.1% 22.8 17.1 26.2% 14.2 13.3 $5,450 Honduras (1996) 62.9% 33.4 22.3 35.0% 16.3 10.6 $2,270 LAC Average (1996) 36.7% 16.9 10.7 16.1% 7.4% 5.1 $6,620 Sources: GNI (gross national income) per capita estimates for 1999 in PPP US$, from World Bank, World Development Indicators 2001. Guatemala poverty estimates calculated by tNE-SEGEPLAN-URL with technical assistance from World Bank using the ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Nicaragua poverty estimates from World Bank (2001a). Panama poverty estimates from World Bank (2000a). Honduras estimates from World Bank (2001c). Other countries: from Wodon (2001). a. Incidence of poverty or headcount index (% of population whose total consumption or income falls below poverty line, FPL or XPL). All poor includes extreme poor (throughout study). b. The Poverty Depth Index (PI) represents the amount needed to bring all poor individuals up to the poverty line (FPL or XPL), expressed as a percent of the poverty line taking into account the share of the poor population in the national population. c. The Poverty Severity Index (P2) is a derivation of PI that takes into account the distribution of total consumption among the poor. In other words, it is a measure of the degree of inequality among the population below the poverty line. CAVEATS: International comparisons of poverty are always difficult due to various methodological differences (welfare measures, poverty lines, survey samples). Annex 4, Page 9 Table A4.8 - Income Inequality and Distribution Population Share (%) Income Share (%) Gini (%) All Guatemala 100 100 57 Area Urban 39 66 54 Rural 61 34 47 Region Metropolitana 22 45 54 Norte 8 4 50 Nororiente 8 7 47 Suroriente 9 6 50 Central 11 9 47 Suroccidente 26 19 51 Noroccidente 13 7 51 Peten 3 2 53 Ethnic group' Indigenous 43 23 46 Non-indigenous 57 77 56 Language ability' Monolingual Spanish 62 80 56 Monolingual indigenous 9 3 39 Bilingual 29 17 48 Income quintiles I (Low) 20 3 22 2 20 6 9 3 20 11 7 4 20 18 11 5 (High) 20 62 38 Based on household definition of ethnicity. bBased on household head's language ability. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de EstadIstica - Guatemala. Annex 4. Page 10 Table A4.9 - Consumption Inequality and Distribution Population Share (%) Consumption Share (%) Gini (%) All Guatemala 100 100 48 Area Urban 39 63 44 Rural 61 37 35 Region Metropolitana 22 43 44 Norte 8 4 38 Nororiente 8 8 39 Suroriente 9 6 37 Central I 10 38 Suroccidente 26 20 40 Noroccidente 13 7 36 Peten 3 2 39 Ethnic group a Indigenous 43 24 36 Non-indigenous 57 75 47 Language ability' Monolingual Spanish 62 79 46 Monolingual indigenous 9 3 29 Bilingual 29 18 36 Consumption quintiles I (Low) 20 5 13 2 20 9 7 3 20 13 6 4 20 20 9 5 (High) 20 53 30 'Based on household definition of ethnicity. b Based on household head's language ability. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Annex 4. Page II Table A4.10 - Consumption Quintiles: levels and distribution National Turban Rural K'Iche Q'ueqchl Kagchiguel Man Other Mayan Non-lnd-Igenous Alli Ma-yan All Indigenous Value in Quetzales 0.1 1,579.92 2,742.81 1,383.42 1,506.13 1,139.60 1,673.58 1,152.45 1,296.81 2,054.91 1,313.15 1,313.44 Q.2 2,638.52 4,633.97 2,123.30 2,418.41 1,706.94 2,503.60 1,764.82 1,868.20 3,548.78 2,016.68 2,018.45 0.3 3,883.52 6,930.03 2,903.94 3,409.92 2,250.62 3,566.45 2,305.85 2,479.54 5,181.96 2,740.08 2,741.54 0.4 6,074.55 10,593.92 4,014.67 4,713.51 3,157.54 4,760.47 3,054.48 3.285.28 8,225.38 3.851.17 3,851.36 0.5 16,626.48 25,704.65 7,916.11 8,851.28 6,371.75 9,526.52 5,149.32 6,994.22 21,322.54 7,703.71 7,735.95 National average 6,161.00110,121.85 3,668.1214,179.06 2,919.701 4,407.09 2,684.361 3,183.34 8,065.97 3,524.85 3.531.181 % of 5th quintile 0.1 9.5% 10.7% 17.5% 17.0% 117.9%/ 117.6%/ 22.4% 18.5% 9.6% 17.1% 17.0%/ 0.2 15.9% 118.0%/ 26.8% 27.3%/ 26.8%/ 26.3%/ 34.3%/ 26.7%/ 16.6% 26.2%/ 26.1% 0.3 23.4% 27.0%/ 36.7% 38.5% 35.3% 37.4% 44.8% 35.5% 24.3% 35.6%/ 35.4% 0.4 36.5% 41.2% 50.7% 53.3%/ 49.6%/ 5 0. 0% 59.3% 47.0%/ 38.6%/ 50.0%/ 49.8% 0.5 100.0%/ 100.0%/ 100.0%/ 100.0% 100.01% 100.0%/ 100.0% 1100.0%/ 100.0% 100.0% 100.0% National average 37.1% 39.4% 46.3% 47.2%/ 45.8%/ 46.3%/ 52.1% 45.5%1 37.8% 45.8% 45.7%/ % National average 0.1 5.1%/ 5.4% 7.51% 7.'2%/ 7.8%/ 7.6%/ 8.6%/ 8.2% 5.1% 7.5% 7.4% 0.2 .86% 9. 2%1 11.6%6 11.6% 11.7% 11.4% 13.2% 11.7% 8.8% 111.4%/ 11.4%/ 0.3 812.6%6 13.7%/ 15.8% 16.3% 15.4 116.2%/ 17.2%/ 15.6% 12.9% 115.6%/ 15.5% 0.4 19.7% 20.9% 21.9%/ 22.6% 21.6% 21.6%/ 22,8% 20.6% 20,4% 211.9%/ 21.W. 0.5 54.0% 50.8%. 43.2%/ 42.4%/ 43.7% 43.2%/ 38.4% 43.9% 52.9% 43.7% 43.8% National average 100.0% 100.01% 100.01% 100.0% 1100.0%/ 100.0%/ 100.0% 100.0% 100.0% 100O.0%/ 100.0%/ Ratios 0.210.1 1.67 1.69 1 53 1.61 1.50 1.50 1.53 1.44 1.73 1.54 1.54 0.3/0.2 1.47 1.50 1.37 1.41 1.32 1.42 1.31 1.33 1.46 1.36 1.36 0.4/0.3 1.56 1.53 1.38 1.38 1.40 1.33 1.32 1.33 1.59 1.41 1.40 Q.5/0.4 2.74 2.43 1.97 1688 2.02 2.00 1.69 2.13 2.59 2.00 2.01 10.5/0.1 10.52 9.37 5.72 5.88 5.59 5.69 4.47 5.39 10.38 5.87 5.89 Table A4.11 - Income Quintiles: levels'and dlistribution National Urban Rua K Iche O'ueqchm -Kaqchiquel Man Other Mayan Non-indigenous All Mayan All Indigenoua Value in Quetzalea 0.1 655.00 1,348.60 558.92 885.01 401.24 759.62 445.39 599.61 855.74 581.83 585.24 0.2 1,813.14 3,280.76 1,377.48 1,696.34 1,091.62 1,881.33 947.51 1,260.70 2,479.06 1,316.17 1,322.12 0.3 2,986.43 5,361.18 2,157.37 2,546.41 1,670.32 2,968.19 1,503.04 1,835.05 4,022.18 2,072.34 2,076.09 0.4 5,110.92 9,215.44 3,343.75 3,763.83 2,515.92 4,228.64 2,337.07 2,855.85 6,913.64 3,197.96 3,205.39 Q.5 17,326.64 28,083.08 8,243.51 8,960.39 6,056.60 9,664.25 5,255.46 7,257.54 22,922.18 7,766.79 7,822.60 National average 5,578.571 9,458.37 3,136.7113,569.50 2,344.481 3,896.83 2,097.58, 2,763.00 7,444.01 2,987.16 3,001.101 % of 5th quintile 0.1 3.8% 4.8% 6.8% 9.7%/ 6.6%/ 7.9%/ 8.5% 8.3% 3.7%/ 7.5% 7.5% 0.2 10.5% 11.7% 16.7% 18.9%/ 118.0%/ 19.5% 18.0% 17.4% 10.8% 17.0% 116.9%/ 0.3 17.2% 19.1% 26.2%/ 28.4% 27.6%/ 30.7% 28.6% 25.3% 17.6%. 26.7% 26.5% 0.4 29.5% 32.86% 40.6%/ 42.0%/ 41.5% 43.8% 44.5% 39.4% 30.2% 41.2% 41.0% 0.5 100.0% 100.0% 1100.0%/ 100.0%/ 1100.0%/ 100.0%/ 100.0%/ 100.0% 100.0% 100.0%/ 100.0% Naioa aerg 32.2% 33.7%. 38.1% 39.8% 38.7% 40.39% 39.9%/ 38.1%. 32.5% 38.5% 38.4%, 0.1 2.4% 2.9% 3.6%/ 4.9%/ 3.4% 3.9%/ 4,3%/ 4.3%/ 2.3% 3.9%/ 3.9%/ 0.2 6.5% 6.9%/ 8.8% 9.5% 9.3%/ 9.7%/ 9.0%/ 9.1% 6.7%/ 8.8%/ 8.8% 0.3 10.7%/ 11.3%/ 13.8%/ 114.3%/ 114.3%/ 15.2%/ 14.3%/ 13.3% 10.8% 13.9%/ 13.8% 0.4 18.3% 19.5% 211.3%/ 21.1% 21.5% 21,7%/ 22.3%/ 20.7% 18.6% 21.4% 21.4% 0.5 62.1% 59.4% 52.6%/ 50.2% 51.7% 49.6% 50.1% 52.5% 61.6%/ 52.0%/ 52.1% National average 1100.0%/ 100.0% 100.0%/ 100.0% 100.0% 100.0% 10(0. 0%/ 100.0% 100.0% 100.0%, 100.0%/ Ratios 0.2t0. 1 2.77 2.43 2.46 1.96 2.72 2.48 2.13 2.10 2.90 2.26 2.26 Q.3/0.2 1.65 1.63 1.57 1.50 1.53 1.58 1.59 1.46 1.62 1.57 1.57 0.4/0.3 1.71 1.72 1 55 1.48 1.51 1.42 1 55 1.56 1.72 1.54 1.54 0.5/0.4 3.39 3.05 2.47 2.38 2.41 2.29 2.25 2.154 3.32 2.43 2.44 10.5/0.1 26.45 1 20.82 114.75 10.368 15.09 12.72 111.80 1 12.10 26.79 13.35 13.37 Annex 4 Page 12 Table A4.12 - Income Sources, by Income Quintiles Income quintiles | Total 1 2 3 4 5 _ Income per capita (Q) 653 1,810 2,983 5,110 17,319 5,578 Labor Income (%) 43 72 74 78 73 73 Agricultural 20 44 30 17 5 13 Salaries 28 21 15 8 2 6 Formal sector 9 10 8 5 1 3 Informnal sector 19 11 7 3 1 3 Net inc. from production -8 23 15 9 3 7 Non-Agricultural 23 28 44 61 68 60 Salaries 12 18 33 43 47 42 Formal sector 6 10 19 30 42 35 Informal sector 6 8 14 13 5 7 Own business 11 10 I 1 18 21 18 Formal sector 0 1 1 2 6 4 Informal sector I 1 9 10 16 15 14 Non-labor income (%) 55 27 26 22 27 27 Return to capital 31 14 12 10 14 14 Donations, gifts 22 12 12 10 6 8 Remnittances 5 3 4 6 4 4 Private I 1 1 1 I 1 Public 16 8 7 3 1 3 Pensions, indemnizaciones 2 1 2 2 4 3 Other b O 0 0 0 3 2 Percentages may not add up to 100 due to rounding. ' As interest received was negligible, the return to capital includes: income from rental of equipment, rental of property and the interest b For example, inheritance or lottery winnings. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Annex 4. Page 13 Table A4.13 - Income Sources, by Consumption Quintiles Consumption quintiles | Total _ 1 2 3 4 5 Income per capita (Q) 1,429 2,408 3,487 5,064 15,503 5,578 Labor Income (%) 77 78 77 76 70 73 Agricultural 49 38 24 14 3 13 Salaries 30 18 11 6 1 6 Formal sector 13 9 6 4 1 3 Informal sector 17 9 5 2 0 3 Net inc. from production 19 20 13 8 2 7 Non-Agricultural 28 40 53 62 67 60 Salaries 17 25 39 47 46 42 Formal sector 8 13 26 36 42 35 Informal sector 9 12 13 11 4 7 Own business 11 15 14 15 21. 18 Formal sector I I I 1 7 4 Informal sector 10 14 13 14 14 14 Non-labor income (%) 22 22 24 26 30 27 Return to capital 10 8 10 11 16 14 Donations, gifts 11 12 12 10 6 8 Remittances 3 4 5 5 4 4 Private 1 2 1 2 1 1 Public 7 6 6 3 1 3 Pensions, indemnizaciones 1 1 1 3 5 3 Other b 0 1 1 2 3 2 Percentages may not add up to 100 due to rounding. a As interest received was negligible, the return to capital includes: income from rental of equipment, rental of property and the interest received. bFor example, inheritance or lottery winnings. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Annex 4, Page 14 Table A4.14 - Income Sources, by Poverty Classification Extreme Poor All Poor Non-Poor Income per capita (Q) 1,345 2,349 9,721 Labor Income (%) 78 77 71 Agricultural 49 34 6 Salaries 31 17 3 Formal sector 13 8 2 Informal sector 18 9 1 Net inc. from 18 17 3 production Non-Agricultural 29 43 65 Salaries 17 29 46 Formal sector 9 17 40 Informal sector 8 12 6 Own business 12 14 19 Formal sector 1 1 5 Informal sector 11 13 14 Non-labor income (%) 20 22 27 Return to capital a 9 9 15 Donations, gifts 10 11 6 Remiittances 3 4 4 Private 1 I 1 Public 6 6 1 Pensions, 1 1 4 indemnizaciones Other' 0 1 2 Percentages may not add up to 100 due to rounding. ' As interest received was negligible, the return to capital includes: income from rental of equipment, rental of property and the interest received. b For example, inheritance or lottery winnings. Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Annex 4. Page 15 Table A4.15 - Consumption Pattrns: National _ _ Annulg ocrnsumption Ievels (prco 1 In Ou tles) ____Percent ot total anmual conmimaUn pucspta 2008) T __d___,g P .. r 94 Eoods ____ I H.th D-2Wi. OdI -M h. W T8d8 Fd %) H-____P_- gOd. 3%) Ed.. ",M ." _____________ T1oO 3%)M TW NACIONAL TOTAL 2488 8aD 577 320 225 244 383 883 Ill 8.181 40.4% 14.0% 9.4% 8. 3.7A 40 8.2 IS5 I0 IOD.0 POB REZA Elrams 845O 149 Ise 21 20 8 28 255 3 5.48 67.8- 102 128 1.4- 1.4A 0l 1. 141%A 0.2 100.0 Pobre (SodOo) I143 287 280 88 45 33 80 388 8 2588 84.9 102% I0.8A 2.7- 1.8A 13 I . 149 0.3A 188.0 No 138183 6 .2A0 857 842 458- 818 771 J,BI 243 1278,1% 8.84 .0 42A 42 7.2A 50 ---- 1.0 1 811 18 184 25 22 10 31 221 4 1.58 87.8% 10.3% 12.3% 1.8% IS- 0fi 30 140 188.0 2 1.477 273 28 84 44 31 78 378 8 2.83 58.0 10.3% I1% 2.4 I 12 2. 14~2% 0 108.0 3 2258 All 380 128 88 88 Is8 817 18 3.884 62.2 105% Sal 3.3 2. 1 A 15.0 04 100. 4 3245 788 572 251 58 157 228 88 49 8317 48.% 13.4% 8.4- 4.1% 28 22 8.4 16.8 080 18.0 13 ~~~~~~~~~5.187 2888 ,447 1. 12D 825 948 1,235 228a48 122 31.2 18.2% 8. . 0 ..8 1538 18 A.REA LUrbana 3.442 1.817 82 857 408 480 708 1.581 303 12.122 34.0 18.2% 82 8.8% 4.5 47 7.0 18.6 I. 18.O Rum1 1m 388 107 98D 179 887 22 3A 51.6 10.4% 8.7 2.9 ____ ______ 49 is. 0. 188.0 M8tropoISAam 0.88 2887 1.112 812 824 058 837 1287 337 12254 31.7 17.0% 8.1% 8.8% 51i 4.6 7.8 15.4 237 t88.0- Norte 5258 28 325 105 88 88 188 492 25 3,078 80.n 82% 10.7A 3.4% 2Mz 3.8- 18.0 0. 188.0 140881185t8 2087 889 818 18 Ise 214 334 870 so 5.41 47.0 122% 8O 3.2 2.9 VIA 5.8 152 1.0 188.0 Suroriesle 2285 831 41? 128 88 III 218 604 22 4.288 48.8% 122. 92 8 3.0 2.3 . A 8.1 18.3 0.8 108.0 Central Z2.52 733 525 255 158 178 310 885 78 823 44.9 1310% 82 4.% 3.% 328.5% 15.4 1.4- 188.0 SsrOodcdest8 2122 538 445 2381 121 187 247 728 88 42678 45.4 3 11% 8- 5.0 3.81 41D. 1586 1.2 188.0 Noroocidenta 1,241 348 340 188 88 75 148 887 28 3.28 80. IDA%12 3.1% 3.523 48 188- 0- "8. P8588 ~~~~ ~~~~ ~~~~487 381 15 82 127 208 73o 53 4901 48. 10"% 8.% 4.3% 2. 2.8 4. 18W3 I. 188.0 ETNIA (2gr.) ond)gena 1.721 388 371 138 77 8l 757 543 28 32531 48 1%10.5 3.8% 2. 2 4. 1.408 8. No 8ndlesna ,0687 1.201. 728 458 335 308 53 .5 12 818 37..7% 142.% 8 5. 4.1 4.5 8. 15 I'd 188.0 ETNIA (8 gr.) Kmih8 1.839 488 418 181 103 133 351 888 38 4.179 48.4 11.7% I00 4.8% 2.5% 3. 4 18.0 0. 188.0 O'eqchi 1A8 248 323 70 88 42 101 4871 18 2.83 84. 82.% 21 .4 I. IA 38 1.05 188. Kaqc81iuI 1.812 888 487 as8 Il 110 304 585 63 4.407 43.3 13.% 10A 4.3 2.5 2.5 8 1.1 co8 Mam 1 353 308 255 88 38 83. 112 383 15 2834 51.6 I1.4% IIO 3.3- 1.4% 4. 14.8% 08 100.% Otros ind)g 1.758 328 335 115 70 84 132 488 18 3= 53.% 102% 10.4 3.8% 2.2 1.7 4. 14.3 0.8 120.0 No in0ffo0na 3.67 1.0 720 455 330 388 53 I25 17 j.10 37.7 143 O 5.8% 4.1% 426 8.. 2.1 .O....1A. SEXO Mas J1no 2,488 821 884 3D4 28 28'85 84 108 8281 40.9 13.5% 83 5.0 3.77 4. 8. 1. . 18. P881. 0 2,4821 89' 5so 3351 2-28 _284 _380- 81 113 8.6237 38.0 14A% 9.4% 5,4 3.7 39 8.1 1.%I. 18 EDAD 0.8 2282 Sol 42 18 188 13X 248 763 54 4271 45.1- 12.3% 8.3 3.8% 3.5%30 8.4 18.& 1.4 188.0 7-12 2.212 818 408 310 158 183 301 a11 74 5.124 43.2 12.1% 8.8 8.1 3.1% 326 5.0 IS. 1.4% 188. 13-18 3.378 728 525 438 185 224 370 887 88 5.784 41.2% 122% 9.1% 7.5% 2.3 39 . 84A IS. 18 1.0 19-24 Z711 875 848 403 284 30 474 1.147 153 7I88 30.3% 13e% 8 57% 3.7% 43 52 18 21% 188 25-69 2.83 1.057 888 378 300 330 880D 1.141 163 7378 38.4% 142% 8. .1 4.1% 48 10 2. 188.0 _______0 ___ 3280 1.763 oil 20 352 1 3231 384 887 171 82D70 1 37.2% 21.71% 213 .8 4.4%44 1 21 1885% Annex 4, Pagq. 16 Table A4.16 - Consumption PaMrns: Urban Amuel cmauniption lmels (per ca In Ou Pereentoltotalannual nsumpUonpercapfta Urbana Fod HwS.V Pe. gwft Ed.WW M.Mj D-blgoft T-.WIt Oth., S T.Wj F-d H-.ng I(%) P-.Idgod.(%)- Ed..fn (%) H.Wt, (%) T-M.4 (%) 01- (%) S.W- J%) T.W TOTAL 3,442 1,617 920 657 458 706 IASI 262 10.122 34.0% IO1D% 92% 6.5% 4.5% 4.7% 7.0% 16.0% ZIL% P0BRE7A Extroma 1789 170 204 49 27 10 22 220 4 ijum 52.6% 119% 13.6% 3.3% i.9% 0.7% t.6% 14.6% P.3% 00 Pobre (todoo) AN 393 368 I" 63 43 96 493 12 3DS4 40.2% 122% llfi% 4.8% 2.1% IA% &2% 15.6% 0.4% 00.0% No pobre 2,073 1,142 047 am m m ljm 341 12,740 32-7% 1613% 0.0% 6. 4.7% 5.0% 7.3% 15.6% 2.7% i0oo% OUINTIL I am 216 208 62 33 11 31 236 5 lmi 62.0% 1211% 12.4% 3.1% ZD% 0,71A t 4.1% 0.3% 100. 2 I.M4 366 320 120 49 34 84 393 a 2,668 40.1% 13.7% 122% 4.6% I.-1 1.3% 3. 4.7% 0.3% 11.-I 3 1,652 491 436 191 93 fig I" 619 17 3090 47.0% 12,8% 112% 4.9% 2. I.6% 3. 15.9% 0.4% - 4 2.717 604 317 177 ise 330 974 58 8,196 43.9% '39% 0.7% 2.% 2.6 S. 15.7% 0.9% OD. 6208 2. awWa Iusi 1.250 924 I'm 1.307 2.766 563 IIJ12 2O.P% 10.0% 8.7% j 6.7% 7. 16.0% 3.1% 00 5 7.1Z 6. AREA Urbana 3,442 1.617 m 262 10,122 34.0% IOID% 92% 6.5% 4.5% 4.7% 7.0% 15.6% 2.5% 100 Ru'a' rft W. W, W. 'Ift w, -t W. M REGr.. Metropofitana 4.127 2.323 I.2D9 917 703 663 1,010 2XI 355 133W 30. 17A% .8% S. 42% 7.5% 15 4% Z9% I 0% Norte 2,647 "O m 389 las 302 3,0 IASO 116 ext 39. 13.3% :"B%% :.I% 2.0% 4.7% 4,9% 16.6% 1.6% I0D.0% P40foriento 3,509 1.071 726 404 343 422 494 t 064 155 SAN 41.4% 12JM 0.6% 4.9% 4.0% 510% S.7% 18.0% te% 100.0% surorierge Z?W 944 wo 334 197 281 "I 102 70 6,704 40.4% :4.1%i O.O% 4.7% 2.9% 42% 6.6% 16.1% 1.2% 100.0% central 2Xg 1.016 873 393 '98 290 415 1,104 12B 7,Iv 41. 4.3% #A% 5.5% ZS% 3-6% 51 I. 100.0% s ddants Z778 1.014 711 497 251 375 471 1.169 148 7,43D 37.4% I3,6% 0.6% 1 3 SID% S. I I uroc 36. :4,D% 0.5% Sl % 43 S. 21 Noroccidede uw SW 682 368 23D 264 370 OD2 122 6,124 9.1% 7.1% paten 2BM M eao 634 216 I 208 442 1.302 160 I 7,498 30.6% 2.9% 2. 4 6.1 '17'42%% 2.1 100.0%. ETNIA (29r.) 42.1 13S% 5.6% 2 2.0% 5.0% 16.7% 1.4% 100 32. 16A% a %I tndfgana 2XQ 7Q "I W7 Is$ IrA 78 15AN 1092D .7% 4'.:% 6.1% 7.2% 2.7% 100. No tndfaena I 3AM iom 1.073 794 673 W7 066 082 320 1 ETNIA (a of.) Kliche Z430 772 5ge MG in 189 308 Mg 76 5,97 41.9% 132% 103% 5.3% 3. 3.3% 6.3% .2% 1.4% :00. .Gqchi 2,W4 650 550 203 160 824 70 S= 45 5% i2A% G.&% 3.9% 2.0% Z71% 6.1% ::-7% 1.3% 00. 141 270 41 142% 102% 5.6% 2. 2.7% 6.6% 16.9% 1.5% 100.0% Kaqctdquel 2.443 W GM 326 163 162 3of 944 se 5xo 12.6% 7.3% 1.0% Zf% 6.4% 15.8% 1.8% 100.0% Mam 1 465 272 70 77 203 W 67 3,735 38. 4.7% ma 43.6% 2.0% 92% 3.5% 22% 0.0% 14.6% 1.2% 100.0% Otras indfg Z3iO 604 IU 360 too 130 320 773 62 5= 32.6% ISA% OD% 6.7% 4.9% 6.1% 7. is. Z7% 10D.0 N ndloona :12M .,se 1.0731 704 SM_ OD7 ON iM2 320 1 I.M suo MaSCUUnO 3,134 1.541 oil 631 467 462 723 i.679 251 IDAN 34 3% 15,1% 9.1% 8.3% 4.0% 4.6% 7. 16.8% 2.6% 100.0% FemeN 3,449 1,887 Oa ml 460 477 692 IrM 253 Iii'm 331% le.5% O-VQ 6.7% 4.SY4 4,7% S. 15.5% 2.6 100.0%1 EDAO no 0-6 2X2 1.090 ass 377 369 302 417 1.333 64 7,708 37.7% 142% 0.9% 4.9% 4. 3.9% 0. 17.3% ZI% 100.0% 6.11% 8.1% 4 4A% 6.6 16.3% ZI% 10D.0% 7-12 3j= V4 746 6U 337 31 1,377 00 0,469 1132% 0% 34,4% 4A% 01% 9 3.6% 4A% 7. 15.2% 2.1% 102.0% 13' 18 3.IBO I, 337 627 WS 327 4017 672 IA12 193 QXQ 32. Is 8.9% 7.0% 4.4% 610% 7. 10.4% 2.8% 100.0% 19 -24 3.723 1.734 1,0117 793 499 684 B54 1.055 320 imi 33. 1610% 19.1% :,I% 4.9% 7 II 11 "I 25-59 3A24 1 046 594 863 1,795 313 i,647 1 30. 25,4% 1Z% .9% 5.1% 4 5 2.7% i0o.0% -60 3.919 3:2401 ":O., I 6,47 GIs 701 I Annex 4, Page 17 Table A4.17 - Consumption Patterns: Rural Annual consumption levels (per Capita, in Qutne)Percent of total annual consumption per capita____ Rural F. - fI IE IMHWl . gft TIMd Ot Tw F -TOTAL 1.600 383 366 107 8D Q6 179 667 23 3.660 61.6%l 10.4 9. 2.0% 2.2% 2 6% 490% 16 660% 123.0 P0O3REZA Exttmam 843 144 190 19 19 9 20 20 3 1,407 6709% 10.0 12.7% 1.3% 123% 0.9 1.9% 14.0 02% 10003 Pable (todOos 1.366 230 262 01 41 31 7 363 8 2.460 66.6% 0.7 10.6% 2.1% 1.7% 1.3% 3.1% 14.7 0.3% 123.6% No pobre 3.328 906 62n 272 192 207 400 1.124 .67 7.182 46.3% 11.3% 0.71% 3.6% 2.7% 4.0% 6.7% 16.7 0,0% 123.0 OUINTIL 1 913 ISO 162 23 21 9 31 219 3 1.671 68.1% 10.1% 12.3% 1.6% 1.4% 0.9% 2.2% 94.0 0.2% 1 2 1.917 264 261 02 43 33 76 372 0 2.633 67. 0. 10.7% 2.0% 1.0% 1 2.0% 14.1% 0.3% 123.0.7%2.0% 1.6%12% .8% 4.1%1"IO 3 2,116 373 364 O0 70 70 163 616 16 3.690 64.0% 9.6 0.1% - 2.6% 2.0% I.8 4.2% 10.9 0.4% 123.0 4 2.987 630 634 172 120 177 327 037 38 5.030 60.0% to. 0%.0% 2.0% 2.2% 2.6% 6.6% Me6 06.% 103.0 5 4,76 1 1.462 1.623 624 406 701 0.62 I.869 167 112.0541 30.0% 12.1% 0.0% 5.2%1 3.4% 6.0%1 0.6% 156.6 1.1' I23.0 AREA Urbana . o. o. o,a nror 'Ao. r oro Ruralo 1 060 362 356 107 so 96 170 667 23 3.666 5610 10.4 . .7% 2.0% 2.2% 2.J 4.0% 16.J 0.0% 123.0 REGION MetropoUitana 2.619 821 608 202 100 240 641 1,017 70 4.366 4123% 12.9 0.3% 4,0% 2.1% 3a8 9.6% 14.0 1.2% 123.0 Norte 1.381 190 273 02 48 49 70 366 0 2.402 66.3% 7.7 11.1% 2.1% 1.0% 2.0 2.9% 15.7% 0.4% 123.0% NoroItente 2.413 670 442 110 104 141 202 702 26 4.701 60.4% 1.9 6.2 2.3% 2.2% 3.0 6.0% 14.0 0.6% 123.2% Su6211ent0 ;600 239 344 73 70 64 101 826 23 3539 63.0% II 0.7% 2.1% 2.0% i.71 4.3% 14.0 06.% 100.0 0entfa1 2.36 610 403 144 120 116 220 676 40 4.404 40.0% 11.4% 0.% 3. 2.0% 2.6% 6.1% 16.2 0.9% 123.0 Suroccddente 19080 362 346 Ise 73 117 164 664 21 3.650 51.4% 9.9 0.6% 2. 2.0 3.2% 4.6% 16.1% 0.9% 123.0 Norooddente 1.523 243 291 40 66 37 l06 400 7 2.900 66,0% 9.0 I10.% 1.0% I.0 1.4% 2.9% 182 0.2% 123.0 Peten 19W7 1 277 1 262 6 46 62 116 613. 13. 68.% 0.3 0.4% 2.0% 1.3% I.9 2.6% 16.4 0.4% 123.0 ETNIA (2gr.) tndbgeoa 1.621 201 307 77 so 66 124 436 131 2.804 S3.1% --. 10.7% 2.7% 1.7% 2.0 4.3% 16.2 0.0% 123. No lndigena 2.283 402 406 130 117 140 237 468 37 4.520 66.4% 10.9 0.0% 3.1% 2.6% 3.1% 0.2% 16.2 0.71% 123. ETNIA (6 gr.) K'iche 1.712 359 336 139 60 100 162 643 26 3.434 49.0% 10.6% 0.0% 4.0% 2.01% 3.% 4.4% 15.0 0.6% 13 O'eqchi 1,420 166 274 43 30 21 66 383 6 2.433 66.0% 6.6% 11.3% 1.0% 1.6% 0.0% 2.7% 14.1% 0.2% 13 Kaqdhrquel 1,660 304 370 123 94 77 240 636 30 3.417 45.0% 11.6% 11.1% 2.0% 2.% 2.2 7.3% 15.7 0.0% 123 Mam 1.373 240 223 62 33 49 go 362 8 2.624 66.4% 10.1 10.9% 2.6% 1.31% 1.9 2.2% 14.4% 0.3% 160 Otroslodlg 16~~~.271 23320 43 22690 367 6 268 609 .%1.2% 1 .0% I3 1 . 2.9 14.1% 0.2% 1230 = fgandl a 223 4I41 3 112 120 237 40 33 4,529 50.4% 10. 0.9% 31% 20 3.1% 6. 1.07% 123.0% SEXO Mascuino I912 86354 103 at 601 101, 661 24 3.723 61.6% 12.4% 0.e% 2.0% 2.2% 2.7% 4.0% ¶6150. 12. FemeOlno ~~~~I.1 3ee 80 306 160 76 02 176 66 3 362 61.4% 10.0 9.9% .01% 2.2% 2. .9% 10O 4.6 13.0% EWAD 0-6 16663 307 208 62 66 se 133 476 16 3.07 64.0%1 10.0% 2.7% 2.1% 2.1 1.9% 4. 16.6% 0.0% 123.0 7-12 1790 331 303 118 67, 90 167 019B 3.4 .9 9. 0.0% 2.% 2.0% 2. 4.09% 16 8.% 123.0 13-It8 1.6884 367 342 163 67 114 187 670 22 3.607 10% .0.%4.1% 1.0% 31 6.1% 15.4% 0.0% 130 19-24 1.9060 460 382 112 60 IIl I90 611 26 3.869 603 10.0 .8% 2.0% 2.3% 2. 4.9 16.7 0.7% 103.0% 25-50 2. 040 439 402 120 96 123 223 627 28 4.1z 90 14.99 2.9 2.3% 2.6.4% 1563 4.7%, 123.0 2=60 ~~~~~~~2.20 Su2 623 72 128 96 166 669 34 .48 62 13 113 1.6 2.0% ~ 2.2 2.0% 12 8% 1.0 Annex 4. Page 18 Table A4.18 - Consumption Patterns: Indigenous Annual consumption levels (per capitali, in Ouetzales) Percent of total annual consumption ercapita TLOTAL 1.721 399 371 136 77 el 170 4 9 301 47 .% 2%246 0 .% 100 POBREZA Extrema 6l6 144 ISO 20 1 6 6 30 205 3 1,433 071i% 10. 13.1% 1.4% 1.3% 0.6 2.1% 14. 0.2% 106. Pbre (todos) 1.200 244 27 63 36 30 77 246 6 3376 64.0% 10.2 11.0% 2.6% 1.6% 1 3% 16.1 0.3 10D. Nopobr 3,111 691 660 307 202 242 4860 1.130 66 7,208 433% 12.4% 6.4% 0.114 2.8% 3.4 6,7% 15.7% 1.4% 160.5% QUINTIL 1 606 107 164 20 30, 6 33 223 3 1.604 57.1% 10.1% 12.5% 1.7% 1.3% 0.0 2.1% 14.3% 0.2% 160.0% 2 1.450 26 266 67 41 34 61 366 6 2.626 00.1% 10.1% 11.3% 2.0% 1.6% 1.3 3,1% 14.7 0.3% 160.0% 3 1.613 410 420 136 76 71 ISO 633 16 3,661 49.0% 10.0 10.6% 306% 2.0% 1.6 4.7% 164% 0.5% 160.0% 4 2.769 640 063 216 141 166 342 610 64 5.630 47.3% 11i.0% 10.0% 3.7% 2.4% 2.6 5.6% IS.7 1.1% 106.0% 5 4.430 10066 1.020 833 394 499 9441 1.824 213 1161 37.5% 14.0% 60% 7.0% 3.3%1 42% 6.0%l 16.4% 1.6%1 10.0 AREA Urbana 2.309 742 001 307 100 154 322 666 76 6.496 42.1% 13.5% 10.2% 0.0% 2.6% 2.8% 0.0% 10.7% 1.4% 160.0% Rural 121 281 307 77 00o6 124 430 13 2.664 63.1% 9.8 10.7% 2.7% 1.7% 2.0% 4.3% 162% 2.5% 10D.0% REGION Metropottana 2.660 616 003 309 208 120 040 666 63 6.363 41.0% 14.4% 10.4% 4.7% 3.3% 2.0% 0,0 13,9 4.8 100.% Norte 1.306 234 261 73 47 42 72 417 16 2.076 642% 6.7% 11.3% 2.6% 1.6% 1.0 2.6% 162 0.6 100.0% Noror0ente 16800 296 303 40 so 36 127 446 10 3,161 54.6% 9.3% 11.0% 1.4% 2.2% 1.1% 4.0% 14.1% 0.3% 100.0% Surortente 1.930 354 361 60 73 46 too 049 32 3,033 533% 6.7 9.7% 2.6% 2.0 1-3 6.0% 15.1% 0.6% 160.0% Central 1.90 560 432 263 166 133 271 710 70 4.441 43.6 13.11 0.7% 4.8% 24%& 3, 6.1 16.0% 1.0 i00.0% Suroccidente 1.695 4C0 3732 162 761 102 106 00 30 3.056 47.4% I1.0% 10.0% 4.6% 2.3% 2.% 4.0% 10.7 0.0% 100.0% Noroccklente 1.495 250 259 69 30 44 101 411 12 2.727 04,M% 9.0 1.0 2.0% 1.4 I.8 37% I10.1 0.4 160.0% Peten2.16 277 346 64 el 41 1361 631 14 306 7.5% 7.4% 6.3% 2.0% 1 .0 1.1% 3.% 16.8 0.4% 100.0% MrIA (2gr.) I,nd=gan 1.721 308 371 163 TY 61 170 543 26 3,531 48,7% V. 113Ir. 10.6% rIt 3.6% 2.2% V. .3% 4.9% 15.4 0.S616 0.0 ETNIA Is Sr.) Kiche 1.639 409 419 III 163 132 201 669 38 4.176 4,4.4 11.7 10.0% 4.6% 2.0% 3.1% 4.6 16,0 0.6% 100.0- C1eqchl 1.096 246 323 I70 I06 42 1014 467 16 2.630 64.6% 6.0 11.1% 2.4% 1.9% 1.4% 3.0% 160.% 0.5% 160.0% Kaqchiquel 1.610 066 467 166 I1 I13 30 666 63 4,407 43.3% 12.6% 12.6% 4.3% 2.0% 2.0% 6.6% 16.6% 12% 100 Mamn I5 3060 250 60 30 03 112 352 16 3,604 01. 5% 1 14% 11.0% 3,3% 1,4% 2.0% 4.3% 14.6 0.6% 100 Otros lndig 1,708 329 330 115 70 54 132 460 19 3.223 0f 3.0 102% r 10.4%1 3,0% rt 2.2% f1.7% ft 4.1% r 14.3% 5.6% 16.0% SEXO Masouilno 1.730 394 366 3130 02 es 16 44 30 3,63 46.7% 11. 14.% 3,9% 2.3% 2A4 S.1 150.0% 16 Femenino 1.712 402 374 130 7 77 170 0421 26 3,51W2 49. 11.4% 10.6 3,0% 21 2 .%14010 EDAD i.523 327 320 so 7: 56 132 474 24 3,1r0.% 0 10.0 2.9% 2.4% I.9 4.4% IS.7 0.6% 100.0% 7-12 1.060 317 311 127 01 66 133 474 20 3,67 O.%% 10.3% 10.1% 4.1% 1.7% 2.3 4.3% 10.4 07 100 13-18 1,708 374 303 Igo as 63 164 5S3 34 35681 47.7% 10.5 66 5.5 1.% 2.6 5.4% 15.7% 0.0% 160.0% 119-24 1,642 432 411 I5O so 62 221 630 30 3.900 472% 1 1. 10.01% 4.1% 223% 2A4 5.7% 15O% 0.0% 100.0% 25-59 1.699 476 421 152 53 100 216 G00 30 3,691 4723% 1 1. 10.0% 3.6% 2.3% 2.0 5.4%1 1012 0.6% 10.3 _--_____o ____ 3073 060 532 07 124 66 161 001 27 4.20 4A.% 13.6% 12.4% 2.1%1 2.6 1 .6% 3,6% 13 06 106 s 8 _D§o sX0 U. 8~ - ~ ~ ___ _ _ _ _ _ - . , O .s t0s E 8 Q 0 a xb QI S Q0 S 1 t rC .~~~~~~~~ ~ -. - 9 C,!o>ie 8! e8 Y g } _~~~~ h ___ __ __ _ ___ _ _ _ s ______g88 2QhO O _ _ __ __ _ _ _ ___ _ = . _I2L SSSS S__ -g e ' l S 5 5 O$ $00 5;8 50 - s EC _ $ > l t BSit r BI 1; 4 ;t S i R - ;1 at1 Gt _555 t5 S___ 3 7jf 3ti ji iii jf §i> _. $ i D w A t . >-~~~~~~~~ ~ *~.L ______~__ ~ii.Li.LLLL UIŽL iLL _ i > > > BO > B W > > i L W S S > E L R 3 B C ZZQZZAZ ~ *~0 ,-Z _ iiD iAi9__ 2 imrno S Fa -- . . Z,Z Z Z Q * " " QQ D ^ D Q Q_ _ b_ _C _ . _ . y at SbX R S;S w E 33 S R g - 0 .55 SSS S 5 iiLi2PALL -Is~ ' ° f S S ~S8 88 8" S 88 3 t7>a Annex 4. Page 20 Table A4.20 - Perceptions of Welfare Changes During Last 5 years Household Questionnaire (Pe centage of Total Households and Standard Devia ton) Worse Same Better Total Population 21.3 (0.9) 47.6 (1.1) 31.0 (0.9) Regions Metropolitan 28.8 (2.6) 34.9 (2.9) 36.3 (2.3) North 16.5 (2.8) 58.2 (2.8) 25.3 (2.7) North East 23.9 (3.0) 47.0 (2.1) 29.1 (3.0) South East 23.0 (2.1) 51.6 (2.2) 25.4 (2.1) Central 22.8 (1.6) 48.1 (2.0) 29.1 (1.8) North West 11.7 (1.3) 55.7 (2.1) 32.5 (1.8) South West 17.2 (1.4) 52.2 (2.3) 30.6 (2.2) Peten 23.1 (3.3) 48.8 (2.1) 28.2 (3.1) Rural/Urban Area Rural 17.3 (0.9) 55.6 (1.3) 27.1 (1.2) Urban 26.6 (1.6) 37.1 (1.6) 36.2 (1.4) Household Head Male 20.1 (1.0) 48.5 (1.2) 31.4 (1.0) Female 26.8 (1.7) 43.6 (2.0) 29.7 (1.8) Poverty level Non-Poor 23.6 (1.3) 39.5 (1.4) 36.8 (1.3) Poor 18.6 (1.1) 57.2 (1.4) 24.2 (1.2) Extremely Poor 16.2 (2.0) 65.9 (2.4) 17.8 (1.8) Consumption Quintile First 16.9 (1.8) 64.6 (2.1) 18.5 (1.6) Second 16.4 (1.5) 58.4 (1.9) 25.1 (1.7) Third 23.3 (1.8) 48.0 (2.2) 28.6 (1.8) Fourth 21.6 (1.6) 44.6 (2.2) 33.7 (1.9) Fifth 25.0 (1.7) 34.5 (1.5) 40.5 (1.7) Ethnic Group Indigenous 15.3 (1.6) 55.4 (1.6) 29.2 (1.3 K'iqche 18.3 (2.0) 48.3 (3.2) 33.3 (2.7) Q'eqchi 12.6 (2.7) 56.8 (3.9) 30.7 (4.3) Kaqchiquel 21.0 (3.6) 54.3 (3.7) 24.6 (2.3) Mam 10.2 (2.2) 62.8 (3.6) 27.0 (3.2) Non-Indigenous 25.1 (1.1) 42.6 (1.3) 32.3 (1.3) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Note: Numbers may not sum to total because of rounding. Annex 4, Page 21 Table A4.21 - Perceptions of Welfare Changes During Last 5 years Community Questionnaire (Percentage of Total Communitiesa) Household Welfare Community Welfare Worse Same Better Worse Same Better Total Communities 19.8 45.3 34.9 9.6 39.2 51.1 Regions Metropolitan 20.0 28.9 51.1 6.8 25.0 68.2 North 14.3 58.7 27.0 9.5 39.7 50.8 North East 20.3 37.5 42.5 12.5 45.0 42.5 South East 17.6 47.1 35.3 9.8 39.2 51.0 Central 26.2 43.7 30.0 13.7 45.0 41.2 South West 22.4 43.4 34.2 16.2 32.4 51.3 North West 16.7 48.9 34.4 2.2 42.9 54.9 Peten 20.6 47.1 32.3 5.9 41.2 52.9 Rural/Urban Area Rural 22.6 46.4 30.1 9.3 39.7 50.9 Urban 14.1 43.0 42.9 10.3 38.1 51.6 Main Language K'iqche 16.0 40.0 44.0 0.0 28.0 72.0 Q'eqchi 9.8 68.3 21.9 4.9 31.7 63.4 Kaqchiquel 15.1 42.4 42.4 6.1 39.4 54.5 Mam 18.5 70.4 .11.1 7.7 46.1 46.1 Spanish 22.2 40.1 37.7 12.3 40.3 47.3 Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Note: Numbers may not sum to.total because of rounding. a. The sample for the ENCOVI(2000) was based on dwellings, not communities. Results reported at the community level are not representative of communities in Guatemala. Table A4.22 - Perceptions Causes of Welfare Change, Causes of Poverty and Community Problems Household Questionnaire (Percentage of Total Households and Standard Deviation) Causes Welfare Causes Poverty Community Change Problems Unemployment 21.0 (0.9) 39.2 (0.9) 2.6 (0.2) Public servicesa 2.9 (0.2) 1.4 (0.1) 49.7 (1.3) Lower income/salary/profits 23.8 (0.8) 6.2 (0.4) 0.3 (0.1) Healthb 2.1 (0.4) 0.3 (0.0) 11.2 (0.8) Educationc 0.2 (0.0) 6.6 (0.5) 4.6 (0.5) Violence/alcoholism/family problems 1.6 (0.3) 2.4 (0.2) 8.2 (0.9) Corruption/bad government 0.5 (0.1) 10.3 (0.6) 0.3 (0.1) .High Prices 20.1 (0.8) -10.0 (0.6) 0.5 (0.2) Lack of land/land titling/loss yields 5.9 (0.5) 5.0 (0.4) 1.6 (0.3) Lack credits/high interest rates 0.3 (0.0) 0.3 (0.1) 0.5 (0.1) Lack technical assistance/training 0.2 (0.0) 0.8 (0.1) 0.6 (0.1) Other 21.1 (1.0) 17.1 (0.8) 19.6 (0.9) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4. Page 22 Table A4.23 - Perceptions of Causes of Welfare Change Community Questionnaire (Percentage of Total Conumunitiesa) Household Welfare Community Welfare Better Same Worst Better Same Worst Labor market 23.5 33.5 27.6 6.8 12.9 12.2 Public services' 11.8 5.3 3.4 53.4 25.3 24.4 Sources of income 32.0 22.8 20.7 2.3 9.0 14.6 Healthb 1.3 0.0 1.1 1.4 4.5 7.3 Educationc 3.9 3.4 2.3 13.1 6.2 2.4 Violence/alcoholism/family problems 0.6 0.0 0.0 2.7 1.1 9.8 Government performance 1.3 4.8 0.0 0.4 0.0 2.4 Prices 0.0 11.2 23.0 0.0 0.0 0.0 LandAand titlingAloss yields 0.0 4.4 1.1 2.7 6.7 2.4 Credits/Interest rates 0.6 0.0 0.0 0.4 0.0 0.0 Technical assistance/training 0.0 0.5 0.0 0.0 0.0 0.0 Community Cohesion 7.2 0.0 0.0 0.0 14.6 0.0 Other 17.6 14.1 13.8 16.7 19.7 24.4 Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Note: Numbers may not sum to total because of rounding. a. The sample for the ENCOVI(2000) was based on dwellings, not communities. Results reported at the community level are not representative of communities in Guatemala. Annex 4, Page 23 Table A4.24 - Perceptions of Causes of Welfare Change - Household Questionnaire Divided by Regions (Percentage of Total Households and Standard Deviation) Causes Welfare Change Metropolitan North Northeast Southeast Central Southwest Northwest Peten Unemployment 13.0 (2.0) 22.4 (2.3) 25.7 (3.7) 18.5 (2.0) 18.9 (1.7) 25.1 (2.0) 30.6 (2.3) 17.8 (2.8) Public services' 1.4 (0.5) 3.1 (0.8) 3.8 (1.1) 3.3 (0.8) 1.4 (0.4) 3.0 (0.7) 6.6 (1.2) 3.3 (0.9) Lower income/salary/profits 26.0 (2.2) 27.4 (3.3) 19.9 (3.2) 21.5 (2.2) 25.6 (1.7) 22.6 (1.8) 22.8 (1.8) 21.9 (2.4) Healthb 0.6 (0.4) 1.4 (0.6) 1.0 (0.4) 3.2 (1.0) 1.8 (0.5) 4.2 (1.7) 2.2 (0.1) 2.0 (0.8) Educationc 0.0 (0.0) 0.6 (0.4) 0.1 (0.1) 0.7 (0.4) 0.2 (0.1) 0.3 (0.2) 0.2 (0.1) 0.3 (0.2) Violence/alcoholism/family problems 1.7 (1.1) 1.4 (0.5) 2.3 (1.1) 1.1 (0.4) 1.7 (0.5) 1.5 (0.4) 1.3 (0.4) 1.5 (0.6) Corruption/bad government 0.5 (0.2) 0.0 (0.0) 0.7 (0.6) 1.3 (0.6) 0.5 (0.3) 0.3 (0.2) 0.4 (0.2) 0.7 (0.4) High Prices 20.9 (2.0) 22.1 (2.6) 21.0 (3.1) 19.2 (2.5) 26.5 (1.9) 20.3 (1.8) 9.8 (1.5) 19.0 (2.4) Lack of landAand titling/loss yields 0.0 (0.0) 10.5 (2.3) 7.8 (2.3) 12.1 (2.4) 4.1 (0.9) 5.9 (1.4) 8.5 (1.5) 16.0 (2.9) Lack credits/high interest rates 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.8 (0.4) 0.8 (0.3) 0.4 (0.3) 0.3 (0.3) 0.0 (0.0) Lack technical assistance/training 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.2 (0.2) 0.0 (0.0) 0.6 (0.3) 0.3 (0.2) 0.1 (0.1) Other 35.7 (2.7) 10.5 (1.7) 17.6 (2.8) 17.7 (2.5) 18.4 (1.7) 15.9 (1.9) 16.8 (1.8) 17.2 (2.6) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4. Page 24 Table A4.25 - Perceptions of Causes of Poverty Change - Household Questionnaire Divided by Regions (Percentage of Total Households and Standard Deviation) Causes Welfare Change Metropolitan North Northeast Southeast Central Southwest Northwest Peten Unemployment 38.8 (2.4) 31.2 (3.0) 48.0 (4.2) 43.7 (2.6) 34.7 (1.9) 40.3 (2.0) 38.2 (2.2) 34.7 (3.1) Public servicesa 0.2 (0.2) 1.8 (0.6) 0.8 (0.4) 1.7 (0.7) 1.7 (0.4) 2.0 (0.5) 2.9 (0.6) 0.7 (0.4) Lower income/salary/profits 4.6 (1.0) 8.2 (1.3) 3.7 (1.0) 3.9 (1.2) 8.7 (0.9) 7.8 (0.9) 6.9 (1.0) .5.3 (1.2) Healthb 0.2 (0.1) 0.2 (0.2) 1.0 (0.6) 0.5 (0.2) 0.1 (0.0) 0.2 (0.1) 0.3 (0.2) 0.9 (0.6) Educationc 6.4 (1.1) 9.6 (2.4) 6.7 (2.8) 5.4 (1.3) 6.1 (0.9) 6.4 (1.0) 6.8 (1.4) 9.2 (1.9) Violence/alcoholisrn/family problems 1.8 (0.5) 3.1 (0.8) 2.7 (1.0) 3.5 (0.7) 2.1 (0.5) 2.4 (0.5) 2.6 (0.6) 1.7 (0.5) Corruption/bad government 11.8 (1.9) 8.0 (1.5) 11.1 (2.0) 10.8 (1.7) 13.1 (1.1) 8.1 (1.0) 9.4 (1.1) 12.3 (2.1) High Prices 10.0 (1.6) 16.8 (2.6) 9.0 (1.7) 10.1 (1.4) 10.8 (1.0) 9.6 (1.6) 6.0 (1.1) 12.1 (1.5) Lack of land/land titling/loss yields 0.6 (0.3) 9.8 (2.1) 1.8 (0.6) 7.2 (1.6) 4.2 (0.7) 6.8 (1.2) 9.0 (1.3) 6.1 (1.6) Lack credits/high interest rates 0.0 (0.0) 0.4 (0.3) 1.0 (0.6) 0.3 (0.2) 0.3 (0.2) 0.5 (0.4) 0.4 (0.2) 0.0 (0.0) Lack technical assistance/training 0.0 (0.0) 0.9 (0.4) 1.0 (0.9) 1.7 (0.8) 1.6 (0.6) 1.1 (0.3) 1.0 (0.3) 0.4 (0.3) Other 25.7 (2.3) 9.8 (1.4) 13.0 (2.4) 11.3 (1.2) 16.6 (1.4) 14.7 (1.6) 16.4 (1.7) 16.3 (1.6) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4. Page 25 Table A4.26 - Perceptions of Community Problems - Household Questionnaire Divided by Regions (Percentage of Total Households and Standard Deviation) Causes Welfare Change Metropolitan North Northeast Southeast Central Southwest Northwest Peten Unemployment 1.9 (0.7) 2.6 (0.7) 6.2 (1.7) 2.4 (0.7) 1.9 (0.5) 2.6 (0.6) 2.1 (0.6) 2.8 (0.9) Public servicesa 35.8 (3.1) 57.6 (3.8) 45.5 (5.4) 49.4 (3.6) 50.2 (2.3) 57.2 (2.6) 60.0 (2.4) 59.5 (3.7) Lower income/salary/profits 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.7 (0.3) 1.1 (0.3) 0.6 (0.2) .0.2 (0.1) 0.5 (0.3) Healthb 2.4 (0.6) 11.2 (2.8) 18.5 (5.7) 15.7 (2.2) 10.7 (1.5) 13.4 (1.9) 17.1 (2.0) 12.0 (2.1) EducationC 3.3 (1.0) 4.1 (1.2) 3.4 (1.0) 4.2 (1.4) 5.2 (0.8) 7.5 (2.4) 2.4 (0.5) 3.3 (1.2) Violence/alcoholism/family problems 26.3 (3.1) 0.9 (0.4) 1.2 (0.5) 2.0 (0.6) 6.6 (1.3) 1.4 (0.4) 1.1 (0.3) 1.8 (0.6) Corruption/bad government 0.7 (0.3) 0.0 (0.0) 0.0 (0.0) 0.3 (0.2) 0.1 (0.1) 0.4 (0.2) 0.4 (0.2) 0.0 (0.0) High Prices 0.5 (0.3) 0.3 (0.2) 0.0 (0.0) 1.4 (0.7) 1.1 (0.3) 0.2 (0.1) 0.5 (0.4) 0.2 (0.1) Lack of landfland titling/loss yields 1.5 (1.0) 3.8 (1.0) 1.1 (0.5) 2.1 (0.7) 1.6 (0.4) 1.2 (0.4) 1.2 (0.4) 1.4 (0.7) Lack credits/high interest rates 0.0 (0.0) 0.3 (0.2) 1.5 (0.7) 1.9 (0.7) 0.5 (0.2) 0.3 (0.2) 0.3 (0.2) 1.0 (0.7) Lack technical assistance/training 0.0 (0.0) 0.9 (0.4) 0.2 (0.2) 1.4 (0.8) 0.5 (0.2) 1.0 (0.4) 0.6 (0.2) 0.2 (0.2) Other 27.3 (2.8) 18.0 (2.6) 22.2 (4.7) 18.4 (2.4) 20.3 (1.6) 14.1 (1.7) 14.2 (1.4) 17.2 (2.7) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4, Page 26 Table A4.27 - Perceptions of Causes of Welfare Change, Causes of Poverty and Community Problems - Household Questionnaire Divided by Rural and Urban Areas (Percentage of Total Households and Standard Deviation) Causes of welfare change Causes of poverty Community problems Rural Urban Rural Urban Rural Urban Unemployment 23.1 (1.2) 18.0 (1.4) 37.2 (1.3) 41.9 (1.4) 2.2 (0.4) 3.0 (0.4) Public servicesa 3.5 (0.4) 2.2 (0.4) 1.9 (0.3) 0.8 (0.2) 55.4 (1.7) 42.2 (1.8) Lower income/salary/profits 22.6 (1.1) 25.5 (1.4) 7.3 (0.6) 5.0 (0.6) 0.5 (0.1) 0.2 (0.0) Healthb 2.7 (0.8) 1.4 (0.3) 0.4 (0.1) 0.2 (0.0) 16.1 (1.4) 4.8 (0.6) Education' 0.3 (0.1) 0.0 (0.0) 6.4 (0.8) 7.0 (0.7) 5.1 (0.8) 4.0 (0.6) Violence/alcoholism/family problems 1.4 (0.3) 1.8 (0,7) 2.3 (0.3) 2.2 (0.3) 0.7 (0.2) 18.0 (1.8) Corruption/bad government 0.4 (0.1) 0.6 (0.2) 9.3 (0.7) 11.7 (1.1) 0.2 (0.0) 0.6 (0.2) High Prices 18.7 (1.1) 22.0 (1.2) 11.6 (0.9) 8.0 (0..8) 0.3 (0.1) 0.8 (0.2) Lack of landAand titling/loss yields 9.1 (0.9) 1.4 (0.4) 7.4 (0.7) 1.8 (0.3) 1.8 (0.3) 1.4 (0.6) Lack credits/high interest rates 0.4 (0.1) 0.1 (0.0) 0.5 (0.2) 0.0 (0.0) 0.8 (0.2) 0.0 (0.0) Lack technical assistance/training 0.4 (0.1) 0.0 (0.0) 1.1 (0.2) 0.6 (0.2) 0.8 (0.2) 0.3 (0.1) Other 17.1 (1.1) 26.8 (1.8) 14.2 (0.9) 20.9 (1.4) 15.9 (1.2) 24.4 (1.6) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4, Page 27 Table A4.28 - Perceptions of Causes of Welfare Change, Causes of Poverty and Community Problems - Community Questionnaire Divided by Gender of Household Head (Percentage of Total Households and Standard Deviation) Causes of welfare change Causes of poverty Community problems Male Female Male Female Male Female Unemployment 21.8 (1.0) 17.7 (1.7) 38.3 (1.0) 43.4 (1.9) 2.2 (0.3) 4.2 (0.9) Public servicesa 3.1 (0.3) 2.4 (0.6) 1.4 (0.2) 1.4 (0.4) 50.8 (1.3) 44.7 (2.4) Lower income/salary/profits 24.5 (1.0) 20.9 (1.8) 6.1 (0.4) 6.8 (1.2) 0.4 (0.0) 0.2 (0.1) Healthb 2.2 (0.6) 2.0 (0.6) 0.3 (0.0) 0.6 (0.3) 11.6 (0.9) 9.5 (1.1) Educationc 0.2 (0.0) 0.4 (0.2) 6.6 (0.6) 6.7 (1.0) 5.0 (0.6) 3.0 (0.8) Violence/alcoholism/family problems 0.7 (0.1) 5.4 (1.5) 2.3 (0.2) 2.8 (0.6) 7.7 (1.0) 10.8 (2.3) Corruption/bad government 0.5 (0.1) 0.5 (0.2) 11.1 (0.7) 6.8 (0.8) 0.2 (0.0) 1.0 (0.4) High Prices 19.5 (0.1) 22.6 (1.9) 10.1 (0.7) 9.4 (1.2) 0.6 (0.1) 0.3 (0.1) Lack of landAand titling/loss yields 6.6 (0.6) 2.8 (0.6) 5.4 (0.5) 3.0 (0.7) 1.6 (0.3) 1.6 (0.7) Lack credits/high interest rates 0.3 (0.1) 0.2 (0.1) 0.4 (0.1) 0.0 (0.0) 0.6 (0.1) 0.2 (0.1) Lack technical assistance/training 0.2 (0.0) 0.1. (0.1) 0.9 (0.2) 0.7 (0.3) 0.7 (0.1) 0.2 (0.2) Other 20.2 (1.0) 25.1 (2.0) 16.8 (0.8) 18.4 (1.7) 18.5 (0.9) 24.3 (2.2) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadistica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, commnunication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4. Page 28 Table A4.29 - Perceptions of Causes of Welfare Change, Causes of Poverty and Problems in Community -Household Questionnaire Divided by Poverty Level (Percentage of Total Households and Standard Deviation) Causes of welfare change Causes of poverty Community problems Non-Poor Poor Ext. Poor Non-Poor Poor Ext. Poor Non-Poor Poor Ext. Poor Unemployment 18.5 (1.3) 23.6 (1.2) 24.1 (2.1) 42.0 (1.3) 36.0 (1.4) 32.5 (2.5) 3.1 (0.4) 1.9 (0.3) 2.3 (0.9) Public servicesa 2.1 (0.3) 3.8 (0.5) 5.1 (0.9) 0.8 (0.2) 2.1 (0.3) 1.2 (0.4) 45.7 (1.6) 54.5 (1.7) 58.0 (2.8) Lower income/salary/profits 24.6 (1.2) 23.0 (1.2) 22.8 (2.1) 5.1 (0.5) 7.7 (0.7) 7.0 (1.0) 0.3 (0.0) 0.6 (0.1) 0.6 (0.3) Healthb 1.9 (0.7) 2.4 (0.4) 2.6 (1.0) 0.4 (0.1) 0.3 (0.0) 0.7 (0.3) 7.7 (0.7) 15.3 (1.3) 16.7 (2.1) Education' 0.0 (0.0) 0.5 (0.1) 0.7 (0.4) 7.0 (0.7) 6.2 (0.7) 7.0 (1.2) 3.8 (0.5) 5.6 (1.0) 5.6 (1.3) Violence/alcoholism/family 1.8 (0.6) 1.4 (0.3) 1.7 (0.6) 2.2 (0.3) 2.6 (0.3) 1.4 (0.4) 13.9 (1.5) 1.6 (0.5) 0.5 (0.3) problems Corruption/bad government 0.6 (0.2) 0.4 (0.1) 0.1 (0.1) 11.1 (0.9) 9.4 (0.7) 8.6 (1.2) 0.4 (0.2) 0.3 (0.1) 0.4 (0.1) High Prices 22.5 (1.2) 17.6 (1.1) 14.0 (1.7) 9.2 (0.8) 10.9 (0.9) 10.6 (1.7) 0.7 (0.2) 0.3 (0.0) 0.6 (0.1) Lack of land/land titling/loss yields 2.6 (0.4) 9.3 (0.9) 14.5 (2.1) 2.3 (0.3) 8.1 (0.8) 13.5 (1.9) 0.9 (0.3) 2.4 (0.4) 3.1 (0.7) Lack credits/high interest rates 0.2 (0.1) 0.4 (0.1) 0.4 (0.3) 0.2 (0.0) 0.5 (0.2) 1.2 (0.6) 0.3 (0.1) 0.8 (0.2) 0.4 (0.2) Lack technical assistance/training 0.0 (0.0) 0.4 (0.2 0.0 (0.0) 0.6 (0.1) 1.2 (0.3) 1.9 (0.6) 0.3 (0.1) 0.9 (0.2) 0.6 (0.3) Other 25.0 (1.4) 17.1 (1.3) 13.7 (1.8) 19.0 (1.2) 14.9 (1.0) 14.1 (1.9) 22.7 (1.4) 15.9 (1.1) 12.2 (1.6) Source: World Bank calculations using ENCOVI 2000, Instituto Nacional de Estadfstica - Guatemala. Note: Numbers may not sum to total because of rounding. a. Public services include: water, electricity, transportation, communication services, garbage collection and housing b. Health includes: poor health and insufficient quantity of clinics and hospitals. c. Education includes: Insufficient quantity of schools and teachers. Annex 4. Page 29 Table A4.30 - Perceptions of Causes of Welfare Change - Household Questionnaire Divided by Consumption Quintile (Percentage of Total Households and Standard Deviation) First Second Third Fourth Fifth Unemployment 23.3 (1.8) 24.5 (1.8) 22.1 (1.7) 21.2 (1.7) 16.1 (0.9) Public services' 4.9 (0.7) 3.7 (0.6) 2.7 (0.5) 2.0 (0.5) 2.1 (0.3) Lower income/salary/profits 23.3 (1.9) 24.0 (1.9) 21.7 (2.1) 27.2 (1.7) 22.7 (1.8) Healthb 2.5 (0.8) 2.3 (0.5)