Report No. 20531-HO Honduras Poverty Diagnostic 2000 June 29, 2001 Poverty Reduction and Economic Managernent Sector Unit Latin America and the Caribbean Region With contributions from: IFPRI-International Food Policy Research Institute PRAF-Programa de Asignaci6n Familiar Document of the World Bank CURRENCY EQUIVALENTS US$1.0 = Lempiras 14.5 FISCAL YEAR January 1 - December 31 MAIN ABBREVIATIONS AND ACRONYMS CESAMO Centro de Salud con Medico CESAR Centro de Salud Rural CPI Consumer Price Index DGEC Direcci6n General de Estadistica y Censo ENEE National Electricity Provider ENIGH Encuesta Nacional de Ingresos y Gastos de los Hogares EPHPM Encuesta Permanente de Hogares de Propositos Multiples FHIS Fondo Hondureino de Inversi6n Social GDP Gross Domestic Product GRH Government of the Republic of Honduras HDI Human Development Index HIPC Heavily Indebted Poor Countries IADB Inter-American Development Bank IFPRI International Food Policy Research Institute IMF International Monetary Fund I-PRSP Interim Poverty Reduction Strategy Paper K-WH Kilowatt Hour MECOVI Mejoramiento de las Encuestas y la Medici6n de las Condiciones de Vida en America Latina y el Caribe PRAF Programa de Asignaci6n Familiar PROHECO Proyecto Hondureno de Educacion Comunitaria PRSP Poverty Reduction Strategy Paper UNAT Unidad de Apoyo Tecnico USAID United States Agency for International Development Vice President: David de Ferranti Country Director: Donna Dowsett-Coirolo PREM Director: Ernesto May Lead Economist: Ian Bannon Task Manager: Quentin Wodon T'ABLE OF CONTENTS EXECUTIVE SUMMARY .......................................................................... i CHAPTER I. POVERTY AND INEQUALITY .......................................................................... I A. THIS REPORT IS A CONTRIBUTION FOR THE GOVERNMENT'S POVERTY REDUCTION STRATEGY .......................... I B. THERE IS UNCERTAINTY AS TO WBETHER POVERTY HAS DECREASED IN HONDURAS IN THE 1990S ..................... 4 C. THE IMPACT OF MITCH ON THE POOR. IS LARGER THAN SUGGESTED BY THE EPHPM SURVEYS ........................... 8 D. WHO ARE THE POOR? A STANDARD POVERTY PROFILE .......................................................................... 10 E. INEQUALITY HAS INCREASED, AND SOME GOVERNMENT PROGRAMS CONTRIBUTE TO THIS .............................. 13 CHAPTER II. MICRO DETERMINANTS OF POVERTY .......................................................................... 19 A. REGRESSIONS ARE BETTER THAN PROFILES FOR ANALYZING THE DETERMINANTS OF POVERTY ........................ 19 B. HOUSEHOLD STRUCTURE, EDUCATICON, EMPLOYMENT, AND LOCATION ALL AFFECT POVERTY ......................... 20 C. THE RURAL POOR ALSO SUFFER FROMI A LACK OF ACCESS TO LAND, CREDIT, AND TECHNOLOGY ...................... 31 CHAPTER III. NON-MONETARY INDICATORS AND BASIC INFRASTRUCTURE ........... .................... 35 A. MORE PROGRESS HAS BEEN ACHIEVE]D FOR NON-MONETARY INDICATORS THAN FOR POVERTY ........................ 35 B. POVERTY CAN BE REDUCED BY THE PROVISION OF BASIC INFRASTRUCTURE SERVICES ..................................... 37 C. THE ELECTRICITY SUBSIDY IS NOT WELL TARGETED AND IT INCREASES INEQUALITY ........................................ 42 D. HONDURAS' SOCIAL INVESTMENT FUJND (FHIS) IS A KEY PROVIDER OF SOCIAL INFRASTRUCTURE ................... 45 CHAPTER [V. EDUCATION, NUTRI[TION AND HEALTH .......................................................................... 51 A. IN EDUCATION, THERE ARE BOTH ACCESS AND QUALITY ISSUES, ESPECIALLY FOR THE POOR ........................... 51 B. THE COST OF CHLD LABOR IN TERMS OF FORGONE FUTURE EARNINGS IS LARGE ............................................... 57 C. BEYOND PRIMARY SCHOOLS, INTERVENTIONS ARE NEEDED IN SECONDARY SCHOOLS ............. ......................... 60 D. BEYOND ACCESS AND QUALITY, TH]ERE ARE ISSUES OF COST-EFFICIENCY IN HEALTH ....................................... 63 CHAPTER V. GROWTH .......................................................................... 71 A. INSTITUTIONAL REFORMS ARE NEEDED TO PROMOTE FASTER ECONOMIC GROWTIH ........................................... 71 B. GROWTH IMPROVES BOTH MONETARY AND NON-MONETARY INDICATORS OF WELL-BEING .............................. 76 C. ELASTICITIES OF POVERTY AND SOCIAL INDICATORS TO GROWTH CAN BE USED TO SET TARGETS ......... ............ 80 ANNEX I. ROBUSTNESS OF THE POVERTY TREND AND POVERTY PROFILE ............ ....................... 85 A. TESTS FOR ROBUSTNESS OF THE POVERTY TREND PRESENTED IN THE I-PRSP ...................................................... 85 B. STANDARD POVERTY PROFILE .......................................................................... 86 ANNEX II. METHODOLOGICAL ANNEXES .......................................................................... 99 MA. I MEASURING POVERTY, INEQUALITY AND INCOME GROWTH IN THE SURVEYS ............................................... 99 MA.2 ANALYZING THE IMPACT OF VARIOUS INCOME SOURCES ON INEQUALITY ................................................... 100 MA.3 DETERMINANTS OF GROWTH: CATEGORICAL ORLINEARREGRESSIONS?............................... 101 MA.4 EDUCATION, LABOR FORCE PARTICIPATION, AND WAGES ........................................................................ 102 MA.5 WAGES AND LABOR FORCE PARTICPATION: AREA VERSUS INDIVIDUAL EFFECTS ........................................ 103 MA.6 DOES CONSULTATION IMPROVE IARTICIPATION AND USAGE? ...................................................................... 105 MA.7 ESTIMATING THE COST OF CHILD LABOR IN TERMS OF FUTURE EARNINGS .................... .............................. 106 MA.8 MEASURING THE IMPACT OF GROWTH ON POVERTY AND SOCIAL INDICATORS ............................................ 107 MA.9 WHO BENEFITS FROM AN IMPRO VEMENT IN ACCESS TO BASIC SERVICES? ................................................... 108 ANNEX III. AN EXCEL DIALOG BOX FOR SIMULATING THE IMPACT OF POLICIES ON PER CAPITA INCOME AND THE PROBABILITY OF BEING POOR .......................................................... 109 REFERENCES .......................................................... 123 List of Tables Table ES.1: Share of households in poverty according to the Interim-PRSP, 1991-99 .................. ............................. ii Table ES.2: Median value of the losses, costs, and relief for those affected by Hurricane Mitch, 1999 ..................... iii Table ES.3: Trend in inequality for labor income, Theil, Gini, and Atkinson indices, 1991-99 ............. .................... iv Table ES.4: Marginal percentage increase in labor income with more education by level, men only ......................... vi Table ES.5: Trend in the Human Development Index, 1975-97 ........................................................................... ix Table ES.6: Trend in umnet basic needs, share of households, 1990 to 1997 ..............................I............................. ix Table ES.7: Trend in malnutrition: Incidence of stunting for first grade students, 1986-1997 ............. ...................... xv Table 1.1: Trend in Poverty According to the I-PRSP, Headcount Index, 1991-99 .....................................................4 Table 1.2: Populations at Risk During Natural Disasters Such as Hurricane Mitch .....................................................9 Table 1.3: Mean Value of the Losses, Costs, and Relief for Those Affected by Hurricane Mitch, 1999 .......... ......... I 0 Table 1.4: Trend in Inequality for Labor Income, Theil, Gini and Atkinson Indices, 1991-99 .................................. 13 Table 1.5: Decomposition by Source of Gini for per Capita Income, 1998 & 1999 EPHPM Surveys ....................... 16 Table 1.6: Decomposition by Source of Gini for per Capita Income, 1998 FHIS Survey .......................................... 17 Table 1.7: Decomposition by Source of Gini for per Capita Consumption, 1999 PRAF Survey ............................... 17 Table 2.1: Marginal Percentage Increase in per Capita Income Due to Demographic Variables ............................... 21 Table 2.2: Marginal Percentage Increase in per Capita Income Due to Education ..................................................... 22 Table 2.3: Marginal Percentage Increase in Labor Income with More Education by Level, Men Only ..................... 22 Table 2.4: Marginal Percentage Increase in per Capita Income Due to Employment Variables ................................ 24 Table 2.5: Employment, Underemployment, and Unemployment, Percentage of Heads and Spouses ...................... 26 Table 2.6: Reduction in Poverty from an Increase in Employment, with and without Wage Impact ......................... 26 Table 2.7: Marginal Percentage Increase in per Capita Income Due to Geographic Location ................................... 27 Table 2.8: Marginal Impact of Location on Labor Force Participation and Earnings for Adult Men .......................... 28 Table 2.9: Department Variance in Wages and Labor Force Participation: Area vs. Individual Effects .................... 29 Table 2.10: Marginal Percentage Increase in per Capita Income Due to Migration ................................................... 29 Table 2.11: Impact of Land Titles, Credit, and Assistance on Farm Investments and Technology ............................ 32 Table 2.12: Impact of Land Titles, Credit, and Assistance on Farm Investments and Technology ............................ 32 Table 3.1: Trend in Unsatisfied Basic Needs, Share of Households, 1990-97 ........................................................... 35 Table 3.2: Trend in Human Development Index, 1991-99 ........................................................................... 36 Table 3.3: Access to Basic Infrastructure by Income Group (decile), National, 1999 ................................................ 37 Table 3.4: Access to Basic Infrastructure by Income Group (decile), Urban/Rural, 1999 ......................................... 38 Table 3.5: Increase in Rent due to Electricity, Water and Sanitary Installation, 1998-1999 ...................................... 39 Table 3.6: Estimating the value of Access to Basic Infrastructure Services by Income Quintile, 1998 ..................... 40 Table 3.7: Reduction in Poverty with Universal Access to Basic Infrastructure Services, 1998 ................................ 41 Table 3.8: Who Benefits from Access to Basic Infrastructure Services ..................................................................... 42 Table 3.9: Electricity Consumption by Level and Subsidies, Monthly, 2000 ............................................................. 43 Table 3.10: The Impact on Poverty of Electric Subsidies ........................................................................... 44 Table 3.11: Community Participation in FHIS Projects by Type of Project and Income Quintile ............................. 47 Table 3.12: Comparative performance of the FHIS for consultation, contribution, and usage, 1998 ......................... 49 Table 3.13: Relationships between consultation, contribution, and usage in FHIS projects, 1998 ............................. 49 Table 4.1: Trend in Average Years of Schooling and Enrollment Rates, National, Urban, and Rural ....................... 51 Table 4.2: Public Expenditures for Education: Basic Indicators for 1990-97 ............................................................. 52 Table 4.3: Share of Public Funds Allocated to Education and Health by Sub-sector, 1997 ....................................... 52 Table 4.4: Schooling and Child Labor by Gender, Age, Location, and Income in Poor Municipalities ..................... 54 Table 4.5: Cost of Schooling by Gender, Age, Location, and Income in Poor Municipalities ................................... 55 Table 4.6: Repetition and Absenteeism Rates by Education Levels, and Public Cost Thereof, 1980-97 ................... 56 Table 4.7: Market Share Cost, and Efficiency of Private and Public School by Level, 1997 ..................................... 56 Table 4.8: Cost Effectiveness of School Inputs to Increase Achievement in Primary Schools, 1996 ........................ 57 Table 4.9: Extent of Child and Adolescent Labor by Gender and Urban/Rural Areas, 1990-98 ................................ 57 Table 4.10: Incidence of Labor Among 15 to 19 Year Olds, Honduras and Other Countries, Percentages ............... 58 Table 4.11: Characteristics of Children Participating in a Social Project in Tegucigalpa, Circa 1990 ....................... 58 Table 4.12: Estimates of the Cost of Child Labor in Terms of Forgone Future Earnings, 1996 ................................ 60 Table 4.13: Targeting of PRAF: Per capita Income of Beneficiaries and Non-beneficiaries, 1998 ........................... 61 Table 4.14: Trend in Malnutrition: Incidence of Stunting for First Grade Students, 1986-1997 ................................ 63 Table 4.15: Trend in Selected Health Indicators in Honduras, 1980-97 ..................................................................... 66 Table 4.16: Daily Averages of Ambulatory Care, by Attention Level, 1990-97 ......................................................... 66 Table 4.17: Health Care for Children Under Five Years of Age by Location and Income Group, 1999 .................... 67 Table 4.18: Current Structure of Health Care Services Providers and Expenditures, Percentages ............................. 67 Table 5.1: What are the Causes of and Cures for Poverty? (Opinion Survey, Percentages) ....................................... 72 Table 5.2: Selected Indicators of Competitiveness for Honduras and Central American Countries .......................... 74 Table 5.3: Trend in Social Expenditures as a Share of Total Public Expenditures, 1990-98 ...................................... 76 Table 5.4: Social Expenditures in Honduras and Other Countries as Percentage of Total Expenditures ................... 76 Table 5.5: Elasticity of Poverty Reduction to Growth in Honduras: National, Urban, and Rural .............................. 77 Table 5.6: Elasticity of Non-monetary Indicators to GDP Growth and Urbanization, Levels .................................... 79 Table 5.7: Estimates of the Marginal Gainis in Social Indicators for Various Groups of Municipalities .................... 80 Table 5.8: Targets for Poverty: A Hypothetical Illustration with Growth at 2 Percent per Capita ............................. 81 Table 5.9: Targets for Social Indicators: An Illustration of the Growth and Urbanization Model ............ ................. 82 Table Al.1: National Poverty Lines (Lempiras per month per person) and Adjustment Factor ................................. 87 Table A1.2: Welfare Ratios and Poverty M\easures, National ......................................................................... 88 Table Al.3: Welfare Ratios and Poverty M\easures, Urban ......................................................................... 89 Table Al.4: Welfare Ratios and Poverty Measures, Rural ......................................................................... 90 Table A1.5: Inequality Measures, National, Urban, and Rural ......................................................................... 91 Table Al.6: Poverty Profile and Summary Statistics for Urban and Rural Sectors, March 1998 ............................... 92 Table A1.7: Poverty Profile and Summary Statistics for Urban and Rural Sectors, September 1998 ........................ 94 Table A1.8: Poverty Profile and Summary Statistics for Urban and Rural Sectors, March 1999 ............................... 96 List of Figures Figure 3.1: Progressivity of FHIS 2 investments at municipal and household levels .................... ............................. 46 Figure 4.1: Three Ingredients for a Good Education System ......................................................................... 51 Figure A 1.1: Adjustment for Underreporling ......................................................................... 86 List of Boxes Box 1.1: Beyond monetary poverty: The importance of enabling the very poor to help others ............. ......................3 Box 1.2: Data for Poverty Monitoring and Analysis in Honduras .......................................................................... 5 Box 1.3: Comparing Poverty Levels and Trends in Honduras with Latin America ................................................... 18 Box 2.1: From the Determinants of Poverty to Policy: Suggestions from Latin America .......................................... 30 Box 2.2: What does it mean to be poor? The Story of the Cabreras Family ............................................................... 34 Box 3.1: Allocating Infrastructure Funds on the Basis of Need: Mexico's Experience ................. ............................ 50 Box 4.1: Education and Health Account for the Bulk of Public Social Expenditures .................. .............................. 52 Box 4.2: PROHECO: Education with Community Participation in Remote Areas .................................................... 53 Box 4.3: Malnutrition, Poverty and Community-Based Government Programs .......................... .............................. 64 Box 4.4: PROGRESSA: Gender-Conscious Program for Education, Health and Nutrition ....................................... 68 Box 5.1: SimSIP - Simulations for Social Indicators and Poverty ......................................................................... 83 Acknowledgements This report was written by Quentin Wodon, with contributions from Ihsan Ajwad, Carlos Anguizola, Carletto Calogero (IFPRI), Rafale Flores (IFPRI), Gabriel Gonzalez, Humberto Lopez, Judith McGuire, Marcial Antonio Munguia (PRAF), Bernadette Ryan, Saul Morris (IFPRI), and Corinne Siaens. Valerie Mercer-Blackman and Piritta Sorsa from the IMF also contributed to the report. The peer reviewers were Kathy Lindert and Stefano Patemostro. The Lead Economist for Central America, Ian Bannon, the Sector Manager for Human Development, Helena Ribe, and the Sector Manager for Poverty, Norman Hicks, provided overall guidance. The team expresses its deepest appreciation to the staff of the Ministry of the Presidency, the Ministry of Finance, UNAT and PRAF for their support. HONDURAS: POVERTY DIAGNOSTIC 2000 EXECUTIVE SUMMARY A. THIS REPORT PROVIDES A DIAGNOSTIC OF POVERTY AND WELL-BEING IN HONDURAS 1. This report is a contribution to the Poverty Reduction Strategy Paper (PRSP) that is being prepared by the Government. The report uses household surveys to provide a diagnostic of poverty, human development, and access to basic infrastructure. The report is based on analytical work conducted by a team comprising staff from the World Bank, Honduras' PRAF (Programa de Asignacion Familiar) and IFPRI (International Food Policy Research Institute). The objective of the present report is limited. It provides a diagnostic of the state of poverty and other indicators of well-being instead of suggesting an overall poverty reduction strategy or providing detailed policy reform options. The report is intended as an input for the Government's PRSP within the context of the country's participation in the Heavily Indebted Poor Countries (HIPC) Initiative for debt relief. The key findings are as follows: * Uncertainty as to the change in poverty: Different assumptions used for poverty measurement lead to different conclusions as to the trend in poverty in the 1990s. When no adjustments are made for under-reporting of income in the surveys, poverty appears to have decreased by about 10 percentage points between 1991 and 1999. Wlith adjustments for under-reporting, this gain is less certain. More work will be needed to analyze the trend in poverty in Honduras, but the 1998-99 ENIGH survey should help. Mitch has had a significant negative impact on households, and although relief efforts have been relatively well targeted, they have not sufficed to offset the losses and costs for households caused by Mitch. Nationally, inequality may have increased somewhat in the 1990s. • Complex determinants of poverty: The probability of being poor increases with the number of babies and children, the fact of being frorn an indigenous population, and the fact of having a household head unemployed, underemployed, and/or female. Poverty decreases with education and employment in non-agricultural occupations. Geography also affects poverty and migration is poverty reducing. Programs for rural productivity (e.g., extension and technical assistance) help in reducing poverty. * Progress in non-monetary indicators: Progress is suggested by the Human Development Index which increased from 0.524 in 1975 to 0.652 in 1997. Progress also appears in the fact that the share of all households nationally with no unrnet basic needs increased from 33 to 53 percent between 1990 and 1997. There has been improvement in meeting basic needs in both urban and rural areas, although the level of satisfaction is much higher in urban areas. There is some scope for reducing poverty through access to basic public infrastructure services such as water, electricity, and sanitary installations. But the subsidy for electricity is costly, it is ineffective for reducing poverty, and it increases inequality. * Room for impTovement in education, health, and nutrition: Honduras has increased public spending for basic education and health care, and some progress has been achieved. But the country still lags behind in secondary education, and there are pockets of low primary enrollment. Among the poor, affordability remains an issue for both education and health. The opportunity cost of child labor in terms of forgone future earnings is large. The social investment fund (FHIS) and PRAF bring benefits to poor households, but more in-depth evaluations will be needed to better assess their impact. Although many health indicators have improved, child malnutrition has not and remains an issue. * Impact of growth: Nationally, a growth rate of one percent in per capita income reduces the share of the population in poverty and extreme poverty by two fifths of a point. Apart from reducing poverty, economic growth also improves non-monetary indicators of well-being such as infant mortality, under five mortality, enrollment in secondary education, illiteracy, access to safe water, and life expectancy. Finally, empirical work suggests that the poor may benefit as much if not more than the non-poor from an expansion in health services, and that they share equally in the expansion of education and basic infrastructure. Mitch notwithstanding, macroeconomic stability has improved in the second half of the 1990s, and this has contributed to growth. But Honduras will have to improve its competitiveness, strengthen governance, and reform the state if it is to sustain high growth and reduce poverty. ii B. CHAPTER 1: HONDURAS' RECORD FOR POVERTY AND WELL-BEING IN THE 1990S IS MIXED 2. According to the Interim-PRSP prepared by the Government, the share of households living in poverty decreased by about 10 percentage points in the 1990s. Following standard practice, the Interim-PRSP considers two poverty lines for measuring poverty. The extreme poverty line is the cost of a food basket designed to meet basic nutritional needs. The moderate poverty line multiplies this cost by a fixed factor in order to also take into account the cost of basic non-food needs. Using per capita income as the indicator of well-being, the Interim-PRSP provides measures of the household level headcount indeX of (extreme) poverty which is the share of all households with per capita income below the (extreme) poverty line. Table ES. 1 presents the results obtained in the Interim-PRSP. Nationally, the household level headcount index of extreme poverty decreased from 54.2 percent in 1991 to 48.6 percent in 1999. The household level headcount index of poverty decreased in a similar fashion, from 74.8 percent in 1991 to 65.9 percent in 1999. From 1991 to 1998 (before Hurricane Mitch), the reduction in extreme poverty and poverty is large, at about ten percentage points. There is progress in both urban and rural areas, although the negative impact of Mitch appears to be larger in rural areas. Table ES.1: Share of households in PoV rty according to the Interim-PRSP, 1991-99 1991 1992 1993 1994 1995 1996 1997 1998 1999 National Extreme poor 54.2 47.4 45.1 47 47.4 53.7 48.4 45.6 48.6 All poor (extreme + moderate) 74.8 69.9 67.5 67.2 67.8 68.7 65.8 63.1 65.9 Urban Extreme poor 46.7 39.2 31.6 39.8 40.6 38.7 35.2 35.7 36.5 All poor(extreme+moderate) 68.4 61.6 55.5 62.6 62.8 61 59 57 57.3 Rural Extreme poor 59.9 53.9 55.8 52.9 53.1 66.4 60 55.4 60.9 All poor (extreme + moderate) 79.6 76.5 77.1 71.1 71.9 75.3 71.7 69.2 74.6 Source: Interim-PRSP using EPHPM surveys. 3. However, while the above poverty trend is relatively robust to the choice of the poverty line, it is not robust to adjustments for under-reporting in the surveys. The trend suggested in table ES. 1 tends not to be affected very much by the use of alternative poverty lines. However, the trend is affected by adjustments for under-reporting of income in the surveys. Under-reporting refers to the fact that households may not fully remember and/or declare their income when interviewed for the surveys. When the levels of income in the surveys are scaled up to reflect the per capita GDP in the national accounts, there would appear to be almost no decrease in poverty from 1991 to 1999. This is because the so-called adjustment factor for under-reporting decreases over time, and thereby offsets the increase in mean per capita income observed in the surveys. There is no simple answer to the question of whether the estimates of poverty with and without adjustment for under-reporting are better. While some would argue that the measurement errors may be larger in the national accounts than in the labor force surveys, others would argue the reverse. Some uncertainty therefore remains. In other words, while poverty may have been reduced in the 1990s, it may not have been reduced to the extent suggested in Table ES.1. Beyond the analysis of past poverty trends, it will be important for the future to establish a good poverty baseline with the 1998/99 ENIGH survey. Furthermore, in order to be able to assess the impact of the poverty reduction strategy, it will be necessary to generate wide agreement on a robust methodology for measuring poverty in the future, and to apply this methodology using comparable surveys over time. 4. The latest labor force survey suggests a rather small increase in poverty following Mitch, but this increase is likely to be underestimated for a number of reasons. The increase in poverty due to Mitch obtained by comparing poverty measures in the March 1998 and March 1999 EPHPM surveys in Table ES. 1 is small. The actual poverty impact of Mitch is probably larger for at least three reasons. First, the EPHPM does not capture very well the income of small farmers who are those who suffered the iii most from Mitch due to a loss of their crops. Second, while labor income (a flow) may have been sustained after Mitch, many households suffered from a loss in assets (a stock) which has implications for future poverty. Third, a number of populations at risk such as street children and squatters are unlikely to be well represented in the EPHPM survey, so that the impact of Mitch on the income of these populations cannot be measured with the survey. 5. Many households suffered from a loss in their crop, wage, and/or small business following Mitch. Households also incurred direct costs for health, housing, and food. The 1999 PRAF survey of households living in poor municipalities provides information on the losses and costs incurred by households due to Mitch, as well as the relief obtained by households. More than a third of the households report having suffered a direct crop loss due to Mitch, and the proportion is higher for richer households (40-50 percent are affected) than poorer households (30 percent are affected). Close to one in ten households declares a loss in wages, and the proportion is similar for the poor and the non-poor. One household in twenty declares a loss in its small business or self-employment earnings, with a higher probability of loss at higher levels of income. Apart from these losses in assets and earnings, almost one in three households incurred health expenditures. For housing, the proportion is one in ten. For food and other costs, the proportion is one in seven. Households located along the main roads of a municipality suffered more damage to their houses, probably because the roads served as evacuation conduit for the rain and floods. By contrast, households living in the suburbs were more likely to suffer from a loss in their crop. Among those households who suffered a loss or incurred a cost, some of the amounts were relatively large, at several thousand Lempiras (Table ES.2). There is also evidence that the poor have been hit more seriously by Mitch than the non-poor in the case of housing, simply because their houses were much more fragile and thereby vulnerable to the Hurricane. Table ES.2: Median value of the losses, costs, and relief for those affected by Hurricane Mitch, 1999 Median value of losses suffered among those suffering a loss (Lempiras) Crop Wages Small business Other loss Municipal center 4000 840 3000 500 Suburbs 3000 720 1000 400 Median value of costs incurred among those incurring a cost (Lempiras) Hospital/health Housing Small business Food/others Municipal center 500 800 3675 200 Suburbs 400 500 2000 100 Median value of relief received among those receiving relief (Lempiras) Food Clothing/shoes Sheets/blankets Medicaments Municipal center 125 200 100 125 Suburbs 200 100 50 100 Source: PRAF staff using 1999 PRAF survey. 6. Although the relief after Mitch was well targeted within departments, the amount of relief distributed was small and it tendeid to favor the departments in the center of the country. The proportion of households who received emergency relief is much smaller in the 1999 PRAF survey than the proportion that suffered from Mitch. One in ten households received food relief, one in twenty received clothes or shoes, and one in forty received medicines. The amount of relief received is smaller than the losses and costs due to Mitch. On the more positive side, those who benefited from the relief effort were also those who needed it the most, in that the amount of food aid received was positively related to the proportion of pre-Mitch assets lost due to the Hurricane, and negatively related to the pre- Mitch assets value. But the relief effort has tended to favor centrally located departments. 7. Women and men have not reacted in the same way to Mitch, and this has implications for policy. In a study on the gender dimensions of disaster relief in Honduras and Nicaragua, Shrader and Delaney (2000) argue that women reacted differently to Mitch than men. Women tend to give priority to physical and psychological health, while men are more concerned with protecting their assets. Women iv also tend to be less informed than men about evacuation warnings and how to put aside key assets to prevent their destruction. When the disaster strikes, women more often take care of children, the elderly, and the disabled than men. These asymmetries between men and women make it likely that asset losses have been greater for women. Interestingly, women also tend to be more risk averse than men, so that they are more likely to believe early warning messages regarding disasters and take action. Hence women need to be taken into account to improve disaster warning and planning systems. There is typically an overemphasis on rebuilding public infrastructure after a disaster, and women tend to be excluded from the decision-making process following a disaster. Community priorities and social needs such as psychological counseling also tend to be ignored. 8. Inequality in per capita income has increased somewhat in the 1990s. Beyond absolute levels of income which can be used to measure poverty, well-being depends on relative levels of income which can be measured by inequality. According to relative deprivation theory, individuals do not assess their levels of welfare only with respect to their absolute level of income. They also compare themselves with others. Thus, for any given level of mean income in a country, a high level of inequality has a direct negative impact on well-being. Table ES.3 provides three measures of inequality nationally, and in urban and rural areas. All three indices - the Theil, Gini, and Atkinson indices - generally take a value between zero and one (the indices have been multiplied by 100 in table ES.3), with a higher value indicating higher inequality. Nationally, all three indices have increased between 1991 and 1999, by 3 percentage points on average. In urban areas, there has been a decrease in inequality, but there has been an increase in rural areas. Although this is not shown in the table, the inequality between urban and rural areas has also increased, contributing to national inequality. Also, while some Government programs such as PRAF reduce inequality, others such as electricity subsidies do not. Table ES.3: Trend in inequality for labor income, Theil, Gini, and Atkinson indices, 1991-99 NATIONAL URBAN RURAL Theil Gini Atkinson Theil Gini Atkinson Theil Gini Atkinson May-91 61.69 55.31 57.55 58.80 54.26 58.48 50.09 50.59 51.85 Mar-92 61.50 55.07 56.56 56.59 53.69 57.39 48.54 48.82 49.41 Mar-93 67.38 56.36 58.01 60.81 54.65 58.13 59.48 51.56 52.33 Oct-94 65.27 57.12 56.89 58.75 54.22 55.11 62.26 55.95 54.28 Mar-95 67.01 57.53 57.40 58.02 53.90 55.81 67.13 56.80 54.49 Mar-96 71.02 58.33 59.20 62.02 54.18 57.51 59.26 54.40 52.61 Jun-97 62.22 55.32 55.51 53.29 51.64 55.34 62.59 53.92 51.26 Mar-98 69.28 58.96 63.56 54.91 53.18 56.80 72.80 58.79 63.19 Mar-99 64.15 57.76 61.97 51.25 52.05 53.94 58.95 55.23 60.41 Source: World Bank staff using EPHPM surveys. C. CHAPTER II: A LARGE NUMBER OF VARIABLES AFFECT PER CAPITA INCOME AND POVERTY 9. It is more useful for policy to know the impact of household characteristics on per capita income and thereby poverty than to know the probability of being poor of various household groups. Poverty profiles typically give the probability of being poor of various household groups defined according to various characteristics, for example the area in which a household lives or the level of education of the household head. The problem with poverty profiles is that they cannot be used to assess what could be the impact of policy interventions on poverty. For example, the fact that poorly educated households have a higher probability of being poor than better educated households may be in part due to the fact that poorly educated households also tend to live in poor areas where good employment opportunities are scarce. While raising the education level of poor households may help, it may not help them to the extent suggested by a simple poverty profile. Regression analysis is needed to sort out the determinants of poverty and the impact of any one variable on per capita income (and thereby the v probability of being poor) holding constant other variables. The results of such regressions are summarized below. 10. Poverty increases with the number of children in the household. Households with a large number of children have a lower level of per capita consumption, and thereby a higher probability of being poor. By contrast, having a large number of adults in the household helps in most cases to reduce the probability of being poor. The results also suggest that households with younger heads are more likely to be poor, and that urban households whose head has no spouse (most of which are single with no children) are less likely to be poor. The main implication of these results is that policies enabling women to take control of their fertility are likely to reduce the number of children in the household and thus reduce poverty, simply because the resources in the household have to be shared among a lower number of household members. 11. Female headed households have per capita income levels 15 to 30 percent lower than male headed households. The factors leading to female headship differ between urban and rural areas. The negative impact of having a female head is larger in rural areas than in urban areas. Bradshaw (1995) argues that the reasons why women end up being household head in Honduras differ between urban and rural areas. In rural communities, most of the female heads are widowed (this does not take into account cases when a woman is temporarily the head due to the migration of the male partner). Male desertion remains rare, in part because men are often tied to their land (they are reluctant to abandon their wives and children if this entails a sacrifice of their assets.) Another reason for the stability of rural marital unions lies in the "respect" that the man and woman have for each other. This "respect" is in part due to the fact that unions are formalized through religious marriages. At the same time, while the number of female headed households in rural areas due to the departure of the man is smaller than in urban areas, lone rural mothers find it more difficult to support themselves, as indicated in our regressions. In the case of a separation, land rights remain with the man. Once alone, a woman usually remains alone, out of "respect." Younger women, especially those who have separated, have few opportunities apart from migration to urban areas or a return home to live with their parents. In urban areas, male desertion is more common. This can be attributed among others to the lack of ties to the land and the lower importance of the notion of "respect". Female-instigated separation is also more common, at least when the woman has a source of income, in part because the stigma attached to female headship is lower in urban areas. The upshot is that while female headship may be less common in rural areas, it is harsher there. 12. The gains from education are suLibstantial. A household with a head having attended university has twice the expected level of income of an otherwise similar household whose head has no education at all. Completing secondary schooling brings in a 70 to 80 percent gain versus no schooling (Table ES.4). Completing primary schooling brings in a 30 to 40 percent gain versus having no education at all. There are no large differences in the gains for the head in urban and rural areas despite the fact that there may be more opportunities for qualified workers in urban areas (the only systematic difference is at the university level). The gains from a well educated spouse are also large and similar in urban and rural areas, but they are smaller than for those observed for the head. This is not surprising given that the employment rate for women is lower than for men for all levels of education, so that women use their education endowment less than men. Another explanation could be that there is gender discrimination in pay. Some evidence of discrimination against women was found by Bedi and Born (1995) using data for 1990. Results from wage regressions confirm the large impact of education, and the higher gains associated with higher levels of schooling. For example, in urban areas in 1999, an increase from 6 to 7 years of schooling generates an increase in labor income of 9 percent, as compared to 14 percent from 15 to 16 years of schooling. The structure of these gains is similar to that of other Latin American countries, with the marginal gains increasing with the education level. The gains have remained stable over the decade, but they are two percentage points higher than in other Latin America countries (Wodon, 2000). vi Table ES.4: Marginal percentage increase in labor income with more education by level, men only Urban Rural 1989 1992 1996 1999 1989 1992 1996 1999 6to7yearsofschooling 0.12 0.10 0.09 0.09 0.13 0.10 0.11 0.11 9to IOyearsofschooling 0.13 0.11 0.11 0.11 0.15 0.12 0.12 0.14 12to 13yearsofschooling 0.15 0.13 0.12 0.13 0.17 0.13 0.13 0.16 15to 16years of schooling 0.16 0.14 0.14 0.14 0.18 0.15 0.14 0.18 Source: World Bank staff using EPHPM surveys. 13. While a better education clearly helps in escaping poverty, it is not enough if only one household member is working. The higher the education level, the higher the future streams of income. More experience also generates more income. However, it can be shown that over the life cycle, one working adult with primary or even secondary education is not enough to help a household emerge from poverty when a typical increase in family size is taken into account to estimate the poverty line (to compare the projected earnings with the poverty threshold, one needs to multiply the per capita poverty line by the number of persons in the households after a marriage and the birth of children). In other words, in both urban and rural areas, one salary typically does not enable a household to emerge from poverty unless the education level of the working adult is very high. This is why it is important to improve employment, training, and earning opportunities for women. At the same time, there will be a limit to the increase in the labor force participation of women observed over the last decade, so that this increase cannot be the base of a long term sustainable strategy for poverty reduction in Honduras. 14. The inability to escape poverty with only one wage earner does not imply that measures such as minimum wages are useful and beneficial for the poor. Following Hurricane Mitch, inflation reached 12 percent in the first half of 1999. This led the Government to increase the minimum wage by 25 percent as of July 1, 1999 to 45.20 Lempiras per day. With fringe benefits, this translates to 52.73 Lempiras per day. In January 2000, the minimum wage was increased by another 6 percent. In principle, the impact of minimum wage legislation on poverty is uncertain. On the one hand, those who benefit from a minimum wage may enjoy higher salaries, and this may lead to lower poverty. On the other hand, if the level of the minimum wage is higher than the marginal productivity of some workers, these will lose their employment, which may increase poverty. Assessing the impact of Honduras' minimum wage on poverty goes beyond the scope of the present study, but an analysis of wage data suggests that the minimum wage is only partially binding in both the formal and informal sectors, and this would suggest little impact on poverty. The main problem with the minimum wage is that it ends up being fiscally costly because of its ripple effects on the pay of, among others, teachers and physicians. That is, increases in the minimum wage may come at the expense of scarce budgetary resources which could be used for poverty reduction programs. 15. Employment patterns have a large impact on per capita income and thereby on poverty. Having a head searching for employment has a large negative impact on per capita income. The impact is also negative if the spouse is searching foT a job. Having a head or spouse seriously underemployed (i.e., working less than 20 hours per week) also reduces expected per capita income, by 30 percent. The impact of milder underemployment (i.e., working between 20 and 39 hours a week) is weaker. By contrast, households with a head not working (not in the labor force) have higher levels of income, which suggests that those heads who are not in the labor force can afford not to be working. Having a head working in the construction, commerce, or transport sector brings in a gain in per capita income of about 30 percent as compared to working in agriculture. But households with heads working in services and mining, manufacturing, and electricity do not do much better than households with heads in agriculture. In the case of services, this may be because a large part of this sector consists of informal and low paying jobs. Self-employment versus salaried employment is good in urban areas, and bad in rural areas. The difference could be explained by the fact that in urban areas, the self-employed include a larger number of vii professionals. Being an employer also generates a large gain in per capita income, while unpaid family work is associated with poverty. There is no systematic gain from being employed in the public sector as opposed to working in the private sector, but working in a large firm brings a gain in per capita income. 16. Reducing poverty through labor markets will require interventions not only on the quality of the jobs available, but also on the qualifications of workers, and it is difficult to disentangle both. A study by IPEA (1999) quoted in the Interim-PRSP argues that the key problem in the labor market is the quality of the jobs available. The study states that 84 percent of the difference in labor income between Honduras and other Latin American countries is due to the lower quality of the jobs, the rest (16 percent) being due to the lower quality of the workers. It is not clear what must be understood from the above. While there is no doubt that the quallity of many jobs in Honduras is low, the lack of qualification of the labor force may well be at the root of the problem. If it is, the policy option would be to improve the qualification of the workforce, rather than that of the jobs. There are also other factors affecting the quality of the jobs available in Honduras. The key message here is that one should be careful before interpreting the IPEA decomposition as implying that the qualification of the workers is not a key issue. 17. Apart from the quality of the jobs available and the qualification of the workers, the lack of work remains also a problem for 10 to 20 percent of households. The economy has created many jobs in the 1990s, and this has helped in the absorption of a higher number of women in the labor force both in absolute and in percentage terms. However, the fact that the rate of unemployment is low by international standards is in part due to the fact that the poor simply cannot afford not to be working. Moreover, one out of ten households has a head thait is either unemployed or underemployed and willing to work more. The same is true for household spouses. The lack of work remains a problem, and this problems has risen with Hurricane Mitch. There is no easy policy answer to the problem of unemployment and underemployment, but some attention should be brought to the issue. More employment opportunities would not eradicate poverty, but they would help to reduce poverty, provided the rise in employment is demand driven (i.e., driven by growth) and pro-poor. 18. Geographic location and migration also have an impact on income. There are substantial differences in per capita income between departments that are not due to differences in the characteristics (such as education) of the households living in the various departments. Apart from Atlantida, Cortes does well. Comayagua, Choluteca, Intibuca, Lempira, and Yoro tend to be poorer. The importance of geographic location is confirmed by an analysis of wages and labor force participation. Importantly, the impact of location on labor force participation has a sign opposed to the impact of location on earnings. This suggests that labor force participation is not much of a choice. In poorer departments, labor force participation is higher out of necessity. The importance of geography tends to justify policies targeted on poor areas, such as providing better infrastructure in poorer areas or accompanying migration from some areas to others. Migration tends to raise per capita income. Individuals living in households where the household head has migrated since his/her birth have a level of per capita income about 5 to 15 percent higher than other households. Migraltion over the last five years also tends to increase income. 19. Land titling programs have been implemented in Honduras to improve land security for the poor. It has been argued in the dlevelopment literature (e.g., Lopez and Valdes, 1997) that insecure property rights are a source of production inefficiency, due to a disincentive to invest in land that is not securely held, and to credit constraints that small farmers may face (without a legal title, they cannot offer their land as loan collateral). In Honduras, during 1983-94, USAID funded a large land-titling program for small farmers (Proyecto de Titu,acion de Tierra para los Pequenos Productores). The percentage of farmers with legal land titles increased from 11 to 56 percent during this period. According to the Instituto Nacional Agrario, the program was to benefit small to medium sized producers by: (i) granting them more secure property titles, and thereby encouraging higher investments; (ii) providing collateral to improve access to credit; and (iii) providing technical assistance. An additional benefit would be that secure titles would facilitate land transactions and thus improve the functioning of rural land markets. viii 20. There is no consensus on the overall impact of land titling programs in Honduras, but the evidence that exists confirms that much more than land titling is needed to ensure a positive impact on small farmers. Lopez (1996) suggests that the USAID program raised the income of farmers significantly by generating higher investments, especially in coffee trees and coffee drying patios. Other studies, however, point to the importance of complementary factors. Jansen and Roquas (1998), relying on qualitative methods, argue that the impact of the land titling program was limited, and appears to have triggered new sources of land conflict. The problems identified by Jansen and Roquas point to the importance of an appropriate legal framework, and transparent implementation and enforcement mechanisms, including a fair and expeditious judicial system. A study by Larso and Palaskas (1999) covering 235 farms (177 farms with titles in Santa Barbara and 58 farms without title in Ocotopeque) points to the importance of technical assistance and access to credit. The authors argue that land titling has a larger impact on farmers with access to markets, with the means to take advantage of these markets, and with tenure insecurity before titling. While technical assistance matters for the both the adoption of better technologies and the investment in new coffee trees, land titling has a positive effect only on investments in new coffee trees. The lack of impact of titling on access to credit suggests that while titling can, in principle, help smaller farmers by providing collateral, as in the rest of Latin America, small farmers in Honduras rarely have access to formal credit. Rural credit markets in Latin America tend to operate as small clusters of highly localized borrowers and lenders who know and trust each other, as a result of which little or no collateral may be required on loans (Lopez and Valdes, 1997). 21. The evidence on the impact of technical assistance is also mixed. Lopez and Valdes (1997) find that in Honduras (as in Chile and Colombia) technical assistance has no significant effect on per capita income. But Martin and Taylor (1995) argue that technical assistance can help, although the way through which households learn about extension is key. The authors examined the impact of multiple media in Honduras for providing extension services in order to promote a range of technologies that varied across crops and regions. The data used for the analysis were gathered in 1990 in the Comayagua region. Information materials were distributed to farmers using television, pamphlets with self explanatory illustrations, and radio. The study identifies the adoption rates for the new technologies as a function of the primary learning source for the farmers. The authors find that having a personal contact with experts is important in promoting new technology, in that learning through Govermment and sales people has the highest impact on the probability of adoption. Learning from a friend about the new technology leads to adoption only for commercial farmers. Learning through a pamphlet or through the radio does not lead to adoption (while a personal contact with an expert is more likely to lead to adoption, the cost of this information strategy is also higher). The authors also suggest that TV announcements may help through a multiplier effect on the impact of personal contact with friends or with Government and sales experts. D. CHAPTER III: NON-MONETARY INDICATORS HAVE IMPROVED MORE THAN POVERTY 22. Honduras' Human Development Index (HDI) has improved, with the HDI being slightly above expectations given the country's per capita GDP level. The HDI is a weighted sum of three indices dealing with life expectancy, educational attainment, and per capita GDP. The three indices are given equal weights of one third. Table ES.5 provides the trend in the HDI between 1975 and 1997, using data from the Human Development Report 1999. Honduras is compared to other countries that are eligible to participate in the HIPC initiative (Honduras, Bolivia, Guyana, and Nicaragua), and to its Central America neighbors. Honduras has improved its HDI, from 0.515 in 1975 to 0.641 in 1997, but the progress has not been the same in all areas. For example, life expectancy has improved faster than GDP per capita. In 1997, the performance of Honduras is broadly similar to that of other PRSP countries in Latin America, but remains below the level reached by Honduras' neighbors. Honduras's HDI ranking is above its GDP ranking, which suggests a comparatively good performance in health and education. ix Table ES.5: Trend in the Human Development Index, 1975-97 PRSP countries in Latin America Central America countries HO BO GUY NI All CR ES GU PA All HDI index 1975 0.515 0.524 - - 0.520 0.741 - 0.517 - 0.629 1980 0.563 0.558 - - 0.561 0.766 - 0.552 - 0.659 1985 0.595 0.584 - - 0.590 0.784 - 0.563 - 0.674 1990 0.616 0.611 - - 0.614 0.787 - 0.588 - 0.688 1997 0.641 0.652 0.701 0.616 0.653 0.797 0.674 0.624 0.791 0.722 Components of 1997 HDI Life expectancy at birth 69.4 61.4 64.4 67.9 65.8 76.0 69.1 64.0 73.6 70.7 Adult literacy rate (%) 70.7 83.6 98.1 63.4 79.0 95.1 77.0 66.6 91.1 82.5 Combined gross enrollment 58 70 64 63 64 66 64 47 73 63 Real GDP per capita 3,330 2,880 3,210 1,997 2,855 6,650 2,880 4,100 7,168 5,199 Life expectancy index 0.74 0.61 0.66 0.71 0.68 0.85 0.74 0.65 0.81 0.76 Education index 0.66 0.79 0.87 0.63 0.74 0.85 0.73 0.60 0.85 0.76 GDP index 0.52 0.56 0.58 0.50 0.54 0.70 0.56 0.62 0.71 0.65 HDI and GDP ranking GDP ranking 117 108 101 121 112 61 108 85 56 78 HDIranking 114 112 99 121 112 45 107 117 49 80 Source: UNDP (1999). 23. Progress has also been achieved in reducing unmet basic needs. Honduras' method for measuring unsatisfied basic needs described in ]Libro Q (1994), uses six indicators. An urban household is said to have an unsatisfied basic need for water if it does not have access to drinking water within the home or property. In rural areas, the household must have access to drinking water through a well or pipe. For sanitary equipment, in urban areas a household must have access to a sanitary system which is not a simple pit. In rural areas, a household must at least have access to a simple pit. In terns of schooling, in both urban and rural areas, the children of primary school age must be enrolled. In terms of the capacity of the household to make a decent living, in both urban and rural areas, the household head must have at least three years of primary education and be working; if this is not the case, there must be at least one person working for every three household members. In terms of crowding, in both urban and rural areas, there should be no more than three people per room, bathrooms not included. Finally, for housing materials, in urban areas the house should not be ad hoc or built with debris, and it should not have dirt floors; in rural areas, it should not be ad hoc or built with debris. In Table ES.6, which is reproduced from the Interim-PRSP, the share of all households nationally with no unmet basic needs increased from 33 to 53 percent between 1990 and 1997. There has been progress in both urban and rural areas, although the level of satisfaction is higher in urban areas. The FHIS has contributed to the improvement in satisfied basic needs, at least for sanitation and primary education. Table ES.6: Trend in unmet basic needs, share of households, 1990 to 1997 National Urban Rural 1990 1993 1997 1990 1993 1997 1990 1993 1997 No unsatisfied basic needs 33 47 53 50 57 65 20 38 42 One unsatisfied basic needs 25 28 26 24 23 22 26 32 29 Two unsatisfied basic needs 20 15 13 13 11 8 26 19 18 Three or more 22 10 8 13 9 5 28 11 11 Source: GRH (2000). 24. Nevertheless, many among the poor still lack access to basic infrastructure services. Nationally, only 6 percent of the households in the poorest income decile have access to water within their house, as compared to 76.5 percent in the richest decile. One fourth of the households in the poorest decile do not x have water within their property, versus 2 percent in the richest decile. One third of the households in the poorest decile have no sanitary installation, versus less than one percent in the richest decile. Many poor households use holes. Three out of four households in the poorest decile have no access to electricity, versus less than 5 percent in the richest decile. There are also large differences between urban and rural areas. Because of the network nature of many services such as water and electricity, many middle- income households in rural areas have less access to these services than poor households in urban areas. 25. Providing access to basic infrastructure services would reduce the extent of poverty. The value of access to electricity, water, and sanitary installations (as measured through the readiness to pay observed via rents) varies from 4 to 13 Lempiras per month per person for the poor. In absolute terms, the value of access is higher for the rich than for the poor, and this is consistent with the fact that the willingness to pay for these services is higher among the rich than among the poor. But in relative terms, as a percentage of income, the value of access to basic infrastructure services is higher for the poor than for the rich. Providing better access to basic services would help in reducing poverty. 26. There are two large subsidies in Honduras for basic infrastructure services: the first is for electricity consumption nationally, and the second is for bus transportation in Tegucigalpa. The cost of the electricity subsidy was estimated at 259 million Lempiras in 1998, while the cost for the bus transportation subsidy was estimated at 114 million Lempiras. By comparison, the budget for the FHIS in 1998 was 579 million Lempiras, and the budget for PRAF was 188 million Lempiras. The subsidies are thus costly. Both subsidies are in principle self-targeted, but this is not much the case in practice for the electricity subsidy (see below), and data is lacking to make an assessment of the transportation subsidy. The bus transportation subsidy is given to all those who ride certain bus lines in Tegucigalpa. There is self-selection because the buses are not comfortable, so that those who can afford better means of transportation do not use the subsidized buses. Still, because the subsidy is limited to Tegucigalpa where the population is relatively less poor, it remains to be seen how much poverty reduction is achieved. 27. The poverty impact of the electricity subsidy is small in comparison to the public cost. The subsidy for electricity is given to all households who have a level of consumption below 300 kwh per month, and this represents 85 percent of the population with a connection to the grid. There is self- selection in theory because of the consumption ceiling for the subsidy, but the ceiling is too high for the self-selection to work. Even in the zero to 20 kwh consumption bracket, according to the 1999 PRAF survey, half of the households are not poor. Among households who consume more than 100 kwh, more than 80 percent of the subsidy goes to the non-poor. Given that most of the subsidy for electricity is spent on households who consume between 100 and 300 kwh per month, the impact of the subsidy on poverty is small in comparison to its cost. Apart from the fact that the electricity subsidy is not very good at reducing poverty, it also contributes to higher inequality. The electricity subsidy could be reformed so that, for example, only those who consume less than 100 kwh benefit from it. While this may not be easy from a political economy point of view (especially following a recent increase in rates of 16 percent by ENEE), it would bring benefits for the poor if the cost savings are directed to them via other programs. 28. The FHIS provides basic social infrastructure with a focus on education and some level of targeting toward poor rural municipalities. The emphasis of the FHIS is more on the efficiency of infrastructure built than on the generation of employment for the poor. As a result, while the FHIS does generate some employment in poor communities, it should not be considered as providing safety nets through employment creation (it was, however, instrumental in the delivery of emergency safety nets to the communities hit by Mitch). Today, education accounts for over half of FHIS funding (56 percent), followed by water and sanitation, health, and municipal development (12-15 percent each). A small part of the funds (4 percent) is devoted to social assistance and environmental projects. The targeting of FHIS occurs in two steps. First, resources are allocated to municipalities according to population. This takes into account a poverty proxy index so that the greater the poverty level, the more resources per capita are xi allocated to the municipality. Second, resources are allocated within municipalities and decisions made with regards to the exact type of investments to be made. An evaluation of FHIS II by ESA Consultores (1999) suggests that the targeting of the FHIS has improved over time, but it needs to improve further. 29. The demand-driven approach of the FHIS is one of its most important characteristics. Three out of four households are consulted by the FHIS before the implementation of projects, and the proportion is higher for the poor. FH:IS projects match the priorities of households in two cases out of three. Slightly more than one out of two households contributes to FHIS projects, with the type of contribution varying according to projiect type and income level. The most common form of contribution is free labor (34.4 percent), followed by money (16.8 percent), supervision (2.3 percent), material (2.2 percent) and paid labor (1.7 percent). A large majority of households use the FHIS facilities. This can be interpreted as an indicator of community satisfaction even though usage is basically free once the project is implemented. Utilization is almost universal among households in the poorest quintile (93.8 percent), and lower for households in the wealthiest quintile (58.0 percent), which helps for targeting purposes. 30. Promoting consultation and contribution is important because this yields a higher probability of final usage, at least in some cases. In virtually all types of project, FHIS is better at promoting community consultation, contribution, and usage than organizations which have not been identified by households. On the other hand, the FHIS is on par with other organizations that have been identified by the households. These results suggest that while FHIS is not necessarily better at community participation than other agencies which can be identified by households, the program still does a good job at promoting participation overall. This is important because there is some evidence that consultation prior to project implementation increases the contributions made by households at the implementation stage as well as (to a lesser extent) the usage of the facilities once they are completed. There is also some evidence that a higher contribution by households at the implementation stage increases the use of the facilities. Also, before concluding this overview of chapter III, it is important to always keep in mind that programs such as the FHIS (or PRAF, discussed in chapter IV) should not follow an "assistance" model, but should rather enable the poor to emerge from poverty on a sustainable basis. E. CHAPTER IV: PROGRESS HAS BEEN MADE IN EDUCATION AND HEALTH, BUT MORE IS NEEDED 31. Honduras has made efforts to protect public spending on education and health care. As in other countries, health and education accotnt for more than 80 percent of public social expenditures. Other significant social expenditures includes funding for PRAF and FHIS. As a share of GDP, education and health expenditures have remained stable in the 1990s. Within social expenditures, education expenditures have increased somewhat, while health expenditures have remained stable. The good news for the poor is that the share of basic expenditures within sectoral expenditures has increased. As a result, despite a reduction in public deficits and a sustained population growth in the 1990s, basic expenditures per capita have increased. A detailed analysis of public spending priorities is provided in the Public Expenditure Review for Honduras (World Bank 2001). 32. Honduras has been doing better in adult literacy and enrollment in primary education, but it still lags behind in secondary education and there are pockets of low primary enrollment. The number of years of schooling for the population aged 10 and over has doubled over the last 25 years. The rate of illiteracy has been cut in half. 'Enrollment in pre-schools has doubled, and nine out of ten nine year olds is enrolled. While rural areas still lag behind urban areas, they are slowly catching up. Today, Honduras is doing as well as other PRSP and Central America countries for adult illiteracy and enrollment in primary education. However, there are still pockets in poor rural areas where primary school attendance should be increased, and the country still lags behind in enrollments in secondary education. xii 33. The lack of a good transition from primary to secondary school is due to both the high cost of secondary schooling, and the low interest in pursuing an education beyond the primary level. While at least three fourths of children below 14 go to school, only slightly more than one fourth of those aged 15 to 17 do. There are two main reasons for not going to school after 15 years of age. First, some students are simply not interested in pursuing their education. These children appear to consider the pursuit of their studies as irrelevant, which could be worse that wishing to go to school but not going because of a lack of quality. Next, there is an affordability problem, with a third of the children saying that schooling is too expensive or that they must work. Uniforms, materials, registration fees, and also transportation costs in the case of secondary school, make the cost of schooling large for older children. A minority of students receive free books and uniforms, but this is the case more for the poorer children. 34. Apart from the problem of affordability, there is a problem of education quality. Although repetition and absenteeism rates have decreased in primary and secondary education, concerns have been raised regarding the quality of the education system. There are many causes: poor teaching quality, insufficient classroom time, teacher absenteeism, inadequate supervision from the school system and the community, lack of external evaluation, insufficient training, inappropriate curriculum, shortages of books and other teaching materials, lack of funds devoted to investrnents as opposed to teacher pay, etc. In the 1998 FHIS survey of households living in the areas where the FHIS is active, between one fourth and one third of the students say that the reason why they missed school recently was either that the teacher did not show up, or that the school closed. UNAT (2000) has estimated that the total cost of repetition and absenteeism amounted to about 330 million Lempiras per year, or close to 20 percent of the education budget for primary, secondary, and tertiary education. 35. The costs of repetition and absenteeism increase with the education level. At the university level, one third of the budget can be considered as wasted due to high repetition among students (who are unlikely to be from poor backgrounds). World Bank reports typically advocate a shift within education and health away from the tertiary sub-sectors in order to channel more resources to the primary and secondary sub-sectors. The basic idea is to implement cost-recovery mechanisms in the tertiary sub- sectors (which are more in demand among the non-poor) in order to channel more funds to the primary and secondary sub-sectors (which are more in demand among the poor). This strategy could be adopted in Honduras as well since a non-negligible part of public expenditures still goes to the tertiary sub-sectors. 36. Increasing teacher quality may be cost-effective for improving achievement in primary school. Bedi and Marshall (1999) suggest that a five point increase in teacher quality (as measured through the use of an "active and participative" teaching methodology) is associated with a 4 to 8 point increase on student achievement tests. The other important inputs raising student achievement are the student/teacher ratio and the availability of preschool programs. By contrast, textbooks play a lesser role. The authors run simulations to test which measures might be most cost effective in order to increase achievements: reducing the class size (with the cost of doing so depending on the type of school - single teacher, dual teacher, or multi-teacher), increasing pre-school coverage, or increasing teacher training through seminars given by experts (with various assumptions as to the impact of these seminars on teacher quality). The conclusion is that enabling teachers to participate in seminars in order to improve the quality of their teaching may be the cheapest way to improve test scores among students. While these empirical results could be debated, they have the merit of pointing to less traditional and more qualitative ways to improve test scores than the usual quantitative measures such as student/teacher and textbook/student ratios. 37. PROHECO (Proyecto Hondureno de Educaci6n Comunitaria) is designed to meet the education needs of dispersed rural populations and to promote a high level of community participation. In some of Honduras' departments, the population density reaches 200 inhabitants per square kilometer. But in others such as Gracias a Dios and Olancho, the density is only 2.7 and 14.6 inhabitants per square kilometer. It is difficult to provide cost-effective education services in these areas, xiii and it is estimated that 100,000 rural children are not enrolled in primary school. Initiated in March 1999, PROHECO is active in 500 communities. None of the participating communities had a school teacher before implementation of the program. The total number of students served at the end of 1999 by the program was 8,139. PROHECO focuses on pre-schools and primary schools, and it has a bilingual component for indigenous populations. The program works though Community-based Education Associations (AECOs) whose locally elected officials receive training from the program. The state transfers the funding for the program directly to the AECOs which are in charge of hiring the teacher. The flexibility and low cost of the program is illustrated by the fact that only 17 percent of the classes take place in a school. Most of the classes take place in private homes (72 percent), and some take place in churches (7 percent) or in other places (4 percent). PROHECO is being expanded in order to boost rural pre-school and primary enrollment. 38. Almost 400,000 children and adlolescents are working today in Honduras, and there is strong evidence that child labor leads to less schooling. A first problem with child labor is that many children working may be at risk of being hurt. A second problem is that among working children, street children face hard living conditions and can represent a threat to society (Wright and Wittig, 1993). A third and more widespread problem is that child labor reduces the probability of schooling, thereby perpetuating poverty from one generation to the next. Given that the children have only a given number of hours per day for schooling, labor, and leisure, child labor is likely to lead to less schooling. When this is the case, the likelihood that the child will emerge from poverty when he/she reaches adulthood will be reduced since the human capital of the child is reduced. However, because parents can reduce the time available for leisure when a child is working, the substitution effect between work and schooling is likely to be partial only. It turns out that the difference in the probabilities of going to school when the child is not working, and when the child is working varies from 48 percent to 75 percent depending on the sample (rural boys, rural girls, urban boys, and urban girls). The substitution between work and schooling is thus quite large. 39. The cost of child labor in terms of forgone future earnings is substantial. To assess the impact of child labor on children, one can predict future earnings according to various levels of education. The simplifying assumption is that if a child is working, and if this does not enable him/her to go to school, the child completes only six years of schooling, up to age 12. In contrast, if the child is not working, and if this enables him to go to school, the child completes 9 years of schooling. Thus, in the first three years after the completion of primary school, a working child enjoys a benefit because he receives a wage. But for the rest of the child's life, the earnings are lower because of the lower level of education achieved. Even after using a discount rate of 5 percent for the future stream of income, the cost of child labor in foregone future earnings represents from 7 to 30 percent of lifetime earnings depending on the sample. 40. PRAF runs programs in the areas of education, health, nutrition, and income support. PRAF was created in 1990 within the context of the structural adjustment program that was put in place at that time. In 1998, the program had 318,000 beneficiaries. PRAF's seven sub-projects had a total budget allocation of 130 million Lempiras. This makes PRAF the second largest program targeted to the poor after FHIS. One third of PRAF's budget is devoted to the Bono Escolar de Primero a Tercer Grado and the Bono Escolar Ampliado al Cuarto Grado. Started in 1990, the program gives cash stipends of 50 Lempiras per month for ten months and for up to three children per household provided the children go to primary school (first to third grade). In 1998, the program was extended to the fourth grade of primary school. The program is targeted to poor areas using municipal data on child malnutrition and unmet basic needs, and to poor households within these areas. Another third of PRAF's budget is devoted to the Bono Materno Infantil. Started in 1991, the program also gives cash grants of 50 Lempiras per month for 12 months to help for the nutrition of pregnant mothers, nursing mothers and their babies, as well as children below five years of age. The rest of PRAF's budget goes to programs for nutrition, old age transfers, school supplies, and women's micro-enterprise investments. xiv 41. The PRAF stipends appear to be relatively well targeted to those in need, at least in rural areas and nationally. According to data from the 1998 FHIS survey, PRAF's targeting is good in the rural sector and nationally, in that the households receiving support tend to have lower income levels than the households who do not benefit from PRAF's programs. The same is observed at the school level, with the households living in areas where PRAF school supplies are given for free being poorer than households living in areas where the schools do not participate. While the targeting is good, it could still be improved, and PRAF has been working with the International Food Policy Research Institute (IFPRI) in order to find more efficient ways of means-testing the program as of the fall of 2000. 42. As part of a broad evaluation effort, PRAF is currently changing the functioning of the Bono Escolar program, and the new rules will be taking effect in the fall of 2000. With funding from the Inter-American Development Bank and analytical support from the IFPRI, PRAF will be conducting a multi-year evaluation of its Bono Escolar project. To this aim, the program will be operating in a different way as of the fall of 2000. The program will still be targeted to poor municipalities using the Censo de Talla, but four different modules will be put in place in different municipalities. The first module will consist of demand-side interventions like the current Bono Escolar; the second module will consist of supply-side interventions (quality of teachers, etc.); the third module will consist of both demand- and supply-side interventions; finally, the fourth module will consist of no interventions at all and serve as control group. This design will help in identifying which mix of interventions is most successful in raising enrollment, attendance, and achievement. Contrary to what is being done in Mexico for the PROGRESA program (also with analytical support from IFPRI), the PRAF stipends will be available to all households with primary school age children living in the participating municipalities. In Mexico's PROGRESA, there is a second level of targeting within poor municipalities, so that only some of the households participate (this is done through means-testing). The rationale for not targeting PRAF within the municipality has to do with the high levels of poverty encountered in participating municipalities (so that targeting might not bring many savings), and the desire to avoid the potentially negative effects of intra-municipal targeting on social cohesion. This new functioning of PRAF's Bono Escolar project is most welcome, and the lessons learned from the evaluation should be especially valuable. This, however, does not detract from the need to also target the stipends to older children in order to facilitate the transition from primary to secondary school. Again, before concluding this section, as was done in the review of FHIS in chapter III, it is important to always keep in mind that programs such as PRAF should not follow an "assistance" model, but should rather enable the poor to emerge from poverty on a sustainable basis. 43. Malnutrition rates suggest a deterioration in the 1990s. Table ES.7 provides information on the percentage of stunted children in the first grade of primary school (stunting occurs when a child has a height at least 2 standard deviations below international standards). Two observations can be made. First, there has been no progress toward reducing the incidence of stunting between 1986 and 1997. In fact, there is a marked deterioration after 1991, when according to the official estimates, 34.9 percent of all children in first grade were stunted, as compared to 40.6 percent in 1997. (Table ES.6 also provides alternative estimates after data cleaning, including not taking into account children who repeat their first grade, so that they are not counted two years in a row, but the trend is very similar.) Second, children who enter primary school at a later age have a higher probability of being stunted. In 1997, 62.3 percent of 9 year olds were stunted, versus 31.6 percent of 6 year olds. This may be because the parents who do not send their children to school until they are relatively old live in isolated locations a long way from schools, or are less educated. These parents are also more likely to be poor and have malnourished children. xv Table ES.7: Trend in malnutrition: incidence of stunting for first grade students, 1986-1997 Official estimates Alternative estimates 6 year 7 year 8 year 9 year Total 6 year 7 year 8 year 9 year Total olds olds olds olds olds olds olds olds 1986 25.1 36.7 47.2 56.0 39.8 NA NA NA NA NA 1991 24.1 33.9 4:2.7 50.2 34.9 NA NA NA NA NA 1993 26.7 36.6 44.8 52.6 35.5 30.1 42.2 52.5 58.7 NA 1994 27.5 39.8 50.0 59.0 38.1 30.3 45.1 57.4 65.3 NA 1995 27.5 41.2 50.9 58.8 38.6 30.4 46.1 57.9 64.9 NA 1996 28.2 40.7 50.9 59.0 38.0 31.2 46.1 58.4 65.8 NA 1997 31.6 43.2 52.4 62.3 40.6 34.7 48.4 59.8 68.9 NA Source: Censo Nacional de Talla for official tabulations and IFPRI staff for altemative estimates. 44. Except for child malnutrition, Honduras has made substantial progress in health indicators. Twenty years ago, Honduras' perforrnance in health indicators was below the Latin American average. Today, despite a lower level of GDP, the country is on par. The vaccination rate for children under five years of age has increased from 49.6 percent in 1980 to 94.2 percent in 1997. This rate is now above the level reached in countries with a high level of human development as measured by the HDI. The percentage of births taking place in hospitals has increased from 37.5 percent in 1980 to 53.8 percent in 1997. The population without access to safe water has decreased from 41.4 percent in 1985 to 25.1 percent in 1997. Finally, the population without sanitary access has decreased from 42.3 percent in 1985 to 26.2 percent in 1997. However, rural areas still trail far behind urban areas. 45. The progress achieved by Honduras in health indicators is in part the result of a large expansion in the public supply of health services, but this has led to over-capacity in some areas. According to a recent report by the World Bank (1998), the number of area hospitals increased by 129 percent between 1990 and 1996. The number of health clinics with doctors (Centro de Salud con Medico, CESAMO) increased by 21 percent.. And the number of rural health clinics (Centro de Salud Rural, CESAR) increased by 41 percent. But the total number of consultations grew by 24 percent only, and the number of consultations per inhabitant increased only by 5 percent. The increase in the supply of health services has lead to over-capacity in some areas. In many hospitals and CESAMOs, the doctors and nurses are contracted for morning shifts only. A typical CESAR takes care only of six patients a day. 46. While the poor use public health posts (CESAMOs and CESARs) more than the non-poor, many among the poor apparently still do not seek health care. When a child is sick, the poor use CESAMOs and CESARs more than the non-poor who rely on public and private clinics and hospitals. But many of the poor still stay at home or with family and friends rather than seeking care when they are sick. This is true for children, pregnant women, and other household members. The lack of demand for health care by some of the poor need not be due primarily to its cost. Many types of consultations are free and when the consultation is not free, the payment can be waived. Independently of cost issues, there is probably a lack of information anmong the very poor on the benefits of seeking professional care. Providing better information might raise the demand of the poor for health care, with no major difficulty for the health posts to meet the increased demand given their current low productivity and usage. 47. The management of public providers of health services could also be improved. In public health institutions, as noted in the World Bank (1998) report, the selection of the staff for key managerial positions is based on medical background. It does not take into account management skills. Another problem is that the Ministry of Health allocates its budget on the basis of past allocations rather than future needs. This system does not provide incentives for improved performance. There are also problems with budgetary execution due to the lack of flexibility for the reallocation of expenditures between sectors and health centers, and the fact that some budgets are frozen and never become fully available. The purchasing procedures are centralized, and the payments may take several months to be xvi made. The institutional framework has also contributed to the concentration of doctors and nurses in cities. Indeed, because there is no financial incentive for doctors and nurses to work in rural areas and small cities, many choose to stay in large cities where it is easier for them to supplement their incomes with a second job. 48. Despite low coverage, the social security system is not financially sustainable. The Ministry of Health is by far the main provider of health services in both urban and rural areas. The Honduran Social Security Institute (IHSS) is not active in rural areas, and its market share is only 16 percent in urban areas. Private providers are three times as large as the IHSS nationally. Most beneficiaries of the IHSS are located in Tegucigalpa and San Pedro Sula. Social security is financed through contributions that are insufficient because the contribution rate (7.5 percent) is applied to an income ceiling (600 Lempiras per month) which has been frozen for 30 years. This ceiling was originally high, but it is today lower than the minimum wage. Hence the system is not financially sustainable. 49. Apart from reforming its social security system, Honduras will need to implement higher cost recovery for tertiary health care given to the non-poor. Today, the co-payments by households for public health care generate only 1.5 percent to 3.5 percent of the total Ministry budget (World Bank, 1998). The standard fee for an ambulatory consultation (including medicines) is one to two Lempiras. Many services are exempt of any fee (e.g., pre-natal consultations, child development clinics, family planning, sexually transmitted diseases, tuberculosis). The prices for basic services are very low compared to their private sector equivalents, and they are waved for patients who cannot pay. This is all fine for the poor, but the current lack of cost recovery leads to insufficient checks on the consumption of subsidized and costly tertiary treatments by the non-poor. Without price signals, the system does not have incentives for health care providers to expand the most valuable services for people. As is the case for tertiary education, better cost-recovery among the non-poor for tertiary health care could be an altemative. F. CHAPTER V: GROWTH IMPROVES MONETARY AND NON-MONETARY INDICATORS OF WELL-BEING 50. Nationally, a growth rate in per capita income of one percent reduces the headcounts of poverty and extreme poverty by almost half a point. A growth rate in per capita GDP of one percent per year would reduce poverty and extreme poverty by about 0.4 percentage points. The elasticity of poverty reduction to economic growth is slightly lower in Honduras than in Latin America as a whole, but when this elasticity is applied to the higher level of poverty observed in Honduras, the expected decrease in percentage points in the headcount of poverty in the country is not necessarily lower than in other countries. Still, the reason for the relatively modest impact of growth on poverty is different in urban and rural areas. In urban areas, growth is associated with higher inequality, which reduces its impact on poverty. In rural areas, there is no correlation between growth and inequality, but the elasticity of poverty to growth remains low. Given the higher level of poverty in rural areas, and the fact that more than half the population is rural, implementing policies that make rural growth more broad-based should be a priority. Note that the estimates of the elasticity of poverty to growth can be used to establish luture targets for poverty reduction. In preparing these simulations, it is important to take into account rural-to- urban migration over time, because it tends to reduce future national estimates of poverty. While simulations of future poverty rates based on the elasticity of poverty to growth and the extent of migration are crude, they give an idea of the gains in poverty reduction that can be expected in the future. To reduce poverty faster, the country would need to increase either its GDP per capita growth rate or its elasticity of poverty to growth. A ten percent increase in per capita GDP growth (to 2.2 percentage points per year) would have the same impact as a ten percent increase (in absolute terms) in the elasticity of poverty to growth. xvii 51. Apart from reducing poverty, growth also improves non-monetary indicators of well-being. Economic growth tends to have a positive impacts on a wide range of non-monetary indicators (the indicators considered in chapter V include infant mortality, under five mortality, child malnutrition, life expectancy, net and gross enrollment in primary, secondary, and tertiary education, illiteracy, access to safe water, access to sanitation, and telephone main lines). For example, for a country such as Honduras, a one percentage point in growth is expected to result in a 0.081 percentage (not percentage point) increase in net primary enrollment. Interestingly, urbanization seems to have a larger positive impact on social indicators than growth. Again, using estimated elasticities of non-monetary indicators to growth, simulations can be done to see the magnitude of the gains which can be expected in the future, taking into account urbanization. For example, infant mortality could be expected to decrease by almost half by 2015. Although this is not discussed in chapter V, there is also evidence that Honduras's performance in terms of the level of its non-monetary indicators of well-being is slightly better than expected for its level of economic development and urbanization. 52. The poor may benefit more from an improvement in health indicators than the non-poor, and they may benefit equally from improvements in education and access to basic infrastructure services. The poor tend to have lower social indicators than the non-poor. For example, the households living in the top third of the country's municipalities in terms of income have one more year of life expectancy than households living in the bottom third of the municipalities. The differences for malnutrition, literacy, and years of schooling between poor and less poor municipalities are larger. While this information is useful, it does not tell us how the various groups of municipalities benefit from an overall increase in the level of non-monetary indicators. It can be shown, however, that poor municipalities benefit significantly more than rich municipalities when there is an increase in life expectancy and literacy. Poor municipalities also benefit proportionately more when there is a decrease in malnutrition. By contrast, there are no differences between poor and non-poor municipalities for an increase in years of schooling. As for basic infrastructure services, the poor tend to benefit in the same proportion as the other groups. 53. Beyond the current focus on reconstruction, Honduras will need to implement important reforms in order to improve growth and make significant progress toward poverty reduction. In its 1998 report on human development in Honduras, UNDP (1998) presents the results of an opinion survey of the elites on how to promote development and poverty reduction. While the survey is small (220 interviews), it does provide a snapshot of the mindset of the country's ruling classes. According to the elites, poverty results more from inadequate growth and employment opportunities than from an inadequate distribution of the available resources in the country. The lack of a good functioning of the economy and the lack of investments, together with the lack of efficient policies and the impact of structural adjustment, are considered as the main factors responsible for the increase in poverty in Honduras. The distribution of income is less of an issue. The solutions proposed by the elites to solve the problem of poverty mirror the causes put forward as being at the root of the problem. Apart from the reform of the state and its development model, higher and better education, employment, productivity, and investments are seen as the keys to improve wages. To a large extent, economic growth is seen as the key for poverty reduction. While this is not surprising since elites do not tend to favor redistribution, it is true that in a country as poor as Honduras, growth is more important than redistribution to achieve long term poverty reduction (but some redistributive policies could actually improve growth). 54. A first condition for faster growth and thereby poverty reduction is the implementation of sound macroeconomic policies. Over the last decade, Honduras has suffered from several macroeconomic shocks. First, in 1990, the Government had to implement a structural reform package, which led to a reduction in per capita GDP. A second recession occurred in 1994. But over the last few years, the country has made significant progress in the management of its economy. Macroeconomic stability has also been maintained after Hurricane Mitch. The report written in February 2000 for the xviii Consultative Group meeting in Tegucigalpa (GRH, 2000b) suggests that macroeconomic indicators have been better than anticipated. The GDP reduction in 1999 has been only 2 percent (instead of an anticipated 3 percent). Inflation has reached 10 to 12 percent, which is below the level observed in 1998. The depreciation of the Lempira versus the US dollar has been less than 5 percent, which is similar to 1998, and the budgetary deficit will be around 5.1 percent of GDP. 55. A second condition for faster growth is an increase in Honduras' competitiveness and in the quality of its institutions. According to INCAE and HIID's measures of general competitiveness, Honduras is less competitive than Costa Rica, El Salvador, Guatemala, and Nicaragua. Although Honduras does relatively well for labor markets, it lags behind in trade openness, technology development, finance, business and marketing, infrastructure, and institutional quality. According to the Heritage Foundation's indicator of economic freedom, Honduras also does well on monetary policy, but poorly on foreign investment, regulation, taxation, trade policy, and black markets. While these indicators are debatable, and while it was expected that Honduras would not match Costa Rica, El Salvador, and Guatemala which are richer countries, Honduras does not seem to do better than Nicaragua. The broad message that emerges is that Honduras needs to improve its competitiveness to support faster growth. 56. A third condition for faster growth is the need to tackle governance and reform the state. Transparency International suggests that the perceived level of corruption is higher in Honduras than in other Central America countries. In the survey of the elites provided in the 1998 UNDP report, fighting corruption and reforming the state are considered priorities, followed by access to education, competitiveness, and income generation. Corruption and regulations hamper the creation of new businesses. Establishing a new company can take three to four months, while in some neighboring countries this can be done in one to two weeks. The lawyer fees for establishing companies are set by a formula and are sometimes very high, even for a simple process. The review of the documents needed for incorporation is inefficient either because the relevant staff do not always understand what needs to be done, or because they extort bribes. Making these processes more automatic and less discretionary would help in reducing corruption. Reforming the customs and tariffs administration would also help. Another impediment to growth is the lack of a good legal system. There is apparently no single publication that contains all Honduran legislation classified by subject matter and that is periodically updated. The courts are slow and sometimes corrupt. Contracts tend to not be enforced in court, and Honduras does not have a dispute settlement system (arbitrator) than can circumvent the slow legal process. The members of the supreme court are appointed every four years following national elections, instead of being appointed for longer periods of time on the sole basis of their credentials as is done in other countries. The training received in law schools is also judged as deficient. More generally, in line with the lack of competitiveness of Honduras suggested by intemational observers, four respondents out of five in the UNDP survey of the country's elites believe that the country is not ready to face of globalization. The same proportion of respondents believes that Honduras's entrepreneurs do not have the ability today to compete in the world economy. 57. Overall, what is lacking to encourage growth and foreign investment in Honduras is perhaps less an identification of the steps that must be taken, than an ability to implement these steps. A draft study by the Foreign Investment Advisory Service (FIAS, 2000) of the World Bank suggests that there are no serious impediments to foreign direct investment in Honduras. There are no serious problems with the foreign exchange system, the tax system, the govermnent interfering with industry, or the application of the rule of law. Yet while foreign direct investment has been flowing at about 2 percent of GDP in the last 2 years (before Mitch), this could and should increase. Why does Honduras have limited appeal to foreign direct investment? Apart from the fact that the country is small, the lack of foreign direct investment is the result of a combination of factors such as those mentioned above. Another obstacle is that the labor force is not particularly skilled or well educated. According to FIAS, between xix 1991 and 1999, the tone of the businessmen was more strongly negative regarding the need for reforms, and there was a lower expectation that things were going to get better. I CHAPTER I: POVERTY AND INEQUALITY A. THIS REPORT IS A CONTRIBUTION FOR THE GOVERNMENT'S POVERTY REDUCTION STRATEGY 1.1. As part of the process for participating in the Heavily Indebted Poor Countries (HIPC) Initiative, the Government of the Republic of Honduras (GRH) is preparing a Poverty Reduction Strategy Paper. Honduras is one of four Latin American countries participating in the HIPC Initiative providing debt relief to highly indebted and poor countries worldwide. In order to participate in the HIPC Initiative, as is the case for other countries, the GRH must prepare a Poverty Reduction Strategy Paper (PRSP). The PRSP must contain a diagnostic of poverty in the country, a strategy for its reduction, and a number of targets that can be monitored over time in order to assess the performance of the country in reaching its goals. The PRSP is to be prepared by the Government in consultation with civil society. While the Government prepares and owns the PRSP, and is responsible for its implementation, international organizations can provide technical assistance. 1.2. Given that experience is lacking in Honduras as elsewhere on how to best write a PRSP, and given that deadlines had to be met for participation in the HIPC Initiative, the GRH first wrote an Interim PRSP (I-PRSP). Having a good PRSP can be a major asset for orienting public policies in a country. But writing a PRSP is no easy task, and experience is still lacking on how to best proceed. Confronted with short deadlines for participation in the HIPC Initiative, the Government of Honduras first wrote an Interim PRSP (I-PRSP). This I-PRSP is being extended into a full PRSP informing the strategic choices to be made by the Govermnent for reducing poverty in the country. This approach of first writing an I-PRSP, and extending the document into a full-fledged PRSP later, has been followed by other countries as well, including Bolivia in Latin America. 1.3. Following a request for technical assistance from the GRH, this report is a contribution to the PRSP process in Honduras. A joint World Bank-IMF mission to launch the PRSP process took place in Honduras in November-December 1999. During this mission, the GRH requested assistance from the World Bank in order to conduct analytical work to support the PRSP. To respond to this request, the World Bank provided two types of technical assistance. First, it made available to the Government consultants based in Tegucigalpa (ESA Consultores). These consultants provided support to UNAT (Unidad de Apoyo Tecnico), the unit cf the Ministry of the Presidency in charge of drafting the I-PRSP. Second, from December 1999 to February 2000, a World Bank-led team conducted analytical work in Washington, D.C. This team came to HIonduras in March 2000 in order to discuss its results with UNAT in a series of meetings. The present report was drafted and delivered to the Government of Honduras in June 2000. It is based on the work of this second team, which included staff from the World Bank, PRAF (Programa de Asignaci6n Familiar) and IFPRI (International Food Policy Research Institute, an international agency working with PILAF, see Box 1.2). The team benefited from the comments of UNAT, World Bank, and IMF staff. The report follows broadly the themes of the Interim PRSP prepared by the GRH. Specifically, the present report consists of five. Chapter 1 provides a diagnostic of poverty. Chapter 2 discusses the determinants cf poverty. Chapter 3 deals with non-monetary indicators of well- being, including access to basic infrastructure services (electricity, water, sanitation, etc.). The chapter also evaluates the electricity subsidy and the impact of the social investment fund. Chapter 4 reviews public expenditures for human capital (education and child labor, nutrition, and health). Chapter 5 analyzes the impact of growth on poverty and non-monetary indicators of well-being, and it briefly discusses policies that could promote faster growth. In the annexes, we provide technical details in an effort to make our methodological assumptions clear and to build analytical capacity in Honduras. 2 1.4. The success of the Government in improving the well-being of Honduras's population should be monitored over time using a battery of indicators rather than poverty measures alone. Both monetary and non-monetary indicators can be used for assessing the impact of government policies. * Monetary indicators: Within the realm of monetary indicators, the level of growth in per capita income or expenditures is a first measure of the performance of a country in increasing well-being. This can be estimated using National Accounts or nationally representative surveys. Beyond growth, in order to take into account distribution issues, analysts have used a range of inequality and poverty measures. Because computing these measures is difficult (requiring good data, good analysis, and many methodological assumptions), there is typically a debate in most countries as to the level of distribution-sensitive monetary indicators and their trend over time. We do provide levels and trends for poverty and inequality in this report, but we believe that our results should be re-evaluated once the data from the 1998-99 Encuesta Nacional de Ingresos y Gastos de los Hogares become available. That survey should provide a better basis for poverty and inequality measurement than existing surveys (reasons for this are mentioned in Box 1.2.) Therefore, many of the results presented in this report are preliminary and subject to modification once better data become available'. * Traditional non-monetary indicators: Many argue that poverty and well-being are multidimensional phenomena that are not well represented by monetary measures of well-being only. We agree with this critique (which does not diminish in any way the need to estimate trends in poverty and inequality), and we therefore analyze several non-monetary indicators of well-being in this report. One well-known indicator is the Human Development Index proposed by the UNDP. Another is the index of umnet basic needs. Other indicators are sector specific such as health (malnutrition, infant mortality, etc.), education (enrollment, assistance, repetition, drop-out, etc.), and basic infrastructure services (electricity, sewerage, sanitary installation, safe water, etc.). Some (but not all) of these other indicators have been analyzed in Honduras under the umbrella concept of unmet basic needs. * Other non-monetar indicators: Traditional monetary and non-monetary indicators still do not fully capture the level of well-being of a population. For example, while domestic violence within the household and social capital within the community matter, they cannot be analyzed using traditional measurement tools. Another example relates to subjective perceptions of welfare, and more generally to work revealing the priorities of the poor. In this report, we do not analyze these issues. - Monetary conversion of non-monetarv indicators: In some cases, it is feasible to put a monetary value on non-monetary indicators, and this can be useful for the analysis of trade-offs between policies. In chapter 2, we analyze the income gains from education and employment. In chapter 3, we estimate the value of having access to basic infrastructure services such as electricity, water, and sanitary installations. In chapter 4, we compute the future income loss for children from working at a young age. These exercises provide valuable information, but they do not capture the full cost or benefit of what is observed. For example, there is an intrinsic value in being well educated, or in having a good job, which goes beyond the monetary income provided by education and employment. This has to be kept in mind in order not to base policy decisions on a monetary cost-benefit analysis only. To give a less obvious example, there is an intrinsic merit in having public policies that promote better access to culture and art for the poor, even if this does not bring monetary benefits. The very poor are not only hungry for food, they are also hungry for creativity and expression. They also need to be recognized as valuable members of society, and they need to be given the opportunity to contribute to others and to society as a whole (see Box 1.1 for an example) instead of always being at the receiving end. Many such issues are not discussed here due to the limited scope of the report. Another alternative to analyze changes in monetary well-being is to use welfare functions taking into account in a flexible way the full distribution of income or expenditures. While it is standard in a poverty assessment to provide measures of growth, inequality, and poverty, and to analyze the relationships between these concepts, it has not yet become standard to use welfare functions. We plan to use welfare functions in subsequent work on Honduras. 3 Box 1.1: Beyond monetary poverty: The importance of enabling the very poor to help others Nod is a 26 year old Honduran. Although a man of small stature, he is very strong because his childhood and his youth were spent doing heavy work. He lives with his mother and grandfather. As a child, he was a shoeshine boy in town. Today you can see him gathering wood in the mountains, for his mother to cook tortillas to be sold by his young sisters in the streets, for the neighbors as well. On Saturday mornings he unloads lorries and vegetable barrows in the peasants' market. Sometimes he sits by the roadside with a vacant look, ruminating on his own personal thoughts. In meetings he keeps to himself and remains silent. Because of his long silences, which can last for several days, because of his singing in the streets or his sudden shouts, just for the pleasure of it, some say he is witless. People avoid him. Nod spent a few years at school where he learned rudimentary reading and writing. In 1990, when staff from the International Movement ATD Fourth World (a grass-roots NGO) arrived in Nod's district in Tegucigalpa, they invited Nod and other young people to help run the Street Library for the children (an educational program for poor children taking place in the streets). Later, Noe said this had changed his life. At first, he was afraid that he would frighten the other children, or that he wouldn't know how to read a book. Nevertheless, he came - j ust to see - like all the other young people who started at the Street Library. And then a child came and snuggled up to him; so he opened a book and began to read out of curiosity. He then realised that he too could give pleasure to a child. He continued to attend the Street Library. As one of the staff members of the NGO explains it, when he arrived in Tegucigalpa, Nod was of great help. He was the only one who came to help the staff member set up a Street Library in a shantytown, outside the district of Rio, on the banks of a river. Later, Noe told the staff member he had met the children from there at the Saturday market where, like him, they unloaded lorries - work much too heavy for them. The staff member also found out later that Nod needed great courage to come with him because it added to the mockery and jokes the others made about him. In the bus, Nod wouldn't sit next to the staff member, and in the street he walked ten meters in front of the staff member. When the staff member left Tegucigalpa some years later, Nod insisted that the Street Library should continue in the shantytown. He went there alone sometimes because he thought that the children there were most in need of it. In 1998, Hurricane Mitch devastated a large part of Central America, and Honduras in particular. The small shantytown on the riverbank in Rio had disappeared completely from the face of the earth; the families who had lived there were scattered around different shelters. The young people who were running the Street Libraries met the families in the shelters and told stories to the children, helping them get over the shock of the hurricane. Noe was there, even though he had just spent several days in the south to help an uncle dig out his house, which had been covered with mud. Some families came back quite quickly to their districts and, lby leaning a few planks against any walls still standing, made makeshift shelters. Upon the visit of the young people, a mother asked them one thing: to hold a party for the children once the Street Library would be in place again in the area. Later on, Nod told that he was going back to work in a job he had occasionally held between long periods of unemployment. But he said that aFter several months of this work with pick and shovel, he would "return to the Street Library, to read books with the children, dance and tell stories... because they need this. I too, need this. It is better than loafing around the streets doing nothing." Source: Provided by the International Movement ATD Fourth World. 4 B. THERE IS UNCERTAINTY AS TO WHETHER POVERTY HAS DECREASED IN HONDURAS IN THE 1990S 1.5. The level of poverty in a country is what matters in real life. But it is the trend in poverty, not its level, which matters for the evaluation of public policies. The role of a PRSP is to help a country in reducing not only its rate of poverty, but also the absolute number of the poor which increases with population growth when the rate of poverty is left unchanged. Reducing the level of poverty is the goal. But the measurement of progress toward that goal is the poverty trend, i.e. the change in level over time. It often happens that different analysts find different poverty levels because they use different methodologies for measuring poverty. This is not a problem as long as they agree on the trend. A poverty level is normatively defined, and therefore subjective. For practical purposes, a poverty trend is neither normative, nor subjective: it is a fact. Below, we focus on the trend in poverty, not its level. 1.6. According to the I-PRSP, the share of all households living in poverty has decreased by about 10 percentage points in the 1990s. Following standard practice, the I-PRSP prepared by the GRH considers two poverty lines for measuring poverty. The extreme poverty line is the cost of a food basket designed to meet basic nutritional needs. The moderate poverty line multiplies this cost by a fixed factor in order to also take into account the cost of basic non-food needs (see annex 1 for details). Using per capita income as the indicator of well-being, the GRH computes the household level headcount index of (extreme) poverty which is the share of all households with per capita income below the (extreme) moderate poverty line. Table 1.1 presents the results obtained by the GRH. Nationally, the share of all households in extreme poverty decreased from 54.2 percent in 1991 to 48.6 percent in 1999. The share of all households in poverty decreased in a similar fashion, from 74.8 percent in 1991 to 65.9 percent in 1999. From 1991 to 1998 (before Hurricane Mitch), the reduction in extreme poverty and poverty is large, at about ten percentage points. There is progress in both urban and rural areas, although the negative impact of Mitch appears to be larger in rural areas. Table 1.1: Trend in poverty according t the I-PRSP, headcount index, 1991-99 1991 1992 1993 1994 1995 1996 1997 1998 1999 National Extreme poor 54.2 47.4 45.1 47 47.4 53.7 48.4 45.6 48.6 All poor (extreme + moderate) 74.8 69.9 67.5 67.2 67.8 68.7 65.8 63.1 65.9 Urban Extreme poor 46.7 39.2 31.6 39.8 40.6 38.7 35.2 35.7 36.5 All poor (extreme + moderate) 68.4 61.6 55.5 62.6 62.8 61.0 59.0 57.0 57.3 Rural Extreme poor 59.9 53.9 55.8 52.9 53.1 66.4 60 55.4 60.9 All poor (extreme + moderate) 79.6 76.5 77.1 71.1 71.9 75.3 71.7 69.2 74.6 Source: GRH (2000) using EPHPM surveys. 1.7. Tests of the robustness of the poverty trend in the I-PRSP can be made using alternative poverty lines and/or alternative indicators of well-being. To compute a poverty measure, three ingredients are needed: a poverty line, an indicator of well-being, and the poverty measure itself. The most widely used poverty measures are the headcount index, poverty gap and squared poverty gap (see annex 2, section MA. 1 for a definition). While the poverty gap and squared poverty gap are better than the headcount index for the evaluation of public policies, the trend in poverty is typically similar for these three measures. This is not necessarily true for the assumptions made for the poverty lines and the indicator of well-being. Changing these assumptions can lead to different trends in poverty over time. In annex 1, we conduct tests for the robustness of the poverty trend in the I-PRSP to the choice of the poverty lines and indicator of well-being. The key results from these robustness tests are presented here. 5 Box 1.2. DATA FOR POVERTY MONITORING AND ANALYSIS IN HONDURAS This box discusses the household surveys used for poverty monitoring and analysis in the country and in this report. The box also outlines initiatives that could be implemented in order to improve data collection and analysis. The key mess,age is that although Honduras has implemented a number of good household surveys over time, there is scope for developing the country's statistical capacity. Census and nationally representative surveys There are three main sources of nationally representative data on households and poverty in Honduras. First, data on non-monetary indicators of poverty are available from the latest national population census of 1988. This information has been used extensively for building poverty maps and targeting government interventions on the basis of Unmet Basic Needs (Necesidades Basicas Insatisfechas). Honduras' Libro Q (1994) provides detailed information on this approach. While the information from the 1988 census is now outdated, a new census will be implemented in 2000 with support from UNDP. Second, the Encuesta Permanente de .Hogares de Propositos Multiples (EPHPM) is a labor force survey conducted every six months with support from USAID. The sample frame consists of 7,200 households, stratified into 4 geographic regions: (1) Tegucigalpa; (2) San Pedro Sula; (3) Other Urban; and (4) Rural. Although concerns have been raised about the quality of the data, the survey is today the sole basis available for poverty monitoring over time. The sample includes a rotating panel, but this feature has never been used to our knowledge, andl work is needed to reconstruct the panel by linking the surveys. Third, from February 1998 to March 1999, also with support from USAID, the Central Bank has implemented a new Encuesta Nacional de Ingresos y Gastos de los Hogares (ENIGH). The survey is nationally representative and the sample consists of 4,000 households evenly distributed across the four regions mentioned earlier (Tegucigalpa, San Pedro Sula, Other Urban, and Rural). The ENIGH is much richer than the EPHPM: it has informnation not only on income and employment, but also on expenditures, housing, education, and health. While the data from the survey were not available at the time this report was written, they could be used for analytical work in a follow-up report. Potential trade-off between the two national surveys There may be a financial trade-off between the two nationally representative surveys for future poverty monitoring and analysis in Honduras. The EPHPM has provided valuable data over time at low cost. The survey is conducted by the Direcci6n General de Estadistica y Censo (DGEC), which has a total budget of about 7 million Lempiras per year. With this budget, the DGEC pays its staff and conducts four surveys: the two rounds of the EPHPM, and two rounds of a survey of basic grain production in rural areas. It can thus be estimated that the cost of each EPHPM survey is in the range of 1.25 million Lempiras. On the other hand, for poverty monitoring and analysis, the EPHPM lacks an expenditure module (expenditures tend to be a better indicator of welfare and poverty than income). More generally, the EPHPM suffers from the limits of a traditional labor force survey in that its short questionnaire does not provide sufficient information for the detailed analyses that are often needed to inform public policy. The ENIGH has an expenditure module and it covers a wide range of topics, but it is expensive. The budget for the 1998-99 survey was 17 million Lempiras. While it is likely that this cost could be reduced in a follow-up survey, the cost of the ENIGH as currently designed would certainly remain much higher 6 than the cost of the EPHPM. If Government and donor support is provided to implement household surveys, the country may be able to maintain its two surveys, for example by fielding the EPHPM twice a year, and the ENIGH once every two years. However, at this time, Government and donor support is lacking, and the March 2000 round of the EPHPM has not been fielded for this reason. If the budget constraint continues to be tight, some choices may have to be made. One possibility would be to replace the EPHPM by the ENIGH, and to use rotating modules in order to reduce the cost of the ENIGH. A detailed analysis of the needs of the country and the cost of the surveys will be needed to make the best decision. Capacity building There is a wide consensus that the country needs to strengthen its statistical capacity in order to better inform its policies. A law has recently been adopted to replace the DGEC by a statistical institute. In the past, the DGEC has been a weak institution, and the few statistics it was generating are not widely used, in part because of a lack of analytical capacity both within the DGEC and in national universities and think tanks. The DGEC was under-financed and poorly staffed. The salary levels were not sufficient to attract staff of high technical ability. The DGEC suffered from poor physical infrastructure, both in terms of its building and computer equipment, and lacked institutional autonomy. Once the Govermment creates the new statistical institute which will replace the DGEC, the country could benefit from participating in the MECOVI (Mejoramiento de las Encuestas y la Medici6n de las Condiciones de Vida en America Latina y el Caribe) coordinated jointly by CEPAL, the Inter-American Development Bank, and the World Bank. This program provides financing and training to improve household survey data collection, dissemination and analysis. Participation in MECOVI would provide an excellent means for supporting poverty analysis and making the best use of limited resources. Beyond the management and analysis of household surveys, a medium- to long-term program addressing deficiencies in other data sources, as well as weaknesses in institutional capacity in Honduras, is required (not all of these activities fall under the guidelines of MECOVI). For example, line ministries should aim at building better data sources, especially after the damage caused by Hurricane Mitch in the Fall of 1998 (data were lost in flooding). Also, a single Government technical unit should be in charge of gathering the various sources of data available in the country. This task was in the past done by UNIS, a unit of SECPLAN, but today the unit and ministry do not exist anymore. In theory, the mission of UNIS has been given to UNAT, but in practice UNAT lacks the means and staff to implement that mission. Surveys by PRAF and FHIS The country (and this report) have also benefited from the surveys conducted by Government agencies: * In 1999, the Programa de Asignacion Familiar (PR4F), with financing from the Inter-American Development Bank (IADB) and assistance from the International Food Policy Research Institute (IFPRI), implemented a survey in the bottom half of the municipalities in terms of malnutrition as measured by the Ministry of Education's annual Censo de Peso y Talla in the first year of primary school. The survey has modules on expenditures, education, health, and the impact of Mitch. This report benefited from a collaboration between PRAF, IFPRI, and the World Bank for the analysis of the survey, and a staff member of PRAF came for training for one month to Washington D.C. * In 1998, the Fondo Hondureno de Inversi6n Social (FHIS), with financing from the World Bank, conducted a survey in its areas of operations. This survey was used in this report as well, in part to analyze the performance of the FHIS. As for the PRAF survey, it has also been used to analyze the determinants of health and education status. The use of both surveys compensated for the fact that the national income and expenditure survey was not yet available at the time of analysis. 7 Improving future poverty analysis All poverty measures provided in this report should be considered preliminary, because once the 1998-99 Encuesta Nacional de Ingresos y Gastos de los Hogares (ENIGH) will be available, it should be feasible to conduct better poverty analysis. The main advantages of the ENIGH survey are as follows: * Use of expenditures instead of income as the preferred indicator of well-being: Expenditures (including the value of auto-consumption) is a better indicator of well-being than income for a number of reasons. One is that it tends to be better measured, at least when the expenditures questionnaire in the survey is well designed. Another reason is that expenditures takes into account the smoothing strategies used by household to cope with shocks. It also takes into account the fact that households behave differently (e.g., through savings) at different periods of their life in order to maximize their discounted welfare over time. * Use of the survey to compute the cost of basic food and non-food needs: There can be a discrepancy between food prices observed on markets and prices paid by poor households. Therefore it is better to compute the cost of basic food needs (which typically corresponds to the extreme poverty line, see annex 2, MA. 1) using survey data rather than data external to the survey. When doing so, the "unit values" obtained at the household level should be corrected for differences in household characteristics. Basing poverty estimates on extreme poverty lines obtained within the survey rather than with external data can make quite a difference in the assessment of poverty trends over time. There are also methods for estimating the cost of basic non-food needs from survey data in order to compute moderate poverty lines. Again, doing so can make a difference for poverty monitoring over time, and between regions (see Wodon, 1997, for detailed applications of the above methods). None of the above could be done with the data available for this report, but it could and should be done with the Encuesta Nacional de Ingresos y Gastos de los Hogares. Thus, the levels (and trends) in poverty presented in this report should Ibe re-evaluated when the survey becomes available. 1.8. Broadly speaking, the poverty trend in the I-PRSP is robust to the choice of the poverty line. Our first test for robustness consists of using a poverty line that differs from that used by the GRH. For this, we simply scale up the extreme and moderate poverty lines obtained in a previous poverty report for Honduras (World Bank, 1995) in order to provide 2,200 kcal per person per day. We use a single national poverty line. Using urban and rural poverty lines would affect the levels of poverty in both urban and rural areas, and therefore the comparison of the two sectors, but it would not affect the trend in each sector by much. We then adjust on a monthly basis the poverty line over time using the CPI instead of the cost of the food basket. There are thus two sources of differences in the choice of the poverty line: its level in the reference period, and its adjustment over time. What matters the most is the adjustment over time, through the use of the CPI rather than the Canasta Basica. The results are given in annex 1, and the conclusion is that the poverty trend in the I-PRSP is robust to the choice of the poverty line2. 1.9. But the trend is not robust to adjustments for underreporting. Labor force surveys such as the EPHPM are not very good at capturing income. Because this type of survey is often the only source of information available for poverty measurement in Latin American countries, it has become a standard practice to test whether the income growth recorded in the surveys matches the national accounts. The assumption is that the surveys suffer from under-reporting of income, and therefore it may be necessary to adjust the income levels recorded in the surveys so that they match the national accounts. The problem of underreporting for poverty trends is that the level of underreporting in the surveys may change over time. 2 The GRH trend is also similar to the trend in the IPEA (1999) report which uses a methodology similar to the GRH. IPEA suggests that the headcount of poverty decreased from 72.9 to 62.4 percent from 1990 to 1999. 8 In Honduras, the surveys may have improved over time, in that at the end of the period the mean level of income in the survey is closer to the national accounts than in the early 1 990s. If the surveys have indeed improved, the reduction in poverty observed in table 1.1 may be due to the improvement in the surveys rather than to an improvement in living standards. As discussed in annex 1, with adjustments for underreporting based on the national accounts, it turns out that there is no decrease in poverty in the 1990s. The reason for this is simple: since the growth in per capita GDP in the national accounts has been fairly limited in the 1990s, when the per capita income data in the surveys is adjusted to reflect the national accounts, there are very limited gains over time. Taking into account a small increase in inequality over the same period, this leaves poverty unchanged (see annex 1 for methodological details). A similar lack of poverty reduction is found by CEPAL (1999) with a more sophisticated adjustment procedure. According to CEPAL, the share of all households below the poverty line (the headcount index) in Honduras was almost the same in 1997 at 74 percent as in 1990 at 75 percent. 1.10. In other words, if there has been a reduction in poverty in the 1990s, the magnitude of this reduction may be lower than indicated in the March 2000 I-PRSP. There is no simple answer to the question of whether the estimates of poverty with and without adjustment for under- and over-reporting are better. While some would argue that the measurement errors may be larger in the national accounts than in the surveys, others would argue the reverse. Still, it is worth mentioning that two key factors account for the lack of poverty reduction with adjustments (annex 1). First, the rate of per capita income growth in the surveys is much higher than the rate of per capita GDP growth in the national accounts. Second, in constant terms, the rate of growth of per capita GDP in the national account is higher than the rate of growth in per capita consumption, in part because the CPI has risen faster than the GDP deflator. The first factor plays a much larger role than the second in the lack of poverty reduction observed with adjustments for under-reporting. The bottom line is that while some poverty reduction has probably been achieved in the 1990s, it may not have been achieved to the extent suggested by table 1.1. Some uncertainty as to the poverty trend in the 1990s will remain, but this need not be the case for future poverty monitoring. If the ENIGH survey is fielded regularly, it should provide a much better instrument for poverty measurement than the EPHPM (see Box 1.2 for a discussion). 1.11. It is also worth noting that if all sources of income are used for measuring poverty in the later rounds of the EPHPM surveys, poverty measures are lower. For the full period under review (1991 to 1999), the EPHPM surveys can be used to compute labor income at the individual level (income from first occupation, a second occupation, and self-employment). Individual-level incomes are then added into a measure of per capita household income. For the last few years (1997-1999), the EPHPM surveys also include information on other sources of income: pensions; subsidies; rent; allowances, grants and stipends; remittances from abroad; support from relatives; and support from other individuals; and other sources of incomes. When the more comprehensive measure of per capita household income available in the later surveys is used, the share of the population (or households) below the extreme and moderate poverty lines is reduced by about five percentage points. C. THE IMPACT OF MITCH ON THE POOR IS LARGER THAN SUGGESTED BY THE EPHPM SURVEYS 1.12. The March 1999 EPHPM survey suggests a small increase in poverty following Mitch, but this increase is likely to be underestimated for a number of reasons. The increase in poverty due to Mitch obtained with the EPHPM survey by comparing poverty measures in March 1998 and March 1999 in table 1.3 is relatively small. The actual impact of Mitch is probably larger for at least three reasons. First, the EPHPM does not capture very well the income of small farmers who are those who suffered the most from Mitch due to the loss of their crops. Second, while labor income (a flow) may have been sustained after Mitch, many households suffered from a loss in assets (a stock) which has implications for 9 future poverty. Third, a number of populations at risk such as street children, squatters, and banana plantation workers are unlikely to be well represented in the EPHPM survey, so that the impact of Mitch on the income of these populations cannot be measured with the survey. More generally, table 1.2 provides examples of populations at riskc during natural disasters together with vulnerabilities. Table 1.2: Populations at risk duri ng natural disasters such as Hurricane Mitch Population at risk Example of vulnerability The illiterate (majority female) Unable to read early warnings and instructions in temporary shelters Small agricultural producers Located on eroded hillsides; lost insurance in the form of seeds Street children (e.g. Tegucigalpa) Flooded out of living space; not likely to go to temporary shelters Squatters Liive in high risk flood planes; not likely to request assistance Banana plantation workers Dependant on private sector social services; not reached by public relief Female-headed households Likely to loose household possessions; slower to return to work Indigenous populations Linguistic, cultural, other obstacles for early warnings and access to relief Source: Adapted from Shrader and Delaney (2000) 1.13. Many households suffered froin a loss in their crops, wages, and/or small business following Mitch. Households also incurred direct costs for health, housing, and food. The 1999 PRAF survey of poor municipalities has information on the losses and costs incurred by households due to Mitch, as well as the relief obtained by households. More than a third of the households report having suffered a direct crop loss due to Mitch, and the proportion is higher for richer households (40-50 percent are affected) than poorer households (30 percent are affected). Close to one in ten households declares a loss in wages, and the proportion is similar fior the poor and the non-poor. One household in twenty declares a loss in its small business or self-employment earnings, with a higher probability of loss at higher levels of welfare. Apart from these losses in assets and earnings, almost one in three households incurred health expenditures. For housing, the proportion is one in ten. For food and other costs, the proportion is one in seven. Households located along the main roads of a municipality suffered more damage to their houses, probably because the roads served as evacuation conduits for the rain and floods. By contrast, households living in the suburbs were more likely ito suffer from crop losses. Among those households who suffered a loss or incurred a cost, some of the amounts were relatively large, at several thousand Lempiras (table 1.3). There is also evidence that the poor have been hit more seriously by Mitch than the non-poor, at least in the case of housing, simply because their houses were much more fragile and thereby vulnerable to the Hurricane. 1.14. Although relief was well targeted within departments, the amount of relief distributed was small and it tended to favor departments in the center of the country. The proportion of households who received emergency relief is muchr smaller than the proportion that suffered from Mitch. One in ten households received food relief, one in twenty received clothes or shoes, and one in forty received medicines. The amount of relief received is also much smaller than the losses and costs due to Mitch (table 1.3). On the more positive side, an econometric (tobit) analysis suggests that those who benefited from the relief effort were those who needed it the most. Controlling for geographic location, the amount of food aid received was positively related to the proportion of pre-Mitch assets lost due to the Hurricane, and negatively related to the logarithm of the pre-Mitch assets value (i.e., the wealth of the household). But the relief effort has tended to be in favor of centrally located departments. 10 Table 1.3: Median value of the losses, costs, and relief for those affected by Hurricane Mitch, 1999 Median value of losses suffered among those suffering a loss (Lempiras) Crop Wages Small business Other loss Municipal center 4000 840 3000 500 Suburbs 3000 720 1000 400 Median value of costs incurred among those incurring a cost (Lempiras) Hospital/health Housing Small business Food/others Municipal center 500 800 3675 200 Suburbs 400 500 2000 100 Median value of relief received among those receiving relief (Lempiras) Food Clothing/shoes Sheets/blankets Medicaments Municipal center 125 200 100 125 Suburbs 200 100 50 100 Source: PRAF staff using 1999 PRAF survey. 1.15. Women and men may not have reacted in the same way to Mitch, and this has implications for policy. In a study on the gender dimensions of disaster relief in Honduras and Nicaragua, Shrader and Delaney (2000) argue that women reacted differently to Mitch than men. While this is not surprising given that women and men have different priorities and responsibilities in society, it has implications (see Box 1.3 for details and the role of women in offsetting part of the losses from Mitch). - Loss of assets: Women tend to give priority to physical and psychological health, while men are more concerned with protecting their assets. Women also tend to be less informed than men about evacuation warnings and how to protect key assets from destruction. When the disaster strikes, women rather than men often take care of children, the elderly, and the disabled. These asymmetries between men and women make it likely that asset losses have been greater for women. - Violence: Men are more involved in search and rescue operations than women, and thus less likely to evacuate to shelters. El-Bushra (1998) observes that women and girls are vulnerable to physical and sexual violence in shelters, while men are vulnerable to alcoholism and aggressive behavior, possibly due to the frustration they feel from being unable to contribute to the families' well being. * Risk aversion and disaster planning: Women tend to be more risk averse than men, so that they are more likely to believe early warning messages regarding disasters and take action (Blaikie et. al., 1994). Women could be more involved in improving disaster warning and planning systems. * Relief priorities: There is typically an overemphasis on rebuilding public infrastructure after a disaster, and women tend to be excluded from the decision-making process following a disaster. Community priorities and social needs such as psychological counseling also tend to be ignored. D. WHO ARE THE POOR? A STANDARD POVERTY PROFILE 1.16. It is standard practice to provide a poverty profile in a report on poverty, but it is important to emphasize the limits of such a profile for policy analysis. A poverty profile is a set of tables giving the probability of being poor according to various characteristics, such as the area in which a household lives or the level of education of the household head. Such a poverty profile is provided in annex 1, and briefly summarized below. Poverty profiles are informative, but they also have limits. Their main problem is that they cannot be used to assess what are the determinants of poverty. This is because the fact that some household group is poor (say agricultural laborers) may be due at least in part to the characteristics of this group (say their education level). If this is the case, the policy implication is to improve education levels, rather than promoting services related to agriculture. In chapter 2, we provide detailed regression results which avoid for the most part the problems associated with poverty profiles. This being said, the following results were obtained for the last three EPHPM surveys. The surveys are for March 1998, September 1998, and March 1999, and the poverty estimates are based on poverty lines that differ from those used by the GRH in the I-PRSP (see annex I for details on the poverty profile): I1 * Geographv: There are large differences in the extent of poverty by department. In general, the departments bordering Salvador and the Gulf of Fonseca tend to be poorer. It should be noted that there are some large variations in the poverty estimates at the departmental level over time. This is due to the limited sample size of the survey which is not large enough to be fully representative at the department level. This does not mean that departmental level data cannot be used for investigations into the geography of poverty, but it implies that one should be careful in interpreting the results. * Household size: On average, rural households tend to be larger than urban households, which is not surprising. In March 1998 for example, the mean rural household consists of 1.24 infants (0-5 year old), 1.99 children (6-14 year old), and 3.44 adults (more than 15 year old), for a total household size of 6.67. By contrast, the mean household size in urban areas is only 5.89, consisting of 0.96 infants, 1.49 children, and 3.44 adults. The numbers are similar in September 1998 and March 1999. Although this is not shown in detail in annex 1, it can be demonstrated that poverty is higher among larger households. But because this finding may be sensitive to the choice (among others) of the equivalence scale for poverty measurement, which reflects the ability of households to save on their per capita expenditures such as rental housing when they have several members, no policy implications can be drawn (see chapter 2). The fact that rural household are larger due to a higher number of infants and children contributes to the higher probability of being poor in rural areas. * Age: The probability of being poor decreases as the individual gets older. In the tables provided in annex 1, urban and rural infants have a probability of being extremely poor of 15.58 and 51.44 percent in March 1998. Urban and rural children have a probability of being extremely poor of 16.08 and 53.44 percent. By contrast, urban and rural adults have lower probabilities of being poor, at 9.01 and 40.93 percent respectively. Note however, as is the case for the poverty profile by household size, that this finding may be sensitive to the choice of equivalence scale (large households tend to have young members, i.e. infants and children; therefore if the choice of equivalence scale overestimates the incidence of poverty among large household, it also leads to an overestimation of poverty among infants and children living in these large households). * Gender of the head and presence or absence of a spouse: In urban areas, households with a female head have a lower probability of being poor than households with a male head. The reverse is observed in rural areas, where households with a female head have a higher probability of being poor. Another finding is the fact that households where the head does not have a spouse (the head is single, divorced or widowed) are less likely to be poor in urban areas than households where the head has a spouse, and more likely to be poor in rural areas than the same comparison group. * Migration: This category is typically not considered in poverty profiles, yet the information provided is instructive. Two types of migration are considered: whether the household head lives in a different place than his/her place of birth (this is the case for half of the households in urban areas, and a third in rural areas in the March 1998 survey), and whether the head has been living in his/her current place of residence for less than five years, in which case we also know if he/she came from a rural or an urban area. Recent migration affects seven percent of the households living in urban areas, and four percent in rural areas. o In urban areas, individuals living in households where the head has migrated tend to do well. However, as for any other result in the poverty profile, the fact that migrant households are doing well in urban areas does not mean that migration to or within urban areas necessarily leads to a gain in per capita income (migrant individuals and households may be better endowed in assets such as human capital, which vwould account for their success). Still, the results are a positive sign, and they are of course coherent with the patterns of migration observed in the country. O A different situation is observed in rural areas where in most cases, individuals living in households where the head has migrated are at least if not more likely to be poor than individuals living in households where the head has not migrated. There is one exception to this finding: in some years, urban households migrating to rural areas are less poor than their rural counterparts, but they remain poorer than their urban non migrant counterparts. It could be that migration to or 12 within rural areas is a last resource strategy for households who do not manage to meet their basic needs, while migration to urban areas is more likely to provide new opportunities for migrating households. But again, testing the impact of migration on per capita income and the probability of being poor must be done through regressions (chapter 2), not a profile. * Education: In urban areas, one out of every six household heads has no education at all, and only one head out of three has an education level beyond the completion of primary school. In rural areas, a third of the household heads have no education, and less than one out of fifteen has gone beyond primary schooling. The corresponding proportions for household spouses are similar. These low levels of education are one of the reasons why poverty is so high in Honduras. Indeed, the lower the level of education of the head or the spouse, the higher the probability of being poor. In March 1999, in urban areas for example, individuals living in households with a head having no education at all had a probability of being poor of 46.1 percent, versus only 2.4 percent when the head has education level at the superior (university) level. In rural areas, the corresponding probabilities of being poor are 79.1 percent for a head with no education, as compared to 12.4 percent for a head with superior education. Similar pattems are observed for the poverty profiles according to the spouse's education. * Employment: Individuals living in households where the head is unemployed and searching for work are much poorer than individuals in households where the head is working. The differences in the probability of being poor are very large. In March 1999 in urban areas, the headcount index of poverty when the household head is working is 25.9 percent, as compared to 64 percent when the head is searching for work. In rural areas, the respective headcount indices of poverty are 48.1 and 95.7 percent, which implies that a rural household with a head searching for work is almost certain to be in poverty (the same differences are observed in the September 1998 survey, but not in the March 1998 survey). In the case of spouses, the differences are less striking. The fact that in many cases individuals living in households where the head is not working are not poorer than other individuals reflects the possibility that some heads are not working due to income sources other than labor. * Sector of employment: Not surprisingly, individuals living in households where the head works in the agricultural sector have a higher probability of being poor than individuals in households whose head works in other sectors. This is true in both urban and rural areas. Heads working in the transport sector tend to have the lowest probability of being poor. * Type of employment: Individuals living in households with heads working as employers (i.e. business owners) have the lowest probability of being poor, followed by heads working as employees or blue collar workers. The same is true for spouses. Having a head working as an unpaid family worker leads to higher poverty, but this is not the case for the spouse. Finally, the self-employed tend to be slightly less poor than others in urban areas, and slightly more poor than others in rural areas. This reflects the better opportunities for surviving on the earnings from one-person small informal businesses in urban as compared to rural areas. * Public sector and size of the firm: Households whose head (or spouse) works in the public sector have a very low probability of being extremely poor in comparison with other households, at 3.15 percent in urban areas, and 8.30 in rural areas. This may be due to better pay in the public sector, but it may also be due to the stability of the pay (by contrast, self-employed heads face a higher risk of having low earnings in any given month, which would show up in the poverty statistics provided here), or to the better education endowment of public sector workers. Hence the reason why public sector workers are better off may not necessarily be due to a specificity of public sector employment per se. Households whose head (or spouse) is employed in a firm with more than ten workers also face a much lower probability of poverty than the average in both urban and rural areas. * Underemployment: Households with a head (and to a lesser extent a spouse) underemployed and working less than 20 hours a week face a higher probability of being poor than the average in both urban and rural areas. The desire to work more, and the impossibility to work more for health or family reasons are also associated with higher poverty measures in rural areas (but not in urban areas), for both the household head and the spouse. Although the share of the population with a 13 seriously underemployed head (working less than 20 hours per week) is similar in urban and rural areas, the extent of serious underemployment for spouses is much larger in rural than in urban areas. Together with the larger incidence of poverty for the underemployed, this points to a problem of underemployment in rural areas more severe than in rural areas where non-farm employment opportunities may be lacking. Indigenous populations: One drawlback of the EPHPM surveys is that they do not have any question which permits the identification of indigenous populations. There is substantial evidence in Central and Latin America that indigenous groups have high probabilities of being poor, and that they suffer from discrimination in labor markets. It would be important to include a question on indigenous status (or languages) in subsequent surveys to test whether indigenous populations suffer more from poor living conditions and opportunities as is the case in other Central and Latin American countries. E. INEQUALITY HAS INCREASED, AND SOME GOVERNMENT PROGRAMS CONTRIBUTE TO THIS 1.17. Beyond absolute levels of income (which can be measured by poverty), well-being depends on relative levels of income (which can be measured by inequality). According to relative deprivation theory, individuals do not assess their levels of welfare only with respect to their absolute level of income. They also compare themselves with others. Thus, for any given level of mean income in a country, a high level of inequality has a direct negative impact on well-being (this is a different argument from the fact that at any given level of economic development, higher inequality implies higher poverty). 1.18. Inequality in per capita income appears to have increased by about three percentage points in the 1990s. Table 1.4 provides three measures of inequality nationally, and in urban and rural areas. The three measures - the Gini, Theil, and Atkinson indices - are defined in annex 2 (section MA.2). All three indices take a value between zero and one (the indices have been multiplied by one hundred in the tables), with a higher value indicating higher inequality. Nationally, all three indices have increased between 1991 and 1999, by 3 percentage points on average. In urban area, there has been a decrease in inequality, but in rural areas, there has been an increase. Although this is not shown in the table, the inequality between urban and rural areas has also increased, contributing to national inequality. However, as is the case for poverty, all these results should be interpreted with caution because changes in the magnitude of under-reporting in the labor force surveys may affect inequality measures. Table 1.4: Trend in inequality for labor income, Theil, Gini, and Atkinson indices, 1991-99 NATIONAL URBAN RURAL Theil Gini Atkinson Theil Gini Atkinson Theil Gini Atkinson May-91 61.69 55.31 57.55 58.80 54.26 58.48 50.09 50.59 51.85 Mar-92 61.50 55.07 56.56 56.59 53.69 57.39 48.54 48.82 49.41 Mar-93 67.38 56.36 58.01 60.81 54.65 58.13 59.48 51.56 52.33 Oct-94 65.27 57.12 56.85) 58.75 54.22 55.11 62.26 55.95 54.28 Mar-95 67.01 57.53 57.40 58.02 53.90 55.81 67.13 56.80 54.49 Mar-96 71.02 58.33 59.20 62.02 54.18 57.51 59.26 54.40 52.61 Jun-97 62.22 55.32 55.51 53.29 51.64 55.34 62.59 53.92 51.26 Mar-98 69.28 58.96 63.56 54.91 53.18 56.80 72.80 58.79 63.19 Mar-99 64.15 57.76 61.97 51.25 52.05 53.94 58.95 55.23 60.41 Source: World Bank staff using EPHPM surveys. 1.19. While some Government programs such as PRAF help in reducing inequality, others such as electricity subsidies do not. Different sources of income (or consumption) have a different impact on the inequality in total per capita income (or consumption). This can be illustrated by decomposing the Gini index of inequality in income (or consumption) according to income (or consumption) sources. The 14 methodology is described in annex 2 (section MA.3). In table 1.5, we use data from the last three EPHPM surveys for the decomposition. In table 1.6, the data is from the 1998 FHIS survey. In table 1.7, the data is from the 1999 PRAF survey. The decomposition is for income in the EPHPM and FHIS surveys, and for consumption in the PRAF survey. The income sources in the EPHPM survey are: wage earnings from a primary or secondary occupation; earnings from self-employment income; pensions; subsidies (it is not exactly clear what is accounted for by this source, but it may cover subsidies for electricity and urban transportation); rents (probably from houses or apartments); transfers ("bonos", again it is not fully clear what is covered, but programs such as PRAF provide cash stipends); remittances from abroad; income support from family members, income support from other individuals; and other income sources. The income sources are similar in the FHIS survey (this was done on purpose when preparing the FHIS survey so that it could be compared to the EPHPM), but transfer income from PRAF is identified separately so that the impact of the program on inequality can be assessed directly. The consumption sources in the PRAF survey have been aggregated in the following categories: food, non durable goods, clothing, telephone, electricity, water, gas, other utilities, health, education, housing. rents, insurance, social security, taxes, and assets depreciation. The following comments can be made: * Income shares: The first column in table 1.5 provides the share of total per capita income accounted by the specific income source. In March 1998, wage earnings from a primary occupation represent 43.4 percent of total income nationally, as compared with 2.5 percent for wage earnings from a secondary occupation, and 42.5 percent for earnings from self-employment. Remittances from abroad, support from family members, and income from rents account for, respectively, 5.5, 3.0, and 1.9 percent of total income that year. All other income sources represent less than one percent of total per capita income. The situation is fairly similar in other years. Not surprisingly, wage earnings are more important in urban areas, while the share of income from self-employment is larger in rural areas. The same comments can be made for the income sources in table 1.6 from the FHIS survey. • Gini indices and Gini correlations: The second and third columns in tables 1.5 and 1.6 provide the Gini indices and Gini correlations of the income sources. The contribution of an income source to inequality depends on the product of the Gini index and the Gini correlation, rather than on the Gini index of the source per se. This is important because income sources which are small in terms of share - which is the case for most income sources - tend to be distributed highly unequally in part because only a small share of the population benefits from them. Yet these sources can contribute to a reduction in inequality when they are not highly correlated with total per capita income. This is the case for public transfers ("bonos") in table 1.5 in the March and September 1998 surveys (but curiously it does not hold in the March 1999 survey), and for the support received from relatives or other individuals in the EPHPM surveys. It is also the case for PRAF transfers in table 1.7. * Gini elasticities: The third and fourth columns provide the absolute contribution of the income sources to inequality and their Gini elasticity. The absolute contribution depends in large part on the income share. The Gini elasticity is independent from the income share since it is the product of the Gini and the Gini correlation of a source divided by the overall Gini. For policy, the key parameter is the Gini elasticity. As explained in annex 2 (section MA.3), a percentage increase in the income from a source with a Gini elasticity smaller (larger) than one will decrease (increase) the inequality in per capita income. The lower the Gini elasticity, the larger the redistributive impact of an income source. The findings for the Gini elasticities suggest that Government transfers reduce inequality, while subsidies increase inequality. This is confirmed by an analysis of the inequality in per capita consumption, in that goods which are subsidized such as electricity contribute to higher inequality. * Decomposition for consumption: The same decomposition can be applied to per capita consumption and its sources, and this is done in table 1.7 using the PRAF survey. The fact that the Gini elasticity for electricity consumption is much larger than one suggests that the subsidies granted by the Government for electricity consumption are inequality increasing instead of being inequality reducing (this is analyzed in more details in chapter 3). If the Government wanted to subsidize certain goods in order to decrease inequality (and reduce poverty), it would be much better to target the goods whose 15 Gini elasticities are lower than one. This is the case for food, gas, or even better, water. This is the main policy message from table 1.7. The rest of the information in the table makes common sense. For example, food represents a bit more than half of the total consumption of households, and the proportion is larger in rural than urban areas. Expenditures on non-durable goods and clothing represent another 20 percent of total expenditures. Health expenditures (8 percent of total consumption), rents (7 percent), and assets depreciation (5 percent) account for much of the rest (our measure of per capita consumption does not include the purchase of durable goods at the time of purchase, but rather a ten percent depreciation for all the durable goods owned by the household). 16 Table 1.5: Decomposition by source of Gin for per capita income, 1998 & 1999 EPHPM surveys March 98 September 98 March 99 Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. Sk CGk Rk S4RkGk RkGk/G Sk Gk Rk SkRaGk RkGk/G Sk Gk Rk SkR&Gk RkGk/G National Primary 0.434 0.708 0.752 0.231 0.922 0.451 0.706 0.754 0.240 0.943 0.461 0.713 0.767 0.252 0.978 Secondary 0.025 0.978 0.730 0.018 1.236 0.017 0.981 0.659 0.011 1.144 0.022 0.977 0.654 0.014 1.142 Self EmpL. 0.425 0.791 0.749 0.252 1.026 0.410 0.777 0.729 0.232 1.004 0.402 0.771 0.723 0.224 0.997 Pensions 0.006 0.993 0.664 0.004 1.142 0.007 0.993 0.713 0.005 1.254 0.008 0.993 0.739 0.006 1.313 Subsidies 0.000 1.000 0.912 0.000 1.580 0.000 1.000 0.753 0.000 1.334 0.001 0.997 0.777 0.001 1.385 Rents 0.019 0.992 0.858 0.016 1.475 0.027 0.996 0.914 0.024 1.612 0.017 0.989 0.816 0.014 1.442 Transfers 0.000 0.997 0.240 0.000 0.415 0.003 0.980 0.289 0.001 0.502 0.003 0.995 0.687 0.002 1.222 Remittances 0.055 0.974 0.738 0.039 1.245 0.049 0.970 0.706 0.033 1.214 0.046 0.964 0.659 0.029 1.136 Family 0.030 0.943 0.423 0.012 0.691 0.034 0.950 0.487 0.016 0.820 0.037 0.939 0.443 0.015 0.743 Individuals 0.002 0.998 0.679 0.001 1.175 0.001 0.997 0.007 0.000 0.013 0.001 0.998 0.315 0.000 0.562 Other 0.005 0.999 0.863 0.004 1.492 0.003 0.998 0.843 0.002 1.491 0.003 0.998 0.677 0.002 1.207 Total 0.577 0.565 0.559 Urban Primary 0.484 0.620 0.705 0.212 0.856 0.512 0.611 0.707 0.221 0.876 0.513 0.627 0.720 0.231 0.913 Secondary 0.026 0.974 0.783 0.020 1.494 0.017 0.982 0.766 0.012 1.524 0.020 0.981 0.748 0.014 1.487 Self Empl. 0.353 0.807 0.699 0.199 1.105 0.343 0.785 0.676 0.182 1.075 0.349 0.790 0.683 0.188 1.093 Pensions 0.008 0.987 0.490 0.004 0.947 0.008 0.988 0.574 0.004 1.149 0.008 0.990 0.639 0.005 1.282 Subsidies 0.000 1.000 0.849 0.000 1.662 0.000 0.999 0.575 0.000 1.165 0.001 0.992 0.589 0.000 1.183 Rents 0.027 0.985 0.792 0.021 1.527 0.039 0.993 0.884 0.034 1.779 0.023 0.980 0.752 0.017 1.494 Transfers 0.000 0.997 0.201 0.000 0.393 0.003 0.993 0.359 0.001 0.722 0.003 0.995 0.722 0.002 1.455 Remittances 0.065 0.963 0.661 0.041 1.245 0.042 0.953 0.550 0.022 1.063 0.045 0.947 0.498 0.021 0.956 Family 0.030 0.932 0.333 0.009 0.607 0.032 0.945 0.416 0.013 0.796 0.035 0.935 0.346 0.011 0.656 Individuals 0.002 0.997 0.610 0.001 1.191 0.000 0.997 0.092 0.000 0.185 0.001 0.998 0.395 0.000 0.799 Other 0.004 0.996 0.680 0.003 1.325 0.004 0.997 0.774 0.003 1.564 0.003 0.998 0.697 0.002 1.410 Total 0.511 0.494 0.494 Rural Primary 0.337 0.735 0.673 0.167 0.861 0.335 0.735 0.655 0.161 0.856 0.352 0.727 0.667 0.171 0.901 Secondary 0.023 0.973 0.616 0.014 1.044 0.017 0.974 0.512 0.008 0.886 0.027 0.968 0.610 0.016 1.097 Self Empl. 0.565 0.761 0.811 0.348 1.075 0.538 0.753 0.793 0.321 1.060 0.513 0.726 0.768 0.286 1.036 Pensions 0.003 0.998 0.791 0.002 1.375 0.005 0.996 0.779 0.004 1.379 0.007 0.996 0.815 0.005 1.507 Subsidies - - - - - - - - - - 0.001 1.000 0.925 0.001 1.718 Rents 0.002 0.996 0.698 0.001 1.210 0.003 0.995 0.735 0.002 1.301 0.004 0.995 0.620 0.003 1.145 Transfers 0.000 0.996 0.075 0.000 0.130 0.004 0.954 0.067 0.000 0.114 0.004 0.995 0.709 0.003 1.312 Remittances 0.035 0.981 0.731 0.025 1.248 0.061 0.983 0.821 0.049 1.433 0.047 0.977 0.763 0.035 1.385 Family 0.029 0.947 0.389 0.011 0.642 0.035 0.950 0.469 0.016 0.791 0.040 0.935 0.431 0.016 0.749 Individuals 0.001 0.999 0.581 0.000 1.011 0.001 0.998 -0.24 0.000 -0.42 0.001 0.997 0.116 0.000 0.215 Other 0.006 1.000 0.974 0.005 1.696 0.000 0.999 0.697 0.000 1.237 0.004 0.997 0.611 0.002 1.132 Total 0.574 0.563 0.538 Source: World Bank staff using EPHPM surveys. The estimates of the Gini index in this table differ from the estimates in table 1.4 because in table 1.4 we used labor income only, while in this table we use total income. 17 Table 1.6: Dec omposition by source of Gini for per capita income, 1998 F IS survey National Urban Rural Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. Sk Gk Rk SkRkGk RkGk/G Sk Gk Rk SkR4Gk RkGk/G Sk Gk Rk SkRkGk RkGk/G Wages 0.48 0.79 0.84 0.32 1.16 0.56 0.70 0.79 0.31 1.09 0.33 0.88 0.82 0.24 1.23 Self empl. 0.43 0.72 0.66 0.20 0.83 0.36 0.75 0.58 0.16 0.84 0.56 0.68 0.79 0.30 0.91 Pensions 0.01 0.99 0.68 0.00 1.18 0.01 0.97 0.58 0.00 1.09 0.00 1.00 0.75 0.00 1.28 Subsidies 0.01 0.99 0.68 0.00 1.18 0.01 0.97 0.58 0.00 1.09 0.00 1.00 0.75 0.00 1.28 Rents 0.01 0.99 0.65 0.00 1.12 0.01 0.99 0.60 0.00 1.14 0.00 0.99 0.57 0.00 0.96 PRAF transfers 0.01 0.93 0.10 0.00 0.16 0.00 0.97 0.40 0.00 0.76 0.01 0.87 -0.16 0.00 -0.23 Remittances 0.03 0.98 0.58 0.02 0.99 0.03 0.98 0.71 0.02 1.35 0.03 0.97 0.37 0.01 0.60 Family 0.03 0.97 0.47 0.01 0.80 0.01 0.96 0.20 0.00 0.37 0.06 0.97 0.64 0.03 1.05 Individuals 0.00 0.99 -0.02 0.00 -0.04 0.00 0.99 0.33 0.00 0.65 0.00 0.99 -0.28 0.00 -0.47 Other 0.01 0.99 0.78 0.01 1.35 0.01 0.98 0.76 0.01 1.45 0.00 0.99 0.26 0.00 0.45 Total 0.57 0.51 0.59 Source: World Bank staff using 1998 FHIS survey. Table 1.7: Decomposition by source of Gini for per capita consumption, 1999 PRAF survey National Urban Rural Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. Share Gini Cor. Abs. Elas. ___________ Sk Gk Rk SkRkGI, RkGk/G Sk Gk Rk SkRkGk RkGk/G Sk Gk Rk SkRkGk RkGk/G Food 0.55 0.38 0.93 0.20 0.84 0.47 0.36 0.91 0.15 0.76 0.58 0.39 0.94 0.21 0.88 Non durables 0.14 0.59 0.89 0.07 1.22 0.16 0.62 0.90 0.09 1.30 0.13 0.55 0.87 0.06 1.17 Clothing 0.05 0.56 0.69 0.02 0.91 0.05 0.55 0.69 0.02 0.88 0.06 0.56 0.68 0.02 0.92 Telephone 0.00 0.99 0.86 0.00 1.99 0.01 0.95 0.77 0.00 1.70 0.00 . 0.00 0.00 0.00 Electricity 0.01 0.93 0.69 0.00 1.50 0.01 0.81 0.75 0.01 1.41 0.01 0.96 0.61 0.00 1.42 Water 0.00 0.48 0.50 0.00 0.57 0.00 0.52 0.54 0.00 0.65 0.00 0.41 0.38 0.00 0.38 Gas 0.01 0.69 0.37 0.00 0.60 0.01 0.75 0.41 0.00 0.71 0.01 0.66 0.33 0.00 0.53 Other Utilities 0.01 0.85 0.54 0.01 1.09 0.02 0.71 0.29 0.00 0.48 0.01 0.89 0.60 0.01 1.30 Health 0.08 0.84 0.78 0.05 1.53 0.09 0.83 0.79 0.06 1.51 0.08 0.83 0.78 0.05 1.58 Education 0.02 0.71 0.58 0.01 0.97 0.03 0.66 0.52 0.01 0.79 0.02 0.71 0.54 0.01 0.94 Rents 0.07 0.57 0.67 0.03 0.89 0.09 0.59 0.76 0.04 1.04 0.06 0.53 0.58 0.02 0.75 Insurance 0.00 0.99 0.93 0.00 2.18 0.00 0.99 0.89 0.00 2.05 0.00 1.00 0.94 0.00 2.30 Social security 0.00 1.00 0.62 0.00 1.44 0.00 0.99 0.35 0.00 0.81 0.00 0.99 0.67 0.00 1.63 Taxes 0.00 0.90 0.69 0.00 1.47 0.00 0.87 0.67 0.00 1.34 0.00 0.91 0.68 0.00 1.51 Assets deprec. 0.05 0.74 0.79 0.03 1.38 0.06 0.76 0.83 0.04 1.47 0.04 0.72 0.76 0.02 1.35 Total 0.43 0.43 0.41 Source: World Bank staff using 1999 PRAF survey. 18 Box 1.3. COMPARING POVERTY LEVELS AND TRENDS IN HONDURAS WITH LATIN AMERICA According to estimates for Latin America based on 13 countries (Wodon et al., 2001), poverty affected a third of the Latin America population in 1998. Extreme poverty, defined as the inability to pay for food needs, affected one of every six people. More precisely, using per capita income-based poverty measures adjusted to per capita consumption in the National Accounts to correct for underreporting, it is estimated that 34.62 percent of the Latin America population was poor in 1998. The share of the population with per capita income below the extreme poverty line was 16.14 percent in 1998. Honduras' own poverty levels are much higher, which is not surprising given the fact that in terms of per capita GDP, Honduras is one of the poorest countries in the region. It is interesting to point out that for the region as in Honduras, urbanization contributes to the reduction in poverty. In Latin America as a whole, the population weighted headcount index of poverty increased in urban and rural areas between 1986 and 1998 by respectively two and five percentage points, while it increased at the national level over the same period by only slightly more than one percentage point. The same is observed for extreme poverty. This apparently surprising result is due to the fact that the share of the population living in urban areas in the region has increased over time, from 68.7 percent in 1986 to 74.6 percent in 1998, and poverty is much lower in urban than in rural areas. It is fair to say that a household migrating from rural to urban areas faces a lower probability of being poor at its place of destination than at its place of origin, so that urbanization contributes to poverty reduction over time. Whether the poverty estimates for Latin America in 1998 are encouraging or not depends on one's time horizon. The headcount indices of poverty observed in 1998 are significantly below those observed in 1992, which suggests progress in the 1990s. This progress is due in part to Brazil where poverty reduction has been substantial between 1992 and 1996. The progress would have been stronger without the 1995 crisis that hit Mexico, where a dramatic increase in poverty was observed in 1996. Still, despite the progress achieved in the 1990s, the shares of the population living in poverty and extreme poverty in 1998 remain high, and the region is only now back to the poverty levels observed in 1986 (at 33.35 and 14.40 percent, respectively). This indicates that the economic recovery of the 1990s and the associated reduction in the share of the population in poverty has been just large enough to compensate for the "lost decade" of the 1980s. A reduction in the number of the poor and extreme poor in the 1990s is observed, but this reduction is small due to population growth. If the comparison is made with 1986, using the estimates based on 13 countries, the number of the poor has increased. In 1998 there were 37 million more poor people than in 1986, and 22 million more people in extreme poverty. A different picture emerges using non-weighted poverty measures in which countries such as Bolivia or Honduras receive the same weight as countries such as Brazil or Mexico. When all countries receive the same weights, one observes a more consistent reduction in poverty throughout the period in review. Thus the number of countries for which there has been progress (and the extent of this progress) is larger than the number of countries for which there has been a deterioration. Still, the reduction in the magnitude of the poverty measures remains limited. The poverty gap and squared poverty gap are better measures of poverty for evaluation purposes than the headcount. If one wants to pay more attention to the poorest of the poor, the squared poverty gap should be preferred as a measure of poverty. But the conclusions reached with these alternative measures of poverty are similar to those reached with the headcount index. 19 CHAPTER II: M[ICRO DETERMINANTS OF POVERTY A. REGRESSIONS ARE BETTER THAN PROFILES FOR ANALYZING THE DETERMINANTS OF POVERTY 2.1. While it is standard to provide a poverty profile in a report on poverty, it is better to provide regressions that give insights into the determinants of poverty. As was mentioned in the previous chapter, a poverty profile is a set of tables giving the probability of being poor according to various characteristics, such as the area in which a household lives or the level of education of the household head. The problem with a poverty profile is that while it gives information on who are the poor, it cannot be used to assess with any precision what are the determinants of poverty. For example, the fact that households in some areas have a lower probability of being poor than households in other areas may have nothing to do with the characteristics of the areas in which the household lives. The differences in poverty rates between areas may be due to differences in the characteristics of the households living in the various areas, rather than to differences in the characteristics of the areas themselves. To sort out the determinants of poverty and the impact of various variables on the probability of being poor, regressions are needed (see annex 2, section MIA.3). In this chapter, we provide the results of such regression. 2.2. To assess the impact of various characteristics on the probability of being poor, it is better to rely on linear rather than categorical regressions. Many analysts use categorical regressions such as probits and logits to analyze the deterrninants of poverty. These regressions assume that the (per capita) income of households is not observed: the analyst only knows whether a household is poor or not. There are three problems with these regressions. First, the analyst is throwing away relevant information (the distribution of income). Second, the regression coefficients are more likely to be biased with categorical regressions than with linear regressions. Third, when categorical regressions are used, it is not possible to predict the change in the probability of being poor following a change in the poverty line. In our linear regressions, the dependant variable is the logarithm of per capita nominal income divided by the poverty line, so that a value of one indicates that the household is at the level of the poverty line. Separate regressions are provided for the urban and rural sectors. Apart from a constant, the regressors include: (a) geographic location according to Honduras' departnents (because some departments are not represented in the survey, this results in 15 departmental dummies, Atlantida being the omitted reference area); (b) household size variables and their square (number of infants, children, and adults), whether the household head is a woman, the age of the head and its square, whether the head has a spouse or not, and the migration status of the head (migration since the birth of the head and/or over the last five years, as well as from where); (c) characteristics of the household head, including his/her level of education; whether he/she is employed, unemployed and searching for work, or not working; his/her sector of activity; his/her position; whether he/she works in the public sector; the size of the firn in which he/she works; whether he/she is underemployed; ancl whether he/she has not been able to work due to health or family reasons; and (d) the same set of characteristics for the spouse of the household head, when there is one. Throughout the next section, we will lpresent results for the last three EPHPM surveys in order to ensure that the regression results are robust to the choice of the survey. 2.3. Annex 3 provides a user-friendly Excel® dialog box that simulates the impact of a change in household characteristics on the expected per capita income and probability of being poor. Below, only statistically significant coefficients in the regressions are reported, and the regression results are presented in small blocks according to the variables discussed in the text. For the interested reader, annex 3 also contains the full set of regressions (coefficients and standard errors) together with a user-friendly software (the diskette is available upon request) which can be used for poverty simulations'. l Our regressions can be considered as a reduced form model. For example, the impact of the household head education on per capita income may come not only from a labor income for the head, but also from the ability of households with a well educated head to save and invest, thereby generating higher capital income. Since there is 20 B. HOUSEHOLD STRUCTURE, EDUCATION, EMPLOYMENT, AND LOCATION ALL AFFECT POVERTY 2.4. With the exception of the impact of geographic location on poverty, the results presented in this section are independent of the choice of the poverty lines used for poverty measurement. As already mentioned, one advantage of using linear regressions for measuring poverty is that when the poverty lines are region-specific as they typically are (for example, one may have a different poverty line for urban and rural areas, or by department within the urban and rural sectors), only the constant and/or the coefficients of the regional dummy variables in the regression will change (this happens in a straightforward way). With linear regressions, it is thus feasible to predict poverty for any poverty line chosen by the analyst without having to rerun a new regression for each poverty line chosen (this is not the case with probits or logits where a new regression is needed for each new poverty line). We focus below on the percentage increase in per capita income associated with household characteristics, rather than on the impact on poverty because this impact depends on the initial position of the household. For example, the impact of a better education on the probability of being poor will be lower for a household who is further below from the poverty line than for a household who is closer to the poverty line (this is also the case with categorical regressions). The fact that we concentrate on the impact on per capita income also means that the results in this section do not depend on the choice of the poverty line. The reader wishing to calculate the impact on poverty for any change in household characteristics given a set of initial conditions for the household can use the Excel® dialog box provided in annex 3. 2.5. Poverty increases with the number of infants and children in the household. It decreases with the age of the head, and it is lower for households whose heads are without a spouse in urban areas. Controlling for other variables, households with a larger number of infants and children have a lower level of per capita consumption, and thereby a higher probability of being poor. This is indicated in table 2.1 by the negative coefficients in the regressions for these variables (the negative impacts are decreasing at the margin since the quadratic variables have a positive sign). By contrast, having a larger number of adults in the household helps in most cases to reduce the probability of being poor. While these results make common sense, they are to some extent sensitive to the methodological choices made for poverty measurement2. Table 2.1 also indicates that households with younger heads are also more likely to be poor, and that urban households whose head has no spouse are less likely to be poor. This is probably because controlling for female headship, a large number of urban heads without spouse are single males whose per capita income does not have to be shared with (many) other family members. From a policy no attempt here in our regressions to model the structure and dynamics of income generation, we should be careful in the interpretation of the coefficient estimates because the percentage increase in per capita income that they represent may capture a number of different factors. Nevertheless, the regression results do provide a feel for the principal factors affecting income and thereby poverty, and they can be used to provide insights for public policy. 2 By using per capita income as our indicator of well being, we do not allow for economies of scale in the household, nor for differences in needs between household members. By ruling out economies of scale, we consider that the needs of a family of eight are exactly twice the needs of a family of four. With economies of scale, a family of eight having twice the income of a family of four would be judged better off than the family of four. Thus, not allowing for economies of scale over-estimates the negative impact of the number of infants and children on poverty. Moreover, by ruling out differences in needs between household members, we do not consider the fact that larger households with many children may not have the same needs per capita than smaller households because the needs of infants and children tend to be lower than those of adults. In other words, our poverty line measures the cost of basic needs for an "average" individual, but very large families do not consist of average individuals because infants and children are over-represented in them. Not considering differences in needs also leads to an overestimation of the impact of the number of infants and children on poverty. Nevertheless, even if corrections were made to take into account both differences in needs and economies of scale within the household, a larger number of infants and children would still lead to a higher probability of being poor, so that a reduction in fertility will still reduce poverty. 21 point of view, the main implication of table 2.1 is that policies enabling women to take control of their fertility are likely to help in reducing fertility. This is discussed further in chapter 4. Table 2.1: Marginal percentage increase in per capita income due to demographic variables [The excluded reference categories are a household with a male head and a spousel March 1998 September 1998 March 1999 Urban Rural Urban Rural Urban Rural Number of infants -0.26 -0.30 -0.27 -0.25 -0.27 -0.25 Number of infants squared 0.02 0.03 0.03 0.02 0.03 0.03 Number of children -0.33 -0.31 -0.33 -0.28 -0.29 -0.27 Number of child squared 0.04 0.03 0.04 0.03 0.03 0.02 Number of adults NS NS NS NS -0.07 NS Number of adult squared 0.01 0.01 NS 0.01 0.01 NS Female head -0.14 -0.21 -0.14 -0.29 -0.15 -0.21 Age of the head 0.01 NS 0.01 0.01 0.02 0.02 Age of the head squared 0.00 NS 0.00 NS 0.00 NS No spouse for the head 0.31 NS 0.38 NS 0.50 NS Source: World Bank staff using EPHPM. NS mieans not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 2.6. Female headed households have per capita income levels 15 to 30 percent lower than male headed households. The factors leading to female headship differ between urban and rural areas. The negative impact of having a female head is larger in rural areas than in urban areas (table 2.1). Bradshaw (1995) argues that the reasons why women end up being household head in Honduras differ between urban and rural areas. In rural communities, most of the female heads are widowed (this does not take into account cases when a vvoman is temporarily the head due to the migration of the male partner). Male desertion remains rare, in part because men are often tied to their land (they are reluctant to abandon their wives and children if this entails a sacrifice of their assets). Another reason for the stability of rural marital unions lies in the "respect" that the man and woman have for each other. This "respect" is in part due to the fact that unions are formalized through religious marriages. At the same time, while the number of female headed households in rural areas due to the departure of the man is smaller than in urban areas, lone rural mothers find it more difficult to support themselves, as indicated in our regressions. In the case of a separation, land rights remain with the man. Once alone, a woman usually .emains alone, out of "respect." Younger women, especially those who have separated, have few opportunities apart from migration to urban areas or a return home to live with the woman's parents. In urban areas, male desertion is more common. This can be attributed among others to the lack of ties to the land and the lower importance of the notion of "respect". Female-instigated separation is also more common, at least when the woman has a source of income, in part because the stigma attached to female headship is lower in urban than in rural areas. The upshot of this analysis and of our regression results is that while female headship is a drawback, it may be harsher in rural areas. 2.7. The gains from education are suijbstantial. A household with a head having gone to the university (superior level in table 2.2) has twice the expected level of income of an otherwise similar household whose head has no education at all. Completing secondary schooling brings in an 70 to 80 percent gain versus no schooling. Completing primary school brings in a 30 to 40 percent gain. There are no large differences in the gains for the head in urban and rural areas despite the fact that there may be more opportunities for qualified workers in urban areas (the only systematic difference is at the university level). The gains from a well educated spouse are also large and similar in urban and rural areas, but they are smaller than for those observed for the head. This is not surprising given that the employment rate for women is smaller than for men for all levels of education, so that women use their education endowment less than men. Another explanation could be that there is gender discrimination in pay. Some evidence of discrimination against women was found by Bedi and Born (1995) using data for 1990. 22 Table 2.2: Marginal percentage increase in per capita income due to education [The excluded reference categories are a household head and a spouse with no education at all] March 1998 September 1998 March 1999 Urban Rural Urban Rural Urban Rural Household head Primary partial 0.25 0.20 0.18 0.18 0.21 0.19 Primary total 0.34 0.38 0.31 0.26 0.38 0.40 Secondary partial 0.53 0.56 0.52 0.26 0.59 0.50 Secondary total 0.69 0.81 0.66 0.79 0.73 0.88 Superior (university) 0.96 0.83 0.88 0.52 1.06 0.76 Household spouse Primary partial 0.18 0.20 0.11 0.10 0.15 0.15 Primary total 0.17 0.19 0.20 0.20 0.14 0.16 Secondary partial 0.28 0.43 0.19 0.33 0.27 0.21 Secondary total 0.39 0.54 0.38 0.58 0.36 0.52 Superior (university) 0.65 0.68 0.75 0.68 0.60 0.56 Source: World Bank staff using EPHPM. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 2.8. Results from wage regressions confirm the large impact of education, and the higher gains associated with higher levels of schooling. Another way to measure the impact of education consists of running Heckman regressions for labor income as a function of education and experience (see annex 2, section MA.4 for details). To look at the trend over time in the retums to education, we ran Heckman regressions for 1989, 1994, and 1996 using the EPHPM. From these regressions, rates of return to (or more precisely marginal gains from) education were computed. Those are given in table 2.3. For example, in urban areas in 1999, an increase from six to seven years of schooling generates an increase in labor income of nine percent, as compared to 14 percent from 15 to 16 years of schooling. The structure of these gains is similar to that of other Latin American countries in that the marginal gains increase with the education level. The gains have remained stable over the decade, but they are about two percentage points higher than in other Latin America countries (Wodon, 2000)3. Table 2.3: Marginal percentage increase in labor income with more education by level, men only Urban Rural 1989 1992 1996 1999 1989 1992 1996 1999 6 to 7 years of schooling 0.12 0.10 0.09 0.09 0.13 0.10 0.11 0.11 9to lOyearsofschooling 0.13 0.11 0.11 0.11 0.15 0.12 0.12 0.14 12to 13 years ofschooling 0.15 0.13 0.12 0.13 0.17 0.13 0.13 , 0.16 15 to 16 years of schooling . 0.16 0.14 0.14 0.14 0.18 0.15 0.14 0.18 Source: World Bank staff using EPHPM. The results are derived from statistically significant coefficients. 2.9. While a better education clearly helps in escaping poverty, it is not enough if only one household member is working. As explained in annex 2 (section MA.4), we also used the results from the Heckman labor income regressions to estimate the projected earnings of a household with only one 3 Using data from the 1990 EPHPM survey, Bedi and Born (1995) have extended the analysis further to assess the impact of the level of schooling and type of experience on wages, and to investigate the credibility of the screening argument according to which apart from the number of years of schooling, what matters is the degree obtained by an individual. They find that the rate of return to schooling increases with higher levels of education (as is also found in our own estimations in table 2.3), with the rate of return to primary schooling being usually lower than the rate of return to other levels of schooling. Their results support the human capital theory and suggest that investment in education are not merely used as a screening device by employers. Nevertheless, they suggest that employers place a premium on graduation from elementary school, versus mere attendance. After the elementary level, years of education are seen as productivity enhancing rather than as a screening mechanism. The authors also suggest that there is some discrimination against women in the labor market. 23 male working adult as a function of the education level of that adult and his/her accumulated work experience over time. The higher the education level, the higher the future streams of income. More experience also generates more income. However, it can be shown that over the life cycle, one working adult with primary or even secondary education is not enough to help a household emerge from poverty when a typical increase in family size is taken into account to estimate the poverty line (to compare the projected earnings with the poverty threshold, one needs to multiply the per capita poverty line by the number of persons in the households after a marriage and the birth of children; for this, some assumptions are needed). In other words, the message is that in both urban and rural areas, one salary typically does not enable a household to emerge froim poverty unless the education level of the working adult is very high. This is why it is important to improve employment, training, and earnings opportunities for women. At the same time, there will be a limit to the increase in the labor force participation of women observed over the last decade, so that this increase cannot be the base of a long term sustainable strategy for poverty reduction in Honduras. 2.10. The inability to escape poverty with only one wage earner does not imply that measures such as minimum wages are useful and beneficial for the poor. Following Hurricane Mitch, inflation reached 12 percent in the first half of 1999. This led the Government to increase the minimum wage by 25 percent as of July 1, 1999 (Decreto No. 00-94) to 45.20 Lempiras per day (1,356 Lempiras per month). With fringe benefits, this translates to 52.73 Lempiras per day (1,582 Lempiras per month.) In January 2000, the minimum wage was increased by another six percent. In principle, the impact of minimum wage legislation on poverty is uncertain. On the one hand, those who benefit from a minimum wage may enjoy higher salaries, and this may lead to lower poverty. On the other hand, if the level of the minimum wage is higher than the marginal productivity of some workers, these will lose their employment, which may increase poverty. Assessing the impact of Honduras' minimum wage on poverty goes beyond the scope of the present study, but there is one question which can be answered. For any one or both of the above effects to be observed, the minimum wage must be binding, and there is no certitude a priori that it will be because countries such as Honduras lack the capacity to enforce their minimum wage legislation. One might think that due to enforcement constraints, minimum wages would tend to protect formal workers, while many of the poor are employed in the informal sector. But this could be a fallacious argument, because informal workers might adjust to formal minimum wages. It turns out that in the case of Honduras, the minimum wage does not appear to be highly binding despite the fact that the minimum wage is set at a very high level in Honduras in comparison with other countries. However, the minimum wage ends up being costly for public expenditures because of its ripple effects on the pay of public employees (e.g., teachers and physicians). That is, increases in the minimum wage may wipe out scarce budgetary resources which could be used for poverty reduction. 2.11. Employment patterns for the head and spouse also have a large impact on per capita income and thereby on poverty. The regression specification enables us to look at various issues (table 2.4): Unemployment: Not surprisingly, having a head searching for employment has a very large negative impact on per capita income in both urban and rural areas. On average, if a value of zero is given to the coefficient which is not statistically significant, the household suffers from a drop in income of 65 percent as compared to the case wrhen the head is fully employed (this is excluded reference category in the regression). The impact is also negative if the spouse is searching for a job, although it is less often statistically significant. These results probably overstate the impact of unemployment on income, because households use smoothing strategies in order to cope with unemployment (the volatility of consumption expenditures is lower than the variability of income because households save and borrow). Still, the fact that unemployment can lead to serious consequences for income is clear. By contrast, households with a head not working have higher levels of income, which suggests that those heads who are not in the labor force can afford not to be working. To some extent, the same is true for the spouse, in that in most cases not being in the labor force does not reduce income. 24 Table 2.4: Marginal percentage increase in per capita income due to employment variables [The excluded reference categories are a household head and a spouse fully employed (at work and not underemployed), and working as wage earners (as opposed to self-employment) in the agriculture sector] March 1998 September 1998 March 1999 Urban Rural Urban Rural Urban Rural Employment of head Available (unemployed) NS NS -0.27 NS NS NS Searching (unemployed) -0.63 -0.48 -0.76 -1.29 -0.73 NS Not working 0.30 NS 0.30 NS 0.42 NS Employment of spouse Available (unemployed) -0.09 -0.23 -0.18 NS -0.08 -0.06 Searching (unemployed) NS NS NS -0.89 -0.29 NS Not working NS -0.75 NS NS NS NS Sector of activity of head Mining/Manufacturing/Electricity NS 0.18 0.14 0.16 NS 0.20 Construction 0.16 0.31 0.28 0.37 0.18 0.49 Commerce NS 0.42 0.28 0.39 0.17 0.46 Transport 0.23 0.45 0.36 0.50 0.34 0.49 Services NS 0.16 0.14 NS NS 0.20 Sector of activity of spouse Mining/Manufacturing/Electricity NS -0.27 NS NS NS NS Construction NS 1.07 NS 0.97 NS 1.10 Commerce NS NS 0.22 NS 0.37 0.34 Transport NS NS NS -0.39 NS 0.92 Services NS -0.46 NS NS NS NS Type of employment of head Self-employed 0.10 -0.24 0.07 -0.21 0.11 -0.26 Employer 0.58 0.93 0.53 0.78 0.63 0.54 Unpaid family work -0.70 -1.03 NS NS NS -0.78 Public sector 0.09 NS NS NS NS NS Size of firm> 1O people 0.12 0.22 0.19 0.18 0.17 0.17 Type of employment of spouse Self-employed 0.12 -0.37 NS NS NS -0.17 Employer 0.33 NS NS NS NS NS Unpaid family work NS -0.38 NS NS -0.32 -0.44 Public sector NS NS NS NS NS NS Size offirm> 1O people 0.16 NS NS NS NS NS Underemployment of head Hours of work per week < 20 -0.33 -0.30 -0.19 -0.25 -0.34 -0.35 20 to 39 hours of work per week -0.15 -0.15 NS NS -0.17 -0.19 Want to work more NS NS -0.11 NS -0.17 -0.22 Not available for health or famnily reasons 0.23 NS NS NS NS NS Underemployment of spouse Hours of work per week < 20 -0.17 NS -0.32 -0.33 -0.18 -0.38 20 to 39 hours of work per week -0.18 NS -0.13 -0.17 NS -0.24 Want to work more NS -0.36 -0.19 NS NS 0.18 Not available for health or family reasons NS NS NS -1.57 0.24 NS Source: World Bank staff using EPHPM. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. * Underemployment: Having a head or spouse seriously underemployed (i.e., working less than 20 hours per week) reduces expected per capita income by about 30 percent in both urban and rural areas. Counting as zeroes the coefficients which are not statistically significant, milder underemployment (i.e., working between 20 and 39 hours a week) reduces per capita income by 25 about 10 percent. Those who would like to work more tend to be poorer, although the impact is not always significant. Controlling for all the other variables, there is no definite direction of the impact of wanting to work more, but not being able to do so for health or family reasons, although in one case (spouses in rural areas in the September 1998 survey), the negative impact is very large. * Sector of activity: Having a head belonging to the construction, commerce, or transport sector brings in a gain in per capita income of about 30 percent as compared to working in agriculture (the excluded reference category). By contrast, households with heads working in services and mining/manufacturing/electricity do not do much better than households with heads in agriculture. In the case of services, this may be because a large part of this sector consists of informal and badly paid workers (in other words, the sector is highly heterogeneous). In the case of mining, manufacturing, and electricity, the result is more surprising. The impact of the spouse's sector of activity tends to be smaller than that of the household head (many coefficients are not significant). * Position held: Self-employment versus salaried employment (the excluded category in the regression) is good in urban areas, and bad in rural areas. The difference can be explained by the fact that in urban areas, the self-employed include a larger number of professionals. As expected, being an employer generates a large gain in per capita income of about 70 percent. Also as expected, unpaid family work is associated with poverty. There is no systematic gain from being employed in the public sector as opposed to workcing in the private sector. A head working in a firm that has more than ten workers brings in a gain in per capita income of close to 20 percentage points. The results are again similar for heads and spouse, but with a lower level of statistical significance for spouses. 2.12. Reducing poverty through labor markets requires interventions not only on the quality of the jobs available, but also on the qualifications of workers, and it is difficult to disentangle both. A study by IPEA (1999) quoted in the draft I-PRSP of the GRH (2000) argues that the key problem in the labor market is the quality of the jobs available. The study states that 84 percent of the difference in labor income between Honduras and other Latin American countries is due to the lower quality of the jobs, the rest (16 percent) being due to the lower quality of the workers. It is not clear what must be understood from the above. While there is no doubt that the quality of many jobs in Honduras is low, the lack of qualification of the labor force may well be at the root of the problem. If it is, the policy option would be to improve the qualification of the workforce. There are certainly other factors affecting the quality of the jobs available in Honduras (some are reviewed in chapter 3). But one should be careful before interpreting the IPEA decomposition as implying that the qualification of the workers is not a key issue. 2.13. Apart from the quality of the jobs available and the qualification of the workers, the lack of work remains also a problem for 10 to 20 percent of households. As noted in IPEA (1999) study and in the draft I-PRSP (GRH, 2000), Honduras has created many jobs in the 1990s, and this has helped in, among others, the absorption of a higher number of women in the labor force both in absolute and in percentage terms. However, the fact that the rate of unemployment is low by international standards is in part due to the fact that the poor simply cannot afford not to be working. In Honduras as in other low- income countries, there is no unemployment insurance and there are few cash transfer programs. Despite the fact that the poor cannot afford not to work, one out of ten households has a head that is either unemployed (and available or searching for work) or underemployed and willing to work more (table 2.5)4. The same is true for household spouses. Given the negative impact of these situations, the lack of work remains a problem, and this problem has risen with Hurricane Mitch. Between March 1998 and March 1999, the increase in unemployment and the desire to work more for those who are working is about two percentage points on average across the samples in table 2.5. There is no easy policy answer to the problem of unemployment and iunderemployment. Some countries have implemented public works programs at wages below minimum pay to ensure that the very poor can have some earnings during bad 4 In general, the measures in table 2.5 are provided for the labor force rather than for heads and spouses; we provide the measure for heads and spouses because this provides a link with the regression estimates. 26 times. It is likely that some of the public funds now used for subsidies would be better employed in this type of programs, provided the self-selection mechanism through low wages is effective. More wvork is needed, however, to test whether this would be an appropriate policy for Honduras, given that there are other types of programs that could have a higher impact on poverty reduction. Table 2.5: Employment, underemployment, and unemployment, percentage of heads and spouses Household head Household head's spouse Urban Rural Urban Rural 1998 1999 1998 1999 1998 1999 1998 1999 Wanttoworkmore(1) 5.54 8.06 4.00 6.95 1.97 4.66 1.65 2.11 Unemployed and available (2) 1.45 2.43 0.91 0.99 5.59 6.52 7.48 6.93 Unemployed and searching (3) 3.43 2.25 0.76 0.93 1.06 0.47 0.55 0.27 Total (1)+(2)+(3) 10.42 12.74 5.67 8.87 8.62 11.65 9.68 9.31 Source: World Bank staff using EPHPM. 2.14. More employment opportunities would help to reduce poverty, provided the rise in employment is demand driven and pro-poor. While unemployment (and underemployment) is a key determinant of poverty in Honduras at the household level, it need not be at the aggregate level. To assess what would be the impact of an increase in employment on aggregate poverty, we run simple simulations whose results are reported in table 2.6. Among the urban adult (aged 25 to 60) male population that is not earning labor income in the survey, we select individuals to whom we give jobs. We give the jobs to either the poorest or the richest (according to their per capita household income) unemployed individuals in the sample. For these individuals, we predict earnings corresponding to their education and experience. The predicted earnings are obtained using Heckman regressions as mentioned in annex 2 (section MA.4). The total number of individuals put to work in the simulations is equal to five percent of the urban adult male population at work in the survey. The simulations are done assuming no change in aggregate wages. That is, we assume a demand-driven expansion in which both the demand for and supply of labor move to the right in a classic supply and demand diagram. The values given in table 2.5 are the percentage point reduction in the measures of poverty obtained with the simulation'. A demand driven expansion that helps the poor land jobs leads to a large decrease in extreme poverty (-2.52 points for the headcount) and poverty (-3.24 points). The impact is similar for the poverty gap and squared poverty gap. Since these poverty measures are smaller in absolute terms than the headcount index, this indicates a larger relative impact in terms of proportionate gains. However, if those who are comparatively richer get the jobs rather than the very poor, there is no reduction in extreme poverty because none of those who get the jobs is extremely poor, and there is a small reduction in poverty because some are moderately poor. These results are rough and indicative at best, but they help to highlight two basic conditions for employment generation to be poverty reducing: it has to be demand driven (i.e., it should not lead to a decrease in the aggregate level of wages), and it should be pro-poor. Table 2.6: Reduction in poverty from an increase in employment without a decrease in wages Extreme poverty Poverty Po PI P2 Po PI P2 Poorest 5% individuals -2.52 -1.55 -1.23 -3.24 -2.35 -1.83 Richest 5% individuals 0.00 0.00 0.00 +8.17 0.00 0.00 Source: World Bank staff using March 1999 EHPHM. 5 As a reminder, the headcount index P0 captures the shares of those with household per capita income below the poverty line; the poverty gap Pi measures the distance separating the poor from the poverty line; and the squared poverty gap P2 measures the square of this distance. If more weight is given to the poorest of the poor, the square poverty gap is a better measure than the poverty gap, and the poverty gap is a better measure than the headcount index. A policy which helps the very poor will not reduce the headcount index if those who are helped do not cross the poverty line, but it will reduce the square poverty gap and (typically to a lesser extent) the poverty gap. 27 2.15. Controlling for household characteristics, geographic location also has an impact on income. Differences in per capita income remain between departments even after controlling for a wide range of household characteristics. In the regressions, the impact of geography is measured with dummy variables for all departments except Atlantida which is the reference department. In table 2.7, the coefficient -0.21 for urban areas in Copan in March 1998 means that an urban household in Copan has an expected per capita income 21 percent below an otherwise similar urban household in Atlantida. By contrast, since the coefficient for rural areas in Copan is not significant for that survey, a rural household in Copan has the same expected per capita income as an otherwise similar rural household in Atlantida. The coefficients are thus measures of how various departments fare versus Atlantida. Many of the results are as expected. Apart from Atlantida, Cortes (where San Pedro Sula is located) does well. Departments such as Comayagua, Choluteca, Intibuca, Lernpira, and Yoro tend to be poorer. There are a few surprises, such as the low performance of Francisco Morazan where the capital Tegucigalpa is located. This may be due to the lack of representativity of the EPHPM data at the departmental level within urban and rural areas. This lack of representativity suggests caution in interpreting the results department by department. But the message that geography does matter even after controlling for observable household characteristics remains valid and important. It also gives a rationale for so-called poor areas policies (e.g., investments in infrastructure), because if geographic effects matter for poverty reduction, the characteristics of the areas in which households live must be improved alongside the characteristics of the households themselves. More work is needed, however, to assess exactly which types of poor areas policies to adopt. Table 2.7: Marginal percentage increase in per capita income due to geographic location [The excluded reference category is the clearp ent of Atlantidal March 1998 September 1998 March 1999 I Urban Rural Urban Rural Urban Rural Colon NS NS NS NS NS -0.26 Comayagua NS NS -0.36 -0.30 -0.17 -0.24 Copan -0.21 NS -0.44 -0.64 NS NS Cortes 0.28 0.24 NS NS 0.19 NS Choluteca -0.22 -0.48 -0.70 -0.83 -0.21 -0.41 El Paraiso NS NS NS -0.45 NS -0.30 Francisco Morazan NS -0.36 -0.22 -0.51 NS -0.41 Intibuca -0.22 -0.59 -0.23 -0.78 -0.33 -0.74 La Paz -0.23 -0.31 -0.41 -0.55 -0.30 -0.45 Lempira NS -0.21 -0.43 -0.48 NS NS Ocotepeque -0.19 0.33 -0.27 NS NS NS Olancho NS -0.30 NS -0.35 NS -0.25 Sta Barbara NS NS -0.45 -0.61 -0.64 -0.28 Valle NS -0.35 -0.35 -0.60 NS -0.36 Yoro NS NS NS -0.19 NS -0.40 Source: World Bank staff using EPHPM. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 2.16. The importance of geographic location is confirmed by wage and labor force participation regressions. To provide an additional test for the impact of geography on standards of living, we ran Heckman regressions (see annex 2, section MA.5) with a full set of geographic dummies in both the log wage and the labor force participation regressions. This was done for men aged 15 to 65 in the EPHPM surveys from March 1998, Septembier 1998 and March 1999. Labor income includes not only wages from a principal occupation, but also earnings from a secondary occupation and from self-employment. Table 2.8 gives the geographic effect when the full sample is used (i.e., not separating urban and rural areas). There is no excluded departrnent in the table, so that the coefficients measure the perforrnance of a department versus the national mean (as opposed to a comparison with a reference department). Several findings stand out. First, and not surprisingly, the direction and magnitude of many of the marginal effects for individual level earnings in table 2.8 is similar to what was observed in table 2.7 for per capita 28 household income. Moreover, some of the "surprises" observed in table 2.7 vanish in table 2.8; this is the case for Francisco Morazan for example, where expected earnings are higher than nationally. Second, in many instances, the impact of location on labor force participation has a sign opposed to the impact of location on earnings. This suggests that labor force participation is not much of a choice: in poorer departments, controlling for individual level characteristics, labor force participation is higher out of necessity . Table 2.8: Marginal impact of location on labor force participation and earnings for adult men [There is no excluded dummy; the coefficients are estimates of differences versus the national mean] March 1998 September 1998 March 1999 Earnings Work Earnings Work Earnings Work Atlantida NS NS 0.30 -0.32 0.20 -0.29 Colon NS NS 0.24 -0.28 NS NS Comayagua NS 0.24 -0.10 NS -0.16 NS Copan NS NS NS 0.48 0.16 0.63 Cortes 0.42 -0.13 0.40 -0.17 0.39 -0.1.3 Choluteca -0.25 0.28 -0.26 -0.22 -0.19 NS El Paraiso NS NS -0.09 0.26 -0.16 0.26 Francisco Morazan 0.06 -0.17 0.17 -0.24 0.19 -0.17 Intibuca -0.32 NS -0.43 0.38 -0.41 NS La Paz NS NS -0.15 NS NS NS Lempira -0.08 NS -0.16 0.35 NS 0.42 Ocotepeque 0.29 NS 0.20 NS 0.21 NS Olancho -0.13 NS NS NS NS NS Sta Barbara 0.16 -0.19 -0.17 NS -0.13 -0.18 Valle -0.24 NS NS NS NS NS Yoro 0.08 NS 0.17 -0.16 NS -0.15 Source: World Bank staff using EPHPM. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 2.17. The differences in labor force participation and wages between departments are due more to differences in the characteristics of the departments than to differences in the characteristics of the households living in the various departments. Using a methodology outlined in annex 2 (section MA.5), we tested whether differences in labor force participation and labor income between departments are due to differences in the characteristics of the individuals living in the various departments (such as education, experience, and demographics), or to differences in the characteristics of the departments themselves (which are captured by departmental dummy coefficients). Summary results in the fonn of the variance between departments in labor force participation and wages under various simulations are presented in table 2.9. In the March 1998 survey for example, using the full sample, the variance in labor force participation between departments when only differences in individual characteristics are taken into account is 0.73, which is much smaller than the variance of 2.73 when only differences in area characteristics are taken into account. This means that differences in area characteristics are more important than differences in individual characteristics in explaining labor force participation differentials between departments. The same holds for wages, and the results are robust to the choice of survey. This confirms the importance of geography in determining labor income, and it helps in justifying poor area policies. Note also that when both the individual and area effects are taken into account, the variance in labor force participation and earnings is even larger. This shows that as expected, the departments with good characteristics are also those whose inhabitants have good characteristics (e.g., a better education). 6 The signs of the departmental coefficients for earnings and labor force participation in table 2.7 can be compared, but the magnitude of the coefficients cannot because one of the equations is a probit while the other is log linear. 29 Table 2.9: Variance in departm nt wages and labor firce participation: area vs individual effects March 1998 September 1998 March 1999 Indiv. Area Both Indiv. Area Both Indiv. Area Both effects effects effects effects effects effects Effects effects effects Whole department (urban+rural) Laborforceparticipation 0.73 2.73 3.91 1.14 8.78 13.76 1.32 5.37 7.91 Wages 149.51 349.54 706.67 128.45 441.46 774.22 157.06 349.34 679.58 Source: World Bank staff using EPHPM. The numbers shown in the table are variances of differences in expected earnings and labor force participation between departments under different scenarios. The individual (area) effects scenario takes into account only the impact of differences in individual (area) characteristics between departments. The scenario with both effects takes into account both types of impacts when computing variances. See annex 2. 2.18. Finally, even after controlling for the impact of geographic location and observable household characteristics, migration is still likely to raise per capita income. The last set of variables used for the regressions for per capita income relates to migration (table 2.10). Individuals living in households where the head has migrated since his/her birth have a level of per capita income about 5 to 15 percent higher than other households. There is also an indication that migration over the last five years increases income. This is because the fact that the coefficients are not statistically significant indicates that at the place of destination, those who have migrated in the recent past do as well as those who have lived there for more than five years. Since migration typically takes place from poorer to richer areas, this suggests that the migrants are likely to do better at their place of destination than at their place of origin. More work would be needed, however, to compute the wage gains that can be expected from migration. Table 2.10: Marginal percentage increase in per capita income due to migration [The excluded reference categories are no migration since birth, or over the last five years] March 1998 September 1998 March 1999 Urban Rural Urban Rural Urban Rural Migration since birth 0.05 0.14 0.05 0.13 0.04 NS Migration to rural areas in last 5 years NS NS NS NS NS NS Migration to urban areas in last 5 years 0.13 NS NS NS NS NS Source: World Bank staff using EPHPM. NS rneans not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 30 Box 2.1: FROM THE DETERMINANTS OF POVERTY TO POLICY: SUGGESTIONS FROM LATIN AMERICA The analysis conducted for Honduras in this section was also conducted for eight other countries in Latin America, with very similar results. As suggested in Wodon et al., (2001), the analysis has a numnber of implications in terms of public policy. Some of these implications are briefly reviewed here. The analysis suggests that programs enabling women to take control of their fertility are likely to help in reducing poverty (better education for girls should help in this respect). Programs promoting earning opportunities for female heads should also have a positive impact. In Chile for example, using household survey results, the government identified in the early 1990s youths and women heads of households as target groups in need of training. This lead to the creation of two training programs: one for women (Capacitacion para Mujeres Jefes de Hogar), and one for youths (Chile J6ven). When asked whether the program improved their conditions for a job search, 61 percent of the women interviewed answered positively. The unemployment rate among program participants was found to be 15 percentage points lower after training in the program, from 58 percent to 43 percent. And the quality of employment also appeared to have improved after the training: a larger share of the women were employed as salaried workers with open-ended contracts. Salary levels and numbers of hours worked also improved. This evaluation was based on a sample of women who participated in the program from 1995 to 1997, but the analysts did not use an adequate treatment and control group methodology, so that it is not clear whether the good results obtained for the program are due to the self-selection of the participants into the program. Still, the evidence available at this stage on the program is encouraging. The large impact of education on per capita income and poverty justifies the implementation of programs such as Mexico's PROGRESA (or Honduras' PRAF). Although a majority of the funds in the program are devoted to stipends for poor rural children in primary and secondary school, the program integrates education interventions with health and nutrition interventions. The program started in 1997, and it now covers 2.6 million families, which represents 4 out of every 5 families in extreme poverty in rural areas and 14 percent of Mexico's population. The results of an evaluation conducted by PROGRESA staff and the International Food Policy Research Institute are encouraging. Female enrollment rate in secondary- level schools increased, and overall school attendance also increased, on average by one year, which should translate in future gains in labor income when the children reach adulthood. The program also improved health outcomes, and reduced morbidity rates among children 0 to 2 years of age. The fact that unemployment and underemployment can severely affect income also provides a justification for workfare and training programs which function in part like safety nets. Trabajar in Argentina is one example of a workfare program that works through public works. In this program, projects are identified by local governments, NGOs and community groups, and can provide employment for no more than 100 days per participant. Project proposals are reviewed by a regional committee, and projects with higher poverty and employment impacts are favored. Workers hired by the project are paid by the Government, specifically the Ministry of Labor. The other costs are financed by local authorities. Example of eligible projects include the construction or repair of schools, health facilities, basic sanitation facilities, small roads and bridges, community kitchens and centers, and small dams and canals. The projects are often limited to poor areas as identified by a poverty map. Wages are set al low levels, so that the workers have an incentive to return to private sector jobs when these are available. Thus, the program involves self-targeting apart from geographic targeting. 31 C. THE RURAL POOR ALSO SUFFER FRLOM A LACK OF ACCESS TO LAND, CREDIT, AND TECHNOLOGY 2.19. In this section, we review findings from the literature on the impact of programs for land titling, access to credit, and extension on rural productivity in Honduras. Although the incidence of poverty is higher in rural than in urban areas, the data used for this report (see Box 1.2 for a description of the surveys) does not enable us to look at the impact on farmers of policies for land titling, access to credit, or technology adoption'. To compensate for this weakness, we review in this section what the literature has to say on these topics in the case of Honduras. Since the empirical results we cite are not ours, we cannot test for their robustness to the specification chosen by the authors. In order to ensure quality, we report only those findings which appear to be both reasonable and based on good analysis. 2.20. Land titling programs have been implemented in Honduras to improve land security for the poor. It has been argued in the development literature (e.g., Lopez and Valdes, 1997) that insecure property rights are a source of production inefficiency, due to a disincentive to invest in land that is not securely held, and to credit constraints that small farmers may face (without a legal title, they cannot offer their land as loan collateral). In Honduras, during 1983-94, USAID funded a large land-titling program for small farmers (Proyecto de Titulacion de Tierra para los Pequenos Productores). The percentage of farmers with legal land titles increased from 11 to 56 percent during this period. According to the Instituto Nacional Agrario, the prograrn was to benefit small to medium sized producers by: i) granting them more secure property titles, and thereby encouraging higher investments; (ii) providing collateral to improve access to credit; and (iii) providing technical assistance. An additional benefit would be that secure titles would facilitate land transactions and thus improve the functioning of rural land markets. 2.21. There is no consensus in the literature on the overall impact of land titling programs in Honduras, but the evidence that exists confirms that much more than land titling is needed to ensure a positive impact on small farmers. Lopez (1996) suggests that the USAID program raised the income of farmers significantly by generating higher investments, especially in coffee trees and coffee drying patios. Other studies, however, point to the importance of complementary factors. Jansen and Roquas (1998), relying on qualitative methods, argue that the impact of the land titling program was limited, and appears to have triggered new sources of land conflict. The problems identified by Jansen and Roquas point to the importance of an appropriate legal framework, and transparent implementation and enforcement mechanisms, including a fair and expeditious judicial system. A study by Larso and Palaskas (1999) covering 235 farms (1 77 farms with titles in Santa Barbara and 58 farms without title in Ocotopeque) points to the importance of technical assistance and access to credit. The authors argue that land titling has a larger impact on farmers with access to markets, with the means to take advantage of these markets, and with tenure insecurity before titling. A set of regressions produced by the authors and reproduced in table 2.11 suggests that while technical assistance matters for the both the adoption of better technologies and the investment in new coffee trees, land titling has a positive effect only on investments in new coffee trees. The lack of impact of titling on access to credit suggests that while titling can, in principle, help smaller farmers by providing collateral, as in the rest of Latin America, small farmers in Honduras rarely have access to formal credit. Rural credit markets in Latin America tend to operate as small clusters of highly localized borrowers and lenders who know and trust each other, as a result of which little or no collateral may be required on loans (Lopez and Valdes, 1997). 7 Honduras has a recent survey of agricultural farmers that could be used for assessing the impact of a number of policies. Time was lacking to analyze this survey for this report, but it could be analyzed in a follow up report. 32 Table 2.11: Impact of Land Titles, Credit, and Assistance on Farm Investments and Technology New Investments Use of Fertilizer Use of Pesticide Use of Improved (log of coffee trees per mz) (Yes/No) (Yes/No) Seeds (Yes/No) Constant NS -3.121 NS NS Title 1.039 NS -0.591 -0.536 Education NS 0.067 0.074 0.094 Credit 0.135 0.149 0.082 0.078 Technical assistance NS 0.593 0.4929 0.618 Off-farm income NS -0.059 -0.038 NS Age NS NS NS NS 1988 NS 1.501 0.733 1.063 1993 NS 1.486 0.757 NS Sample selection correction NS NS NS NS Source: Larson, Palakas, and Tyler (1999: 375). NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 1O% level. Coefficients not underlined are significant at the 5% level. 2.22. The evidence on the impact of technical assistance is also mixed. Lopez and Valdes (1997) find that in Honduras (as in Chile and Colombia) technical assistance has no significant effect on per capita income. But Martin and Taylor (1995) argue that technical assistance can help, although the way through which households learn about extension is key. The authors examined the impact of multiple media in Honduras for providing extension services in order to promote a range of technologies that varied across crops and regions. The data used for the analysis were gathered in 1990 in the Comayagua region. Information materials were distributed to fanners using television, pamphlets with self explanatory illustrations, and radio. The study identifies the adoption rates for the new technologies as a function of the primary learning source for the farmers. The farmers are broken into two groups: those who produce for the market (tomatoes and rice for example), and those who produce for their own subsistence (maize and beans for example). Table 2.12 presents the main findings. For each type of farmer, the column shows the ways in which the farmers hear about new technologies. The second column indicates the marginal impact of learning about the new technology on the probability to adopt the technology (logit model). The authors find that having a personal contact with experts is important in promoting new technology, in that learning through Government and sales people has the highest impact on the probability of adoption. Learning from a friend about the new technology leads to adoption only for commercial farmers. Learning through a pamphlet or through the radio does not lead to adoption (while a personal contact with an expert is more likely to lead to adoption, the cost of this information strategy is also higher). Finally, the authors suggest that TV announcements may help through a multiplier effect on the impact of personal contact with friends or with Government and sales experts. Table 2.12: Impact of Land Titles, Credit, and Assistance on Farm Investments and Technology Subsistence crops Commercial crops How did Marginal How did Marginal farmer learn? impact on farmer learn? impact on (percentage) adoption (percentage) adoption Radio 1.1 NS 0.4 NS Government extension agents 12.3 2.15 18.2 1.74 Pamphlet 0.0 NS 1.6 NS Sales person 1.1 3.25 2.77 1.47 Friend 13.0 NS 26.6 0.89 PVO extension agent 3.8 NS 0.8 NS Family member 10.3 NS 2.8 NS Uses technology as a matter of custom 60.4 43.5 Source: Martin and Taylor (1995). The sum of the percentages need not sum to 100. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 33 2.23. Simulations also suggest that the adoption of new technologies can have positive effects on income. Lopez-Pereira and Sanders (1992) describe the fanning environment in Honduras as characterized by highly variable rainfall, steeply sloping hillsides, and poor access to capital and credit. In part due to a lack of infrastructure, smrall farmers are unable to sell their cereals at prices offered by marketing agencies, nor can they secure loans at the market interest rate. The authors simulate the impact of the adoption of new sorghum technologies that have two main advantages: they can be used to make tortilla when maize supplies are depleted, and they can serve as animal feed. New sorghum technologies help farmers increase their stock of pigs and chickens, which leads to supplemental income (the demand for meat is increasing in Honduras). The simulations suggest that larger gains in income as a result of the adoption of new sorghum technologies are obtained in farms which already have soil conservation technology. When cereal price collapses are prevented and credit conditions are improved, the expected income effects of the new technologies are also improved. The authors discuss some constraints that prevent the adoption of the new technologies. They suggest that low producer prices could be avoided through the construction of storage bins where farmers hold their grain for a few weeks after harvest. Farmers usually sell their grain immediately after the harvest, which is the period of low grain prices. With storage space, the farmers might be able to extract higher margins by selling weeks after the harvest. However, the farmers need cash to meet credit obligations and the only source of cash is after the sale of grain. Thus, storage might not help these farmers if deadlines for payments are fixed or if access to more flexible credit is limited. Still, more work would be needed to assess whether which policies provide the largest productivity gains among the various types of investments that can be made. 2.24. To conclude, the determinants of poverty are complex. This complexity implies that the problem of poverty cannot be solved with a few "magic bullets" or policies. The issues are even more complex than suggested above when the multidimensional nature of the living conditions of the poor (i.e., non-monetary dimensions of well-being) is taken into account. In the chapters that follow, we analyze only a few of the policies that would make a difference. More work would be needed to identify the many trade-offs explicit or implicit in any comprehensive strategy for poverty reduction. 34 BOX 2.2: WHAT DOES IT MEAN TO BE POOR IN RURAL AREAS? THE STORY OF THE CABRERAS FAMILY The Cabreras' family lives in the Sierra Mountains. Their farm sits at an altitude of 1,400 meters, 14 kilometers up the path from the hard-topped road that runs to the closest town, Marcala. Ricardo and Reina Cabreras' live in a hut, simple, stark, without windows, some two meters wide by three meters long with their seven children. The walls of the house are constructed of tree limbs tied together with vines, and the roof is made from zacate grass. The interior of the house is adorned with the daily laundry of faded t-shirts and little girls' dresses; a hornilla, a wood-burning stove made of clay, sits in the corner with two low and narrow benches along the wall. The only piece of store-bought furniture is a cut-glass cabinet, which holds the cups and dinner plates of plastic and tin and Ricardo's battery-powered radio. A water pump and a separate sleeping hut just beyond the main house are the only conveniences afforded to the family. They have never owned a car; walking has always been the only means of transportation. Reina's father, an impoverished farmer, left her, her mother and two sisters when she was eight. Reina, 27, married Ricardo when he was 21 and she was 14. By that age, she had attended school for only fifteen days, because her father, later other relatives, wanted her working in the fields. So she could neither read nor write. Ricardo who completed second grade, taught her to write her name. Ricardo grew up in the Sierra mountains with his parents, three brothers and a sister; two brothers died of fever because there was no money for a doctor. According to Ricardo, poverty was not the only channel of segregation for his family: "We were Lenca, pure Lenca. Lenca first andforemost. True Hondurans, native people, we were struggling to hold onto our land and our own language." At age 21, just weeks after he was married, Ricardo was drafted into the army for five years. A combat infantryman, he was paid US$4 a month. After paying for food, clothes, and medicine, he could send home only 25 cents each month. When he returned from the army, his family gave him a quarter of their two hectares of land. Now Ricardo works six months of each year for large landowners and six months on his own farm, growing corn, bananas, plantains and chilies for his family, plus coffee as a cash crop. At age 61, Ricardo's father continues to clear the steep fields of his farm by slashing at tree roots with his machete and yanking up weeds that threaten his coffee, that is when he is not working on the large plantations of the rich. A full day hoeing and tilling another's land will earn them 25 Lempiras per person, about two US dollars. Ricardo and his father own 8,000 coffee trees, which will produce, weather permitting, 20,000 pounds of berries in a season. But growing coffee is difficult. Each tree costs them more than US$2 in labor and materials before it bears fruit, and there is no fruit for the first three to four years. Despite the consumer prices of US$8 -US$14 for premium coffee in the United States, buyers, known as "coyotes", pay farmers only 25 cents per pound for their coffee beans. Ricardo is a member of a cooperative of small growers and a leader of the Indigenous Council of the Lenca People, a union of farmers from the mountains. He values hard work and education: "My people and I don't want any sweet music. We want our children to be educated, we want to know how to farm better. We don't want to be cheated." Despite the struggle of the Cabreras family, there are other families who are still worse off. Further up the steep mountain path, these families do not have access to doctors nor safe drinking water. When someone is sick, the local medicine man, a former soldier, prescribes indigenous herbs and plants for most ailments and, in extreme cases, a shot of penicillin if the family can afford it. When cholera infected the area several years ago, two people in the small mountain community died from lack of adequate medical assistance. Large families of nine or ten children live in one-room huts of mortar and sticks, causing the eyes of the children to be red and watery from the smoke of the fire. They are the "forgotten ones." Source: Adapted from Richards (1998). 35 CHAPTER III: NON-MONETARY INDICATORS AND BASIC INFRASTRUCTURE A. MORE PROGRESS HAS BEEN ACHIEVED FOR NON-MONETARY INDICATORS THAN POVERTY 3.1. In Honduras as in other Latin American countries, non-monetary indicators have improved more than monetary indicators. This is the case for unsatisfied basic needs. Honduras' method for measuring unsatisfied basic needs is described in the Libro Q (1994). It uses six indicators. * Water: An urban household is said to have an unsatisfied basic need for water if it does not have access to drinking water within the home or property. In rural areas, the household must have access to drinking water through a well or through a pipe. * Sanitarv installation: In urban area, a household must have access to a sanitary system which is not a simple pit. In rural areas, a household must at least have access to a simple pit. * Primary education: In urban and rural areas, the children of primary age must be enrolled. * Subsistence capacity: In urban and rural areas, the household head must have at least three years of primary education and be working; if this is not the case, there must be at least one person working for every three household members. * Crowding: In urban and rural areas, there should be no more than three people per room, bathrooms not included. * Housing: In urban areas, the house should not be ad hoc or built with debris, and it should not have dirt floors; in rural areas, it should not be ad hoc or built with debris. "Ad hoc" means improvised, i.e. the house is typically a collection of materials put together on a temporary basis. Table 3.1, which is reproduced from the draft I-PRSP by the GRH (2000) indicates that the share of all households nationally with no unmet baLsic needs according to the above criteria increased from 33 to 53 percent between 1990 and 1997 (these estimates are based on the EPHPM). There has been progress in both urban and rural areas, although as expected, the level of satisfaction is higher in urban areas. Honduras' Social Investment Fund (Fondo Hondureno de lnversi6n Social, FHIS hereafter) has been a major contributor to the improvement in satisfied basic needs related to sanitation and primary education. Table 3.1: Trend in unsatisfied basic needs, share of households, 1990 to 1997 National Urban Rural 1990 1993 1997 1990 1993 1997 1990 1993 1997 No unsatisfied basic need 33 47 53 50 57 65 20 38 42 One unsatisfied basic need 25 28 26 24 23 22 26 32 29 Two unsatisfiedbasic needs 20 15 13 13 11 8 26 19 18 Threeormore 22 10 8 13 9 5 28 11 11 Source: GRH (2000). 3.2. Another non-monetary indicator of well-being showing progress over time is UNDP's Human Development Index (HDI). The HDI is a weighted sum of three indices based on underlying indicators. The three underlying indicators deal with life expectancy, educational attainment, and per capita income. Because per capita income is included in the HDI, the HDI is a mixed indicator rather than a purely non- monetary measure of well-being. Denoting by X the value of any one of the three underlying indicators, the corresponding index is computed using a formula taking into account the actual value of the indicator and fixed minimum and maximum values. For any given country, the indices are computed as Index = (Actual X - Minimum X)/(Maximum X - Minimum X). This formula is such that for each country, the value of the indices is between zero and one. The higher the value for the index, the better the performrance of the country. The indicators and corresponding indices are: * Life expectancy, with the maximum and minimum values set at respectively 25 and 85 years; * Educational attainment, which is itself a weighted average of two components. The first component is the adult literacy rate index for which the minimum and maximum values are 0 and 100 percent. The second component is the combined gross enrollment ratio index for primary, secondary, and 36 tertiary education, with minimum and maximum values also fixed at 0 and 100 percent. In the HDI calculation, the adult literacy index and the combined gross enrolment ratio index are given equal weight, so that the educational attainment index is simply the arithmetic mean of its two components. The per capita income index is based on the logarithm of real per capita GDP measured using Purchasing Power Parity values in U.S. dollars, with the minimum and maximum values set at log(100) and log(40,000). According to UNDP, income enters into the HDI as a proxy for a decent standard of living, i.e. a proxy for "the dimensions of human development not reflected in a long and healthy life and in knowledge." It is worth noting that the way in which income enters in the HDI index has been modified for the UNDP's 1999 report as compared to previous reports. The HDI index is an arithmetic mean of the above three indices. Real GDP, life expectancy, and educational attainment are thus given equal weights of one third in the HDI. As indicated in table 3.2, progress has been achieved in raising the level of the HDI. In the table based on data from the Human Development Report 1999 (UNDP, 1999), Honduras is compared to two groups of countries. The first group consists of the four Latin American countries that are likely to participate in the HIPC debt relief initiative (Honduras, Bolivia, Guyana, and Nicaragua). The second group consists of Honduras' neighbors in Central America. Honduras has improved its HDI, from 0.515 in 1975 to 0.641 in 1997, but the progress has not been the same in all areas. For example, life expectancy has improved faster than GDP per capita. In 1997, the performance Honduras is broadly similar to that of other PRSP countries in Latin America, but remains below the level reached by Honduras' neighbors. Honduras's HDI ranking is above its GDP ranking, which suggests a comparatively good performance in health and education'. Table 3.2: Trend in the Human Development Index, 1975-97 PRSP countries in Latin America Central America countries I HO BO GUY NI All CR ES GU PA All HDI index 1975 0.515 0.524 - - 0.5195 0.741 - 0.517 - 0.629 1980 0.563 0.558 - - 0.5605 0.766 - 0.552 - 0.659 1985 0.595 0.584 - - 0.5895 0.784 - 0.563 - 0.6735 1990 0.616 0.611 - - 0.6135 0.787 - 0.588 - 0.6875 1997 0.641 0.652 0.701 0.616 0.6525 0.797 0.674 0.624 0.791 0.7215 Components of 1997 HDI Life expectancy at birth 69.4 61.4 64.4 67.9 65.775 76.0 69.1 64.0 73.6 70.675 Adult literacy rate (%) 70.7 83.6 98.1 63.4 78.95 95.1 77.0 66.6 91.1 82.45 Combined gross enrollment 58 70 64 63 63.75 66 64 47 73 62.5 Real GDP per capita 3,330 2,880 3,210 1,997 2,854.25 6,650 2,880 4,100 7,168 5199.5 Life expectancy index 0.74 0.61 0.66 0.71 0.68 0.85 0.74 0.65 0.81 0.7625 Education index 0.66 0.79 0.87 0.63 0.7375 0.85 0.73 0.60 0.85 0.7575 GDP index 0.52 0.56 0.58 0.50 0.54 0.70 0.56 0.62 0.71 0.6475 HDI and GDP ranking GDP ranking 117 108 101 121 111.75 61 108 85 56 77.5 HDIranking 114 112 99 121 111.5 45 107 117 49 79.5 GDP-HDI ranking +3 -4 2 0 0.25 16 1 -32 7 -2 Source: UNDP (1999). l Additional indicators to compare Honduras' performance with that of other countries are given in chapters 4 and 5. 37 B. POVERTY CAN BE REDUCED BY THE PROVISION OF BASIC INFRASTRUCTURE SERVICES 3.3. Despite progress in recent years, the poor still lack access to basic infrastructure services. Table 3.1 documented an improvement in the satisfaction of basic needs in Honduras. Using the Libro Q definitions, the share of all households nationally with no unmet basic needs increased from 33 to 53 percent between 1990 and 1997. Nevertheless the poor still lack access to basic infrastructure services. Basic access statistics are provided nationally, as well as in urban and rural areas, in tables 3.3 and 3.4. * Water: Nationally, only 6 percent of the households in the poorest income decile have access to water within their house, as compared to 76.5 percent in the richest decile. One fourth of the households in the poorest decile do not have water within their property, versus 2 percent in the richest decile. * Sanitarv installation: Nationally, one third of the households in the poorest decile have no sanitary installation, versus less than one percent in the richest decile. Many poor households use holes. * Electricitv: Nationally, three out of four households in the poorest decile have no access to electricity, versus less than five percent in the richest decile. * Differences between urban and rural areas: The differences are large (table 3.4). Because of the network nature of many services (water and electricity), many middle-income households in rural areas have less access than poor households in urban areas. This is also apparent in the 1999 PRAF survey of poor municipalities. In mnunicipal centers, 70 percent of households have electricity, versus 24 percent in the suburbs (and 33 percent in the whole sample). In the suburbs, only the households in the richest decile of the population have a probability of having access to electricity above one half. Table 3.3: Access to basic infrastructure services by income group (decile), national, 1999 Income decile 1 2 3 4 5 6 7 8 9 10 Access to water Public source 19.74 28.52 38.27 49.81 60.02 65.49 68.99 75.50 77.24 83.69 Private or collective 54.12 49.41 48.32 39.38 30.07 27.42 24.29 20.38 17.65 14.13 Well 12.88 9.07 6.68 5.78 5.41 2.83 3.55 2.29 3.28 2.12 River 12.93 11.26 5.44 4.02 3.60 2.97 2.38 1.41 1.54 0.00 Other 0.33 1.74 1.29 1.01 0.91 1.28 0.79 0.42 0.30 0.07 Inside the house 6.32 7.92 9.99 17.63 24.95 32.16 37.44 50.97 55.36 76.50 Inside the propriety 68.30 70.97 75.61 70.67 65.25 59.57 56.15 46.10 40.29 21.94 Outside of property 25.38 21.11 14.40 11.70 9.80 8.27 6.41 2.93 4.35 1.56 Sanitary installation None 32.15 30.80 23.45 13.97 13.43 9.04 6.55 2.58 1.22 0.85 "Inodoro" 22.14 25.34 30.70 41.18 54.39 59.76 60.41 73.57 80.54 90.79 Latrine 45.71 43.86 45.85 44.85 32.18 31.20 33.04 23.85 18.24 8.36 Piped connection 4.52 6.35 13.52 18.49 35.22 41.13 43.87 58.32 62.87 79.30 Septic tank 21.17 22.418 19.26 27.17 20.60 21.91 19.92 16.29 19.54 13.08 Hole 42.16 4036 43.77 40.37 30.76 27.91 29.66 22.81 16.37 6.76 Electricity None 75.04 65.25 46.48 33.49 25.92 20.35 16.12 9.71 6.55 4.71 ENEE 24.46 34.68 53.20 65.71 73.91 78.68 82.12 89.88 91.93 93.26 Collective 0.00 0.07 0.32 0.47 0.00 0.73 0.69 0.33 0.40 0.50 Individual 0.50 0.00 0.00 0.33 0.17 0.24 1.07 0.09 1.13 1.52 Source: World Bank staff using March 1999 EPHPM survey. ENEE is the national electricity provider. 38 Table 3.4: Access to basic infrastructure services by income group (decile), urban/rural, 1999 1 2 3 4 5 6 7 8 9 10 Urban areas Access to water Public source 69.95 80.46 84.62 86.30 90.96 91.50 91.98 95.83 93.61 95.81 Private or collective 25.03 11.06 10.87 9.22 6.54 6.43 6.83 3.36 5.65 4.11 Well 3.09 3.71 1.51 2.46 1.07 0.00 0.37 - 0.23 0.33 0.00 River 0.61 0.84 0.93 0.25 0.29 0.00 0.00 0.00 0.00 0.00 Other 1.31 3.93 2.07 1.76 1.13 2.07 0.82 0.58 0.41 0.08 Inside the house 26.42 27.68 24.38 24.01 35.88 39.69 47.51 61.90 63.90 84.43 Inside the propriety 60.58 57.27 65.44 67.16 59.39 55.41 48.45 36.07 33.65 15.12 Outside of property 13.00 15.06 10.18 8.84 4.73 4.89 4.04 2.03 2.45 0.45 Sanitary installation None 20.26 16.98 8.24 5.47 4.02 3.25 2.18 1.12 0.43 0.49 "Inodoro" 33.61 39.09 47.15 49.07 69.17 72.63 71.35 82.50 86.11 96.51 Latrine 46.13 43.93 44.61 45.47 26.80 24.12 26.47 16.38 13.46 3.00 Piped connection 23.12 25.69 37.62 34.90 55.53 60.29 62.42 72.41 77.93 90.34 Septic tank 11.34 19.09 11.03 19.41 13.64 13.18 11.18 10.28 8.70 6.09 Hole 45.28 38.25 43.11 40.22 26.80 23.29 24.22 16.19 12.94 3.08 Electricity None 30.95 16.57 9.95 9.01 4.01 1.87 1.38 0.39 0.23 0.41 ENEE 69.05 83.43 90.05 90.71 95.99 98.13 98.62 99.57 99.65 99.38 Collective 0.00 0.00 0.00 0.29 0.00 0.00 0.00 0.03 0.12 0.21 Individual 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Rural areas Access to water Public source 13.91 16.27 21.84 23.99 26.01 28.23 30.71 22.89 33.83 24.22 Privateorcollective 57.50 58.46 61.60 60.72 55.91 57.51 53.36 64.43 49.45 63.28 Well 14.01 10.34 8.51 8.12 10.17 6.89 8.84 7.62 11.09 12.51 River 14.36 13.72 7.04 6.68 7.23 7.22 6.35 5.06 5.63 0.00 Other 0.22 1.22 1.01 0.48 0.68 0.15 0.73 0.00 0.00 0.00 Inside the house 3.98 3.26 4.88 13.11 12.93 21.36 20.66 22.67 32.72 37.63 Inside the propriety 69.20 74.21 79.22 73.16 71.69 65.53 68.98 72.08 57.89 55.41 Outside of property 26.82 22.53 15.90 13.73 15.38 13.10 10.37 5.25 9.39 6.96 Sanitary instalUation None 33.53 34.06 28.84 19.98 23.76 17.35 13.81 6.38 3.33 2.62 "Inodoro" 20.81 22.10 24.87 35.60 38.15 41.31 42.20 50.45 65.77 62.73 Latrine 45.66 43.84 46.30 44.42 38.10 41.34 43.99 43.18 30.90 34.65 Piped connection 2.36 1.79 4.98 6.89 12.89 13.68 12.99 21.84 22.93 25.17 Septic tank 22.31 23.28 22.18 32.66 28.24 34.43 34.46 31.84 48.28 47.40 Hole 41.80 40.86 44.01 40.47 35.11 34.54 38.73 39.94 25.46 24.81 Electricity None 80.16 76.74 59.43 50.82 49.99 46.83 40.65 33.81 23.30 25.80 ENEE 19.28 23.18 40.13 48.03 49.64 50.80 54.66 64.77 71.45 63.27 Collective 0.00 0.09 0.44 0.59 0.00 1.78 1.83 1.09 1.12 1.94 Individual 0.56 0.00 0.00 0.56 0.37 0.58 2.86 0.33 4.12 8.99 Source: World Bank staff using March 1999 EPHPM survey. ENEE is the national electricity provider. 39 3.4. Simple methods can be used to assess the impact on poverty of policies promoting access to basic infrastructure services for the poor. Traditionally, poverty measures and access to basic infrastructure services have been presented as alternative measures of well-being, as if there was no way through which the impact on poverty of access to basic services could be measured. Yet the poverty reduction impact of basic services can be measured by estimating the gain in the implicit rental value of owner-occupied houses when access to a basic infrastructure service is provided. This gain can then be added to the income of the household in order to have a rough measure of the impact on poverty of access. To estimate the gain in rental value due to access to basic services, we use hedonic semi-log rental regressions with the logarithm of the rent (for those households paying rent) depending on the characteristics of the house and its location. Using the parameter estimates from the regressions, the impact of an electricity connection on the rent for those who pay a rent (and on the imputed rental value of the house for those who do not pay a rent) can then be computed as the expected percentage increase in the rent paid. Table 3.5 gives the coefficient estimates in the rental regressions for the access to electricity and water, and for sanitary installations. The parameter estimates depend on the survey used, but access to services clearly has a positive impact on rental value. Below, we will use for simulations the estimates obtained for the September 1998 EPHF'M survey because they appear the most reasonable2. Table 3.5: Percentage increase in rent due to electricity, water and sanita installation, 1998-99 March 98 September 98 March 99 Access to electricity (ENEE) 28.32 31.17 60.13 Access to water inside the house NS 41.32 34.60 Piped sanitary installation NS 35.96 NS Source: World Bank staff using EPHPM surveys. ENEE is the national electricity provider. 3.5. The value of access to electricity, water, or sanitation varies from 4 to 13 Lempiras per month per capita for poor households in Honduras. In a semi-log regression setting, the impact of access to basic services will be proportional to the expected rent computed using all housing characteristics except the services. For example, the value of access to electricity is going to be larger for the non-poor (who pay higher rents) than for the poor. In relative terms however, when compared with the level of per capita income of the households, the impact of access to electricity may be higher for the non-poor than for the poor. Table 3.6 provides income levels and expected rents by income quintile using the September 1998 EPHPM. All figures are given per cajpita (thus the rent is divided by the household size). The value of access to basic services is computed at the household level as the parameter 3 times the expected rent without access. If we consider as being poor those households in the bottom three quintiles, the value of access to electricity, water, and sanitaiy installations varies from 4 to 13 Lempiras per capita per month. In absolute terms, the value of access is higher for the rich than for the poor (because the rich have higher expected rents), and this is consistent with the fact that the willingness to pay for these services is higher 2 There are two important caveats in using the hedonic method for assessing the value of a connection, and both caveats may reduce the actual value of a connection. First, for those households who are tenants and pay a rent, the method may not apply simply because the value of a connection is a benefit for the owner rather than for the tenant. In a competitive rental market, an owner niay increase the rent after receiving a connection, in which case the tenant (who is more likely to be poor than the owner) has no gain of its own. In practice however, especially in poor rural areas, a good number of the poor are owners, even if their house is very modest. Second, for owners, while the value of a connection is received at once at the time of connection, the benefit is continuous. In other words, one could compute the one-shot value of the connection as the discounted stream over time of its benefits, and this one- shot value could be realized if the owner were to sell the house and move. At the same time, if the price of electricity includes a fixed term, this fixedl term may have been computed so as to offset the cost of the connection for the utility over time. In this case, there is no additional benefit from the connection, apart from the fact that there is no more rationing for the household for that good. Thus, if the fixed term of the tariff structure is taken into account, the value of the connection is likely to be lower than what has been estimated. [There may be other limits to the methodology, for example if most rental units are under rent control and rent control is binding.] 40 among the rich than among the poor. But in relative terns, as a percentage of the income of the people, the value of access to basic infrastructure services is higher for the poor than for the rich. Table 3.6: Estimating the value of access to basic infrastructure services by income quintile, 1998 Quintile Actual per Expected per Expected per Expected gain Expected gain as capita capita rent capita rent from access share of per income w/o access with access capita income All households 1 94.37 19.76 - - - 2 280.26 34.46 - - - 3 516.29 59.75 - - - 4 915.60 90.89 - - - 5 2553.00 163.55 - - - No electricity 1 91.17 10.56 14.42 3.86 7.22 2 271.29 14.37 19.62 5.26 1.98 3 502.96 18.35 25.06 6.71 1.33 4 864.66 24.59 33.58 8.99 1.05; 5 1960.15 32.47 44.34 11.88 0.64 No water inside the house 1 88.43 8.02 12.13 4.10 7.32 2 280.59 13.70 20.70 7.01 2.55 3 503.21 16.91 25.56 8.65 1.67 4 861.00 28.29 42.77 14.48 1.68 5 1697.67 38.50 58.20 19.70 1.27 No piped sanitary inst. 1 92.96 13.71 19.64 5.93 11.01 2 273.53 22.76 32.61 9.85 3.64 3 507.43 29.89 42.82 12.93 2.56 4 877.12 44.31 63.48 19.17 2.20 5 2194.06 65.48 93.82 28.34 1.57 Source: World Bank staff using September 1998 EPHPM survey. NA = not applicable. Lempiras pe. month. Note that the expected gain as a share of per capita income in last column is not equal to the mean expected gain from access divided by the mean per capita income because the average of a ratio is not the ratio of the averages. 3.6. The results obtained with this methodology are similar to those obtained with more complex methods. It could well be that our estimate of the value of a connection to electricity is too high. We have only a limited number of housing characteristic in the regression, so that there may be an omitted variable bias, which in our case would typically result in an over-estimation of the parameters. Still, the fact that the value of a connection in percentage terms of a household's income is higher for the poor than for the non-poor is likely to be true even if there is a bias in the parameter estimates. Moreover, the value of an electricity connection in the bottom half of the population is worth about two to three percent of the poverty line (if we set the poverty line at, say, 400 Lempiras per month in 1998), and this does not sound unrealistic. Another, more complex method for estimating the value of access to basic services consists in estimating an Almost Ideal Demand System (AIDS) with a number of expenditures censored (thus the system has both linear and tobit regressions). Following this route, Raygoza (1998) found that for Mexico in 1994, the market price of a connection to electricity had a value of about 2.5 percent of the Mexican poverty line, which is fairly close to our own estimates for Honduras using the simpler hedonic method. [Raygoza argues that the average actual benefit of the connection is only half its market price, in part due to the fact that the compensating variation from access to electricity is below the market price in about the same way that a transfer in kind has a lower value than an equal amount provided in cash.] 3.7. Because of the value of basic infrastructure services, providing access helps in reducing poverty. Table 3.7 provides the reduction in poverty obtained when all those households who lack access to one of the basic services get access. The initial poverty measures "without access" for those who do not 41 have access could be debated (see chapter 1), but this does not change the basic message. Nationally, if access to electricity is provided to all those who do not have access today, and if our method for valuing access is accepted, the headcount index of extreme poverty would be reduced by 0.28 percentage points only (a percentage change of -0.86 percent) because the value of the access is not large enough to lift many households above the poverty line. The square poverty gap would be reduced by 0.36 percentage points (a percentage change of -3.45 percent). The reduction in poverty with universal access to water within the home would be lower, while the reduction of poverty with universal piped sanitary installations would be higher, with the differences being essentially due to the number of households without access. If we consider only the households without access for the poverty comparisons, the impact on poverty is (as expected) larger. Importantly, all these estimates do not take into account the indirect benefits of infrastructure such as its impact on the creation of small businesses or health and education outcomes. Table 3.7: Reduction in poverty with universal access to basic infrastructure services, 1998 National sample Households without access Without With Percentage Without With Percentage access access change access access change Universal access to electricity Headcount, extreme poverty 31.91 31.63 -0.86 59.15 58.38 -1.31 Poverty gap, extreme poverty 15.77 15.41 -2.28 31.60 30.59 -3.21 Squared poverty gap, extreme poverty 10.51 10.15 -3.45 21.55 20.52 -4.75 Headcount, poverty 49.08 49.03 -0.10 77.47 77.34 -0.18 Poverty gap, poverty 26.09 25.78 -1.18 47.13 46.26 -1.84 Squared poverty gap, poverty 17.63 17.29 -1.94 33.86 32.90 -2.85 Universal water within the home Headcount, extreme poverty 31.91 31.81 -0.32 54.64 53.82 -1.51 Poverty gap, extreme poverty 15.77 15.64 -0.80 30.75 29.72 -3.35 Squared poverty gap, extreme poverty 10.51 10.39 -1.22 21.44 20.39 -4.90 Headcount, poverty 49.08 49.00 -0.17 74.44 73.75 -0.92 Poverty gap, poverty 26.09 25.97 -0.44 44.97 44.02 -2.11 Squared poverty gap, poverty 17.63 17.50 -0.70 32.73 31.73 -3.07 Universal piped sanitary installation Headcount, extreme poverty 31.91 31.20 -2.20 46.62 45.18 -3.09 Poverty gap, extreme poverty 15.77 15.17 -3.78 23.73 22.50 -5.16 Squared poverty gap, extreme poverty 10.51 9.95 -5.40 16.07 14.90 -7.27 Headcount, poverty 49.08 48.78 -0.62 67.98 67.35 -0.93 Poverty gap, poverty 26.09 25.44 -2.48 38.06 36.73 -3.49 Squared poverty gap, poverty 17.63 17.02 -3.47 26.26 25.00 -4.79 Source: World Bank staff using September 1998 IEPHPM survey. 3.8. The poor appear to benefit as much as the non-poor from an increase in access to basic infrastructure services. It is an empirical question whether the poor benefit more or less than the non- poor from a national increase in access to basic infrastructure. Typically, if the access rate is low, the non-poor are likely to benefit more frorn an increase in access. But once the middle class and the rich have been served, the poor may benefit more from additional increases in access. Estimates of so-called marginal benefit incidence in Honduras are provided in table 3.8 using municipal level data (a description of the methodology is provided in annex 2, section MA.9). Three groups are considered: those living in poor municipalities, those living in (comparatively) rich municipalities, and those living in municipalities with middle-range income levels. The ranking of the municipalities is computed within Honduras's 16 departments for which we have data, raither than nationally. One thus compares how poor, middle, and rich municipalities fare within a given geographic area, and the definition of which municipalities are 42 poor, middle, or rich is specific to each department. On average, the marginal benefit incidence estimates for the three groups of municipalities in a given area must be one, since the increase in the mean access for a department as a whole must be allocated to the three groups of municipalities. The question is whether (comparatively) poorer municipalities benefit more or less than other municipalities from a departmental increase in access. In table 3.8, it can be seen that in most cases, the poor, the middle group and the rich benefit in the same proportion from a national increase in access (most of the differences in the marginal benefit incidence estimates are not statistically significant in the right part of the table). The only statistically significant differences are observed for sanitary installations where the poor did less well than other groups in 1988 but not in 1996 (as mentioned earlier, we would expect the poor to benefit more from access over time), and for unmet basic needs where the poor did better than other groups in 1988. For increases in access to safe water in 1988, no differences are observed between groups. Table 3.8: Who benefits from an expansion in access to basic infrastructure services? Estimates of the marginal benefit incidence Tests of differences in the marginal benefit by municipal income group incidence estimates (p-values) Poor Middle Rich Poor versus Middle versus Poor versus Middle Rich Rich Water 1988 1.096 1.090 0.814 NS NS NS Sanitary installation 1988 0.833 0.889 1.278 NS 5% level 5% level Sanitary installation 1996 0.986 1.068 0.947 NS NS NS Unmet basic needs 1988 1.178 0.977 0.845 10% level NS 5% level Source: World Bank staff municipal level data. C. THE ELECTRICITY SUBSIDY IS NOT WELL TARGETED AND IT INCREASES INEQUALITY 3.9. There are two large subsidies in Honduras for basic infrastructure services: the first is for electricity consumption nationally, and the second is for bus transportation in Tegucigalpa. The cost for the electricity subsidy was estimated at 259 million Lempiras in 1998, while the cost for the bus transportation subsidy was estimated at 114 million Lempiras. By comparison, the budget for the FHIS (see the last section of this chapter for an evaluation) in 1998 was 579 million Lempiras, and the budget for PRAF (see chapter 4 for an evaluation) was 188 million Lempiras. The subsidies are thus costly. 3.10. Both subsidies are in principle self-targeted, but this is less the case for the electricity subsidy and data is lacking to make an assessment of the transportation subsidy. The subsidy for electricity is given to all households who have a level of consumption below 300 kwh per month, and this represents 85 percent of the population with a connection to the grid. There is self-selection because of the ceiling for the "lifeline", but the ceiling is too high for the self-selection to be restricted to the poor. The bus transportation subsidy is given to all those who ride certain bus lines in Tegucigalpa. The self-selection process takes place because the buses are not comfortable, so that those who can afford other means of transportation do. Still, because the subsidy is limited to Tegucigalpa where the population is comparatively less poor, it remains to be seen how much poverty reduction is actually achieved. Below, we provide an evaluation of the electricity subsidy for which we have data on beneficiaries. Due to a lack of appropriate data, we do not provide a comparable evaluation for Tegucigalpa's bus subsidy. 3.11. Most of the subsidy for electricity is spent on households who consume between 100 and 300 kwh per month. Table 3.9 gives the current structure of electricity consumption by level in Honduras, together with the existing subsidy. For example, the number of households with a monthly consumption below 20 kwh is 115,723, which represents 20.5 percent of all households connected to the grid. Their total consumption is 388,626 kwh, or 3.36 kwh per household. Without the subsidy, they would have to pay a total bill of 929,256 Lempiras, but this bill is reduced to 303,282 Lempiras. Overall, 80 percent of 43 the subsidy is spent on households who consume more than 100 kwh. The question is whether this represents a good use of resources from the point of view of poverty and inequality reduction. Table 3.9: Electricity consumption by level and subsidies, Monthly, 2000 Consumption Number of Cumulative Total Average Total bill Existing Total bill level in kwh clients percentage of consumption consumption without subsidy up to after subsidy clients of clients per client subsidy 250 kWh deduction (kwh) (kwh) (Lempiras) (Lempiras) (Lempiras) 0-20 115,723 20.5% 388,626 3.36 929,256 595,973 333,282 20-100 129,285 43.4% 7,584,510 58.67 5,095,693 2,716,580 2,379,113 100-150 71,967 56.2% 9,002,165 125.09 7,387,370 3,761,883 3,625,488 150-200 63,623 67.4% 11,156,098 175.35 10,314,338 5,148,710 5,165,628 200-250 52,733 76.8% 11,840,873 224.54 11,618,468 5,746,445 4,872,023 250-300 42,343 84.3% 11,676,795 275.77 11,895,559 5,850,730 6,044,559 Total 475,674 51,649,067 108.58 47,240,684 23,820,321 22,420,093 Source: Information provided by ENEE. 3.12. The impact on poverty of the electricity subsidy is small in comparison to the public cost. Table 3.10 provides an estimate of the poverty reduction achieved with the subsidy when the poverty line is set at 400 and 600 Lempiras per person per month (we use round numbers rather than actual poverty lines for illustrative purposes; the conclusions remain robust with other reasonable poverty lines). To construct the table, we used the data frorm the 1999 PRAF survey and computed poverty measures under the current situation (i.e., with the subsidy) by electricity consumption bracket (this is feasible because the PRAF survey has a consumption module with information on expenditures for electricity, and the consumption level for electricity can be found from the expenditure level using the price structure of ENEE). Using the variables common to the PRAF and EPHPM surveys, we also made a prediction of the electricity consumption that could be expected for all households in the EPHPM survey given their characteristics, and we computed similar poverty measures by electricity consumption levels under the current situation3. Next, we computed poverty measures without the electricity subsidy by subtracting from the per capita income (in the EPHF'M survey) or consumption (in the PRAF survey) the value of the subsidy for those households receiving the subsidy. The poverty measure without the subsidy minus the poverty measure with the subsidy is an upper bound estimate for the poverty reduction impact of the subsidy (due to substitution effects, the actual poverty reduction from the subsidy must be lower). * Leakage to the non-poor: In the zero to 20 kwh consumption bracket, according to the PRAF survey, 55 percent of the households are not poor (39 percent in the EPHPM survey). Among households who consume more than 100 kwh, more than 80 percent of the subsidy goes to the non-poor. * Small impact on poverty: the impact on poverty is small in comparison to the cost of the program, and this is a direct consequence of the lack of good targeting in the subsidy. For example, if the subsidy was perfectly targeted to the poor, it could be shown that the reduction in the poverty gap obtained with the subsidy would be four to five times larger than the current reduction observed in table 310. 3.13. Apart from the fact that the electricity subsidy is not very good at reducing poverty, it contributes to higher inequality. As was mentioned in chapter 1, different sources of consumption have a different impact on the inequality in total per capita consumption. Decomposing the Gini index of inequality in consumption according to consumption source can help in measuring the impact of price 3The advantage of doing so is that the EPH[PM survey is nationally representative while the PRAF survey is not. One curious outcome in comparing the two surveys is that the poverty measures based on per capita consumption in the PRAF survey of poor municipalities are lower than the poverty measures based on per capita income in the nationally representative EPHPM survey. This suggests a good reporting of consumption in the PRAF survey, and a less good reporting of sources of income in the EPHPM survey. Our findings are not affected by this. 44 subsidies on the inequality in total per capita consumption. This impact is proportional to the Gini elasticity (see annex 2, section MA.2). If the Gini elasticity of a good is larger (smaller) than one, a subsidy for that good will be inequality increasing (decreasing). Using the 1999 PRAF survey, it can be shown that except for the lowest level of consumption in municipal centers (less than 20 kwh), a marginal reduction in pricing (i.e., an increase in the subsidy) is inequality increasing, rather than decreasing. 3.14. The electricity subsidy could be reformed so that only those who consume less than 100 kwh benefit from it. Reforming the electricity subsidy may not be easy from a political economy point of view (especially following a recent increase in prices of 16 percent by ENEE), but it would bring benefits for the poor if the cost savings are better directed to them than the current subsidy. One possibility would be to give the electricity free to all households who consume less than 20 kwh per month, because many of these households are poor (this would also reduce the collection costs of ENEE). Then, the subsidy could be phased progressively between 20 kwh and 100 kwh. Above 100 kwh, there would be no more subsidy. This would provide large cost savings for the budget and the resulting resources could be put to better use elsewhere (e.g., for new connections to the grid, or for other sectors in the economy). Table 3.10: Im act on poverty of electricity subsidies, 1999 Kwh consumed PRAF survey (poor municipalities) EPHPM survey (national sample) Headcount Poverty gap Sq. pov. gap Headcount Poverty gap Sq. pov. gap Poverty Line at 400 Lempiras per person per month, with subsidy 0 - 20 44.93 12.29 5.92 20- 100 35.66 10.19 4.11 61.24 28.31 18.22 100- 150 16.82 5.60 2.35 40.16 17.02 10.18 150 -200 10.98 2.24 0.72 25.23 8.80 4.82 200-250 15.64 4.56 1.51 24.78 9.01 5.24 250 -300 17.09 2.38 0.79 15.89 3.84 1.75 300- 10.15 2.19 0.87 5.22 1.86 0.99 Poverty Line at 400 Lempiras per person per month, without subsidy 0 - 20 44.93 12.42 5.99 20- 100 36.00 10.45 4.26 61.57 29.08 18.94 100- 150 20.57 6.06 2.61 40.39 18.12 11.13 150 - 200 10.98 2.67 0.93 26.63 9.92 5.60 200 - 250 15.64 5.32 2.00 27.23 10.42 6.34 250 - 300 17.09 3.06 1.02 25.62 5.03 2.37 300- 10.15 2.70 1.12 6.30 2.27 1.29 Poverty Line at 600 Lempiras per person per month, with subsidy 0 - 20 71.01 28.85 14.44 20 - 100 63.47 23.36 11.36 79.85 42.53 28.22 100 - 150 43.39 13.29 6.34 62.03 28.20 17.31 150 - 200 27.45 7.47 2.95 47.19 17.73 9.60 200 - 250 34.16 11.88 5.21 43.10 17.17 9.74 250 - 300 29.15 9.36 3.73 41.42 12.81 5.57 300- 17.97 5.65 2.60 13.50 4.26 2.12 Poverty Line at 600 Lempiras per person per month, without subsidy 0 - 20 71.01 29.00 14.56 20- 100 63.47 23.70 11.60 80.27 43.22 28.94 100- 150 44.74 13.97 6.75 63.25 29.38 18.37 150 - 200 31.26 8.19 3.35 47.92 19.10 10.66 200 - 250 35.80 13.15 6.05 45.29 18.83 11.08 250 - 300 29.15 10.41 4.37 42.59 14.55 6.71 300 - 17.97 6.33 3.03 14.08 4.95 2.58 Source: World Bank staff using 1999 PRAF and EPHPM surveys. 45 D. HONDURAS' SOCIAL INVESTMENT FUND (FHIS) IS A KEY PROVIDER OF SOCIAL INFRASTRUCTURE 3.15. The Fondo Hondurenlo de Inversi6n Social (FHIS) was originally established in part to generate employment within a context of structural adjustment, but its mission has evolved since then. The FHIS was first created on a temporary basis in 1990 in a context of structural adjustment in order to generate employment among the poor while at the same time extending the country's network of basic social infrastructure (education, health, water, and sanitation) into areas with high levels of unmet basic needs. Today, while the FHIS still finances basic social infrastructure with some level of targeting toward poor rural municipalities, the emphasis is more on the efficiency of the infrastructure built than on the generation of employment for the poor. As a result, while the social fund does generate some employment in poor communities, it should not be considered as providing safety nets through employment creation. The FHIS, however, has been instrumental in the delivery of emergency safety nets to the communities hit by Hurricane Mitch in October 1998. 3.16. The FHIS finances a variety of projects, but more than half of the funding is devoted to education. Since its inception, the FHIS has implemented more than 8,000 projects in Honduras. After the first phase of the program (FHIS I), FHIS II was implemented between 1994 and 1998, and this is the phase of the program evaluated in this paper. The third phase of the program (FHIS III) will be implemented between 1998 and 2002. Education accounts for over half of the resources (56 percent), followed by water and sanitation, health, and municipal development at roughly equal shares of 12-15 percent. A small part of the funds (4 percent) is devoted to social assistance and to environmental projects. A large proportion of the budget of the FHIS is donor-financed. 3.17. The targeting performance of the FHIS has improved over time, but it needs to improve further. Successful targeting of the poor is a key policy goal of social investment funds in Latin America. Yet there is a feeling that while social funds have succeeded in reaching the poor, at times with a better targeting performance than other programs, they could do better. In Honduras, the targeting of FHIS investments occurs in two steps. In the first step, resources are allocated to each municipality according to population, and then adjusted by means of a poverty proxy index so that the greater the poverty level, the more resources per capita are allocated to the municipality. In the second step, resources are allocated within the municipality and decisions are made with regards to the exact type of investments to be made. The procedures in this second step have evolved over time, from a highly centralized system (FHIS 1) to one with increasing community participation in decision-making (FHIS II and III). An evaluation of FHIS II by ESA Consultores (1999) based on a survey conducted in 1998 suggests that FHIS II has been better targeted at the municipal level than FHIS I (which was not well targeted), and that the targeting of FHIS II has been better at the household than at the municipal level. * The first concentration curve in Figure 3.1 provides the cumulative share of the FHIS funds allocated to municipalities as a function of the cumulative share of the population living in these municipalities, with the municipalities ranked by increasing order of income. There is some progressivity at work in that the poorer municipalities have a higher share of FHIS funds. But the extent of redistribution is limited (the diagonal line represents an egalitarian distribution of the funds between municipalities). * The second concentration curve provides the distribution of the benefits of FHIS at the household level. The households in the areas of influence of the FHIS projects have been regrouped in five income quintiles which do not correspond to the position of the households in the distribution of income of the FHIS survey, but ratlher to their position in the overall national distribution of income. This has been feasible because the FHIS survey includes an income module that is comparable to data collected in the nationally-representative EHPMs carried out two times a year by the Government's Census and Statistics Bureau. Thus, each household interviewed in the FHIS survey can be placed within the national income distribution. Though the FHIS survey itself is not nationally representative, this technique allows for an evaluation of the targeting performance of the program on 46 a national basis. Clearly, the targeting of FHIS investments is better at the household than at the municipal level. This may be because within municipalities, FHIS targets poorer areas. * The third concentration curve provides the distribution of the benefits of FHIS according to who uses the facilities provided by the program. The level of redistribution is higher for usage than for access, which implies that there is some degree of self-selection at work in the use of the facilities. In other words, a number of FHIS projects tend to be more in demand among the poorer segments of the population. Thus, overall there is some degree of targeting in FHIS projects, but progress could still be made by making the allocation of funds more pro-poor at both the municipal and household levels. Figure 3.1: Progressivity of FHIS 2 investments at municipal and household levels PItbQf M iHIS 2 at unapdl lewve Progressivity of FHIS 2 at household level: access ,g s .---< 70 70.0 C go~~~~~9l myiaiRlo 2 0 0 0 so Shh ODlShateo FH5~~~~~~~~~~rgesiiyo FHI thueodlvl sg 20 20 s h ze 0 20 40 go go lo Shareaau(51 Share Population Progressivity of FHII S 2 at household level: usage 700 so Share wo 0 2D 40 go go i Share Population Source: ESA Consultores (1999) 3.18. The demand-driven approach of the FHIS is one of its most important characteristics. The reviews of social funds point to several positive features: an emphasis on rural areas, where the incidence of poverty is typically higher and where other govemment expenditures have difficulties in reaching the poor; a separation between financing and implementation entities, i.e., local bodies, non-govemrnental organizations, and private contractors; and demand-driven orientation, such that the projects are in tune with local needs and allow experimentation with new ideas. This last feature is perhaps the most important one. Despite the presumption that social funds are demand driven, there has been surprisingly 47 little effort devoted to asking communities whether the funds indeed meet their priorities4. In the case of the FHIS however, the demand-driven approach can be evaluated using data from the 1998 survey. Specifically, it is feasible to analyze: i(a) whether households have been consulted during the project design phase (consultation rate); (b) whether the projects selected at the community level match the priorities of the household (priority matching rate); (c) whether the households have contributed to the implementations of the projects through contributions in cash or in kind (contribution rate); and (d) whether the households have used the facilities provided by the projects (usage rate). This is done below by type of FHIS project (education, health, water, sewerage, latrine) and by income quintile (table 3.1 1). Because of the comparability of the income module in the FHIS and EPHPM surveys, the quintiles correspond to the position of households in the overall national distribution of income (in other words, the number of households from the FHIS survey in each of the five quintiles need not be the same.) 3.19. Three out of four households are consulted by the FHIS before the implementation of projects, and the proportion is higher for the poor. The FHIS survey questionnaire asks whether the community benefited from an investment in a school, health center, water project, sewerage system or latrines, and for all positive answers., asks whether the households were consulted during project preparation (for example, invited to a town meeting to discuss the project, or asked to sign a statement approving the project). About three-quarters of all households in the project area of influence (72.9 percent, not shown in table 3.11) report being consulted during the project design phase. The rates of consultation are highest among households with per capita income levels that place them in the poorest national income quintile (81.5 percent). In terms of project type, the rates of consultation are the highest for sewerage and latrines, at 85-86 percent in both cases. Table 3.11: Community participation in FHIS projects by type of project and income quintile, 1998 Educ. Health Water Sew. Latrine Q1 Q2 Q3 Q4 Q5 Consultation 67.77 74.75 64.96 85.03 85.91 81.5 70.9 75.1 59.1 70.2 Contribution 53.26 53.06 52.25 63.52 74.49 65.87 53.87 65.65 47.93 47.97 Free Labor 25.84 37.46 26.20 12.17 64.67 44.8 41.0 36.2 20.5 9.2 Money 20.55 13.22 25.13 28.92 3.29 13.3 8.0 20.0 22.8 34.4 Material 6.39 0.73 3.19 4.55 9.45 4.9 3.7 4.9 7.8 11.1 Supervision 0.86 2.28 8.79 18.33 0.89 0.4 2.5 2.9 3.8 3.8 Paid Labor 2.21 2.06 0.64 1.98 0.32 4.2 0.6 1.2 0.3 0.2 Usage 80.28 92.33 97.31 90.31 82.59 93.8 78.8 88.8 80.9 57.8 Source: World Bank staff using 1998 FHIS survey. All figures are percentages. 3.20. FHIS projects match the priorities of households in two cases out of three. The FHIS survey asks respondents to identify which of the five main project types offered by FHIS (school, health center, water system, sewerage system, or latrine) they would have chosen as their first priority. These priorities can be matched with the actual projects implemented by FHIS. Although this is not shown in table 3.11, the priority matching rates are good overall, even though there are large variations in the priority matching rates by project type. Results also not shown in the above table reveal that FHIS projects match with community priorities at a higher rate in rural areas (69.3 percent) than in urban areas (59.6 percent). 4 Although social funds have been in existence for over 15 years, evaluations of their effectiveness drawing on representative household surveys remain rare. The evaluation of the FHIS by ESA Consultores (1999) is an exception. There is also a concern that in some cases, the investments carried out by social funds have not been adequately maintained over time, particularly when the funds did not involve sufficient community participation through consultation, training, and capacity building. 48 3.21. Slightly more than half of the households contribute to the FHIS projects, and the type of contribution varies according to both project type and income leveL For the projects carried out in the community since 1994, the survey asks whether the households contributed to the project, and if so, how they contributed. Overall, the most common form of contribution to FHIS projects is free labor (34.4 percent), followed by money (16.8 percent), material (6.1 percent), supervision (2.2 percentl, and paid labor (1.7 percent). The contribution rates are higher among the poor (first three quintiles) and the type of contribution also varies by quintile. Almost half of all households in the poorest quintile contribute free labor (44.8 percent) and far fewer make other forms of contribution. Those in the wealthiest quintile are more likely to contribute money (34.4 percent), and are also those most likely to supervise projects (3.8 percent) as a form of contribution. Paid labor on FHIS projects is more prevalent among the poor, but the share of poor households working for pay on FHIS project is very small, so that FHIS is not an important source of employment in the communities where projects are being carried out. As mentioned earlier, this is due to the fact that the mandate of the fund has evolved away from employment generation toward a stronger focus on generating quality investments for improving access to human capital among the poor. Contribution rates are highest for FHIS sewerage and latrine projects, as would be expected since these projects provide private goods that require household labor or capital investments before they can become operational. Results by geographic location (not shown in table 3.11) suggest that contribution rates are higher in urban (61.4 percent) than in rural areas (55.0 percent). 3.22. A large majority of households use the FHIS facilities, and usage among the poorest households is near universal. Project utilization rates reflect how useful FHIS investments are perceived to be by the community. They can be interpreted as indicators of community satisfaction with project outcomes (but this has limitations since usage is basically free once the project is implemented). The survey asks whether or not the households have used, are using, or in the case of pipeline projects will be using the facilities. Overall, the households report high utilization rates for FHIS facilities, at 83.4 percent. The results by income quintile are particularly striking, revealing that utilization is almost universal among households in the poorest quintile (93.8 percent), and far lower for households in the wealthiest quintile (57.8 percent). This indicates self-selection, whereby the services offered by the FHIS are more in demand among poorer households. Utilization is virtually universal for water (97.3 percent) and health (92.3 percent) projects, but it is lower for education and latrine projects. The results by geographic area (not shown in table 3.11) suggest higher utilization rates for FHIS projects in rural areas (90.3 percent) as compared to urban areas (75.1 percent). 3.23. FHIS projects have higher community participation than projects whose sponsor cannot be identified, and it is on par with projects by other agencies that can be identified by households. As indicated in table 3.12, controlling for household level characteristics, in virtually all types of project, FHIS is better at promoting community consultation, contribution, and usage than organizations which have not been identified by households. On the other hand, the FHIS is on par with other organizations that have been identified by the households. In education projects for example, the consultation rate for FHIS project is not statistically significantly different from the consultation rate for non-FHIS projects, which is itself 28 percent higher than for projects whose sponsor is unknown. These results suggest that while FHIS is not necessarily better at community participation than other agencies which can be identified by households, the program still does a good job at promoting participation overall. 49 Table 3.12: Comparative p rformance of the FHIS for consultation, contribution, and usage, 1998 Consultation Contribution Usage Education FHIS (versus non-FHIS) NS -15.76% NS Not known (versus non-FHIS) -28.12% -22.77% 9.49% Health FHIS (versus non-FHIS) NS NS NS Not known (versus non-FHIS) - 35.43% NS NS Water FHIS (versus non-FHIS) -8.43% NS NS Not known (versus non-FHIS) -44.57% -18.63% NS Used Water FHIS (versus non-FHIS) NS NS NA Not known (versus non-FHIS) -71.59% NS NA Latrines FHIS (versus non-FHIS) NS NS NS Not known (versus non-FHIS) _ -33.71% -34.67% 19.43% Source: World Bank staff using 1998 FHIS survey. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. 3.24. Promoting consultation and contribution is important because this yields a higher probability of final usage, at least in some cases. There is some evidence that consultation prior to project implementation increases the contributions made by households at the implementation stage as well as (to a lesser extent) the usage of the facilitiies once they are completed. There is also limited evidence that a higher contribution by households at the implementation stage increases the use of the facilities. Establishing these results empirically i's not straightforward because of a so-called endogeneity problem in that the households who are consulted (in part because of their own drive to be consulted) about the fund projects prior to their implementation in the community are almost by definition more likely to contribute to the implementation, andi those that contribute may do so because they plan to use the facilities. Using an instrumental variable to control for this endogeneity problem (as discussed in annex 2, section MA.6), it was found that consultation before the implementation increased the probability of contribution for health, water, and sewerage projects (table 3.13). Prior consultation also increases final usage for some projects (education and latrines). Finally, in one case (latrines), a contribution by households was found to increase usage. These results suggest that consultation of the potential beneficiaries of a project early on may indeed improve outcomes, since outcomes (such as an increase in school enrollment or a decrease in the incidence of illnesses) ultimately depends on the usage of the facilities provided by the beneficiaries. Also, before concluding this chapter, it is important to always keep in mind that programs such as the FHIS (or PRAF, discussed in chapter 4) should not follow an "assistance" model, but should rather enable the poor to emerge from poverty on a sustainable basis. Table 3.13: Relationshi ps between consultation, contribution, and usage in FHIS projects, 1998 Marginal impact of Marginal impact of Marginal impact of consultation on contribution consultation on usage contribution on usage Education NS 14.83%o NS Health 22.37% NS NS Water 13.42% NS NS Sewerage 18.63% NA NA Latrines NS 18.62% 25.96% Source: World Bank staff using 1998 FHIS survey. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. NA means not available due to low sample size. 50 Box 3.1: ALLOCATING INFRASTRUCTURE FUNDS ON THE BASIS OF NEED: MEXICO'S EXPERIENCE In Mexico, as noted in the World Bank poverty assessment, allocations for new basic social infrastructure are based on need and rely on a formula. The allocation of funds from the federal entity to states is based on a weighted index of well-being called the Masa Carencial Municipal (MCM). MCM takes into account five indicators of well-being: the household per capita income (with a weight equal to 0.462), the average level of education per household (0.125 weight), a measure of the living space (0.239 weight), a measure of the availability of drainage (0.061weight), and a measure of access to electricity-fuel combustion (0.114 weight). MCM is calculated first at the household level, then at the municipality level, and finally at the state level. The federal entity makes the transfers to the states on the basis of the state-level aggregate MCM. Then the allocation is made from the state to the municipalities along similar lines. States who do not have the necessary information to apply the formula for their allocations to municipalities may use a simpler rule based on the arithmetic mean of the shares of the economically active population earning less than two minimum wages, the adult illiterate population, the population living in houses without drainage, and the population in houses without electricity. To cushion smaller and/or richer states from their reduction in basic infrastructure funding, one percent of the funds was allocated to each state equally in 1998. In 1999, each state still received 0.5 percent of the funds. Thereafter, only the formula will rule. The formula has increased infrastructure funding for the poorest states. The six poorest states have increased their share of these transfers from 29 percent in 1988 to 49 percent in 1999. In 2000, with the elimination of the fixed 0.5 percent share provision, this share will further increase to 54 percent. While the FAIS formula might be improved by finding a better way to define the weights of the five indicators on the basis of their elasticities of substitution, the current formulas are probably good enough. Additional relevant household-level information (such as direct measures of access to education and health facilities) could be incorporated into the formula, but for policy purposes, the allocation between states would not be affected much by such improvements because the various indicators are highly correlated with each other. What is more important is to find mechanisms to monitor the allocation of funds within municipalities. In this respect, the decision to apply similar formulas for the allocation within states is sound. The majority (90 percent) of funds are transferred to a municipal fund. The rest (10 percent of the relevant budget) goes into a state municipal fund. This 90/10 repartition is intended to promote responsiveness to local needs and priorities. Moreover, as of 1998, the allocation formula (or its simpler equivalent) must be used for the allocation of funds between municipalities so as to ensure redistribution within states as well as between states. The experience of 1997 during which states could allocate their funds to municipalities as they wished shows that the imposition of federal rules for within state allocations may be needed. In the states of Guerrero and Tlaxcala, the allocations between municipalities in 1997 were almost uniform, without regard for the relative state of deprivation of the municipalities. The changes made to the Law for fiscal coordination in 1999 should help in ensuring that the resources go to poor communities. One remaining challenge ahead is to design appropriate institutional management and control mechanisms. Many local governments lack the expertise and personnel to manage the funds, and resources have not yet been made available to help them increase their operating budgets, hire new staff or train existing staff, and modernize their administration. Another potential danger lies in the short-term assignments in the local political system. Municipal elections are held every three years and municipality Presidents can only serve for one term, which may imperil the continuity of the municipal policy. But on the other hand, while longer terms or re-election may help for stability, they can also create fiefdoms when there is no control. Civil society will have a role to play here in ensuring that the decentralization/devolution be pro-poor. 51 CHAPTER IV: EDIUCATION, NUTRITION, AND HEALTH A. IN EDUCATION, THERE ARE BOTH ACCESS AND QUALITY ISSUES, ESPECIALLY FOR THE POOR 4.1. Three ingredients are needed for a good education system: access, quality, and delivery. Here we focus on access and quality. As discussed in the World Bank's (1999) education strategy (Figure 4.1), access means that the students must be able to go to a school which is not too distant, and that they must have the means to afford the cost of schooling. Beyond access, if schooling is to be of use for the children, quality is key. Finally, delivery relates to issues of governance, resources, and evaluation. In this section, we focus on access, and especially on affordability for the poor, as well as on quality. Figure 4.1: Three Ingredients for a Good Education System ACCESS |4QALlY Students ready to learn -Relevant-curriculum R Supportive learning environment istaff Access to provision Teachiig and learnIng,process: A GOOD EDUCATION SYSTEM T DELIVERY Good governance Adequate resources Sound evaluation 4.2. Despite substantial progress over the last 25 years, Honduras lags behind for enrollment in secondary education, and there are pockets of low primary school enrollment in poor rural areas. As indicated in table 4.1, the number of years of schooling for the population aged 10 and over has doubled over the last 25 years. The rate of illiteracy has been cut in half. Enrollment in pre-schools has doubled, and nine out of ten nine years old are enrolled. While rural areas still lag behind urban areas, they are slowly catching up. This progress means that today, Honduras is doing as well as other PRSP and Central America countries for adult illiteracy and enrollment in primary education. However, the country still lags behind for enrollment in secondary education, and there are pockets of low primary school enrollment. This suggests that the main problems of access in Honduras are in primary school enrollment in poor and remote areas, and the lack of a good transition from primary to secondary school. [Honduras is reforming its primary and secondary education, with the number of years needed to complete primary education increasing from six to nine. Since the reform will take some time to be implemented nationally, we continue t:o refer to the transition from primary to secondary school, rather than to the completion of the new system of primary school. But the age group we have in mind is the one to which the reform of the education system applies.] Table 4.1: Trend in average years of schooling and enrollment rates, National, urban, and rural National Urban Rural 1974 1988 1L990 1997 1974 1988 1990 1997 1974 1988 1990 1997 Years of schooling 2.3 3.4 3.9 4.6 4.1 3.9 5.6 6.1 1.4 2.8 2.6 3.3 Illiteracy rate 40.4 32.0 :21.7 16.7 19.1 17.0 11.1 9.5 51.2 42.0 29.7 22.6 Pre-school enrollment 22.0 32.5 :25.7 44.1 33.1 46.4 38.1 60.3 18.0 12.5 19.4 33.5 Enrollment 9 years old 69.0 70.0 85.1 91.8 80.1 84.2 92.8 96.1 61.0 61.9 81.0 89.6 Source: UNAT (2000). The years of schooling is for the population aged 10 and over. 52 Box 4.1: EDUCATION AND HEALTH ACCOUNT FOR THE BULK OF PUBLIC SOCIAL EXPENDITURES This box provides a brief overview of the trend in public social expenditures. We do not go into a detailed analysis because the World Bank has prepared a Public Expenditure Review(World Bank 2001). In the 1990s, public social expenditures have represented about a third of total Government expenditures. Health and education account for more than 80 percent of public social expenditures, which justifies the focus on these two sectors in this chapter. The other categories are social security (0.1 percent of total expenditures in 1998), housing (0.6 percent), and others (5.0 percent). The weight of the "other" category has increased over time, from 2.0 percent in 1990 to 5.0 percent in 1998 chiefly because of the growth of special programs such as PRAF and FHIS. Apart from these two programs, the category also includes expenditures from the Department of Labor, JNBS, INJUPEMP, and electricity and transport subsidies. Table 4.2 reveals the trend in expenditures for education and health by using the following ratios: expenditures/GDP; expenditures/social expenditures; basic expenditures (essentially primary education and primary health care)/social expenditures; expenditures/total expenditures of the central Government; basic expenditures/total expenditures of the central Government; expenditures per capita (in Lempiras per year); and basic expenditures per capita (also in Lempiras). Key findings are as follows: * As a share of GDP, education and health expenditures have remained stable in the 1990s. As a share of social expenditures, education expenditures have increased a bit, while health expenditures remained stable. The same trend is observed for the comparison with total Government expenditures. * The good news for the poor is that the share of basic expenditures within sectoral expenditures has increased. As a result, despite a reduction in public deficits and a sustained population growth in the 1990s, basic expenditures per capita have increased. Table 4.2: Public exp enditures for education: basic indicators for the 1990-97 Education expenditures Health expenditures 1990 1993 1994 1997 90-97 1990 1993 1994 1997 90-97 Exp./GDP 4.2 4.9 3.9 4.1 4.2 2.8 2.8 2.5 2.0 2.6 Exp./Social Exp. 49.6 54.8 49.6 54.4 54.5 35.7 31.4 33.7 29.2 33.5 Basicexp./SocialE 25.3 32.1 25.0 26.8 29.1 8.7 12.7 11.1 13.1 12.6 Exp./CentralGvtexp. 16.2 18.3 16.2 18.7 18.0 11.0 10.0 11.0 10.0 11.1 Basic exp./Central Gvt 8.3 10.7 8.2 9.2 9.6 3.1 4.9 4.1 4.9 4.7 Exp. Per capita (Lps) 345.7 423.1 328.9 352.5 358.0 226.4 237.6 212.7 175.1 220.3 Basic exp. per capita 176.4 247.7 166.2 173.4 192.3 65.4 112.5 84.3 92.1 92.5 Source: Adapted from UNAT (2000). World Bank reports typically advocate a shift within education and health away from the tertiary sub- sectors in order to channel more resources to the primary and secondary sub-sectors. The basic idea is to implement cost-recovery mechanisms in the tertiary sub-sectors (which are more in demand among the non-poor) in order to channel more funds to the primary and secondary sub-sectors (which are more in demand among the poor). This strategy could be adopted in Honduras as well since a non-negligible part of public expenditures still go to the tertiary sub-sectors. In the case of education for example, it could be argued that subsidizing the public university (UNAH) does not yield substantive research externalities because the university is plagued by inefficiencies. Moreover, repetition rates are high, graduation rates are low, and the returns to an education at UNAH are apparently far below those at private universities. Table 4.3: Share of public funds allocated to education and health by sub-sector, 1997 Education (1997) Health (1994-97) Primary Secondary Tertiary Primary Secondary Tertiary Share of funds 60.8 19.5 19.7 53.0 23.5 23.5 Source: Adapted from UNAT (2000). 53 Box 4.2: ATTACKING POCKETS OF LOWV ENROLLMENT IN PRIMARY SCHOOLS: PROHECO Many countries are confronted with the difficult task of serving the education needs of populations living in remote rural areas. The task is difficult because the dispersion of the population makes it costly to supply traditional services. Honduras, with an overall population density of 51 inhabitants per square kilometer, faces the same problems. In some departments, the population density reaches 200 inhabitants per square kilometer. But in others such as Gracias a Dios and Olancho, the density is only 2.7 and 14.6 inhabitants per square kilometer. It is estimated that 160,000 children are not enrolled in primary school. Of these, 100,000 live in rural areas, in many cases in poor and geographically isolated communities. The new program PROHECO (Proyecto Hondurefio de Educaci6n Comunitaria) is funded by the World Bank and designed to meet the education needs of dispersed rural populations, while at the same time promoting a high level of community participation. Initiated in March 1999, the program was active in 500 communities by the end of that year (PROHECO, 1999). The majority of participating communities are not accessible by Jeep; 60 percent do not have access to piped water; 96 percent do not have access to electricity; and 55 percent of the parents are illiterate. None of the participating communities had a school teacher before the implementation of the program. The total number of students served at the end of 1999 by the program was 8,139 (5,032 in first grade; 1,573 in second grade; 996 in third grade; 349 in fourth grade; 189 in fifth grade). The population living in the program's areas of operation was 87,651. PROHECO focuses on pre-schools and primary schools, and it has a bilingual component for indigenous populations. The program works though Community-based Educative Associations (AECOs for Asociaciones Educativas Comunitarias) whose locally elected officials receive training from the program. The state transfers the funding for the program directly to the AECOs which are in charge of hiring the teacher. The flexibility (and low infrastructure cost) of the program is illustrated by the fact that only 17 percent of the classes take place in a schlool. Most of the classes take place in private homes (72 percent), and some take place in churches (7 percent) or in other places (4 percent). 4.3. The lack of a good transition from primary to secondary school is due to both the high cost of secondary schooling, and the low interest in pursuing an education beyond the primary level. Table 4.4 provides basic statistics on schooling, child labor, and the reason for not going to school. The table is based on the 1999 PRAF survey representative of households living in Honduras' poor municipalities . * Differences by age group: While at least three fourths of children below 14 go to school, only slightly more than one fourth of those aged 15 to 17 do. There are two main reasons for not going to school after 15 years of age. First, some of the students are simply not interested in pursuing their education. This is stated in the reasons for not going to school, but it also appears in the fact that some drop-outs consider that having finished the primary cycle is a reason good enough not to go forward. Some children appear to consider the pursuit of their studies as irrelevant, which could be worse that wishing to go to school but not going because of a lack of quality. The lack of interest may be responsible for half of the drop-outs. Next, there is an affordability problem, with a third of the children saying that schooling is too expensive or that they must work. * Differences by gender: As expected, more boys undertake paid work than girls, while more girls undertake unpaid (domestic) work than boys. Having to work is cited as a reason to leave school 'To illustrate the problem of the transition from primary to secondary school by giving data on enrollment patterns by age group, we separated the sample in three groups: 6-9 years, 10-14 years, and 15-17 years. In principle, a 12 year old child could attend secondary education, but in practice, this is rarely the case because many children enroll in primary school late, and many others repeat one or more years while in primary school. Therefore, the drop in enrollment between the 10-14 years and 15-17 years groups can be considered as representative of the transition problem between primary and secondary school. Changing the years in the table would not alter the conclusions. 54 more by boys than by girls, while "other reasons" (which may include domestic tasks as well as marriage at a young age for the older children) are cited more by girls than by boys. * Differences by location: Enrollment and attendance are a bit higher for children living in the municipal centers than for those living in surrounding areas, but there are exceptions. Overall, there are no clear patterns of differenciation between the children according to their geographic location in the municipality. Note that since the PRAF survey was implemented in poor municipalities only, many of the municipal centers should be considered as rural areas, rather than urban areas. * Differences by income group: The children have been classified in three income groups according to the level of per capita expenditures in the household (Lempiras per month). School attendance rises with income, and child labor decreases. The problem of affordability and the necessity to work are more prevalent in the poorer income groups, at least for the household living in the municipal centers. Table 4.4: Schooling and child labor b gender, age, location, and income in poor municipalities Boys by age group Girls by age group 6to9 10to14 15to17 6to9 10tol4 15to17 Schooling and child labor (%) Went to school in the last 12 months 83.90 71.92 27.44 83.58 76.86 32.53 Is going to school right now 82.88 69.58 25.60 82.96 74.80 30.10 Is working (paid or unpaid) NA 32.96 76.06 NA 27.65 71.11 Unpaid work NA 23.70 45.30 NA 25.29 58.79 Paid work NA 9.26 30.76 NA 2.35 12.32 Reasons for not going to school (%) Not interested 11.92 38.46 36.14 3.68 23.35 28.32 Finished the cycle 0.00 6.35 10.40 0.00 9.73 10.40 Too expensive for the family 15.89 18.39 18.32 14.71 29.18 23.99 Has to work 2.65 17.73 23.27 0.00 4.67 11.56 Relatives do not allow 6.62 3.68 0.99 8.09 7.78 4.05 Bad quality of education 0.66 0.00 0.74 1.47 1.95 1.16 Too old to go to school 0.00 0.00 0.50 0.00 0.78 0.58 Other reason 62.25 15.38 9.65 72.06 22.57 19.94 Municipal center by income group Surrounding areas by income group Y<200 200-500 Y>500 Y<200 200-500 Y>500 Schooling and child labor (%) Went to school in the last 12 months 65.09 63.78 89.78 60.12 75.48 73.21 Is going to school right now 63.68 61.39 89.57 57.42 74.76 70.89 Is working (paid or unpaid) 47.12 44.86 14.37 50.92 29.66 37.74 Unpaidwork 34.62 35.92 11.19 42.33 20.00 32.68 Paid work 12.50 8.94 3.17 8.59 9.66 5.06 Reasons for not going to school (%) Not interested 22.08 16.40 35.35 18.31 25.24 19.57 Finished the cycle 1.30 5.40 1.01 5.28 4.76 9.42 Too expensive for the family 31.17 12.00 17.17 19.01 23.33 15.94 Has to work 14.29 8.40 5.05 7.39 9.05 7.25 Relatives do not allow 3.90 6.80 3.03 5.28 3.81 2.90 Bad quality of education 0.00 1.00 1.01 0.70 1.43 0.00 Too old to go to school 0.00 0.40 1.01 0.00 0.00 0.00 Other reason 27.27 49.60 36.36 44.01 32.38 44.93 Source: World Bank staff using 1999 PRAF survey. NA indicates that the survey does not provide information on the work of children below 1 0. Y is the level of per capita expenditures of the household in Lempiras per month. 55 4.4. Uniforms, materials, registration fees, and also transportation costs in the case of secondary school, make the cost of schooling high for older children. For the children aged 15 to 17 who are enrolled, the cost of schooling is twice the cost observed among the children aged 10 to 14 (table 4.5). The largest expense for those enrolled is by far the cost of uniforms and materials (this is also true at lower levels of schooling). The expenditures per child increase rapidly with the level of total per capita expenditures (proxy for income) of the household. Only a small number of students receive books and uniforms for free, but this is the case mnore for the poorer children. The number of students who receive schooling supplies from the Government's program PRAF is low, and there are few differences between income groups. Also included in the direct cost of schooling is the cost of public transportation, especially for older children. By contrast, at a lower age, going to school by foot is an almost universal rule. The average number of hours of classroom per week is slightly above 20 hours at all grade levels. Table 4.5: Cost of schooling by gender, age, location, and income in p or municipalities Boys by age group Girls by age group _ 5-10years 10-15years 15-18 5-l0years 10-15years 15-18 Direct cost of schooling (Lempiras) Inscription fee (peryear) 5.50 13.60 70.51 3.18 15.91 81.61 Books (per year) 8.50 18.15 52.04 8.60 32.69 75.37 Uniforms and material (per year) 158.69 231.01 429.92 131.63 237.10 411.16 Other costs, incl. transport (last month) 23.62 32.63 89.45 23.52 38.51 96.86 Assistance received (percentage) Books 6.84 4.68 7.91 5.44 4.59 8.05 Uniforms 10.94 10.82 5.76 10.57 8.65 9.40 Other types of assistance 2.46 3.22 0.00 3.32 2.36 2.01 PRAF bag ("bolson") 3.42 2.92 0.00 3.63 2.36 0.00 Opportunity cost in terms of time Hours of class per week 22.36 22.85 22.12 21.93 23.43 22.42 Time needed to go to school (minutes) 14.60 14.84 16.83 14.26 14.16 16.16 Municipal center by income group Surrounding areas by income group Y<200 200-500 Y>500 Y<200 200-500 Y>500 Direct cost of schooling (Lempiras) Inscription fee (per year) 1.78 3.75 29.12 0.93 7.45 12.00 Books (per year) 2.08 7.73 46.41 3.48 8.64 19.40 Uniformns and material (per year) 54.96 127.43 341.40 53.69 187.85 248.42 Other costs, incl. transport (last month) 9.81 16.12 64.46 4.65 22.16 45.81 Assistance received (percentage) Books 5.19 5.66 5.06 7.05 4.50 4.76 Uniforms 18.52 8.43 6.00 19.58 11.90 5.36 Other types of assistance 1.48 4.78 1.29 3.92 2.41 1.79 PRAF bag ("bolson") 1.48 4.78 1.18 2.61 2.89 2.68 Opportunity cost in terms of time Hours of class per week 19.96 21.74 23.13 21.99 22.66 21.96 Time needed to go to school (minutes) 9.61 17.35 9.78 23.10 11.61 15.44 Source: World Bank staff using 1999 PRAF survey. Y is the level of per capita expenditures of the household in Lempiras per month. 4.5. Apart from the problem of affordability, there is a problem of quality in education, especially at higher levels of schooling. This represents a waste of resources with no benefit for the poor. Although repetition and absenteeism rates have decreased in primary and secondary education (table 4.6), concerns have been raised regarding the quality of the education system. There are many causes to the 56 problem: poor teaching quality, insufficient classroom time, teacher absenteeism, inadequate supervision from the school system and the community, lack of external evaluation, insufficient training, inappropriate curriculum, shortages of books and other teaching materials, and lack of funds devoted to investments as opposed to teacher pay. In the 1998 FHIS survey of households living in the areas where Honduras' social fund is active, between one fourth and one third of the students say that the reason why they missed school recently was either that the teacher did not show up, or that the school closed. UNAT (2000) has estimated that the total cost of repetition and absenteeism amounted to about 330 million Lempiras per year, or close to 20 percent of the education budget for primary, secondary, and tertiary education. The cost associated with repetition and absenteeism increases as a proportion of the budget with the level of education. At the university level, one third of the budget can be considered as wasted because of repetition and absenteeism (for students who are unlikely to be from poor backgrounds). Table 4.6: Repetition and absenteeism rates by education levels, and public cost thereof, 1980-97 1980 1985 1989 1993 1997 Cost in 1997 Share of (Lempiras) budget (%) Primary school Repeating 16.2 15.6 12.3 11.7 9.7 105,553,688 14%oftotal Absenteeism 4.9 4.5 4.3 3.3 3.5 35,859,876 budget Secondary school Repeating 22.9 16.8 19.0 9.3 10.3 35,919,252 20% of total Absenteeism 14.1 23.5 22.7 19.3 19.2 33,597,060 budget Superior level (UNAH) Repeating 8.0 17.1 19.6 14.0 17.0 64,287,450 34% of total Absenteeism 18.3 10.6 12.4 12.0 15.6 56,724,675 budget Source: Adapted from UNAT (2000). 4.6. The problem of quality in public schools can also be made apparent by a comparison with private schools. Table 4.7 provides basic statistics for public and private schools. The private sector is present at all levels of schooling in Honduras, and especially at the secondary level (although its market share at that level has decreased over time as the supply of public secondary education has increased). While the cost of private schools is higher than that of public schools, the quality is also higher as measured by lower repetition rates and higher returns to education (at least at the university level). Table 4.7: Market share, cost, and efficiency of private and public school by level, 1997 Share of public sector Cost per student, 1996 Repetition rate Academic return Public Private Public Private Public Private Public Private Primary 93.9 6.1 1,006 1,962 9.7 2.9 39.0 n.a. Secondary 61.8 38.2 1,916 4,759 10.3 8.7 n.a. n.a. Superior 83.7 16.3 7,725 10,168 17.0 8.6 37.3 79.4 Source: Adapted from UNAT (2000). The cost per student is in Lempiras per year. 4.7. Research suggests that increasing teacher quality may be the most cost-effective way to improve achievement in primary school as measured by test scores. Bedi and Marshall (1999) analyze the deterninants of school attendance and student achievement using data from 33 randomly selected rural schools in the department. of Valle in southern Honduras. They estimate education production functions for achievement (test scores) in mathematics and Spanish for grades one to three. One of the inputs in the production function is a measure of teacher quality, and more specifically the use by the teacher of an "active and participative" teaching methodology. This measure of teacher quality was obtained by having the teacher take a test with 25 situations indicative of a cooperative versus a teacher- centered style of the teacher. The authors find that a five point increase in the teachers' score is associated with a 4 to 8 point increase on student achievement tests. The other important inputs raising student achievement are the student/teacher ratio and the availability of preschool programs. By contrast, 57 textbooks play a lesser role. The authors run simulations to test which measures might be most cost effective in order to increase achievernents: reducing the class size (with the cost of doing so depending on the type of school - single teacher, dual teacher, or multi-teacher), increasing pre-school coverage, or increasing teacher training through seiminars given by experts (with various assumptions as to the impact of these seminars on teacher quality). The results of the cost effectiveness analysis are given in table 4.8 in the form of a cost in Lempiras per point increase in student achievement. The conclusion is that enabling teachers to participate in serninars in order to improve the quality of their teaching may be the cheapest way to improve test scores among students. While these empirical results could be debated, they have the merit of pointing to less traditional and more qualitative ways to improve test scores than the usual quantitative measures such as student/teacher ratio and textbook/student ratios. Table 4.8: Cost effectiveness of school inputs to increase achievement in primary schools, 1996 Grade I Grade 2 Grade 3 Math Spanish Math Spanish Math Spanish Reducing class size (for mean class size) 57.48 129.65 By school type: Single teacher 35.85 80.86 2 teacher 59.29 133.71 Multi teacher 59.66 134.55 Increasing preschool coverage: per student 59.64 45.61 41.85 115.3 Increasing teacher training Assumption 1 14.21 28.76 28.91 20.90 18.35 Assumption 2 4.74 9.59 9.64 6.97 6.12 Assumption 3 1.90 3.74 3.85 2.79 2.45 Source: Bedi and Marshall (1999) B. THE COST OF CHILD LABOR IN TERMS OF FORGONE FUTURE EARNINGS IS LARGE 4.8. Almost 400,000 children and adolescents are working today in Honduras. Table 4.9 from UNDP (1999) gives estimates of the number of children (14 years old or less) and adolescents (15 to 19 years old) working in Honduras according to the EPHPM surveys. In 1998, out of a labor force of 2.2 million people, 100,000 were children and 279,000 were adolescents. The number of children and adolescents actually employed (versus in the labor force) at the time of the survey were slightly lower, at 98,000 and 261,000 respectively. Most of the children and adolescent working are males, but this may be in part because domestic work is not as well captured in the surveys. As expected, the incidence of child and adolescent labor is higher in rural than in urban areas. Table 4.10, also from UNDP (1999) suggests that the incidence of labor (in the labor market or at home through domestic work) among those 15 to 19 years old is substantially higher in Honduras than in other Latin America countries. This relates to the transition problem between primary and secondary school and to the low income levels in the country. Table 4.9: Extent of child and adolescent labor by gender and urban/rural areas, 1990-1998 1990 1994 1998 1990 1994 1998 Labor force 1,605,917 1,825,438 2,222,658 Female labor force 473,814 555,418 766,387 Children 89,532 79,614 100,213 Children 16,884 18,955 29,511 Adolescents 193,601 223,528 279,309 Adolescents 43,671 57,409 73,891 Employment 1,525,137 1,775,162 2,134,992 Urban labor force 709,898 83,213 1,065,891 Children 85,819 78,894 97,845 Children 22,022 22,427 29,190 Adolescents 181,009 214,785 261,032 Adolescents 66,734 81,610 109,545 Male labor force 1,132,103 1,270,020 1,456,271 Rural labor force 896,019 993,308 1,156,767 Children 72,648 60,659 70,702 Children 67,512 57,187 71,023 Adolescents 149,929 166,11]9 205,418 Adolescents 126,866 141,918 169,764 Source: UNDP (1999). 58 Table 4.10: Incidence of labor among 15 to 19 years old, Honduras and other countries, percentages Urban areas Rural areas Not in school and in Not in school and Not in school and in Not in school and labor force domestic work labor force domestic work Honduras Male 49 5 75 2 Female 19 21 20 51 LAC countries Male 33 1 60 1 Female 16 12 21 33 Source: UNDP (1999). 4.9. A first problem with child labor is that many children working may be at risk of being hurt. Using the EPHPM, UNICEF (1999) reports that out of the 98,000 children working in Honduras in 1998, 48.0 percent were employed in agriculture, mining, and other similar activities; 10 percent were employed in industries such as graphics, chemistry, food and beverage, leather, tobacco, and ceramics; 2 percent were employed in occupations involving hauling, retail selling, and packaging; and 6 percent were employed as domestics. According to UNICEF, all these occupations involve some level of risk. In this case, two thirds of the working children (and a similar proportion of working adolescents) are at risk of being hurt. Even though this estimate of the share of working children at risk may be on the high side, the concern for the hazard involved in many activities performed by children and adolescents remain. Table 4.11: Characteristics of children participating in a social project in Tegucigalpa, circa 1990 "Market" children (N=909) "Street" children (N=l 10) Mean age 10.3 years 12.9 years Gender 54% boys 95% boys Migrants 17% 40% Orphans 2% 5% Parents married 15% 8% Good relations with their family 78% 32% Living with one or both parents 86% 57% Attending school About half 10% Methods of support Money from informal sector employment Income source: 42% begging; 15% is shared with family carrying things; 12% theft Meals per day 59% eat 3 meals 6% eat 3 meals, but nutrition is ok Have been arrested by police Almost none About half; 20% in street gang; more than half use drugs (sniffing) Sexually active 5% 44% If sexually active, treated for STD 40% 85% Main source of problems Extreme poverty of families Being family-less in the first place source: Wright and Wittig (1993) 4.10. A second problem with child labor is that among working children, "street children" face very hard living conditions and can represent a threat to society. Wright and Wittig (1993) use data from Proyecto Alternativas, a health, education, and social services organization in Tegucigalpa, to provide a description of the characteristics of street children as opposed to market children (table 4.1 1). Street children consider the streets as their home; market children use the street for their work, but have somewhere else to go home. Although street children tend to be older, they remain young with a mean age of 12.9 years. They are almost exclusively boys, and more likely to have migrated to Tegucigalpa from another area. They do not have good relationships with their family; some are orphans. Only half live with a parent. One in ten is attending school. Street children do not hesitate to beg or to steal to make a living. One out of five street children belongs to a gang, and one out of two has been arrested by the police. One out of two is sexually active, and among those, a majority has been treated for a sexually 59 transmitted disease. Clearly, in the case of street children as opposed to working children in general, apart from poverty, a source of problems is the lack of a family and guidance. Special interventions are needed for street children. 4.1 1. A third and more widespread problem is that child labor reduces the probability of schooling, thereby perpetuating poverty from one generation to the next. Given that the children have only a given number of hours per day for sclhooling, labor, and leisure, child labor may lead to less schooling. When this is the case, the likelihood that the child will emerge from poverty when he reaches adulthood will be reduced since the human capital of the child is reduced. Because parents can reduce the time available for leisure when a child is working, the substitution effect between work and schooling is likely to be partial only. As explained in anmex 2 (section MA.7), bivariate probit regressions can be used to estimate the expected probability of going to school when a child is working or not, and thereby the substitution effect. These probabilities are given for urban boys, urban girls, rural boys, and rural girls in table 4.12. The analysis is performed for children aged 12 to 15 years, and for the purpose of comparability with a regional study having the same information for other countries (Wodon, 2000), paid child labor as opposed to unpaid child labor is taken as the reference. The probability of going to school when working is low. It varies from 6 percent to 34 percent depending on the sample. The probability of going to school when the child is not working is much higher, at 60 percent to 83 percent. The difference in the probabilities of going to school when the child is not working, and when the child is working, provides an estimate of the substituticn effect between work and schooling. The estimates vary from 48 percent to 75 percent. While substitution effects are not unitary (child labor can take place after schooling, or the parents can reduce the time allocated to leisure when children work), they are large. 4.12. As a result of the substitution between child labor and schooling, and of large returns to education in Honduras, the cost of child labor in terms of forgone future earnings is high. The next step in estimating the cost of child labor consists in predicting future earnings according to various levels of education. The assumption is that if a child is working, and if this does not enable him to go to school, the child completes only the primary level of education (six years of schooling, up to age 12.) In contrast, if the child is not working, and if this enables him to go to school, the child completes the lower secondary level (9 years of schooling.) Thus, in the first three years after the completion of primary school, a working child enjoys a benefit because he receives a wage. But for the rest of the child's life, the earnings are lower because of the lower level of education achieved. Computing the net actualized value (with a five percent discount rate) of the difference in the future streams of income with only primary education, and with 3 years of secondary education provides the "difference in income" figures in table 4.12. These represent the net monetary loss due to the lower level of education achieved for working children as a percentage of t]he life-time income that the children could have expected had they remained in school instead of working. The figures take into account the expected probability of working and the expected wage when working for various education levels as obtained from standard Heckman regressions (see annex 2, Section MA.4 for a discussion of the model). Taking the difference between the net discounted earnings with two levels of schooling, dividing this difference by the expected life-time earnings when children receive 9 years of schooling, and multiplying the result by the substitution effect between child labor and schooling gives the estimate of the cost of child labor in terms of foregone lifetime earnings. The cost is large in table 4.12, at 7 to 29 percent of lifetime earnings depending on the sample. The cost in percentage terms is larger for girls essentially because of the larger impact of a better education on the probability of working. An alternative measure of the cost of child labor is to divide the discounted loss in future earnings by the yearly poverty line in order to get an estimate of the number of equivalent additional years out of poverty that a child (not his whole family when the child reaches adulthood) could hope for if he/she was not working. Table 4.12 indicates that in Honduras, this cost varies from 6 to 13 "poverty years." Whichever measure of the cost of child labor is used, the cost is higher in Honduras than in most other Latin American countries because of the large substitution effects between child labor and schooling on the one hand, and the high returns to education on the other hand. 60 Table 4.12: Estimates of the cost of child labor in terms of forgone future earnings, 1996 Urban boys Urban girls Rural boys Rural girls Probability of schooling if working (1) 0.34 0.07 0.06 0.06 Probability of schooling if not working (2) 0.82 0.83 0.60 0.60 Difference in probability (3)=(2)-(1) 0.48 0.75 0.54 0.55 Difference in income (4) 15.26% 23.07% 19.82% 53.23% Cost of child labor (5)=(3)*(4) 7.29% 17.42% 10.66% 29.10% Cost in "poverty years" 9.63 13.02 11.08 6.40 Source: World Bank staff using EPHPM. C. BEYOND PRIMARY SCHOOLS, INTERVENTIONS ARE NEEDED IN SECONDARY SCHOOLS 4.13. To discourage child labor and promote schooling, governments fund programs that reduce the cost of schooling by providing benefits in cash or in kind conditional on school assistance. School-based transfers in cash or in kind reduce the opportunity costs for poor parents of keeping their children in school. This opportunity cost is essentially the loss in child wages or in the value for the parents of fne domestic work done by the children which cannot be done when the children go to school, plus the cost of supplies, transportation, etc. In many cases, the opportunity cost of schooling is difficult to estimate, and it is not obvious that the transfers should be equal to the opportunity cost for the parents to send their children to school. Indeed, it is reasonable to think that the parents have an intrinsic interest in having their children go to school, either for altruistic motives, or for the future benefits that intergenerational transfers provide once the children reach adulthood. In other words, a relatively small transfer may have a large impact. At what level of schooling should the grants be provided? This depends on the characteristics of the country. In Brazil and Argentina, the programs focus on the children in secondary school, since these are the children more likely to be pulled out of school during a crisis. In Mexico, the program covers the end of primary schooling, and the lower secondary school cycle. In Venezuela, the program covers primary school children. In some cases, these programs are tied not only to attendance, but also to school performance, including passing on to the next grade. While this may provide valuable incentives for the students to perform, one has to make sure that such conditions do not exclude the poorest which have more difficulties than others in being successful in school. 4.14. In Honduras, PRAF (Programa de Asignaci6n Familiar) runs a number of programs in the areas of education, health, nutrition, and income support. PRAF was created in 1990 in order to serve Honduras' poor within the context of the structural adjustment program that was put in place at that time. In 1998, the program had 318,000 beneficiaries. PRAF's seven sub-projects had a total budget allocation of 130 mnillion Lempiras. This makes PRAF the second largest program targeted to the poor after the FHIS (Fondo Hondureino de Inversi6n Social) . The seven sub-projects are as follows (PRAF, 1999): * Bono Escolar de Primero a Tercer Grado and Bono Escolar Ampliado al Cuarto Grado: Started in 1990, the project gives cash stipends of 50 Lempiras per month for ten months and for up to three children per household provided the children go to primary school (first to third grade). In 1998, the program was extended to the fourth grade of primary school. The program is targeted first to poor areas using municipal information on child malnutrition and unmet basic needs, and next to poor households (in principle, to households with per capita income below 600 Lempiras per montht). The cost of the Bonos Escolar in 1998 was 37.7 million Lempiras for the first three grades, and 6.5 million Lempiras for the fourth grade. * Bono Materno Infantil: Started in 1991, the project also gives cash grants of 50 Lempiras per month for 12 months. Its main objective is to help for the nutrition of pregnant mothers, nursing mothers 2 The electricity subsidy has a larger budget than PRAF, but it is not well targeted to the poor (see chapter 5). 61 and their babies, as well as children below five years of age. Apart from the targeting of young children and pregnant women, the criteria for means-testing are similar to those of the Bono Mujer de Familia. The conditionality for the grant is a regular consultation visit for basic health care. The women also must avoid having more than one child every two years. Each family can have up to three beneficiaries of the grant. The total budget for this sub-project was 42.8 million Lempiras in 1998. * Bono Nutricional: Started in 1998 in the departments of Santa Barbara, Intibuca, and El Paraiso, the project gives cash stipends of 50 Lempiras per month for 12 months to households with children below five years of age at risk of malnutrition. The targeting criteria and conditions for participation are similar to the Bono Maternal Infantil. The total cost of the project is 6.5 million Lempiras. * Bono para la Tercera Edad: Started in 1993, the project gives cash grants of 50 Lempiras per month for 12 months to individuals at least 65 years old. The prograrn is targeted to 63 municipalities with high levels of poverty. The program is in principle means-tested as beneficiaries must have a level of income below 400 lempiras per month, and have three unmet basic needs. There were 11,500 beneficiaries in 1998, for a total cost of 6.7 million Lempiras. * Bolson Escolar: Started in 1992, the project consists of distributing school material to poor children. In 1998, 76,000 school bags were distributed, for a total cost of 6.8 million Lempiras. * Desarrollo Integral de la Mujer: Started in 1991 as a training component for women, the project now provides support for investments, training, micro-credit, entrepreneurship, and retail shop. This is a comparatively smaller program whLich reached 2,235 families through 8 projects in 1998. 4.15. The PRAF stipends appear to be relatively well targeted to those in need, at least in rural areas. To analyze the targeting of PRAF, we use the information on the program given in the 1998 FHIS survey. The survey is not nationally representative: it covers the areas where the FHIS is active. Table 4.13 provides the mean per capita income of PRAF beneficiaries versus non-beneficiaries, with three different ways of identifying PRAF beneficiaries. The first part of the table compares the income levels of the households living in areas where schools participate in PRAF's Bolson Escolar project versus the income levels of the households living in areas where the schools do not participate in the project. In the rural sector and nationally, targeting is good. In urban areas, it is not. The rest of the table uses two other criteria to identify at the household rather than at the school level those who benefit from PRAF. The first criteria is whether somebody in the household has received a cash transfer from PRAF. The second criteria is whether a child under 10 years of age or a pregnant woman in the household has received PRAF support. With the first criteria, the findings are similar to those obtained at the school level. With the second criteria, in all cases the income of PRAF beneficiaries is below that of non-beneficiaries. Apart from the evidence in table 4.13, it may also be recalled that in chapter 1, a source decomposition of the Gini index of inequality indicated that PRAF transfers were inequality reducing. Overall, the targeting of the program is thus relatively good, but it could still be improved, and PRAF has been working with the International Food Policy Research Institute (IFPRI) in order to find more efficient ways of means-testing the program in the next incamation of the program in the fall of 2000 (see below). Table 4.13: Targeting of PRAF: Per cpita income of beneficiaries and non-beneficiaries, 1998 PRAF in-kind support ("bolson") At least one household member A child under 10 or a pregnant is distributed in a school in the has benefited from a cash transfer woman in the household has living area of the household from PRAF received PRAF support % of Yes No, %of Yes No % of Yes No people people people National 17.84% 158.96 479.53 15.50% 317.17 449.09 15.44% 184.72 473.17 Urban 1.01% 640.76 574.92 11.24% 625.38 567.33 6.55% 314.50 592.02 Rural 35.93% 144.35 321.19 19.81% 139.95 316.46 24.44% 149.50 324.19 Source: World Bank staff using 1998 FHIS sunrey. 62 4.16. There is limited evidence that PRAF may have a positive impact on primary school enrollment. The 1998 FHIS survey can be used to provide a tentative evaluation of the impact of PRAF on school enrollment using matching methods3. As was the case in table 4.13, two criteria are used at the household level to identify beneficiaries. For the first criteria (whether a child under 10 or a pregnant woman in the household has received PRAF support), we need to use the sample of children aged 6 to 9 to measure the impact of PRAF on enrollment because the question is not asked for older children. For the second criteria (whether the household as a whole has benefited from PRAF support through a cash transfer), we can use the larger sample of children aged 6 to 12. The impact on school enrollment has been estimated separately for boys and girls, and for urban and rural areas. Although the results tend to depend on the specification of the model used to analyze the determinants of participation in the program, there is some indication that PRAF may increase enrollment. Apart from enrollment, better attendance will be one of the goals of the next version of the program that will be implemented in the fall of 2000. 4.17. Beyond the current focus of PRAF (and PROHECO) on primary school, it would be valuable to promote programs increasing enrollment in secondary schools. As was mentioned in Box 4.2, there are still 160,000 primary school age children who are not enrolled in Honduras, and these belong to some of the poorest families in the country. Programs such as PRAF and PROHECO are helping these families, and these programs should be continued. At the same time, given that a key problem in Honduras is the lack of a good transition from primary to secondary school, other initiatives should be implemented at that level. Given the debt relief which will be received by Honduras under the HIPC initiative, it may be feasible to implement new programs for the transition to secondary schools without endangering the funding in place for programs targeting primary school aged children in poor areas. 4.18. As part of a broad evaluation effort, PRAF is currently changing the functioning of the Bono Escolar program, and the new rules will be taking effect in the fal of 2000. With funding from the Inter-American Development Bank and analytical support from the International Food Policy Research Institute (IFPRI), PRAF will be conducting a multi-year evaluation of its Bono Escolar project. To this end, the program will be operating in a different way as of the fall of 2000. The program will still be targeted to poor municipalities using the Censo de Talla, but four different modules will be put in place, in different municipalities. The first module will consist of demand-side interventions like the current Bono Escolar; the second module will consist of supply-side interventions (quality of teachers, etc.); the third module will consist of both demand- and supply-side interventions; finally, the fourth module will consist of no interventions at all and serve as control group. This design will help in identifying which mix of interventions is most successful in raising enrollment, attendance, and achievement. Contrary to what is being done in Mexico for the PROGRESA program (also with analytical support from IFPRI; See Box 4.4 at the end of this chapter), the PRAF stipends will be available to all households with primary school age children living in the participating municipalities. In Mexico's PROGRESA, there is a second level of targeting within poor municipalities, so that only some of the households participate (this is done through means-testing). The rationale for not targeting PRAF within the municipality has to do with the high levels of poverty encountered in participating municipalities (so that targeting might not bring many savings), and the desire to avoid the potentially negative effects of intra-municipal targeting on social 3 Each child participating in the program is characterized by a vector of variables X Ideally, we might want to match each participating child with a non-participating child having the same characteristics. But this is not feasible if X is large. Fortunately the matching can be made on the probability of participating only (Rosenbaum and Rubin, 1983, 1985). Instead of matching on X one can construct the control group on the basis of the conditional program participation probabilities Prob (P=l I A) where P stands for program participation. For each child in the sample, we compute Prob (P=l j X) using a standard logit model. Then, for each participating child, we find a non-participating child whose probability of participating is closest to that of the participating child, where closeness is measured by the difference between the probabilities. Doing this for all participants generates the control group. Then, school enrollment can be compared in the treatment and control groups, and attributed to the impact of the program. 63 cohesion. This new functioning of PRAF's Bono Escolar project is most welcome, because the lessons leamed from the evaluation should be valuable. But this does not detract from the fact that it would still be better to target the stipends to older children in order to facilitate the transition from primary to secondary school. Also, before concluding this section, it is important to always keep in mind that programs such as PRAF (or the FHIS, discussed in chapter 3) should not follow an "assistance" model, but should rather enable the poor to emerge from poverty on a sustainable basis. D. BEYOND ACCESS AND QUALITY, THIERE ARE ISSUES OF COST-EFFICIENCY IN HEALTH4 4.19.' Malnutrition rates suggest a deterioration in the 1990s. Honduras has a good measure of child malnutrition thanks to an annual census done in the first year of primary school. Table 1.14 provides information on the percentage of stunted children in the first grade of primary school (stunting is itself measured as the share of children among a given age group having a height at least 2 standard errors below international standards for that age). Three observations can be made. * First, there has been no progress toward reducing the incidence of stunting between 1986 and 1997. As a matter of fact, there is a deterioration after 1991, when according to the official estimates, 34.9 percent of all children in first grade were stunted, as compared to 40.6 percent in 1997. Table 4.14 also provides altemative estimates after data cleaning, including not taking into account children who repeat their first grade (so that they are not counted two years in a row). The trend is very similar. -- Second, children who enter primary school at a later age have a higher probability of being stunted. In 1997, 62.3 percent of 9 year olds were stunted, versus 31.6 percent of 6 year olds. This may be because the parents wait for the children to be tall and strong enough before sending them to school. It may also be due to a choice by parents to first send healthier children to school, which results in late enrolment for those children who have suffered from malnutrition in their infancy, and thereby are stunted. A third and more straightforward explanation could be that the parents who do not send their children to school until they are relatively old live in isolated locations a long way from schools, or are less educated. These parents are more likely to be poor and have malnourished children. - Although this is not shown in table 4.14, the incidence of stunting according to the Censo de Talla is much higher in rural areas (at 47.6 percent in 1997) than in urban areas (at 28.5 percent in 1997). There have also been national nutrition surveys in 1987, 19934 and 1996. The 1987 sample is by health region and not directly comparable, but the other two samples are. The results show a decrease in malnutrition in Tegucigalpa (frorn 34.2 percent to 24.9 percent), San Pedro Sula and medium cities (from 28.6 percent to 17.4 percent), and the rural south (from 40.3 percent to 35.3 percent). But malnutrition increased in small cities (from 30.1 percent to 34.7 percent), the rural west (from 60.0 percent to 62.3 percent) and the rural north and center (from 41.0 percent to 41.9 percent). Table 4.14: Trend in malnutrition: Incidence of stunting for first grade students, 1986-1997 Official estimates Alternative estimates 6 year 7 year 8 year 9 year Total 6 year 7 year 8 year 9 year Total olds olds oldLs olds olds olds olds olds 1986 25.1 36.7 47.2 56.0 39.8 NA NA NA NA NA 1991 24.1 33.9 42.7 50.2 34.9 NA NA NA NA NA 1993 26.7 36.6 44.8 52.6 35.5 30.1 42.2 52.5 58.7 NA 1994 27.5 39.8 50.0 59.0 38.1 30.3 45.1 57.4 65.3 NA 1995 27.5 41.2 50.9 58.8 38.6 30.4 46.1 57.9 64.9 NA 1996 28.2 40.7 50.9 59.0 38.0 31.2 46.1 58.4 65.8 NA 1997 31.6 43.2 52.4 62.3 40.6 34.7 48.4 59.8 68.9 NA Source: Censo Nacional de Talla for official tabulations and IFPRI staff for alternative estimates. 4 A report on health was written by the World Bank (1998) recently. The health sector is also discussed extensively in the review of basic social expenditures prepared by UNAT (2000). Therefore we restrict ourselves to reporting some of important findings together with a few new empirical results from the PRAF and FHIS surveys. 64 Box 4.3: MALNUTRITION, POVERTY, AND COMMUNITY-BASED GOVERNMENT PROGRAMS Explaining the possibility of a divergence between poverty and malnutrition trends In Honduras, there appears to have been more improvement in poverty than in malnutrition in the 1990s. While poverty and malnutrition tend to coincide, neither is a necessary nor sufficient condition for the other. The main reason why malnutrition rates may not track poverty rates is that malnutrition is the result of a complex process involving interactions between food intake, disease, and child caring behaviors. Food intake is affected by food supply, access (purchasing power), and utilization, which are only partially related to poverty. Disease is a function of exposure, susceptibility, and recovery rates, most of which are not related to income in the short term. Child caring behaviors (e.g., breast feeding, psychosocial stimulation, and affective behaviors) are functions of culture, socioeconomic status, education, and social environment, all variables which are again not directly related to income. Given the complex process leading to malnutrition, it is not surprising that a decrease in poverty need not improve nutrition, at least in the short term. While the family may be able to buy more food, the food may not be given to the young child who needs it most because of existing cultural beliefs and practices. Or the child may be given the food, but the continued use of contaminated water means that diarrhea and parasites will rob the child of the extra nutrients as soon as he ingests them, and the family may still not have access to health care services to rapidly address the disease. Furthermore, if the improvement in earnings is achieved through the employment of the mother, breastfeeding and child care may be jeopardized and the alternatives may be inferior nutritionally, micro-biologically, and emotionally, so that the child's nutrition status will suffer. In other words, malnutrition is rather sticky: while it will ultimately improve with increased income, this may take a long time if family feeding and caring behaviors and the health environment are slower to change than income itself. Another explanation for the possibility of short term divergence between poverty and malnutrition trends has to do with the measurement of malnutrition. Malnutrition (i.e., stunting) is measured as height for age less than two standard deviations below the internationally accepted reference mean. This is generally agreed to be the most robust and internationally comparable measure of general malnutrition among young children. In Honduras, as in most developing countries, this stunting occurs during the first three years of life, after which it stabilizes at a high prevalence. Measuring stunting among children under three years of age is thus more sensitive to changes in income than measuring stunting among children under five years of age. Because in Honduras the data are gathered for children under 5, there is a built-in inertia in the indicator which may make it less rapidly responsive to changes in poverty. Designing community-based nutrition programs: Atenci6n Integral a la Nihez (AIN) Malnutrition may be the underlying cause of over half of the deaths in children under five in developing countries (Pelletier, 1994). Well-nourished children rarely die of diarrhea or pneumonia. By contrast, poor growth is as serious a 'danger sign' for risk of death as vomiting or rapid breathing. To deal with malnutrition, Governments have implemented a wide range of programs. Atencion Integral a la Niniez (AIN), the preventive health program for children started in 1990 in Honduras, has achieved encouraging results. The program focuses on children under two, because this is the age when most of the process of malnutrition takes place. AIN started out as a clinic-based program but has evolved into a community-based health and nutrition promotion program, Through a learning and adaptation process, the Government gradually perfected the program. In AIN as implemented today, all children under two are weighed and counseled monthly. The attention is targeted on the children and families that need the most help, as indicated by the fact that the 65 children fail to gain adequate weight in the previous month. AIN is fundamentally a client-driven program. A great deal of effort has been spent understanding the perceptions and problems of poor families with respect to the health and nutrition of their children. Extensive qualitative research and home trials were carried out to make sure that the program messages were adapted to the constraints and motivations of the target population. This research was then turned into practical tools for the community volunteers, such as counseling cards, a simple information system, practical training, and a supportive supervision system. The program is based on a team approach to community health. In every community, at least three volunteers are selected by the community to implement the program. This has created "esprit de corps" allowed efficient division of labor, and reduced the costs of turnover. Three volunteers are needed for every 25 children under 2 and the job is designed to accommodate their available time. Each volunteer or "monitora" is given a manual which consists of step-by-step instructions on each aspect of the job, with monitoring indicators, tables of weight gain, examples of the use of counseling cards, and a guide for action. The main task of the monitoras is negotiating behavioral change with the mothers. They are assisted in this task by the counseling cards which identify key messages based on the child's age, health, growth performance, and lactation status. In addition, the monitoras can treat minor illnesses, distribute micronutrients, and refer sick children to the health clinics. When a child fails to gain weight, the monitoras visit the family and try to trouble-shoot in order to find feasible solutions to the problems presented. The program has a simple and practical information system based on five indicators: the total number of children under 2, the number attending the monthly weighing session, the number gaining adequate weight, the number failing to gain adequate weight, and the number failing to gain adequate weight for two or more consecutive months. This information system is used for monitoring from the community level up to the national level, especially to target supervision and resolve operational problems. The training is based on the manual: it is job based, hands-on, and practical, which is much better than lectures. Finally, the program attacks the most difficult problems first. In each health clinic area, two of the poorest communities are chosen to be the first to receive the program. This allows for more intensive supervision of these poorer communities and greater opportunities for learning. By mid 1999, an evaluation of the new AIN was carried out, based on baseline data in AIN-C and control communities. The results were impressive: malnutrition had been reduced to less than 10% in most communities. The qualitative assessments were also positive. The program now working in 9 health areas, was scheduled to be extended to up to 20 health areas in 2000. Source: Villalobos, McGuire, and Rosenmoller (2000) 4.20. Except for child malnutrition, ]Honduras has made substantial progress in health indicators. Twenty years ago, Honduras' performance in health indicators was below the Latin American average. Today, despite a lower level of GDP, the country is on par. Table 4.15 indicates that the vaccination rate for children under five years of age has increased from 49.6 percent in 1980 to 94.2 percent in 1997. This rate is now above the level reached in countries with a high level of human development as measured by the UNDP's Human Development Index (HDI). The percentage of births taking place in hospitals has increased from 37.5 percent in 1980 to 53.8 percent in 1997. The population without access to safe water has decreased from 41.4 percent in 1985 to 25.1 percent in 1997. Finally, the population without sanitary access has decreased from 42.3 percent in 1985 to 26.2 percent in 1997. In all these indicators, rural areas trail behii.d urban areas, but progress has been achieved in both and can be expected to continue with the trend toward economic development and urbanization. As mentioned, one indicator where there is no progress is child malnutrition. A strengthening of targeted programs with a nutrition component may be needed in order to reduce malnutrition (see Box 4.3 for a discussion of child malnutrition). 66 Table 4.15: Trend in selected health indicators in Honduras, 1980-1997 1980 1985 1989 1990 1993 1997 High HDI countries Vaccination rate among 0-4 years old 49.6 69.0 79.0 81.8 96.2 94.2 81.0 Malnutrition among 0-4 years old N.D. N.D. 39.1 39.4 N.D. 38.5 14.0 Percentage of births in hospital 37.5 37.5 N.D. 37 N.D. 53.8 85 Population without access to water N.D. 41.4 38.0 34.0 30.0 25.1 18 Population without sanitary access N.D. 42.3 41.0 38.0 34.0 26.2 22 Source: UNAT (2000). 4.21. The progress achieved by Honduras in health indicators is in part the result of a large expansion in the public supply of health services, but this has led to over-capacity in some areas. According to a recent report by the World Bank (1998), the number of area hospitals increased by 129 percent between 1990 and 1996. The number of health clinics with doctors (Centro de Salud con Medico, hereafter CESAMO) increased by 21 percent. And the number of rural health clinics (Centro de Salud Rural, hereafter CESAR) increased by 41 percent. At the same time, the total number of consultations grew by 24 percent only, and the number of consultations per inhabitant increased only by 5 percent. The sharp increase in the supply of health services has lead to over-capacity in some areas. In many hospitals and CESAMOs, the doctors and nurses are contracted for morning shifts only (there is no or less activity during the afternoon). As shown in table 4.16, a typical CESAR takes care of only six patients a day. Table 4.16: Daily avera ge of ambulatory care, by attention level, 1990-97 1990 1993 1994 1997 Hospital National 303 291 309 278 Hospital Regional 199 183 192 224 Hospital Area 121 90 100 109 CESAR 6 5 5 6 CESAMO 27 25 26 26 Source: UNAT (2000). 4.22. While the poor use public health posts (CESAMOs and CESARs) more than the non-poor, many among the poor apparently still do not seek health care. Table 4.17 suggests that for children under five years of age, the incidence of illnesses is not higher among the poor than among the non-poor (if anything, the non-poor report more illnesses than the poor). When a child is sick, the poor tend to use CESAMOs and CESARs more than the non-poor who rely more on public and private clinics and hospitals. But many of the poor still stay at home or with family and friends rather than seeking care when they are sick. The same observation can be made for pregnant women and other household members. The lack of demand for health care by some of the poor is not necessarily due to its cost. Many types of consultations are free and when the consultation is not free, the payment can be waived. In table 4.17, the health expenditures for the poor when sick tend to be low (but they are not zero, and of course for the poorest, even a small cost can be large). Independently of cost issues, there is probably a lack of information among the very poor on the benefits of seeking professional care. Giving better information might raise the demand of the poor for health care, with no major difficulty for the health posts to meet the increased demand given their currently low productivity and below capacity usage. 67 Table 4.17: Health care for children Un der five years of age by location and income group, 1999 Municipal center by income group Surrounding areas by income group Y<200 200-500 Y>500 Y<200 200-500 Y>500 No illness 23.58 20.54 18.59 23.81 16.34 13.15 Diarrhea 34.96 36.01 39.20 35.12 41.22 43.66 Other illnesses (flu, respiratory, etc.) 70.73 72.32 74.87 70.54 78.78 81.22 Health expenditures if sick 18.10 27.12 147.77 7.78 34.53 147.51 Place where care is requested Use of public clinic or hospital 3.19 2.25 8.33 3.13 3.65 5.95 Use of private clinic or hospital 1.06 4.87 17.90 0.39 3.28 16.76 Use of CESAMO 30.85 37.08 30.56 17.19 18.43 11.35 Use of CESAR 25.53 27.34 18.21 27.34 25.18 28.11 Use of private consultation 2.13 1.50 6.48 1.17 1.28 4.32 Stay at home or with family/friends 41.49 32.96 24.07 51.95 54.93 35.68 Person from whom care is requested Visit a nurse when sick 31.91 38.58 26.54 32.42 29.20 31.35 Visit a doctor or dentist when sick 31.91 34.83 53.70 17.19 22.99 35.14 Count on parents and family when sick 41.49 33.33 23.15 51.17 53.47 34.05 Source: World Bank staff using 1999 PRAF survey. Y is the per capita expenditures of the household (Lps/month). 4.23. The management of public providers of health services could also be improved. In public health institutions, as noted in the World Bank (1998) report, the selection of the staff for key managerial positions is based on medical backgroumd. It does not take into account management skills. Another problem is that the Ministry of Health allocates its budget on the basis on past allocations rather than future needs. This system does not provide incentives for improved performance. There are also problems with budgetary execution due to the lack of flexibility for the reallocation of expenditures between sectors and health centers, and the fact that some budgets are frozen and never become fully available. The purchasing procedures are centralized, and the payments may take several months to be made. The institutional framework has also contributed to the concentration of doctors and nurses in cities. Indeed, because there is no financial incentive for doctors and nurses to work in rural areas and small cities, many choose to stay in large cities where it is easier for them to supplement their incomes with a second job. 4.24. Despite low coverage, the social security system is not financially sustainable. As indicated in table 4.18, the Ministry of Public Health (Ministerio de Salud Publica, MSP hereafter) is by far the main provider of health services in both urban and rural areas. The Honduran Social Security Institute (Instituto Hondureno de Seguridad Social, IHSS) is not active in rural areas, and its market share is only 16 percent in urban areas. Private proviiers are three times as large as the IHSS nationally. According to the World Bank report (1998), the IHSS claims a total coverage of 1.24 million (22 percent of the population). Half a million are direct beneficiaries and the rest are their qualifying dependents. But the NHES survey suggests that less than 10 percent of the population are effectively covered by IHSS health insurance. Most beneficiaries are concentrated in Tegucigalpa and San Pedro Sula. Social security is financed through contributions that are insufficient because the contribution rate (7.5 percent) is applied to an income ceiling (600 Lempiras per month) which has been frozen for 30 years. This ceiling was originally high, but today it is much lower than the minimum wage. As a result, the social security system is not financially sustainable. Table 4.18: Current structure of health care services providers and ex enditures, percentages Providers National Urban Rural Expenditures 90-93 94-97 90-97 MSP (Health Secretary) 61.7 47.3 81.3 Primary 50.3 53.0 51.6 IHSS (Social Security) 10.0 16.3 0.0 Secondary 23.8 23.5 23.6 Private Sector 28.3 36.3 18.7 Tertiary 33.0 23.5 24.8 Source: UNAT (2000). 68 4.25. Apart from reforming its social security system, Honduras will need to implement higher cost recovery for tertiary health care given to the non-poor. Today, the co-payments by households for public health care generate only 1.5 percent to 3.5 percent of the total MSP budget (World Bank, 1998). The standard fee for an ambulatory consultation (including medicines) is one to two Lempiras. Many services are exempt of any fee (e.g., pre-natal consultations, child development clinics, family planning, prevention and cures for sexually transmitted diseases and tuberculosis). The prices for basic services are very low compared to their private sector equivalents, and they are waived for patients who cannot pay. As a result, according to the NHES survey, the cases of sick people discouraged from seeking attention by MSP charges are rare. This is all fine for the poor, but the problem with the current lack of cost recovery is that there are insufficient checks on the consumption of subsidized and costly tertiary treatments by the non-poor. Because of a lack of price signals, the system itself does not have incentives for health care providers to expand the most valuable services. As is the case for tertiary education, better cost-recovery among the non-poor for tertiary health care is needed in order to reduce the cost and improve the quality of public health care. Box 4.4: PROGRESA: A GENDER-CONSCIOUS PROGRAM FOR EDUCATION, HEALTH AND NUTRITION The new social program PROGRESA in Mexico provides means-tested conditional transfers to encourage investment by the poor in their human capital. The program was introduced in 1997 in response to the rising poverty after the 1995 macroeconomic crisis which affected Mexico. It has become the largest poverty alleviation tool of the Government, and it is reaching today 2.6 million rural households. The program is geared towards improving high-school enrollment and attendance, especially among girls. It is also trying to decrease pre-schoolers' and pregnant and/or lactating mothers' malnutrition, and to provide incentives for family preventive health care. The program seeks to integrate these objectives so that children's learning is not affected by poor health, malnutrition, or necessity to work, and parental ability to pay for increased nutrition and education is not a constraint on children's development. The main components of the program consist of: (a) Educational grants to foster enrollment and regular school attendance; continued receipt of these grants is conditional on individual child attendance reports by school teachers; (b) Basic health care for all household members, with a strengthening of preventive medicine through health sessions; attendance to the sessions is required to receive full payment of food monetary transfers; and (c) Monetary transfers and food supplements to improve family's food intake, particularly of children and women, but also of older individuals (who benefit from a substantial share of the financial transfers, a fact that is often overlooked when discussing the program). Food supplements are given for malnourished children and pregnant and lactating mothers. The program follows a two-step targeting procedure. The first step consists in a geographical targeting of marginal communities (a "marginality" index is built from census and health/education ministries data, but communities without have access to basic health and primary education infrastructure cannot participate). In eligible communities, a survey questionnaire is applied to all households in order to determine socio-economic status. A principal component analysis is used to classify households as "poor" (eligible) or "non-poor". A listing of eligible households is then presented to the community, which has an opportunity to adjust it for exclusion or inclusion of households. Eligible households can decide to take-up the program and eligibility cards are then supplied to mothers when the household is eligible to receive all three benefits, or to the household head when the household includes no woman or is only eligible for food transfers. Registration takes place during a community assembly. In 1999, at the time of the program evaluation, PROGRESA's budget was US$ 777 million (0.2 percent of Mexico's GDP). Administrative costs represent 8.9 percent of total costs (including 2.67 percent for targeting costs at the household-level and 2.31 percent for conditioning costs). How effective is the program in contributing to development targets? Apart from its immediate impact on poverty through the 69 cash transfers given to households, PROGRESA has been found to reduce child mortality by 12 percent. It has also been found to increase the number of years of schooling of the children. Because enrollment in primary school is already high in Mexico, the increase for years of primary school was relatively low, at 76 years of schooling for a cohort of 1,000 girls, and 57 years for a cohort of 12,000 boys. The increase in years of secondary school was much larger, at 479 hours for girls and 249 hours for boys. The cost of generating an extra year of schooling was found to be around US$ 5,550 for primary education and US$ 1,000 for secondary education. Several features of PROGRESA have a gender focus. First, PROGRESA targets women as beneficiaries to address family needs. The mechanisms PROGRESA uses to deliver its resources may be one of the most innovative features of the program. The program's main focus is on women, as the "key to household food security" and health. This anti-poverty strategy recognizes that mothers effectively and efficiently use resources to address the most immediate needs of their families, especially of the children. As it delivers the benefits mostly to women, PROGRESA has the potential of changing the intra- household decision-making processes, at least on children's related outcomes. These questions were examined both through quantitative analysis of three rounds of survey data about decision-making process and expenditure shares dynamics, and through focus groups discussions of these issues in 1999. Being a beneficiary of PROGRESA decreases the probability that the husband takes decisions alone in five of the eight decision-making categories. Over time, men have been less likely to take decisions alone, especially when they affect children, and women have been more likely to decide by themselves on the use of their extra income. The qualitative results show that by giving money to women, the state has forced recognition among men and in the communities as a whole of the contribution and role of women in caring for the family. Most men do not have problems with their wives participating in PROGRESA since they see the benefits extending to the whole family. In addition, participation in group discussions and tasks is reported to have developed women's awareness, knowledge and confidence and control over their movements. The fact that the government is providing recognition to women is noticed by beneficiaries and non-beneficiaries alike in the selected communities. Women report that they wander out of the house more often, they have more opportunities to share concerns and problems faced by their households, they are more comfortable in speaking in front of others, they are gaining health knowledge and may better gear household expenditures. In general, men do not take the PROGRESA income from their wives and continue giving the same amount of money for household expenditures: the extra income is used for needs that could not be addressed before and has relieved some of the stress of making each day's expenditures. Second, PROGRESA is focusing on girls' education. The economic returns to secondary education are relatively large and provide children with opportunities to escape poverty as adults. While primary school enrollment is relatively high in rural Mexico, around 93 percent, for poor children, access to secondary school is a major hurdle and the enrollment rate declines to 55 percent after children complete primary school. Girls tend to enroll less than boys in secondary school and drop out earlier. In order to reverse this tendency, the grants' amount increases faster for girls than for boys in high school. The evaluation of the impacts of PROGRESA grants used a combination of statistical methods to control for community and family-level effects and different samples of children. In primary school, where enrollment rates reached 90 to 94 percent, the program increases girls' enrollment by 0.96 to 1.45 percentage point and boys' by 0.74 to 1.07 point. In secondary school, as original rates were 67 and 73 percent for girls and boys respectively, the proportional increase have been 11 to 14 percent for girls and 5 to 8 percent for boys. PROGRESA helps reducing the dip between primary and junior secondary schooling, as it boosts enrollmei1t rates among those who have completed sixth grade by 14.8 percentage points for girls and 6.5 points for boys. An important group of girls are therefore extending their schooling past primary school. One of the main results is to equalize the chances for school attainment of girls and boys. 70 One of the premises of PROGRESA is that better education for girls can improve their future status in their households and in the labor market, their living standards, and participation in the society at large. The qualitative evaluation showed that women themselves support this assumption, despite the fact that many of them are actually not participating in the formal labor market. Women are convinced that (1) education will help girls to find better paid and less exploitative jobs, which will enable them to better withstand failure of their marriages and possible single motherhood, (2) education helps girls to have a better life in general, delay their marriage and improve their standing in their families, (3) education helps girls and women to better defend themselves vis a vis men and in public, and (4) education makes women build their self-esteem. Mothers were more confident about the future positive effects of PROGRESA for their daughters than about the ones the benefits grant them at the present. Most empowerment effects of PROGRESA might therefore come in the long-term through education rather than in the short-term through income. While women did not report that PROGRESA has modified men's attitude lowards girls' education, it seems that the program has been successful in counteracting biases since girls' enrollment has increased. The defacto presence of girls in schools will likely raise awareness about girls' education and may change the norrns, but employment opportunities for these young women will have to increase for education to be valued. Most women explain that the cash transfers are higher for girls because girls have higher expenses than boys for clothing and cosmetics. Even if the incentives work, there might be some value in educating "promotoras" and, in turn, beneficiaries about the ideas behind the program. Some "promotoras" have understood these ideas and are successful at generating discussions among beneficiaries about the value of girls' education. Finally, PROGRESA is focusing on health care for pregnant and lactating mothers. As seen above, among the basic health services promoted by PROGRESA are pre-natal care, infant delivery and baby care, family planning, nutrition and growth monitoring of infants as well as detection and control of cervical cancer. In 1998, survey results indicated that 44 percent of 12-36 month old children were stunted (low height for age, a major form of Protein Energy Malnutrition), a result of early infancy and in utero malnutrition, which has potential long-term impacts on developmental outcomes and income generation. Pre-natal care visits have increased by 8 percent in the first trimester of pregnancy, which in turn decreased the percentage of first visits in the second and third trimester of pregnancy. This behavioral change is documented to have a significant effect on the health of babies and pregnant mothers. While it was too early to detect any fertility behavior changes among the beneficiaries of PROGRESA, better women's education, care of pregnant women and health of infants are likely to yield changes in birth spacing and reproductive health decision-making in the medium to long-run. Source: Based on various publications by PROGRESA, including PROGRESA (2000). 71 CHAPTER V: IMPACT OF GROWTH A. INSTITUTIONAL REFORMS ARE NEEDED TO PROMOTE FASTER ECONOMIC GROWTH 5.1. This chapter suggests avenues for faster growth in Honduras, gives estimates of the impact of growth on poverty and other indicators, and it proposes future targets for these indicators. The first section of the chapter briefly considers some policies and reforms which could be implemented in Honduras to promote faster economic growth. We do not provide new results in this section: we simply summarize what we believe to be the conventional wisdom reached by observers of the country's policies and institutions. The second section gives estimates of the (positive) impact of economic growth on both monetary and non-monetary indicators of well-being. Finally, the third section uses these estimates to suggest future targets for these indicators. This is done in some detail because establishing targets for poverty reduction and other indicators is one of the mandates of the PRSP to be prepared by the GRH. 5.2. Over the last eighteen months, the GRH has focused its attention on the reconstruction effort needed after the damage caused by Hurricane Mitch. Since an evaluation of the reconstruction effort has been prepared by the GRH (2000b), there is no need to give details here. Although the disbursement of international aid has been a bit slow, with only a third of the commitments made available in 1999 (US$ 492 million), and although much remains to be done, the reconstruction effort has paid off in many direct and measurable ways. But the reconstruction effort may also have been beneficial for poverty reduction in a number of indirect ways. First, the mobilization for the reconstruction, including the dialogue between the GRH and civil society, should hopefully carry on and thereby facilitate the consultations needed for the PRSP. Second, programs such as the social investment fund have been instrumental in the reconstruction. These programs have been reinforced, which may be good for the poor. Third, the country has been made conscious of the need to reduce its vulnerability (especially among the poor) to natural disasters. 5.3. Beyond the current focus on reconstruction, Honduras will need to implement important reforms in order to improve growth and make significant progress toward poverty reduction. In its 1998 report on human development in Honduras, UNDP (1998) presents interesting results from an opinion survey of the elites on how to promote development and poverty reduction. While the survey is small (220 interviews), it does provide a snapshot of the mindset of the country's ruling classes. Table 5.1 gives the opinion of the respondents regarding the causes and cures for poverty. According to the elites, poverty results more from inadequate growth and employment opportunities than from an inadequate distribution of the available resources in the country. The bad functioning of the economy and the lack of investments, together with the lack of efficient policies and (for some) the impact of structural adjustment, are considered as the main factors responsible for the increase in poverty in Honduras. The distribution of income is less of an issue. The solutions proposed by the elites to solve the problem of poverty mirror the causes put forward as being at the root of the problem. Apart from the reform of the state and its development model, higher ,and better education, employment, productivity, and investments are seen as the keys to improve wages. The fact that economic growth is seen as the key for poverty reduction should not be surprising, since the elites are unlikely to favor redistributive policies. Still, while poverty depends on both the available pool of income in an economy and on the distribution of this pool in the population, it should be emphasized that in a country as poor as Honduras, growth is more important than redistribution (even if redistribution can help for growth, see paragraph 1.7). This point has been made in many studies, including a report by IPEA (1999), and in the I-PRSP of the GRH (2000a). 72 Table 5.1: What are the causes of and cures for poverty? (Opinion survey, percentages) Causes of poverty Structural adjustment 15 Unjust distribution of income 6 Lack of efficient policies 8 Population growth 10 Subtotal: Policy factors 23 Low productivity 6 Lack of employment and investments 10 Rural to urban migration 2 Bad fiuctioning of the economy 6 Subtotal: Structural factors 24 Cost of living and low wages 19 Poverty more visible 5 Subtotal: Business factors 35 Other factors 14 Solutions _ Income redistribution 7 Increase in employment 16 Reform of the State 9 Investment and productivity 20 Change in development model 6 Increase in investments 3 Other development actions 13 Sub total economic growth 39 Sub total development model 35 Education 20 Other actions 6 Source: Opinion survey of the elites by UNDP (1998). 5.4. A first condition for faster growth and thereby poverty reduction is the implementation of sound macroeconomic policies. In the first half of the 1990s, Honduras suffered from macroeconomic shocks. But over the last few years, the country has made significant progress in the management of the economy. Macroeconomic stability has also been maintained after Hurricane Mitch. The report written in February 2000 for the Consultative Group meeting in Tegucigalpa (GRH, 2000b) suggests that macroeconomic indicators have been better than anticipated. The GDP reduction in 1999 has been only 2 percent (instead of an anticipated 3 percent). Inflation has reached 10 to 12 percent, which is below the level observed in 1998. The depreciation of the Lempira versus the US dollar has been at less than 5 percent, which is similar to 1998, and the budget deficit will be around 5 percent of GDP. 5.5. A second condition for faster growth is an improvement in Honduras's competitiveness and in the quality of institutions. Table 5.2 provides indicators of competitiveness, microeconomic freedom, and corruption in Honduras and a comparison with its neighbors. The indices are quoted in FIAS (2000) and were proposed by INCAE-HIID, the Heritage Foundation, and Transparency International. * General competitiveness: The first indicator, provided by INCAE (Instituto Centro Americano de Administracion de Empresas) and HIID (Harvard Institute for International Development) measures general competitiveness. Honduras has the lowest average index (30) of the five countries presented. The country lags behind for trade openness, technology development, finance, business and marketing, infrastructure, and institutional quality. It does better for labor markets (in part because low wages constitute a comparative advantage, but at the same time there are important rigidities) and governance. * Economic freedom: The second indicator, provided by the Heritage Foundation, attempts to measure economic freedom. Honduras is fourth out of five. In comparison with its neighbors, Honduras does well on monetary policy, but poorly on foreign investment, regulation, taxation, trade policy, and black markets. Extensive business regulations reduce foreign direct investments. Establishing a new company can take three to four months, while in some neighboring countries this can be done in one to two weeks. The fees for establishing companies are high, even for a simple process. The review of the documents needed for incorporation is inefficient. The Honduran Commercial Code does not provide a procedure for recording liens on personal property. Thus, financial institutions cannot grant loans on cars or equipment because they cannot record a security interest in the property. For land property, title registration is slow and the offices in charge of the program are not organized well enough. In rural areas, only about half of the farmers have a legal title to their property. The Government enacted in 1992 an investment law to promote both internal and international 73 investment. With a few exceptions., investors are free to invest and the law provides equal treatment of both locals and foreigners. Moreover, in 1993, Honduras adopted a modem legal structure for the protection of intellectual property (patents, trademarks and copyrights). Nevertheless, copyright protection remains uncertain. And other aspects of corporate law do not give foreign investors an incentive to incorporate in Honduras. * Microeconomic competitiveness: The third indicator, again provided by INCAE-HIID, measures the microeconomic competitiveness of the country within a sample of 59 countries with medium to high competitiveness. Honduras is at the bottom of the group, with an overall rank of 58. The country is last among the five countries presented for each of the indices used to measure overall microeconomic competitiveness. One of many obstacles to the competitiveness of Honduras is the lack of flexibility for labor contracts. As in many other Latin American countries, formal employees are over-protected since it is difficult for an employer to fire a worker without having to pay large compensation costs in terms of severance payments. This makes employers reluctant to hire permanently, and it creates a pattemn of reliance on temporary jobs with high turnover that is not conducive to investments in worker productivity. Four respondents out of five in the UNDP survey believe that the country and its entrepreneurs are not ready to face globalization. * Corruption: The fourth indicator, provided by Transparency International, measures the perceived level of corruption in Honduras. Again, Honduras has the lowest rating in the five countries considered. The country's elites are aware of the problem. According to the UNDP survey, fighting corruption and reforming the state come first in terms of priorities for development, before access to education, competitiveness, and income generation. Making regulatory processes more automatic and less discretionary would help in reducing corruption. Reforming the customs and tariffs administration is also essential. Finally, a better legal system would help. There is apparently no single publication that contains all Honduran legislation classified by subject matter and that is periodically updated. The courts are slow and sometimes corrupt. Contracts tend not to be enforced in court, and Honduras does not have a dispute settlement system (arbitrator) that can circumvent the slow legal process. The members of the supreme court are elected every four years following the general election, instead of being appointed for longer periods of time on the basis of their qualification, as is done in other countries. The training received in law schools is also deficient. 5.6. What is lacking in Honduras is perhaps less an identification of the steps that must be taken, than an ability to implement these steps. The indices presented in table 5.2 are certainly debatable, and it was to be expected that Honduras would not compare well with Costa Rica, El Salvador, and Guatemala which are richer countries. But Honduras does not seem to do better than Nicaragua either. The broad message that emerges is clear: Honduras needs to improve its competitiveness by reforming its institutions and implementing. The draft study by the Foreign Investment Advisory Service (FIAS, 2000) of the World Bank suggests that there are no serious impediments to higher levels of foreign direct investment in Honduras. There are no serious problems with the foreign exchange, the tax system, the government interfering with industry, but the application of the rule of law could be improved. Yet while foreign direct investment has been flowing at about 2 percent of GDP in the last 2 years (before Mitch), this could increase. Why does Honduras have limited appeal to foreign direct investment? Apart from the fact that the country is small, the lack of foreign direct investment is the result of a combination of factors such as those mentioned above. These constraints could be fixed. According to the FIAS study, between 1991 and 1999, while the tone of the businessmen has become more insistent on the need to implement reforms, there has a lower expectation that things are going to get better in the near future. 74 Table 5.2: Selected indicators of competitiveness for Honduras and Central American countries Costa Rica El Salvador Guatemala Honduras Nicaragua INCAE-HIID General competitiveness (O - worst; 100 - best) Overall index 52 XX 37 30 36 Trade openness 76.73 56.36 62.55 45.82 63.64 Labor market 86.26 33.65 39.81 63.98 42.65 Technology development 48.18 0.0 12.04 5.47 15.69 Finance 33.05 46.44 25.92 22.18 25.94 Govemance 40.08 63.89 66.27 55.56 55.16 Business and marketing 54.76 22.22 31.22 11.11 32.80 Infrastructure 33.96 38.43 43.28 22.39 23.51 Institutions quality 42.91 31.76 15.54 15.20 30.41 Heritagefoundation economicfreedom (5 - worst; I - best) Overall index 2.8 2.25 2.75 3.15 3.5 Foreign investment 2 1 3 3 2 Regulation 3 3 4 4 4 Taxation 2 2.5 2.5 3.5 3 Trade policy 4 3 3 4 4 Government intervention 2 1 1 2 2 Wages and prices 2 2 3 3 3 Black market 3 3 3 4 5 Banking 3 2 2 3 3 Property rights 3 2 3 3 4 Monetary policy 3 3 3 2 5 INCAE-HIID Microeconomic Competitiveness (59 countries) Overall index 36 52 51 58 55 Business operation and strategy 37 44 45 54 47 Business climate 38 47 47 54 48 Supply conditions 37 49 45 51 45 Demand conditions 31 45 48 53 48 Related industries 45 49 48 58 49 Structure, strategy and rivalry 37 42 46 52 50 among finns Corruption Perception Index s (O- worst; 100 - best) Corruption index 5.1 3.9 3.2 1.8 3.1 Rank among 99 countries 32 49 68 94 70 Source: INCAE-HIID, Heritage Foundation, and Transparency International, as quoted by FIAS (2000). 5.7. The fact that in Honduras economic growth is important for poverty reduction does not mean that growth should be promoted independently of redistribution. Two arguments can be made for the role of redistribution in relationship with growth. First, higher initial inequality may result in lower subsequent growth, and thereby in lower poverty reduction over time. This is in part because under high inequality, access to credit and other resources is concentrated in the hands of the privileged, thereby preventing the poor from investing or protecting themselves from shocks. Second, higher levels of inequality reduce the benefits from growth for the poor. This is because a higher initial inequality reduces the share of the gains from growth that goes to the poor. At the extreme, if a single person has all the resources, then whatever the rate of growth, poverty will never be reduced through growth. In other words, a high level of inequality may reduce (in absolute terms) the elasticity of poverty reduction to growth. These two arguments suggest that instead of hampering growth, well designed redistributive policies may promote growth and increase the benefits from growth for the poor. 75 5.8. Moreover, while the growth of some sectors will benefit Honduras greatly, there is always a risk that the poorest will not be able to participate in the new opportunities being created. This risk can be illustrated with the example of shlimp production employment patterns. Under decree 968 of 1980 and agreement 229 of 1991, the Ministry of Tourism (SECTUR) and the General Directorate of Fishing (DIGPESCA) transferred coastal land use rights to exporters. By 1993, more than 31,000 hectares in Southern Honduras had passed from communal gathering use to export development, and 80 farms operated 11,500 hectares of ponds. The most commonly cited figure in terms of employment generation is that 12,000 families live from the Honduran shrimp industry. Much of the work involves the gathering of larva for shrimp seed to stock the ponds. Stanley (1999) argues that a gatherer's product is influenced by his/her effort and by the total larvae available at any point in time, which is itself a function of stochastic variables such as water temperature, salinity, and the lunar cycle. Although Stanley does not present his findings in relationship with poverty, the findings do have a bearing. The key issue is the "moral hazard" faced by the owner (the principal), and the method used to solve the problem. While the owner may want to reward effort, he does not know if the performance of its workers (the agents) is due to their effort or external factors. Stanley suggests that this moral hazard problem is solved by the principal using two different employment systems: tournaments or "piece rates". The tournament system is the better system for the owners. Under tournament rules, the contractors choose their permanent gatherers carefully and job retention is a function of both effort and luck. The gatherers are ranked according to their performance and the winners are retained for the next period and paid a salary with a bonus. The losers are fired. The tournament players remaining for several months with a labor contractor enter into a labor-consumption contract with additional benefits (like those offered in the formal sector). The bottom line is that mariculture firms are willing to pay higher wages for the best employees in order to reduce turnover costs, and this naturally leads to a segmentation of the labor market for shrimp production. Only the most qualified get permanent jobs at high pay. The less skilled are relegated to temporary contracts at best. In 1993, shrimp farms were legally required to pay permanent laborers between 17.50 and 21.00 Lempiras daily. But this did not help the poor because they did not have permanent contracts. And it did not matter for the better skilled gatherers because they were already paid higher amounts in recognition of their skills. There is nothing surprising in this story. The very poor who lack skills are often relegated to second level jobs. The functioning of the shrimp production market gives but one concrete illustration of the mechanisms at work that may prevent the poor from benefiting as much as the non-poor from the growth of higher value added sectors in an economy. 5.9. Naturally, reforming the state does not mean getting rid of the state. Even if there is a consensus on the need to reform the state in Honduras, the elites still believe that the state is responsible for solving problems in the areas of security, education, health-nutrition-employment, and corruption. As a matter of fact, while the share of total government expenditures devoted to social expenditures has fluctuated between 30 percent and 37 percent in the 1990s (table 5.3), this level has remained below what many other Latin American countries devoted to the social sectors (in table 5.4, as a share of total public expenditures, only Jamaica and Nicaragua have lower levels of spending for the social sectors than Honduras). There may be some room fcr increasing selected social expenditures in Honduras in future years if the country participates in the HIPC initiative for debt relief (see chapter 4 for priorities). 76 Table 5.3: Trend in social expenditures as a share of total public expenditures, 1990-98 Total Social Education Health Social Housing Others* Expenditure Security 1990 29.7 16.2 10.4 0.2 0.9 2.0 1991 37.0 20.4 11.6 0.3 0.3 4.5 1992 31.0 17.0 9.8 0.2 0.6 3.4 1993 32.7 18.3 10.1 0.2 0.7 3.4 1994 31.0 16.2 10.3 0.2 0.8 3.6 1995 36.3 19.0 13.7 0.1 0.7 2.7 1996 34.3 18.2 11.9 0.1 0.0 4.1 1997 31.9 18.7 9.2 0.1 0.0 3.9 1998 32.0 17.6 8.6 0.1 0.6 5.0 Source: UNAT (2000). (*) Includes Ministry of Labor, JNBS, INJUPEMP, FHIS, PRAF, and Subsidies Table 5.4: Social expenditures in Honduras and other countries as percentage of total expenditures Belize 39.1 Costa Rica 61.9 Bolivia 37.2 Dominican Republic 36.6 Brazil* 46.2 Honduras 34.0 Chile 61.8 Jamaica 20.1 Colombia 39.4 Nicaragua 29.8 Source: UNAT (2000). B. GROWTH IMPROVES BOTH MONETARY AND NON-MONETARY INDICATORS OF WELL-BEING 5.10. Nationally, a one percentage point increase in per capita income (i.e., a growth rate of one percent) reduces the headcount of poverty and extreme poverty by half a point. Elasticities of poverty reduction to growth were estimated using the EHPHM surveys and the method described in annex 2 (section MA.8). Denote by y the gross elasticities of poverty to growth, i.e., the percentage reduction in poverty obtained with a one percent growth rate holding inequality constant. Denote by , the elasticity of inequality to growth, i.e., the percentage change - this can be a reduction or an increase - in inequality obtained with a one percent growth rate. Finally, denote by o the elasticity of poverty to inequality controlling for growth, i.e., the percentage increase in poverty resulting from a one percent increase in inequality holding growth constant. The net elasticity of poverty to growth, i.e., the percentage decrease in poverty obtained from a one percent growth rate while allowing inequality to change, is X2 y + PS. Table 5.5 provides the elasticities for the headcount index, poverty gap, and squared poverty gap. * Gross impact of growth on poverty: Without changes in inequality (as measured by the Gini index), a one percent increase in per capita income results at the national level in a -0.68 percent (y) decline in the headcount index of poverty. With a national headcount for poverty at, say, 60 percent, this would represent two fifths of a percentage point decline in the headcount (60*-0.68/100=-0.41). For extreme poverty, the elasticity is larger (-0.99), but the initial level is lower at, say, 40 percent. Thus, one percentage point in growth would reduce extreme poverty by 0.40 percentage points as well (40*- 0.99/100=0.40). If the headcounts of poverty and extreme poverty are believed to be higher, as in the I-PRSP prepared by the GRH, the reduction in (extreme) poverty will be about half a point. * Impact of growth on ineguality: Nationally, growth does not result in a higher level of inequality, since the elasticity of inequality to growth is very low (0.03) and not statistically significant. * Impact of inequality on poverty: The elasticity of poverty to inequality (8) is relatively large, and larger for the poverty gap and squared poverty gap than for the headcount index since these measures are sensitive to the inequality among the poor. Yet because the elasticity of inequality to growth is basically zero, this has no bearing on the impact of growth on inequality. [The absence of a 77 systematic correlation between growth and inequality nationally does not detract from the fact that inequality rose in Honduras betweeni 1991 and 1999, thereby contributing to high levels of poverty.] * Net impact of growth on povertv: The net impact (X) of growth on poverty is virtually equal to the gross impact, once again because of the lack of a correlation between inequality and growth. Table 5.5: Elasticity of Poverty Reduction to Growth in Honduras: National, Urban, and Rural Headcount Poverty gap Squared poverty gap Extreme Moderate Extreme Moderate Extreme Moderate National Gross elasticity of poverty to growth y -0.99 -0.68 -0.94 -0.86 -0.73 -0.87 Elasticity of poverty to inequality o 2.19 0.82 3.00 1.92 3.26 2.53 Elasticity of inequality to growth D NS NS NS NS NS NS Net elasticity of poverty to growth X =y + PS -0.93 -0.65 -0.86 -0.81 -0.64 -0.80 Urban Gross elasticity of poverty to growth y -1.43 -1.08 -1.07 -1.16 NS -1.07 Elasticity of poverty to inequality 8 3.04 1.55 3.60 2.65 3.77 3.23 Elasticity of inequality to growth P 0.19 0.19 0.19 0.19 0.19 0.19 Net elasticity of poverty to growth X = y + 138 -0.85 -0.78 NS -0.65 NS NS Rural Gross elasticity of poverty to growth y -0.87 -0.51 -0.97 -0.78 -0.92 -0.88 Elasticity of poverty to inequality 8 1.21 0.41 1.86 1.12 2.13 1.54 Elasticity of inequality to growth B NS NS NS NS NS NS Net elasticity of poverty to growth X = y + ,B8 -0.97 -0.55 -1.13 -0.87 -1.10 -1.00 Source: World Bank staff using EPHPM surveys. NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 100/6 level. Coefficients not underlined are significant at the 5% level. 5.11. Although the elasticity of poverty reduction to growth appears to be lower in Honduras than in Latin America as a whole, the imnpact of economic growth on poverty need not be lower in Honduras because the initial level of poverty to which the elasticity is applied is higher. An exercise similar to that reported in table 5.6 (but in panel form) was performed with data over time for twelve Latin American countries in Wodon (2000). According to this study, the net elasticity of the headcount index of poverty to growth in Latin Anmerica is -0.94, which is higher than that observed for Honduras nationally (-0.65) Similarly, the net elasticity of extreme poverty to growth in Latin America is -1.30, which is again higher than in Honduras (-0.93). The same pattern is observed for the poverty gap and the extreme poverty gap. This does not necessarily suggests that in Honduras growth is not broad-based enough in order to have a large impact on poverty, because the elasticity in Honduras must be applied to a higher initial level of poverty. In other words, since poverty is higher in Honduras than in Latin America as a whole, a smaller elasticity (in absolute value) in Honduras may well yield a level of poverty reduction in percentage points similar to that observed in Latin America as a whole. Still, it is worth noting that the reason for the relatively low net elasticity of extreme poverty in Honduras is different in urban and rural areas according to the estimates in table 5.5. In urban areas, the gross inequality is higher, but the positive correlation between inequality and growth reduces the net elasticity. In rural areas, there is no positive and significant correlation between inequality and growth, but the gross elasticity of poverty to growth is lower. Given the high level of poverty in rural areas, and the fact that more than half the population is rural, one priority for the Government should be to implement policies that make rural growth broad-based. 5.12. Apart from reducing poverty, growth also improves non-monetary indicators of well-being. Economic growth has positive impacts on a wide range of non-monetary indicators including infant 78 mortality, under five mortality, child malnutrition, and life expectancy at birth for the health sector; adult illiteracy, net and gross enrollment in primary, secondary, and tertiary education, as well as illiteracy among the adult population for the education sector; and access to safe water, sanitation, and telephones for the basic infrastructure sector. Table 5.6 provides estimates of the elasticities of these indicators to growth computed from a worldwide panel data set. Although two models were estimated, with the estimation done in so-called "levels" or "differences", only the levels model is displayed in table 5.6. These models are discussed in detail in a manual for SimSIP, a set of simulation tools for Social Indicators and Poverty (see Box 5.1 at the end of the chapter). In each model, the elasticities depend on the level of economic development of the country as captured by real per capita GDP in U.S. dollars (PPP - Purchasing Power Parity method, 1985). In the levels model for example, in a country such as Honduras (with a real per capita GDP below $2,500 at PPP 1985 prices), one percentage point in growth is expected to result in a 0.081 percentage (not percentage point) increase in net primary enrollment. While the magnitude of each elasticity depends on the social indicator and level of development of the country, there is no doubt that economic growth is associated with strong non-monetary benefits in terms of education, health, and basic infrastructure. Yet in some cases we observe no or negative impact. For example, with the levels model, the gross primary enrollment tends to decrease with growth, which may suggest improvements in efficiency. In general, however, when the growth elasticity is negative, real per capita GDP is large, implying that the elasticity is used only for highly developed countries. Also, when growth has no impact on an indicator, this can be interpreted as a sign that special targeted programs may be needed to improve the social indicator under review. Interestingly, urbanization also seems to have a larger impact on many social indicators than growth. While the fact that urbanization has a positive impact is not surprising, the magnitude of the impact is. It could be that urbanization is correlated with omitted variables in the regressions which also have positive impacts. Overall, as explained in the manuals for SimSIP, the model presented in table 5.6 should not be given too much weight in terms of causal interpretation, but they can be used to set targets for social indicators within the framework of a PRSP, and this is done below. 79 Table 5.6: Elasticities of Social Indicators to Growth and Urbanization, Levels Health Indicators Infiastructure Indicators Infant Under 5 Life Malnutrition Access to Access to Telephone Mortality Mortality Expectancy Prevalence Safe Sanitation Mainlines (under 1) at Birth (under 5) Water per 100 persons Per capita GDP Y<1 000 -0.040 NS 0.015 NS 0.666 0.769 0.671 1000<=Y<2500 0.053 NS 0.020 -0.480 NS NS 0.538 2500<=Y<5000 -0.371 -0.467 0.008 NS NS NS 0.904 5000<=Y<10000 -0.354 -0.369 -0.023 -1.100 NS -0.414 0.517 10000<=y -0.184 NS -0.009 NS NS NS -0.528 Urbanization U<0.20 0.046 NS 0.054 0.525 2.102 NS NS 0.20<=U<0.40 -0.137 NS 0.087 NS 2.205 1.285 0.396 0.40<=U<0.60 -0.072 NS 0.017 NS 1.097 NS 1.277 0.60<=U<0.80 -0.646 NS 0.029 -3.300 NS NS 1.244 0.80<=U 0.552 NS -0.313 NS NS -2.043 0.432 Time trend (not in log) Uniform .. .. .. -0.013 0.006 0.022 0.044 Africa -0.015 -0.015 0.005 .. Asia -0.031 -0.030 0.007 .. ECA -0.031 -10.025 0.002 .. LAC -0.031 -0.033 0.005 .. MENA -0.039 -0.043 0.009 .. OECD -0.037 -0.038 0.003 .. .._ .. _ .. Education Indicators Net Primary Net Adult Gross Gross Gross Enrollment Secondary Illiteracy Primary Secondary Tertiary Enrollment Enrollment Enrollment Enrollment Per capita GDP Y<1000 0.314 0.550 -0.060 0.065 0.171 0.253 1000<=Y<2500 0.081 0.357 0.045 -0.050 0.475 0.635 2500<=Y<5000 0.023 0.318 -0.059 -0.077 0.128 0.231 5000<=Y<10000 0.042 0.232 -0.115 -0.083 0.171 0.688 10000<=y NS NS -0.105 -0.102 NS 1.281 Urbanization U<0.20 0.500 1.492 0.231 0.452 0.657 1.907 0.20<=U<0.40 0.132 0.279 NS 0.520 0.716 2.054 0.40<=U<0.60 0.060 0.226 -0.319 0.113 0.528 2.734 0.60<=U<0.80 NS 0.493 -0.635 -0.192 0.661 4.135 0.80<=U -0.232 0.680 -0.446 -0.428 NS 7.083 Time trend (not in log) Unifonn .. .. .. 0.004 Africa 0.001 0.013 -0.024 .. 0.035 Asia -0.001 0.003 -0.031 .. 0.011 ECA NS NS -0.039 .. 0.010 LAC 0.003 0.012 -0.028 .. 0.023 MENA 0.013 0.029 -0.025 .. 0.031 OECD 0.000 0.007 -0.040 .. 0.015 Source: Wodon et al. (2001). Note: NS means not statistically different from zero at the 10% level. Coefficients underlined are significant at the 10% level. Coefficients not underlined are significant at the 5% level. The symbol '..' implies that the parameter was not included in the model. 80 5.13. An analysis of the marginal gains in social indicators suggests that the poor benefit more from an overall increase in these indicators than the non-poor for health, but less so for education. The poor tend to have lower social indicators than the non-poor. This is illustrated in table 5.7 with municipal data on four indicators. The data is for 1996-97, with 289 municipalities in 18 departments (UNDP, 1998). The municipalities have been classified into three groups of identical size. The first group consists of all municipalities that have low levels for any one of the indicators within their department. The second group includes the municipalities whose performance is in the middle range within their department. The third and last group consists of the municipalities that are top performers in their department. Households living in the top third of the municipalities have one more year of life expectancy than households living in the bottom third of the municipalities. The differences for malnutrition, literacy, and years of schooling are larger. While this information is useful, it would be more useful for policy to know how the various groups benefit from an overall increase in the indicators. Estimates of these marginal benefits are given in table 5.7. A marginal benefit larger (smaller) than one indicates that a group benefits more (less) than the other groups from an overall increase in the indicator (the methodology used to estimate these marginal benefits is described in annex 2, section MA.9). It can be shown that poor municipalities benefit significantly more than rich municipalities when there is an increase in life expectancy (at the five percent level) and literacy (at the nine percent level). Poor municipalities also benefit more than the middle group when there is a decrease in malnutrition (at the six percent level). By contrast, there are no statistically significant differences between poor municipalities and the other groups in the marginal benefit incidence for an increase in years of schooling, which is not surprising given that even for the comparatively rich, there is a lot of scope for higher education levels. Table 5.7: Estimates of the marginal gains in social indicators for various groups of municipalities Mean levels for the social indicators by Marginal increase in the indicators group of municipalities (A value of one is the benchmark) Bottom third Middle Top third Bottom third Middle Top third Life expectancy 67.54 68.028 68.791 1.106 0.995 0.899 Malnutrition 56.552 49.413 39.722 1.125 0.883 0.992 Literacy 47.535 61.429 73.735 1.143 0.949 0.908 Years of schooling 1.411 2.011 3.167 0.994 1.049 0.957 Source: World Bank staff usimg data from UINDP (1998). C. ELASTICITIES OF POVERTY AND SOCIAL INDICATORS TO GROWTH CAN BE USED TO SET TARGETS 5.14. The elasticities of poverty to growth can be used to establish future targets for poverty reduction in Honduras. Establishing targets for poverty reduction and for other indicators of well-being is one of the mandates of the PRSP prepared by the GRH. An illustration on how to establish such targets is given in table 5.8. Consider hypothetical initial conditions with the headcounts of extreme poverty in rural and urban areas at respectively 20 and 50 percent. Given the urbanization rate in 1999 of 46.2 percent, the national headcount for extreme poverty is then 36.14. Assuming a growth in per capita income of 2 percent over the full period', the headcount index of extreme poverty is reduced in urban and rural areas to 15.20 percent and 36.55 percent in 2015. Nationally, assuming no change in urbanization, extreme poverty is reduced to 26.76 percent. Taking into account the increase in urbanization (so that the weights for the urban and rural sectors change over time in the estimation of national poverty), extreme poverty is reduced nationally by an additional two percentage points, to 24.64 percent. These simulations are crude, but they give an idea of the gains toward poverty reduction that can be expected in the future. ' To achieve a per capita GDP growth rate of 2 percent, the country would need to have a GDP growth rate slightly below 5 percent in the next few years (given a population growth rate at 2.8 percent today; much later, the population growth rate is expected to decrease below 2 percent). 81 To reduce extreme poverty further, the country would need to increase either its GDP growth rate or its elasticity of extreme poverty to growth. In the simple model presented here, a ten percent increase in per capita GDP growth (to 2.2 percentage points per year) would have the same impact as a ten percent increase (in absolute terms) in the elasticity of extreme poverty to growth (to -1.09). It can then be shown that such an improvement would reduce poverty by close to one percentage point more in 2015. Table 5.8: Targets for poverty: An hypothetical illustration with growth at 2 percent per capita 1999 2000 2001 2002 2003 2004 2005 2010 2015 Urbanization rate 46.20 46.80 47.40 48.00 48.60 49.20 49.80 52.80 55.79 Extreme poverty Urban 20.00 19.66 19.33 19.00 18.67 18.36 18.04 16.56 15.20 Rural 50.00 49.03 48.08 47.15 46.23 45.33 44.46 40.31 36.55 National with urbanization 36.14 35.47 34.81 34.16 33.52 32.90 32.29 29.40 26.76 National w/o urbanization 36.14 35.28 34.45 33.63 32.84 32.06 31.30 27.77 24.64 Poverty Urban 40.00 39.38 38.76 38.16 37.56 36.98 36.40 33.65 31.10 Rural 80.00 79.12 78.25 77.39 76.54 75.70 74.86 70.84 67.02 National with urbanization 61.52 60.72 59.93 59.15 58.38 57.62 56.87 53.27 49.90 National w/o urbanization 61.52 60.52 59.53 58.56 57.59 56.65 55.71 51.20 46.98 Source: Estimates using the elasticities in table 5.5. The initial conditions in 1999 are hypothetical. 5.15. The elasticities of social indicators to growth (and urbanization) can also be used to set targets because they may provide more realistic projections than simple extrapolations. As is the case for poverty targets, the elasticities in table 5.6 can be used to set targets for social indicators, but with one caveat. In the case of poverty, there is no alternative to the use of the elasticities for establishing targets. In the case of social indicators, there is one alternative. Instead of using the model of table 5.6, one could find the curve of best fit for the historical trend in the indicators, and use the forecast for the targets. In most cases however, it could be argued that for Honduras, the model in table 5.6 works as well if not better than time series extrapolations using the line of best fit (whether this line is linear, exponential, logarithmic, or power-based). In any case, examples of targets for the social indicators using the historical trend, the levels model (i.e., the elasticities in table 5.6), with a growth rate of GDP of four percent per year, and the most probable scenarios for future urbanization and population growth are given in table 5.9. These simulations are proviided for illustrative purpose only. (Other simulations could easily be provided using the newly developed :SimSIP; see Box 5.1 for details). 82 Table 5.9: Targets for social indicators: An illustration of the growth and urbanization model 1999 2000 2001 2002 2003 2004 2005 2010 2015 Health Indicators Infant Mortality Trend 31.8 30.6 29.5 28.4 27.3 26.3 25.4 21.0 17.4 Levels 32.0 31.0 30.0 29.1 28.2 27.3 26.5 22.7 19.5 Under-five Mortality Trend 42.2 40.5 38.8 37.2 35.7 34.3 32.9 26.7 21.6 Levels 42.6 41.2 39.9 38.6 37.3 36.1 34.9 29.6 25.1 Life Expectancy Trend 69.8 70.4 71.0 71.6 72.2 72.8 73.4 76.4 79.3 Levels 69.6 70.0 70.4 70.8 71.3 71.7 72.1 74.2 76.5 Malnutrition Trend 31.0 31.3 31.6 32.0 32.3 32.6 32.9 34.6 36.2 Levels 28.4 27.8 27.3 26.8 26.3 25.8 25.3 22.7 20.4 Education Indicators Illiteracy Rate Trend 19.6 19.2 18.8 18.4 18.1 17.7 17.4 15.7 14.2 Levels 19.4 18.8 18.2 17.6 17.0 16.5 16.0 13.7 11.8 NetPrimaryEnrollment Trend 80.5 80.8 81.0 81.3 81.5 81.8 82.0 83.2 84.3 Levels 80.7 81.1 81.5 81.9 82.4 82.8 83.2 85.3 87.5 Net Secondary Trend 29.0 28.5 28.0 27.6 27.1 26.7 26.2 24.1 22.1 Enrollment Levels 31.1 31.7 32.4 33.1 33.8 34.5 35.2 39.0 43.5 GrossPrimaryEnrollment Trend 116.2 117.2 118.2 119.2 120.2 121.2 122.1 126.8 130.0 Levels 113.5 114.0 114.6 115.1 115.7 116.2 116.7 118.9 121.1 Gross Sec. Enrollment Trend 40.8 42.4 43.9 45.6 47.2 49.0 50.7 60.2 71.0 Levels 39.1 40.6 42.2 43.8 45.5 47.3 49.1 58.8 71.2 Tertiary Enrollment Trend 11.9 12.2 12.4 12.6 12.8 13.1 13.3 14.4 15.4 Levels 12.8 13.4 14.2 15.0 15.8 16.6 17.4 20.9 25.8 Infrastructure Indicators Access to Safe Water Trend 73.3 75.0 76.7 78.3 80.0 81.7 83.3 91.7 100.0 Levels 72.3 73.8 75.7 77.6 79.4 81.1 82.8 89.9 98.3 Access to Sanitation Trend 83.3 87.0 90.7 94.3 98.0 100.0 100.0 100.0 100.0 Levels 72.6 74.2 75.8 77.5 79.2 81.0 82.8 92.4 100.0 Telephone Mainlines Trend 4.1 4.5 4.9 5.3 5.8 6.3 6.9 10.5 16.0 Levels 4.1 4.4 4.7 5.0 5.4 5.8 6.2 8.6 12.1 Source: Based on SimSIP. The predictions use World Bank data on initial conditions (latest observation available for each indicator) which may be different from those used by the GOH. The GOH may also have different forecasts for GDP growth, population growth, and urbanization (see text for our own assumptions). These targets are given for illustrative purpose only. Alternative simulations corresponding to the data and growth/population/urbanization forecasts of the GOH could easily be obtained using the simulators in SimSIP. 83 Box 5.1: SIMSIP - SIMULATIONS FOR SOCIAL INDICATORS AND POVERTY Many governments set targets for poverty and social indicators (e.g., in education, health, and access to basic infrastructure services such as safe water and sanitation). Governments then propose policies that will improve their chances of reaching the targets, and they estimate the cost of reaching the targets. The use of targets as a basis of country strategies is common in countries preparing PRSPs, but it also takes place in other, richer countries as well. This box briefly describes user-friendly Excel-based simulators which have been created in order to fa,cilitate the setting of targets for poverty and social indicators and the estimation of the cost of reaching targets. SimSIP has four modules: (a) SimSIP_Goals helps analysts assess whether PRSP targets are realistic; (b) SimSIP_Costs provides estimates of the cost of reaching targets; (c) SimSIP_Incidence analyzes who is likely to benefit from additional social expenditures; and (d) SimSIP_Determinants analyzes the micro-determinants of poverty and other outcomes. Below, the focus is on SimSIP_Goals and SimSIP_Costs. Details on other modules are available upon request. SimSIP Goals. SimSIP_Goals is an Excel Worksheet that can be used for setting targets for education, health, basic infrastructure, and poverty indicators (the list of indicators is as follows: gross primary, secondary, and tertiary enrollment rates; net primary and secondary enrollment rates; rate of illiteracy among the adult population; infant mortality rate, under-five mortality rate, life expectancy, and under five malnutrition rate; access to water, access to sanitation, and telephone main lines; and poverty measures - headcount, poverty gap, and squared poverty gap). At this stage, simulations can be made only for Latin American countries, but the simulator will be adapted to other regions. The indicators in the worksheet correspond roughly to the International Development Goals. For education, health, and infrastructuire services, the indicators are provided at the national level only. Targets can be based on either historical trends or model-based forecasts. For historical trends, projections into the future are based on country-level historical trends observed for each specific indicator. Four different ways of fitting a historical trend at the country level are considered for each indicator. The best fit historical trend among the four functional forms is selected for the simulations. Time is the only exogenous variable. For model-based forecasts, the simulator relies on an econometric model giving elasticities of the indicators to economic growth, per capita, urbanization, and time. The elasticities are estimated with two different specifications using world-wide panel data sets, and they are allowed to vary with a country's level of development (i.e., GDP per capita) and urbanization. For poverty, the indicators are provided at the rural and urban level. This yields national poverty measures when urbanization is taken into account. The simulations for poverty are based on estimated elasticities of poverty to growth, taking into account the impact of growth on inequality. Apart from simulating future levels of poverty as a function of economic growth, population growth, and urbanization growth, the user is provided with the contribution of each of these variables to poverty reduction. Given assumptions for these variables, the user can also assess how income inequality would have to change in order to reduce poverty by the stated objective (say, a reduction in headcount of 50 percent by 2015). SimSIP_Costs. SimSIP_Costs can be used to estimate the cost of reaching education, health, and basic infrastructure targets, and to check whether the overall cost can be funded under alternative scenarios. The simulator has interfaces for education, basic health care, basic infrastructure, and fiscal sustainability. Education. The costing is done for preschool, primary school and two levels of secondary school (as well as general administrative costs) through cohort analysis. Three sets of assumptions must be entered by the user in the simulator: country demographics, the performance of the education system (age at entry in the various schooling cycles, as well as, structure of repetition, promotion, and drop out rates), and costs (supply-side costs, including teacher wages and teacher-student ratios; demand-side costs related to the provision of stipends to part of the student body; and investment costs related to the training of new teachers and the construction of new classrooms). All variables are allowed to change over time. 84 Simulations are provided for education outcomes or targets and for the cost of reaching these outcomes. * Education outcomes: The outcomes or targets can be specified in terms of enrollment rates (net or gross), in terms of completion rates, or in terms of quality variables such as the time it takes to complete a cycle. Rather than specifying a target, the user must propose changes in the indicators of performance of the system (such as entry rates, or repetition-promotion-dropout rates) and assess whether the outcome is realistic or not. For the most important indicators such as net and gross enrollment rates, to check if outcomes are realistic, the user can use the goals module of SimSIP. * Costs of reaching taraets: On the basis of the education outcomes and the cost structure specified by the user, the simulator provides an estimate of the costs of reaching the targets. Three different types of costs are considered: supply-side recurrent costs (consisting mostly of teacher wages and administration costs), demand-side costs (stipends provided to low income students), and supply-side investment costs (mainly construction of new classrooms, training of teachers). Basic health care. The costing is done for the provision of basic health care packages. Three different packages are considered. They differ in terms of the number of services included. The services comprise general mortality reduction programs with emphasis on acute diarrhea and respiratory diseases among babies and young children; immunisation and nutrient deficiency programs; pregnancy care including pre-natal and post-natal assistance; community and environment programs; adult and senior health issues; education on medical drugs use; and occupational health programs. The basic packages are provided by mobile health teams, community teams, and officials of the Ministry of Health. Three sets of assumptions must be entered by the user: country demographics, parameters behind basic health care delivery systems (e.g., exact specification of all members of mobile teams in charge of providing populations with health care; number of villages to be covered by a single team; number of annual visits per village) and costs (e.g., wages, cost of medicines, travel costs, etc.). Simulations are provided for coverage outcomes and the costs of reaching targets. Also presented are the total gains in wellbeing from basic health packages. * Health coverage outcomes: This is the population covered by mobile health teams in targeted areas. * Costs and gains in well-being: Based on the cost structure specified by the user, the simulator yields estimates of total annual costs in the local currency of the country. Annual cost by operating team are also provided along with annual cost per individual reached by the programs (annual cost per capita). The present value of investments in basic health packages is calculated and the cost effectiveness of the programs is estimated in reference to gains in Disability Adjusted Life Years (DALYs). Basic infrastructure. This deals with targets for access to safe water, sanitation, and electricity, and the cost of reaching these targets. Again, three sets of assumptions must be entered by the user for country demographics, coverage levels (information on current coverage and targets), and costs (the costs per beneficiary are separated into investment, operations and maintenance costs). Options for water system technology relate to the type of water supply systems (piped or non-piped), the water distribution mechanism (gravity fed, pump fed and spring protection systems), and the population density served by the systems (high density or concentrated, semi-dispersed and dispersed population). For sanitation, the options include conventional sewage systems, pour-flush latrines, and dry latrines. The various costs per beneficiary (investment, operations, and maintenance) can be shared between the public sector and the households, with the option of including subsidies. The simulator returns coverage rates and overall costs. Fiscal sustainahility. The simulator integrates the information on costs provided by the education, health, and basic infrastructure worksheets into a fiscal sustainability framework. The total resources of the Government are derived from assumptions regarding taxation rates and GDP growth, the overall structure of public spending, and the availability of HIPC debt relief funds. Projections are made about the share of total public spending devoted to social and targeted interventions, so as to suggest the need for adjustments in the budget in order to cover the cost of reaching targets. The simulator includes features which enable policy maker to assess trade-offs within and between sectors (e.g., how much additional coverage for basic health care can be afforded if one reduces net secondary school enrolment targets by 5 percent?). 85 ANNEX 1: ROBUSTNESS OF' THE POVERTY TREND AND POVERTY PROFILE A. TESTS FOR ROBUSTNESS OF THE POVERTY TREND PRESENTED IN THE I-PRSP As was mentioned in Chapter 1, tests of the robustness of the poverty trend in the I-PRSP can be made using alternative poverty lines and/or alternative indicators of well-being. Test for robustness to the choice of the poverty line. Our test for robustness consists in using a poverty line that differs from that used by the GRH. For this, we simply scale up the extreme and moderate poverty lines obtained in a previous poverty report for Honduras (World Bank, 1995) in order to provide 2,200 kcal per person per day. We use a single national poverty line (using urban and rural poverty lines would affect the levels of poverty in both urban and rural areas, and therefore the comparison of the two sectors, but it would not affect the trend in each sector much). We then adjust on a monthly basis the poverty line over time using the CPI instead of the cost of the food basket. This follows previous work on Honduras in an international perspective discussed in Wodon (2000). There are thus two main sources of differences in the choice of the poverty line versus the method used in the I-PRSP: the level of the poverty lines in the reference period, and their adjustment to increases in prices over time (there are also other differences, such as the adjustment to food basic needs in order to account for non-food basic needs, but this does not matter much in the present context). What matters the most is the adjustment over time, through the use of the CPI rather than the Canasta Basica. The resulting poverty lines are given in table Al.1, and the resulting welfare ratios and poverty measures are provided in tables A 1.2 to Al.4, in the columns marked "NON ADJUSTED". In the tables, W is the welfare ratio, PI the headcount index, P2 the poverty gap, and P3 the squared poverty gap (see Annex 2, section MA. 1, for a definition of these terms). All measures are at the individual rather than the household level. Although the levels in tables Al .2 to A1.4 differ from those in table 1.1 in chapter 1, the trend is similar. For example, in table A1.2, the headcount indices of poverty and extreme poverty at the national level decrease by respectively 7.37 and 11.08 percent, which is only a bit larger than the decrease observed in table 1.1 of clhapter 1, at 5.6 and 8.9 percent. Given these results, we can conjecture that the GRH trend appears to be robust to the choice of the poverty lines. Test for robustness to adjustments for underreporting. The problem of underreporting for poverty trends is that the level of underreporting in the surveys may change over time. In Honduras, the surveys seem to improve over time, in that at the end of the period the mean level of income in the survey is closer to the National Accounts than in the early 1990s (The inverses of the adjustment factors are reported in table Al.1; a lower inverse implies a higher adjustment factor, which in turns implies a better measure of income in the survey, i.e. less underreporting). This means that the reduction in poverty observed in table 1.1 in chapter 1 may be due to the improvement in the surveys rather than to an actual betterment of living standards. This is indeed what is suggested by the estimates in the columns "ADJUSTED FOR CONSUMPTION" in tables A1.2 to Al.4. With adjustments for underreporting, there is no decrease in poverty in the 1990s. Figure Al.l helps to understand why. In the Figure, the bold line represents the mean nominal per capita income in the EE'HPM surveys in any year in proportion of its value in 1991 . The dashed line below the bold line is the nominal per capita GDP in the National Accounts, also in proportion of its value for 1991. The other dashed line below per capita GDP represents per capita consumption (we use this variable because private per capita consumption is a better indicator of standards of living than per capita GDP). The ratio of per capita consumption to the mean income reported in the surveys is the adjustment factor that can be used from the National Accounts. This is the plain line going down, whose scale is on the right axis. The adjustment factor decreases from 1.43 in 19'91 to 1.07 in 1998. If we use the simplest adjustment methodology which consists in multiplying all the incomes in the survey by the adjustment factor without changing the poverty lines of the GRH, the decrease in poverty vanishes. The reason is 86 that with the adjustment factor, there is almost no growth in per capita income, and thereby no poverty reduction. A similar lack of poverty reduction is found by CEPAL (1999) with a more sophisticated adjustment procedure. According to CEPAL, the share of all households below the poverty line (headcount index) was the same in 1997 at 74 percent than in 1990 at 75 percent. Figure A1.1: Adjustment for underreporting 5.00 - 1.50 . 41.00 Not adjusted @ 2 00 - .-~ ---- Per capita GDP °~ 2.00 - - - " - -- 0.50 -Per capita Cons. 1 00 - - ~ - - __ _ it05000.00 -Adjustment factor] 0.00 -, , , , , , -0.001 91 92 93 94 95 96 97 98 Year Implication of the second test. The implication of the second test is that if there has been a reduction in poverty in the 1990s, the magnitude of this reduction is probably lower than indicated in the I-PRSP. There is no simple answer to the question of whether the estimates of poverty with and without adjustment for under- and over-reporting are better. While some would be argued that the measurement errors may be larger in the national accounts than in the surveys, others would argue the reverse. Still, it is worth stating again that two key factors account for the lack of poverty reduction with adjustments. First, the rate of per capita income growth in the surveys is much higher than the rate of per capita GDP growth in the national accounts. Second, in constant terms, the rate of growth of per capita GDP in the national account is higher than the rate of growth in per capita consumption, in part because the CPI has risen faster than the GDP deflator. The first factor plays a much larger role than the second in the lack of poverty reduction observed with adjustments. Addendum on the income aggregate. It is also worth noting that if all sources of income are used for measuring poverty in the later rounds of the EPHPM surveys, poverty measures are lower. For the full period under review (1991 to 1999), the EPHPM surveys can be used to compute labor income at the individual level (income from first occupation, a second occupation, and self-employment), and to add individual income into a measure of per capita household income. This is how table A1.2 was constructed. For the last few years (1997-1999), the surveys also includes information on other sources of income: income from pensions; subsidies; income from rent; allowances, grants and stipends; remittances from abroad; support from relatives; and support from other individuals; and other sources of incomes. As indicated in tables A1.2 to Al.4, using total income in the later surveys reduces the poverty estimates for the headcount index by about 5 percentage points. B. STANDARD POVERTY PROFILE Tables A1.6 to A1.8 provide a basic poverty profile using the last three EPHPM surveys (March and September 1998 and March 1999) and the poverty lines in table Al.1. The headcount indices of poverty and extreme poverty are provided, as well as the poverty gap and squared poverty gap for poverty (but not for extreme poverty). The mean values of the variables are also given. In most cases, these mean values are the proportion of households in the relevant categories. 87 Table A1.1: National Poverty Lines (Lempiras per month per person) and Adjustment Factor Extreme poverty line Moderate poverty line Adjustment factors for I I ~ ~~~~~undeffeporting May-90 48.91 82.41 68.41 Sep-90 53.12 89.50 69.34 May-91 68.04 114.65 70.24 Sep-91 72.07 121.43 76.10 Mar-92 74.18 124.99 76.48 Sep-92 76.03 128.11 83.09 Mar-93 79.19 133.42 92.99 Oct-94 107.09 180.44 92.41 Mar-95 122.74 206.80 104.01 Oct-95 137.46 231.61 109.50 Mar-96 148.99 251.04 99.77 Sep-96 169.92 286.31 91.68 Jun-97 193.91 326.73 98.79 Mar-98 210.51 354.69 102.20 Sep-98 229.01 385.87 101.22 Mar-99 241.69 407.24 101.22 Source: World Bank (1994) and author's estimation using CPI for adjustments over time. 88 Table A1.2: Welfare Ratios and Poverty Measures, National NON ADJUSTED NON ADJUSTED ADJUSTED TO ADJUSTED TO CONSUMPTION CONSLTMPTION Extreme Moderate Extreme Moderate W P0 P1 P2 W P0 PI P2 W P0 P1 P2 W P0 P1 P2 LABOR INCOME May-90 1.94 49.09 26.20 18.26 1.15 68.13 39.86 28.45 2.84 35.56 18.54 13.25 1.68 54.81 29.65 20.67 Sep-90 1.96 49.71 26.42 18.13 1.16 69.29 40.28 28.59 2.83 37.09 18.75 12.97 1.68 55.56 30.20 20.86 May-91 1.89 45.09 22.53 15.32 1.12 66.87 36.56 25.09 2.70 31.05 15.79 11.16 1.60 52.93 26.55 17.97 Sep-91 1.94 46.53 23.06 15.57 1.15 67.51 37.11 25.56 2.55 35.22 17.47 12.07 1.51 55.99 29.32 19.80 Mar-92 2.13 39.77 19.43 12.99 1.26 62.44 32.67 21.92 2.78 30.26 14.56 10.07 1.65 50.34 25.24 16.76 Sep-92 2.26 40.71 20.90 14.38 1.34 60.51 33.29 23.08 2.71 34.27 17.57 12.17 1.61 53.70 28.42 19.54 Mar-93 2.57 34.62 16.70 11.47 1.52 56.24 28.74 19.17 2.76 31.41 15.45 10.75 1.64 53.32 26.79 17.82 Oct-94 2.53 36.73 18.28 12.40 1.50 58.04 30.19 20.46 2.74 33.56 16.85 11.51 1.62 54.62 28.05 18.97 Mar-95 2.62 36.59 17.93 12.11 1.55 56.20 29.62 20.10 2.52 37.68 18.67 12.57 1.49 57.61 30.67 20.86 Oct-95 2.59 37.76 19.84 13.67 1.53 55.31 30.93 21.68 2.36 41.20 21.51 14.84 1.40 59.40 33.22 23.41 Mar-96 2.44 39.23 20.54 14.20 1.45 59.04 32.44 22.63 2.44 39.23 20.50 14.17 1.45 58.98 32.39 22.59 Sep-96 2.27 38.65 18.03 11.69 1.35 59.39 30.93 20.46 2.48 35.00 16.32 10.65 1.47 56.20 28.50 18.70 Jun-97 2.56 33.69 17.06 11.64 1.52 53.40 28.24 19.18 2.59 33.33 16.86 11.51 1.54 53.07 27.94 18.96 Mar-98 2.81 35.32 19.71 14.54 1.67 52.49 29.62 21.43 2.75 35.85 20.05 14.76 1.63 53.57 30.12 21.79 Sep-98 2.60 36.08 19.82 14.47 1.55 53.64 30.26 21.68 2.57 36.27 20.01 14.60 1.53 54.08 30.54 21.89 Mar-99 2.53 37.72 20.69 15.04 1.50 55.79 31.34 22.53 2.51 37.81 20.86 15.15 1.49 56.15 31.59 22.71 ALL INCOME SOURCES Jun-97 2.85 29.49 13.00 7.73 1.69 49.44 24.20 15.17 2.88 29.21 12.80 7.60 1.71 49.07 23.89 14.95 Mar-98 3.18 30.93 15.69 10.59 1.88 47.72 25.33 17.34 3.11 31.44 16.02 10.82 1.84 48.83 25.82 17.69 Sep-98 2.97 31.88 15.77 10.52 1.76 49.08 26.08 17.63 2.93 32.19 15.96 10.65 1.74 49.48 26.35 17.83 Mar-99 2.86 32.60 16.11 10.73 1.70 50.92 26.61 17.97 2.83 32.70 16.27 10.84 1.68 51.18 26.85 18.15 Source: Authors' estimation using EPHIPM. 89 Table A1.3: Welfare Ratios and Poverty Measures, Urban NON ADJUSTED NON ADJUSTED ADJUSTED TO ADJUSTED TO CONSUMPTION CONSUMPTION Extreme Moderate Extreme Moderate W P0 PI P2 W P0 PI P2 W P0 P1 P2 W P0 PI P2 LABOR INCOME May-90 3.09 29.03 16.10 12.32 1.84 49.43 25.62 18.10 4.52 19.86 12.18 10.14 2.68 34.36 18.22 13.52 Sep-90 3.04 27.61 14.29 10.57 1.81 49.50 24.33 16.43 4.39 17.78 10.48 8.55 2.61 33.01 16.59 11.87 May-91 2.71 30.00 15.07 10.90 1.61 51.40 25.66 17.31 3.86 19.51 10.89 8.67 2.29 37.24 17.84 12.49 Sep-91 2.78 29.86 16.09 12.13 1.65 50.36 25.78 18.08 3.66 22.49 12.92 10.38 2.17 37.93 20.05 14.48 Mar-92 3.14 23.87 12.17 8.99 1.86 45.35 21.25 14.14 4.10 17.16 9.58 7.65 2.44 32.18 15.78 10.99 Sep-92 3.39 23.19 12.83 9.86 2.01 41.33 20.59 14.41 4.08 19.56 11.16 8.89 2.42 33.97 17.14 12.40 Mar-93 3.70 21.18 11.11 8.52 2.19 38.84 18.79 12.84 3.98 18.73 10.43 8.17 2.36 35.91 17.41 12.01 Oct-93 3.50 18.64 10.43 8.26 2.08 37.04 17.46 11.97 3.00 23.19 11.85 9.04 1.78 45.10 20.96 13.93 Oct-94 3.46 21.37 10.67 7.73 2.05 42.49 19.27 12.54 3.74 18.69 9.84 7.30 2.22 38.92 17.51 11.54 Mar-95 3.60 20.87 10.19 7.58 2.14 39.41 18.10 11.98 3.46 21.74 10.62 7.79 2.05 40.91 18.94 12.48 Oct-95 3.50 19.85 10.43 7.78 2.07 37.59 17.95 12.07 3.19 23.39 11.37 8.30 1.89 42.42 19.86 13.20 Mar-96 3.69 19.15 10.61 8.23 2.19 38.80 17.97 12.27 3.69 19.15 10.59 8.22 2.19 38.74 17.93 12.24 Sep-96 3.29 20.38 10.60 8.05 1.95 38.95 18.29 12.30 3.59 17.68 9.82 7.65 2.13 35.65 16.57 11.33 Jun-97 3.49 19.79 11.22 8.61 2.07 35.86 18.06 12.66 3.53 19.76 11.11 8.55 2.10 35.66 17.84 12.53 Mar-98 4.06 17.59 9.92 7.89 2.41 32.21 15.99 11.26 3.98 18.14 10.09 7.98 2.36 33.37 16.35 11.47 Sep-98 3.83 16.37 8.66 6.62 2.27 32.10 14.91 10.03 3.78 16.46 8.75 6.67 2.25 32.34 15.12 10.15 Mar-99 3.82 19.05 10.36 7.81 2.27 34.29 16.68 11.62 3.78 19.06 10.45 7.86 2.24 34.41 16.86 11.72 ALL INCOME SOURCES Jun-97 3.99 14.47 6.30 3.94 2.37 30.29 12.94 7.75 4.04 14.44 6.21 3.88 2.40 29.97 12.73 7.62 Mar-98 4.71 11.65 4.87 3.10 2.80 25.25 10.37 6.10 4.61 12.10 5.02 3.18 2.74 26.23 10.70 6.29 Sep-98 4.40 11.76 4.78 2.88 2.61 26.14 10.46 6.00 4.34 11.87 4.86 2.92 2.58 26.43 10.64 6.11 Mar-99 4.33 12.95 5.56 3.43 2.57 27.31 11.17 6.70 4.29 12.97 5.63 3.47 2.55 27.37 11.33 6.79 Source: Authors' estimation using EP]PM. 90 Table AI.4: Welfare Ratios and Poverty Measures, Rural NON ADJUSTED NON ADJUSTED ADJUSTED TO ADJUSTED TO CONSUMPTION CONSUMPTION Extreme Moderate Extreme Moderate W PO pi P2 W PO P1 P2 W P0 PI P2 W P0 PI P2 LABOR INCOME May-90 1.14 63.06 33.22 22.40 0.68 81.15 49.77 35.65 1.67 46.49 22.96 15.42 0.99 69.05 37.61 25.64 Sep-90 1.21 65.19 34.92 23.42 0.72 83.16 51.45 37.11 1.74 50.61 24.54 16.06 1.03 71.36 39.75 27.16 May-91 1.32 55.76 27.81 18.44 0.78 77.81 44.28 30.60 1.88 39.22 19.26 12.91 1.11 64.03 32.72 21.84 Sep-91 1.34 58.40 28.03 18.03 0.79 79.74 45.19 30.89 1.76 44.30 20.71 13.28 1.04 68.86 35.92 23.59 Mar-92 1.40 51.21 24.64 15.86 0.83 74.72 40.89 27.50 1.83 39.68 18.14 11.82 1.09 63.39 32.05 20.91 Sep-92 1.43 53.44 26.79 17.69 0.85 74.44 42.54 29.41 1.72 45.01 22.26 14.59 1.02 68.03 36.64 24.76 Mar-93 1.73 44.61 20.85 13.66 1.03 69.17 36.13 23.87 1.86 40.83 19.17 12.67 1.10 66.26 33.76 22.13 Oct-94 1.82 48.37 24.05 15.93 1.08 69.82 38.46 26.45 1.97 44.82 22.15 14.70 1.17 66.50 36.02 24.60 Mar-95 1.87 48.57 23.83 15.56 1.11 68.99 38.41 26.29 1.80 49.83 24.80 16.22 1.07 70.33 39.60 27.25 Oct-95 1.54 58.37 30.66 20.45 0.91 75.69 45.86 32.75 1.41 61.70 33.18 22.36 0.83 78.94 48.58 35.16 Mar-96 1.48 54.62 28.14 18.77 0.88 74.55 43.52 30.56 1.49 54.62 28.09 18.73 0.88 74.49 43.46 30.51 Sep-96 1.48 52.90 23.86 14.56 0.88 75.34 40.79 26.85 1.61 48.54 21.42 13.03 0.96 72.20 37.81 24.48 Jun-97 1.82 44.63 21.66 14.03 1.08 67.21 36.26 24.31 1.84 44.01 21.39 13.84 1.09 66.77 35.88 24.02 Mar-98 1.95 46.28 24.30 16.56 1.07 68.61 40.48 29.53 1.77 49.96 27.98 20.17 1.05 69.63 41.09 30.01 Sep-98 1.62 51.92 28.77 20.77 0.96 70.91 42.59 31.04 1.60 52.20 29.05 20.96 0.95 71.51 42.93 31.32 Mar-99 1.50 52.85 29.07 20.91 0.89 73.15 43.22 31.38 1.48 52.99 29.30 21.08 0.88 73.71 43.52 31.62 ALL INCOME SOURCES Jun-97 1.94 41.32 18.28 10.71 1.15 64.52 33.06 21.01 1.97 40.85 18.00 10.53 1.17 64.11 32.68 20.72 Mar-98 1.95 46.28 24.30 16.56 1.16 65.59 37.24 26.29 1.91 46.84 24.78 16.90 1.13 66.82 37.86 26.77 Sep-98 1.82 48.06 24.59 16.65 1.08 67.48 38.63 26.96 1.80 48.52 24.87 16.84 1.07 67.96 38.97 27.24 Mar-99 1.68 48.51 24.67 16.66 0.99 69.99 39.11 27.12 1.66 48.68 24.90 16.82 0.98 70.41 39.42 27.36 Source: Authors' estimation using EPHPM. 91 Table A1.5: Inequality Measures, National, Urban, and Rural NATIONAL URBAN RURAL Theil Gini Atkinson Theil Gini Atkinson Theil Gini Atkinson LABOR INCOME May-90 76.93 59.87 62.12 73.12 58.38 65.28 49.37 51.52 50.41 Sep-90 82.09 60.82 61.58 71.85 57.02 61.31 71.38 56.33 54.22 May-91 61.69 55.31 57.55 58.80 54.26 58.48 50.09 50.59 51.85 Sep-91 68.50 56.71 58.75 58.83 54.91 62.22 66.39 52.11 50.78 Mar-92 61.50 55.07 56.56 56.59 53.69 57.39 48.54 48.82 49.41 Sep-92 62.89 56.71 57.68 55.40 53.90 59.26 51.00 51.35 49.06 Mar-93 67.38 56.36 58.01 60.81 54.65 58.13 59.48 51.56 52.33 Oct-93 NA NA NA 52.06 52.08 56.70 NA NA NA Oct-94 65.27 57.12 56.89 58.75 54.22 55.11 62.26 55.95 54.28 Mar-95 67.01 57.53 57.40 58.02 53.90 55.81 67.13 56.80 54.49 Oct-95 70.76 58.24 58.60 56.80 52.54 53.99 80.38 58.93 56.46 Mar-96 71.02 58.33 59.20 62.02 54.18 57.51 59.26 54.40 52.61 Sep-96 61.79 54.65 52.88 54.86 51.41 54.30 52.81 50.31 44.64 Jun-97 62.22 55.32 55.51 53.29 51.64 55.34 62.59 53.92 51.26 Mar-98 69.28 58.96 63.5'6 54.91 53.18 56.80 72.80 58.79 63.19 Sep-98 63.36 57.20 61.82 47.45 50.26 50.67 67.85 57.30 63.36 Mar-99 64.15 57.76 61.97 51.25 52.05 53.94 58.95 55.23 60.41 ALL INCOME SOURCES Jun-97 59.03 54.11 45.69 49.22 49.92 41.73 58.28 52.23 42.35 Mar-98 66.47 57.74 54.38 50.90 51.08 41.86 69.19 57.42 55.34 Sep-98 65.91 56.46 53.06 50.38 49.36 38.79 68.72 56.29 54.91 Mar-99 59.63 55.98 51.90 45.88 49.42 39.29 55.49 53.85 51.33 Source: Authors' estimation using EPHPM. 92 Table A1.6: Poverty Profile and Summary Statistics for Urban and Rural Sectors, March 1998 Urban Rural P0 Ext. P0 Mod. P, Mod. P2 Mod. Stat. P0 Ext. P0 Mod. PI Mod. P2 Mod. Stat. Monthly normalized pc income 4.71 1.95 Log normalized pc income 1.10 0.10 Overall poverty measure 11.65 25.25 10.37 6.10 46.28 65.59 37.24 26.29 LOCATION Atlantida 16.05 32.39 13.80 8.47 6.94 29.22 51.79 25.78 17.03 4.40 Colon 15.63 36.19 14.19 7.13 2.63 44.99 69.98 34.65 22.35 4.33 Comayagua 9.55 28.36 10.36 5.40 4.83 40.93 62.95 33.45 23.43 6.51 Copan 26.11 49.32 19.22 11.00 2.76 55.10 80.71 44.42 30.42 6.77 Cortes 4.27 13.97 4.79 2.80 24.60 18.89 39.66 15.99 9.86 7.15 Choluteca 27.93 40.18 23.83 16.57 5.19 53.23 65.56 42.94 32.69 7.79 El Paraiso 9.13 27.59 9.16 5.06 3.11 40.37 64.06 33.29 22.00 7.96 Francisco 9.00 22.21 8.26 4.45 33.66 51.93 74.97 43.58 32.14 6.08 Intibuco 11.29 46.62 14.02 5.61 0.80 68.29 81.23 58.65 47.82 4.50 Lapaz 24.98 44.55 19.92 11.79 1.11 52.58 80.54 44.58 31.82 3.56 Lempira 27.51 46.90 19.98 12.03 0.36 56.82 67.02 44.07 33.02 6.96 Ocotepeque 22.98 32.99 18.96 14.28 0.54 39.44 55.65 30.33 20.34 2.62 Olancho 8.95 21.87 6.81 3.10 3.40 57.44 72.07 43.88 31.46 9.51 Santa Barbara 25.14 38.02 18.73 11.47 2.74 32.43 53.55 24.66 14.62 9.06 Valle. 26.46 55.20 25.57 16.30 1.63 52.97 73.66 44.44 32.71 3.55 Yoro 19.64 27.52 15.01 10.06 5.72 49.44 66.50 38.50 26.37 9.24 HOUSEHOLD Numberofbabies 15.58 33.57 13.57 7.71 0.96 51.44 72.54 41.48 29.26 1.24 Number of children 16.08 33.10 13.93 8.21 1.49 53.44 71.57 41.94 29.84 1.99 Number of adults 9.01 20.23 8.23 4.91 3.44 40.93 60.24 33.44 23.50 3.44 Head female 15.22 30.33 13.23 8.01 26.45 44.76 62.56 36.13 26.01 18.85 Age head 44.97 47.05 No spouse 14.73 29.03 12.77 7.78 28.66 42.90 62.58 35.55 25.56 22.83 Birth migration 9.89 23.52 8.77 4.94 55.00 36.44 57.84 29.35 19.39 31.13 Rural migration 15.59 23.17 11.79 7.62 2.92 35.82 64.18 29.78 18.08 2.57 Urban migration 9.19 29.39 8.76 4.27 3.89 30.09 51.85 26.46 17.49 1.93 HOUSEHOLD HEAD EDUCATION No education 26.07 40.30 20.38 13.19 15.70 58.26 76.03 45.41 32.40 34.32 Primary partial 13.44 32.02 12.29 6.67 27.52 46.47 67.12 37.99 26.98 44.28 Primary total 10.59 28.18 10.30 5.87 24.14 31.71 53.67 26.73 17.93 15.10 Secondary partial 7.79 19.27 7.74 4.66 11.84 19.79 38.29 16.26 9.52 2.83 Secondary total 2.94 7.36 2.69 1.50 11.12 10.18 16.15 9.29 6.81 2.61 Superior 0.58 2.19 0.86 0.61 9.68 11.46 18.63 11.29 8.18 0.86 EMPLOYMENT Working 9.10 22.37 8.28 4.39 80.61 45.76 65.78 36.69 25.54 87.66 Available (unemployed) 15.51 41.27 14.36 9.36 1.45 61.38 74.03 53.05 44.32 0.91 Search (unemployed) 49.68 63.15 42.82 33.28 3.43 62.28 66.95 53.41 45.84 0.76 Notworking 16.41 31.69 13.92 8.92 15.96 50.05 48.04 38.03 35.12 11.58 SECTOR OF ACTIVITY Agriculture 29.87 44.47 23.78 15.55 9.50 53.36 72.87 43.13 30.98 61.46 Mining/Manuf./electricity 8.87 23.85 8.59 4.63 18.18 35.41 60.53 27.12 16.97 6.67 Construction 11.56 26.14 11.51 7.49 7.60 28.88 50.72 21.12 11.55 4.06 93 Commerce 8.13 21.34 7.80 4.12 21.29 29.17 46.27 22.76 14.45 8.54 Transport 4.23 9.84 3.56 1.90 5.10 4.80 26.56 8.50 3.56 0.87 Services 7.72 20.62 7.16 3.64 22.27 26.10 46.44 20.76 12.54 6.82 POSITION/OTHER Employee/worker/cooperative 9.18 23.38 8.56 4.66 45.94 33.19 60.78 26.48 15.45 25.58 Self-employed 15.22 30.06 13.10 7.70 28.15 53.82 71.04 43.26 31.59 58.80 Employer 4.38 9.20 4.81 3.44 9.62 9.51 19.60 7.62 4.33 3.96 Unpaid family work 31.92 31.92 19.15 12.74 0.23 100. 100 66.41 47.29 0.07 Public sector 3.15 8.58 2.70 1.47 9.07 8.30 27.74 8.68 4.18 2.74 Size of firm> 10 people 4.88 16.78 5.33 2.74 33.67 19.44 44.95 17.22 9.63 12.62 UNDEREMPLOYMENT Hours of work < 20 21.73 36.78 18.45 12.66 21.84 50.53 64.90 40.95 30.86 15.51 20< hours of work <39 16.08 29.79 12.44 6.52 8.04 49.08 67.57 39.24 27.43 13.43 Morethan39hours 8.01 21.14 7.61 4.01 70.12 44.85 65.36 36.07 25.11 70.94 Wanttoworkmore 12.28 30.98 11.17 5.65 5.54 55.07 76.34 44.14 31.64 4.00 Cannot workmore health/fain. 10.76 10.76 5.73 3.06 0.65 75.44 82.40 56.07 40.94 0.44 HOUSEHOLD SPOUSE EDUCATION No education 22.07 40.57 18.10 10.39 9.95 60.69 78.94 47.56 33.88 24.29 Primary partial 13.94 31.43 12.73 7.52 18.44 47.80 68.23 38.30 27.02 33.59 Primary total 9.82 26.11 9.13 5.09 19.04 36.05 56.18 29.70. 20.11 14.00 Secondary partial 6.83 15.93 6.40 3.72 8.55 18.41 39.08 16.62 10.44 2.69 Secondary total 1.77 6.89 2.36 1.34 10.91 4.03 8.06 2.88 1.63 2.19 Superior 0.37 0.37 0.37 0.37 4.44 14.81 27.51 9.08 4.32 0.40 EMPLOYMENT Working 7.49 17.30 6.73 3.80 34.17 31.67 51.91 25.42 16.57 21.44 Available (unemployed) 18.96 35.13 15.05 8.63 5.59 65.89 83.65 53.64 40.55 7.48 Search (unemployed) 9.25 31.38 12.43 7.17 1.06 18.25 42.08 15.61 8.30 0.55 Notworking 13.22 29.61 11.85 6.92 36.11 53.64 72.38 42.75 30.55 55.19 SECTOR OF ACTIVITY Agriculture 6.97 23.97 5.91 1.90 1.14 18.46 38.97 18.20 12.10 2.64 Mining/Manuf./electricity 14.09 26.76 11.83 7.02 9.25 44.04 65.15 34.27 22.93 5.59 Construction 1.58 1.58 1.58 1.58 0.47 0.00 0.00 0.00 0.00 0.11 Commerce 5.53 14.93 5.72 3.45 13.25 30.58 52.39 24.88 15.97 9.60 Transport 37.00 37.00 25.92 18.15 0.34 25.13 33.05 16.32 9.13 0.24 Services 3.82 13.02 3.91 1.66 10.63 25.01 39.72 18.50 11.85 3.69 POSITION/OTHER Employee/worker/cooperative 4.22 11.49 4.32 2.37 15.86 7.94 26.58 7.82 3.74 5.10 Self-employed 11.72 26.13 10.23 5.81 15.08 38.99 60.64 31.06 20.49 14.70 Employer 2.06 6.54 2.91 1.37 1.63 0.00 11.50 0.96 0.08 0.35 Unpaid family work 7.37 14.91 6.16 4.01 2.52 42.49 54.36 31.33 22.70 1.73 Public sector 0.00 1.37 0.30 0.07 5.38 3.34 11.94 2.76 1.35 1.82 Size of firn> 10 people 1.91 6.97 2.18 0.98 12.55 4.09 19.16 5.42 2.48 3.42 UNDEREMPLOYMENT Hours of work <20 13.29 30.25 11.93 6.91 40.60 52.50 71.63 41.77 29.70 60.73 20< hours of work <39 10.35 21.36 9.24 5.54 8.97 35.10 57.87 27.71 17.65 7.36 Morethan39hours 10.58 21.94 9.31 5.55 21.74 37.03 55.87 30.81 21.80 9.09 Want to work more 10.00 26.77 10.74 5.57 1.97 65.99 84.33 43.93 28.11 1.65 Cannot work more health/fam. 5.79 21.41 8.29 3.96 0.40 74.10 74.10 42.76 24.68 0.05 Source: Authors' estimation using EPHPM. 94 Table A1.7: Poverty Profile and Summary Statistics for Urban and Rural Sectors, September 1998 Urban Rural Po Ext. P0 Mod. PI Mod. P2 Mod. Stat. Po Ext. P0 Mod. PI Mod. P2 Mod. Stat. Monthly normalized pc income 4.40 1.82 Log normalized pc income 1.05 0.06 Overall poverty measure 11.76 26.14 10.46 6.00 48.06 67.48 38.63 26.96 LOCATION Atlantida 9.96 20.68 7.72 4.81 6.93 21.56 39.59 17.02 9.81 4.40 Colon 8.34 20.90 7.82 3.83 2.64 26.29 54.82 22.93 12.72 4.35 Comayagua 24.66 38.32 17.77 10.96 4.84 44.84 67.45 35.14 22.46 6.53 Copan 22.48 42.95 18.05 9.97 2.76 63.87 80.66 50.72 37.05 6.78 Cortes 7.79 17.64 6.32 3.39 24.55 20.89 44.92 19.27 11.36 7.11 Choluteca 35.38 65.00 32.78 22.85 5.19 64.35 82.54 55.06 42.75 7.79 El Paraiso 7.73 20.51 7.24 3.51 3.11 42.91 68.61 36.91 25.54 7.97 Francisco 7.46 22.05 7.69 4.11 33.65 52.25 73.28 41.98 29.11 5.94 Intibuco 13.89 31.58 12.70 6.79 0.80 76.64 86.83 61.12 46.75 4.52 La paz 21.58 34.80 16.04 8.94 1.12 58.84 73.47 44.62 32.24 3.57 Lempira 23.66 38.11 20.46 14.50 0.37 58.67 67.34 48.03 38.05 6.98 Ocotepeque 9.87 21.76 9.90 5.66 0.54 31.31 49.16 24.95 17.14 2.63 Olancho 12.32 25.36 9.96 5.44 3.42 48.12 69.76 36.27 23.19 9.55 Santa Barbara 26.06 47.24 22.07 12.51 2.74 56.82 77.93 45.45 31.57 9.07 Valle 21.89 52.68 21.26 12.08 1.62 47.20 64.83 37.79 26.47 3.55 Yoro 9.57 25.07 10.34 5.06 5.72 40.77 58.93 29.78 18.40 9.25 HOUSEHOLD Number of babies 14.63 34.16 13.23 7.31 0.99 54.26 74.71 43.43 30.63 1.29 Number of children 15.75 32.35 13.63 8.08 1.47 54.63 73.72 43.46 30.56 1.97 Number of adults 9.56 21.83 8.59 4.91 3.52 42.64 61.87 34.56 23.90 3.41 Head female 14.81 29.14 11.97 6.73 26.63 47.04 66.45 37.87 26.55 18.39 Age head 45.16 47.12 No spouse 13.92 27.05 11.29 6.43 28.14 44.99 64.08 36.46 25.58 21.12 Birth migration 9.44 24.47 9.11 4.95 54.03 36.49 57.98 29.45 18.80 29.79 Rural migration 9.13 36.24 13.17 5.89 2.73 41.55 63.22 31.19 19.29 1.74 Urban migration 1.19 17.26 2.75 0.87 4.06 36.78 51.89 29.73 20.01 1.34 HOUSEHOLD HEAD EDUCATION No education 24.05 44.68 20.14 12.09 16.43 55.88 75.02 45.01 32.39 34.25 Primary partial 17.81 37.37 14.89 8.35 26.50 49.08 68.46 39.41 27.46 43.20 Primary total 8.28 23.94 8.56 4.66 25.60 38.63 61.05 30.90 20.10 16.85 Secondary partial 3.81 13.02 4.03 1.99 11.16 33.51 50.99 26.75 16.90 2.58 Secondary total 2.58 8.53 2.87 1.56 10.57 9.66 18.87 8.46 5.37 2.29 Superior 2.73 4.20 2.64 2.33 9.74 15.26 21.10 11.46 6.96 0.83 EMPLOYMENT Working 8.81 23.28 8.24 4.30 81.46 47.74 67.16 38.17 26.46 86.93 Available (unemployed) 17.10 51.11 19.55 10.34 1.54 59.38 67.32 44.28 32.55 0.92 Search (unemployed) 54.15 63.56 44.90 36.18 3.20 88.48 96.24 82.87 74.49 0.57 Not working 18.58 33.50 15.01 8.77 15.34 48.47 68.39 39.78 28.31 12.50 SECTOR Agriculture 26.70 46.59 23.21 14.77 10.13 55.06 73.48 43.77 31.11 64.63 Mining/Manuf./electricity 11.26 25.68 10.00 5.45 16.30 32.83 63.77 28.22 16.58 5.54 Construction 11.93 29.30 12.07 8.06 8.10 20.65 42.20 17.06 9.13 2.75 95 Commerce 5.64 18.19 5.61 2.62 21.86 26.71 45.07 22.08 13.98 7.45 Transport 3.22 11.34 2.79 1.36 4.84 16.37 43.40 18.75 11.45 1.10 Services 8.66 21.89 7.80 4.22 23.35 30.63 48.07 24.19 15.81 5.99 KIND OF WORK Employee/worker/cooperative 8.13 23.40 8.29 4.57 45.61 33.38 61.28 27.29 15.66 25.10 Self-employed 16.03 30.48 13.31 7.95 30.81 56.53 72.58 45.04 32.85 58.59 Employer 3.28 10.61 3.15 1.51 7.97 12.02 25.41 9.25 4.70 3.58 Unpaid family work 0.00 14.69 0.48 0.02 0.19 16.70 39.53 24.74 19.53 0.20 Public sector 4.61 16.27 4.79 2.54 8.79 12.43 32.30 11.34 6.43 2.55 Size of firm> 10 people 4.90 18.04 5.65 2.90 33.10 22.45 50.97 19.48 10.16 12.61 UNDEREMPLOYMENT Hoursofwork<20 23.24 37.53 19.24 12.75 21.12 50.08 69.44 41.17 29.78 15.79 20< hours of work <39 14.65 32.98 12.76 7.13 9.41 48.39 66.90 37.76 26.40 15.28 More than 39 hours 7.87 21.75 7.48 3.80 69.48 47.52 67.15 38.24 26.44 68.93 Wanttoworkmore 16.77 38.44 15.74 9.06 5.66 58.48 75.56 44.72 31.47 4.04 Cannot workmore health/fan. 10.65 27.52 10.67 5.24 1.12 75.10 78.41 52.58 40.45 0.60 HOUSEHOLD SPOUSE EDUCATION No education 21.74 50.02 21.65 13.19 9.79 60.63 78.77 47.96 34.43 26.43 Primary partial 17.38 37.48 14.48 8.13 19.89 50.03 69.34 40.17 28.00 32.99 Primary total 8.19 21.12 7.93 4.49 18.07 37.01 61.82 30.48 19.88 14.72 Secondary partial 4.89 18.52 5.97 2.90 9.68 18.97 38.68 16.88 9.47 2.43 Secondary total 1.91 4.33 1.56 0.95 10.36 6.00 6.00 4.07 3.04 2.01 Superior 2.53 2.78 2.54 2.42 4.07 0.00 29.83 6.93 1.81 0.31 EMPLOYMENT Working 7.04 17.78 6.58 3.58 33.47 32.30 52.02 26.32 16.92 18.86 Available (unemployed) 18.65 37.34 15.45 8.61 4.63 57.43 80.38 46.54 33.09 3.99 Search (unemployed) 51.76 57.08 45.02 38.43 0.57 100 100 48.17 23.20 0.03 Not working 13.72 32.39 12.75 7.35 37.83 54.06 73.51 43.25 30.60 59.99 SECTOR Agriculture 11.26 22.34 9.41 5.33 0.41 48.04 71.04 36.79 24.71 2.60 Mining/Manuf./electricity 11.37 25.83 10.23 6.09 8.40 46.37 66.10 37.91 26.02 4.58 Construction 0.00 2.27 0.50 0.11 0.49 0.00 15.69 5.52 1.95 0.24 Commerce 6.18 18.79 6.56 3.53 13.32 27.01 48.17 22.17 13.55 8.60 Transport 0.00 0.00 0.00 0.00 0.28 100 100 47.43 22.49 0.06 Services 7.38 13.42 6.11 3.70 11.15 17.11 32.22 14.44 7.74 3.04 KIND OF WORK Employee/worker/cooperative 3.67 8.73 3.30 1.70 16.14 21.97 41.44 18.42 11.02 4.73 Self-employed 12.96 29.61 12.04 7.36 14.17 37.65 59.49 30.40 19.66 12.67 Employer 3.11 4.01 2.02 1.18 1.21 0.00 0.00 0.00 0.00 0.16 Unpaid family work 7.19 24.45 7.58 3.35 2.53 29.02 36.78 22.07 14.78 1.55 Public sector 0.39 2.08 0.55 0.30 5.02 3.85 9.74 3.23 1.32 1.66 Sizeoffirm>10people 1.94 4.12 1.45 0.70 13.06 16.53 33.50 14.43 8.76 3.51 UNDEREMPLOYMENT Hoursofwork<20 14.45 32.76 13.29 7.85 41.19 53.99 73.39 43.13 30.44 63.22 20< hours of work <39 11.70 27.60 11.35 7.03 8.26 35.94 54.52 29.32 19.76 6.06 More than 39 hours 4.13 12.32 3.88 1.72 22.31 23.43 44.20 19.63 11.61 9.61 Wanttoworkmore 23.55 31.96 17.82 11.51 1.72 36.15 74.85 33.14 21.75 0.44 Cannotworkmorehealth/fam. 19.81 23.74 13.82 8.56 0.50 100.00 100.00 91.32 83.39 0.02 Source: Authors' estimation using EPHPM. 96 Table A1.8: Poverty Profile and Summary Statistics for Urban and Rural Sectors, March 1999 Urban Rural P0 Ext. P0 Mod. PI Mod. P2 Mod. Stat. P0 Ext. P0 Mod. PI Mod. P2 Mod. Stat. Monthly normalized pc income 4.33 1.68 Log normalized pc income 1.03 0.03 Overall poverty measure 12.95 27.31 11.17 6.70 48.51 69.99 39.11 27.12 LOCATION Atlantida 17.68 33.65 15.18 10.14 6.93 37.63 62.15 29.65 19.31 4.40 Colon 12.81 34.61 14.01 8.65 2.65 49.71 72.68 40.96 29.54 4.37 Comayagua 22.85 39.70 18.56 11.80 4.86 48.69 75.21 41.18 28.29 6.56 Copan 16.14 32.82 12.86 7.05 2.76 48.44 68.81 37.99 25.26 6.79 Cortes 6.50 18.10 6.41 3.49 24.49 29.41 54.99 25.28 16.30 7.09 Choluteca 38.32 58.54 29.90 18.38 5.18 53.50 79.92 42.94 29.20 7.78 El Paraiso 5.17 17.32 5.09 3.00 3.11 48.71 67.02 37.86 25.46 7.98 Francisco 9.16 21.06 7.88 4.72 33.68 47.98 66.24 37.13 25.83 5.81 Intibuco 23.68 45.61 18.97 11.37 0.80 75.08 89.31 65.82 54.71 4.53 Lapaz 19.66 36.47 16.42 9.74 1.12 60.10 79.48 47.19 33.33 3.58 Lempira 19.45 35.26 17.27 11.03 0.37 51.02 70.75 40.06 27.38 7.00 Ocotepeque 8.14 19.43 8.83 5.28 0.54 36.89 59.90 30.57 20.59 2.63 Olancho 6.45 26.81 6.85 3.07 3.43 50.90 68.48 38.82 26.62 9.59 Santa Barbara 34.18 61.92 30.22 18.96 2.74 43.41 67.43 34.62 21.77 9.08 Valle 22.15 39.39 18.14 11.52 1.62 47.49 70.37 39.33 26.79 3.55 Yoro 16.92 32.57 13.29 7.46 5.73 50.49 70.73 41.96 30.62 9.27 HOUSEHOLD Number of babies 16.09 34.95 13.75 7.99 0.99 55.22 76.82 43.95 30.78 1.29 Number of children 15.89 32.73 13.64 8.34 1.52 54.56 75.62 43.59 30.57 2.00 Numberofadults 11.06 23.34 9.59 5.75 3.56 43.21 64.85 35.23 24.14 3.50 Head female 13.68 32.01 12.63 7.27 26.20 48.61 70.99 39.56 27.39 18.06 Age head 45.55 47.54 No spouse 14.28 28.30 12.32 7.52 28.46 46.81 69.34 38.23 26.52 21.29 Birth migration 13.10 27.41 11.25 6.76 58.50 43.16 64.89 34.98 23.62 34.96 Rural migration 17.90 39.42 18.38 12.75 2.61 53.09 73.31 38.75 25.25 2.47 Urban migration 8.61 26.52 10.37 5.31 4.79 43.06 64.74 34.63 23.16 1.58 HOUSEHOLD HEAD EDUCATION No education 25.65 46.11 20.07 11.93 16.44 56.91 79.12 45.12 31.39 34.84 Primary partial 19.25 33.90 15.62 9.87 26.95 49.52 69.75 39.72 27.68 43.92 Primary total 8.52 29.44 9.43 5.08 24.69 38.65 61.66 32.62 22.59 15.41 Secondary partial 7.53 18.58 6.85 3.80 11.70 26.32 56.05 23.13 13.41 2.93 Secondary total 4.33 8.88 4.12 2.98 10.53 5.40 24.67 7.73 3.93 2.23 Superior 1.23 2.41 1.12 0.72 9.69 12.37 12.37 10.61 9.13 0.66 EMPLOYMENT Working 11.42 25.93 9.96 5.66 83.96 48.13 69.06 38.53 26.61 88.70 Available (unemployed) 11.62 24.53 10.88 7.53 2.43 54.04 81.64 49.58 35.11 0.99 Search (unemployed) 53.07 63.96 41.23 31.00 2.25 59.01 95.71 59.44 48.36 0.93 Not working 15.76 29.73 13.68 9.05 13.79 50.82 75.63 42.30 29.52 10.37 SECTOR Agriculture 27.60 46.73 22.30 14.14 10.39 55.13 75.08 44.02 31.26 64.99 Mining/Manuf./electricity 9.11 25.12 8.49 4.27 18.48 34.77 59.90 28.38 17.62 5.88 Construction 13.64 33.49 11.01 5.85 7.66 27.70 49.13 21.78 12.87 3.67 97 Commerce 10.66 24.90 9.74 5.79 21.97 27.56 53.83 22.79 12.90 7.52 Transport 5.64 11.88 4.53 2.58 4.63 9.48 39.12 15.00 8.36 1.42 Services 11.30 22.47 9.67 5.92 23.05 35.14 54.69 27.97 18.57 6.10 KIND OF WORK Employee/worker/cooperative 11.29 24.39 9.49 5.51' 44.62 37.48 62.73 29.03 17.41 24.65 Self-employed 16.44 35.08 14.33 8.37 33.11 54.08 74.28 43.82 31.37 61.43 Employer 2.76 7.78 3.16 2.29 8.07 16.86 23.66 13.33 9.48 3.20 Unpaid family work 20.17 20.17 13.40 9.55 0.38 83.03 83.03 70.07 61.65 0.29 Public sector 2.10 10.59 3.10 1.81 8.35 17.01 26.83 11.46 6.91 2.34 Size of firm > 10 people 7.73 18.48 6.81 4.09 33.27 30.48 54.61 25.39 15.84 13.64 UNDEREMPLOYMENT Hours of work < 20 22.12 37.79 19.03 13.04 20.01 50.75 73.70 42.79 30.73 15.63 20< hours of work <39 16.59 31.89 13.08 7.27 9.57 52.01 74.40 41.70 28.86 14.60 More than 39 hours 9.86 23.69 8.68 4.82 70.39 47.25 68.17 37.70 25.89 69.52 Want to work more 20.34 33.62 15.71 10.33 8.06 56.64 76.25 43.77 30.56 6.95 Cannot work more health/fam. 12.34 28.23 10.08 4.72 1.29 64.56 69.91 46.73 34.70 0.74 HOUSEHOLD SPOUSE EDUCATION No education 27.90 49.57 22.68 13.71 10.27 58.12 79.83 47.39 33.93 25.71 Primary partial 15.97 34.38 13.92 8.31 19.48 49.48 70.81 38.89 26.39 33.21 Primary total 11.72 29.75 10.55 5.95 17.72 42.98 63.56 34.47 23.71 15.08 Secondary partial 7.70 19.72 6.66 3.99 9.20 22.03 54.84 22.60 13.24 2.47 Secondary total 0.90 2.42 1.07 0.90 10.27 4.04 5.53 2.50 1.58 1.77 Superior 0.74 2.84 0.78 0.36 4.60 15.00 29.73 15.42 10.86 0.45 EMPLOYMENT Working 9.03 21.09 7.82 4.37 37.52 38.33 60.29 30.85 19.81 26.52 Available (unemployed) 15.41 36.13 14.14 8.53 6.52 59.37 76.05 45.05 30.98 6.95 Search (unemployed) 14.55 26.33 13.58 9.59 0.47 24.34 59.55 25.40 12.43 0.27 Notworking 16.19 33.43 13.92 8.56 33.54 54.53 75.26 43.77 31.17 51.92 SECTOR Agriculture 32.99 42.84 22.96 14.55 0.49 55.17 81.51 43.52 29.47 3.25 Mining/Manuf./electricity 13.62 24.56 9.75 5.50 9.43 48.19 74.95 38.65 24.25 6.06 Construction 0.00 23.84 1.35 0.09 0.57 0.00 76.51 16.94 4.21 0.21 Commerce 7.77 20.13 7.33 4.26 15.29 31.64 49.66 25.22 16.44 12.44 Transport 0.00 69.66 13.07 2.72 0.77 0.00 0.00 0.00 0.00 0.08 Services 7.42 15.83 6.62 3.80 11.41 34.35 55.10 28.01 17.21 4.64 KIND OF WORK Employee/worker/cooperative 4.34 10.86 3.89 2.24 15.49 36.71 58.06 28.42 17.03 6.07 Self-employed 13.89 30.14 11.53 6.40 17.41 38.36 60.25 31.41 20.44 17.50 Employer 0.00 11.28 2.32 0.55 1.20 42.81 42.81 22.94 12.29 0.24 Unpaid family work 9.48 25.19 9.38 5.65 3.86 41.72 66.72 33.18 22.26 2.87 Public sector 0.00 1.50 0.06 0.01 4.99 5.78 12.81 4.98 2.57 1.55 Size of firm> 10 people 3.67 7.79 3.24 2.12 13.06 36.68 55.32 27.51 16.59 4.59 UNDEREMPLOYMENT Hours of work<20 16.91 34.24 14.34 8.78 39.76 54.31 74.66 43.31 30.67 59.07 20< lours of work <39 15.08 23.78 10.49 6.26 8.82 37.13 64.02 31.69 19.89 9.41 More than 39 hours 3.67 15.51 4.55 2.26 22.85 29.14 50.01 23.63 14.51 10.18 Wanttoworkmore 13.84 38.69 13.07 6.28 4.66 37.92 67.78 31.50 18.23 2.11 Cannot work more health/fam. 10.53 24.49 8.55 4.23 1.19 81.03 87.84 44.84 24.33 0.28 Source: Authors' estimation using EPHP'M. 99 ANNEX 2: METHODOLOGICAL ANNEXES MA. 1: MEASURING POVERTY, INEQUALITY, AND INCOME GROWTH IN THE SURVEYS To measure poverty, we use the first three measures of the FGT (Foster, Greer, and Thorbecke, 1984) class. Each measure is computed with both extreme and moderate poverty lines. The first measure is the headcount index of poverty, which is simply the percentage of the population living in households with a per capita consumption below the poverty line. This is denoted by Po. The second measure, which captures the depth of poverty, is the poverty gap index P,. It estimates the average distance separating the poor from the poverty line as a proportion of that line (the mean is taken over the whole sample with a zero distance allocated to the households who are not poor.) The third measure, which captures the severity of poverty, is the squared poverty gap index P2. It takes into account not only the distance separating the poor from the poverty line, but also the inequality among the poor. Denoting by Y; the nominal per capita income for household i, by Z the poverty line (extreme or moderate), by N population size, by w; the weight for household i (equal to the household size times the expansion factor, the sum of the weights being N), the three poverty measures are obtained for values of 0 equal to 0, 1, and 2 in: PO =- lciz (wi /N) [(Z - Yi)/Z] While in table I only headcount indices are reported, higher order measures (poverty gap and squared poverty gap) are provided in Appendix 2. We also make use of these higher order poverty measures in subsequent chapters. Note that the above formula gives poverty measures at the individual level since the weight of each household is proportional to its size. By contrast, the GRH estimates in table 1 are household based, with the sum of the weight (expansion factors) w; being the total number of households in the population. Household level poverty measures tend to be lower than individual level poverty measures, because larger households tend to be poorer. It is better to use individual level measures. To obtain a trend for income inequality, we use three different measures: the Gini, Theil, and Atkinson indices. Denoting by Fi the normalized rank (taking a value between zero for the poorest individual and one for the richest) of household i in the distribution of income, and by Y the mean per capita income, and dropping the weights for notational ease, the three indices are defined as follows: G = 2 cov (Y;, Fi)/Y T =-i - E logI ) A = I 4 I = n i. Y ' Y ni= In the Atkinson index, £ measures the aversion to inequality. Note that while poverty measures are sensitive to adjustments for under- or over-reporting in the surveys to reflect the national accounts, inequality measures are typically not sensitive to these adjustments (and when they are sensitive, the impact of adjustments on the inequality rneasure tends to be very small). Finally, apart from poverty and inequality measures, we provide welfare ratios, which are mean levels of per capita income normalized by the poverty line (extreme or moderate). A welfare ratio equal to one indicates that on average households have income at the level of the (extreme or moderate) poverty line. Economic growth in the surveys (as opposed to the growth observed in the National Accounts) is measured by percentage changes in welfare ratios over time. As is the case for poverty, welfare ratios are sensitive to adjustments for under- and over-reporting. The welfare ratios are defined as W = 1i (w; /N) (Yi /Z). The simplest way to make adjustments for underreporting in the surveys consists in multiplying the welfa-e ratio by the per capita GDP or consumption in the National Accounts and then to divide the result by the income per capita as recorded in the surveys (Yi). More sophisticated methods for taking into account underreporting can be used, but the main message will not change in the case of Honduras. 100 MA.2: ANALYZING THE IMPACT OF VARIOUS INCOME SOURCES ON INEQUALITY To analyze the impact of various sources of income on inequality in per capita income, one can use a source decomposition of the Gini index proposed by Lerman and Yitzhaki (1985; see also Garner, 1993 for an application to inequality in consumption rather than income). Denote total per capita income by y, the cumulative distribution function for total per capita income by F(y), and the mean total per capita income across all households by py. The Gini index can be decomposed as follows: Gy = 2 cov[y, F(y)]/py = £j S,R,G, where Gy is the Gini index for total income, Gi is the Gini index for income yi from source i, Si = pi/gy is the share of total income obtained from source i, and R; is the Gini correlation between income from source i and total income. The Gini correlation is defined as R; = cov [y., F(y)] / cov[(y1, F(yj)], where F(y;) is the cumulative distribution function of per capita income from source i. The Gini correlation R; can take values between -1 and 1. Income from sources such as income from capital which tend to be strongly and positively correlated with total income will have large positive Gini correlations. Income from sources such as transfers tend to have smaller, and possibly negative Gini correlations. The overall (absolute) contribution of a source of income i to the inequality in total per capita income is thus SiRiGi. The above source decomposition provides a simple way to assess the impact on the inequality in total income of a marginal percentage change equal for all households in the income from a particular source. As proven by Stark et al. (1986), the impact of increasing for all households the income from source i in such a way that y; is multiplied by (1 + e;) where e; tends to zero, is: " =S, (RiG, - Gy) This equation can be rewritten to show that the percentage change in inequality due to a marginal percentage change in the income from source i is equal to that source's contribution to the Gini minus its contribution to total income. In other words, at the marginal level, what matters for evaluating the redistributive impact of income sources is not their Gini, but rather the product R1G, which is called the pseudo Gini. Alternatively, denoting by m = R,Gi/Gy the so-called Gini elasticity of income for source i, the marginal impact of a percentage change in income from source i identical for all households on the Gini for total income in percentage terms can be expressed as: vGY I d-i SiRiGi _S5 =Si(77-1) Gy Gy Thus a percentage increase in the income from a source with a Gini elasticity r1 smaller (larger) than one will decrease (increase) the inequality in per capita income. The lower the Gini elasticity, the larger the redistributive impact. The same decomposition can be applied to per capita consumption and its sources. 101 MA.3: DETERMINANTS OF POVERTY: C'ATEGORICAL OR LINEAR REGRESSIONS? It has become a standard practice to anal[yze the determinants of poverty through categorical regressions such as probits and logits. When using such categorical regressions, it is assumed that the actual (per capita) income of households divided by the poverty line, which is denoted by the latent variable y*i, is not observed. We act as if we only know whether a household is poor or not, which is denoted by the categorical variable yj, which takes the value one if the household is poor, and zero if the household is not poor. If we denote by Xi the vector of independent variables (including a constant), the model is: y*; = i'X1 + E; with y; = 1 if y*j >O and yi = O if y*j . O Under the hypothesis of a normal standard distribution for the error term £;, this model can be estimated as a probit. The probability for a household with characteristics Xi of being poor is given by Prob[yi* > O]= Prob[P'X; + £; > 0] = Prob [£; >-P'X;] = F (P'X;) where F denotes the cumulated density of the standard normal distribution. The marginal impact of a change in a continuous variable XA on the probability for household i of being poor, all other variables being held constant, is f(4'XO)f3A, where f is the standard normal density. A coefficient DA positive (negative) implies a positive (negative) effect of an increase in the corresponding variable on the probability of being poor. The marginal probability variations can be measured for any particular value of the Xi vector since f(j3'Xi)PA depends upon Xi. The convention is to compute the marginal effects at the sample mean. If XA is discrete, its impact on the probability of being poor can be obtained by comparing the cumulated normal densities at various values. The main problem with such categorical regressions is that the estimates are sensitive to specification errors. With probits, the parameters will be biased if the underlying distribution is not normal. The alternative is to use the full information available for the dependant variable (indicator of well-being), and to run a regression of the log on the indicator (if its distribution is log normal.) Assume that k*j is the normalized indicator divided by the poverty line, so that k*i = y*i/z, where z is the poverty line. A unitary value for w*j signifies that the household has (per capita) income exactly at the level of the poverty line. Then, we can run the following regression: Log k*j = y'X; + si From this regression, the probability of being poor can then be estimated as follows: Prob[log k*j <0 I Xj] = F[-(y'Xi)/o] where as is the standard deviation of the error terms and, as before, F is the cumulative density of the standard normal. This does not mean that probitllogit regressions should never be used. Categorical regressions will typically have better predictive power for classifying households as poor or non-poor. However, to conduct inference on the impact of variables on poverty, it is better to use linear regression. Another advantage of linear regressions; is that probabilities of being poor can be computed for any poverty line the analyst whishes to use without having to rerun a new regression for every poverty line. This is with region-specific poverty lines valid for urban or rural areas as a whole, or for specific departments within the urban and rural sectors, only the constant and/or the coefficients of the regional dummy variables in the regression will change, and this happens in a straightforward way. 102 MA.4: EDUCATION, LABOR FORCE PARTICIPATION, AND WAGES There are different ways to look at the impact of education on wages. The returns to education presented in Table 2.3 were obtained using the standard Heckman model which can be used to capture the impact of education on both the probability of working and the expected wage when working. Denote by log wi the logarithm of the wage observed for individual i in the sample. The wage w; is non zero only if it is larger than the individual's reservation wage (otherwise, the individual chooses not to work.) The difference between the individual's wage and reservation wage is denoted by A*j. The individual's wage on the market is determined by geographic location (separate regressions are run for the urban and rural sectors), years of experience E, and years of schooling S. There may be other determinants of wages but these are not observed. The difference between the individual's wage and his reservation wage is determined by the same characteristics, plus the number of babies B, children C, and adult family members A of the individual (and their square.) The Heckman model is written as: w; = w*i if A*i > 0, and 0 if A*i < 0 Log w*i = cr. + P WIEi + P w2Ei2 +f w3Si + i w4Si2 + Swi ,&*i=axA+PAlEj+PA2Ei +PA3Si+PA4Si +PA5Bi+PA6Bi +0A7Ci+PAgCi2+0A9Ai+PA1oA; +&Ai = mAi + 8Ai The expected value of swi is not zero. Denoting by (p and $ the standard normal density and cumulative density, and noting that ca, the standard error of &Ai, is normalized to one, we have: E[Log w*i IA*i>0] =aw+pw,Ei+p w2Ei2+0w3Si+,PW4SI2+X(p(mAi)/(I(mAi) E[Log w*i IA*i<0] =ow+0,jEj+Pw2Ei2+Pw3Si+pW4Si2_X(p(Mi)l[ 1 -(mDAi)] If k is statistically different from zero, the returns to education will differ between the employed and the unemployed, although the difference will typically be small. The returns provided in Table 4.4 are computed from the above wage regressions by taking the first derivative of the expected wage with respect to the number of years of schooling. Thus the return to education for year of schooling S is aE[Log w*i]/aS = f3w3+2pw4S when X is zero. The returns are increasing (decreasing) with the number of years of schooling if the coefficient Pw4 is positive (negative.) These returns do not take into account the positive impact on the probability of working of education (i.e., the fact that PA3Si+PA4Si2 is typically positive.) The returns also do not include estimates of the costs of schooling for parents and society (which reduce the returns) and of the indirect effects and externalities associated with education (which typically increase the returns, from the point of view of both the society and the household.) In order to take into account the impact of education on the probability of working, the above regressions can be used to compute the product of the expected wage when working times the probability of working as a function of the level of education reached. This was done to test whether households could expect to emerge from poverty with only one adult male member working (the answer in a nutshell is no). A similar procedure was used for estimating the cost of child labor in terms of foregone future earnings, although with a slightly different sample to estimate the regressions (in this case, the sample includes younger individuals and the results of the procedure are reported in the section on child labor). 103 MA.5: WAGES AND LABOR FORCE PARTICIPATION: AREA VERSUS INDIVIDUAL EFFECTS Differences in wages and labor force participation between departments can be due to differences in the characteristics of the households living in the various departments (e.g., differences in education levels, experience, or demographics), or to differences in the characteristics of the areas in which the households live (e.g., infrastructure, regional development, etc.). Siaens and Wodon (2000) extend a methodology proposed by Ravallion and Wodon (1999) to look at these effects. The first step consists in estimating a Heckman model such as the one described in Box 2.2. In order to capture area effects, apart from the education, experience, and demographic variables, the specification includes departmental dummy variables in both the probit for labor force participation and the log wage regression. In other words, if wi is the wage of individual i when working, Li is the categorical variable indicating whether the individual is working or not, Xi is a vector of individual education and experience variables, Di is a vector of geographic dummies, and Z; is a vector of household demographics, we estimate jointly: Log w; = ,BXi + 8D; + &i i= tXi + (pZi + cD1 + ui The coefficient vectors 8 and a can be estimated so as to represent deviations from the national mean rather than deviations from a reference department. In this case, there is no overall constant in the regressions and the sum of all geographic coefficients in each regression is zero, i.e. 11811 = Ilail = 0. (This facilitates the interpretation of the coefficients and the subsequent manipulations for the simulations, but to do so it is necessary to estimate the regressions twice using standard statistical packages.) Using the regression results, simulations are then conducted to estimate whether it is area or individual effects that are driving the differences in labor force participation and wages between departments. Individual effects. The first set of simuilations consists in estimating the predicted wage and labor force participation in each department using as determinants of the differences between departments only the differences in household characteristics between departments. Dropping the selection terms in the wage equation for simplicity in the notation (these correction terms were included in the empirical work), and denoting by Xd and Zd the sample means of the individual characteristics at the departmental level, this leads to estimates of the expected wage W'd and expected labor participation Lcd in department d as: W,d = E [W; I Xi= Xd] = exp(PXd) Ld = E[L; =1 I Xi= Xd and Zi = Zd ]= F(PX +dpZd), where F is the cumulative standard normal density. The "c" subscript in these estimates stands for concentration whereby the impact of the concentration of individual characteristics in some departments versus others leads to differences in the performance at the departmental level. The numbers shown in table 2.11 under the column "individual effects" are the variance across d of the above estimates. Area effects. The second set of simulations consists in estimating the predicted wage and labor force participation in each department using as deterninants of the differences between departments only the differences in the characteristics of the idepartments. Dropping the selection terms in the wage equation for notational simplicity, and denoting by X' and Zn the national sample means for the individual level variables, and by Dd a vector of zeroes except for the dth department, this is obtained as follows: Wgd = E [W; I X; = Xn and D= Dd] = exp(pXn + 8Dd) Lgd = E[Li=l I lXi= XandZi= Zn and Di= Dd] = F(iX + cpZ + aDd) The "g" subscript in these estimates stands for geographic effects whereby controlling for individual effects, the impact of the geographic effects leads to differences in the performance at the departmental level. In table 2.11 under the column "area effects", we have the variance across d of these estimates. 104 Joint effects. The third simulation consists in finding the impact of both individual and area effects, and computing the variance of the resulting simulated departmental measures. This is obtained from: Wj = E [W; I Xi = Xdand Di = Dd] = exp(pXd + 8D d) Lid = E[L; =1 I X; = Xd and Zi= Zd and Di = Dd] = F(gXd+ 9Zd + aDd) The "j" subscript in these estimates stands for joint effects whereby the impact of both concentration and geographic effects is taken into account to analyze differences in the performance at the departmental level. In table 2.11 under the column "area effects", we have the variance across d of these estimates. 105 MA.6: DOES CONSULTATION IMPROVE PARTICIPATION AND USAGE? The FHIS survey can be used to test whether the consultation of project beneficiaries increases their contribution to the implementation of the projects, and their use of the facilities once the implementation has been completed. Denote the consultation of household i living in area j by the latent variable CON*jj. Only a categorical variable is observed, CONij. Consultation depends on a vector of household characteristics Xi (including a constant), a vector of geographic variables Zj, and the sponsor of the project. The sponsor can be the FHIS, another known organization, or an organization unknown to the household (dummy variable UNKNOW). Ihe regression for the probability of consultation is: CON*ij = yc.n'Xi + 45co'2j + aconFHIS*ii + P,.nUNKNOWN*ii + &coni CONij = I if CON*jj > 0 and CONij = 0 if CON* j < 0 Contribution or participation in the implementation is modeled next and denoted by PAR*¶j. Consultation may affect positively participation. However, one cannot simply estimate a model where consultation is exogenous because households may make special efforts to be consulted, or to attend consultation meetings, if they expect to participate (and'or use the facilities). To assess the impact of consultation on participation without bias, one needs to find an instrumental variable which determines consultation, but not participation conditional on consultation. This variable is obtained by taking as an additional determinant of consultation at the household level the rate of consultation among all the other households in the project area. That is, if Nj is the number of households in community j, one computes a rate of consultation among all other households for each household RCON ij = (1kA3, kej CONij)/(Nj -1). With this additional variable, one can estimate the following probit for consultation: CON*; = Y0.'Xiu + 8.7Z + 0c,nFHIS*ij + PconUNKNOWN*ij + XRCONij + Sconij CONj= 1 if CON*ij > O and CONij = O if CON*ij < 0 Then, one can use the index function from this probit, that is the right hand side of the above equation less the error term, which is denoted by CONHij, as the predicted value for CON*ij in the following probit: PAR*1j = yp&'Xij + 8p&'Zj + axpr FHIS*1, + Owpar UNKNOWN* j + Xpar CONHij + epafij PARij = 1 if PAR*1j > O and PARj = 0 if PAR*Oj < O The parameter X ,,, is the unbiased estimate of the impact of consultation on participation. The final step is to model usage. Again, consultation and participation are likely to be endogenous, since households are more likely to participate (i.e., contribute to the implementation of the project) if they expect to use the facilities. Fortunately, following the same procedure, a participation rate in the project area for all households but the one considered can be constructed, and this is denoted by RPAR ij = (Ik,Li, kej PARij)/(Nj -1). Then, one obtains the expected participation rate at the household level by running a first stage regression, and denoting the index value for participation by PARHij, one then estimates: USE*ijfy..e Xij+.se.Zj+O seFHIS*jj+.use.UNKNOWN*jj +kseCONHij+0usePARHii+6u.eJi USE1j = 1 if USE*1; > 0 and USEij = 0 if USE* < 0 As expected, it was found that the impact of consultation on participation, as well as the impact of consultation and participation on usage is positive and in some cases statistically significant, but it is much lower than would have been estimated without controlling for the endogeneity of these variables. 106 MA.7: ESTIMATING THE COST OF CHILD LABOR IN TERMS OF FUTURE EARNINGS To estimate the cost of child labor, we proceed in two steps (Wodon and Siaens, 2000.) First, we analyze the determinants of child labor and schooling using bivariate probits models for urban boys, urban girls, rural boys, and rural girls. Using bivariate probits generates efficiency gains in the estimation because the correlation between the error terms of the work and schooling regressions is taken into account. It also enables us to compute the probability of going to school conditional on working or not. Denoting by S* and L* the latent and unobserved continuous schooling and work variables, by S and L their categorical observed counterparts, and by X the vector of independent exogenous variables, the bivariate probit model can be expressed as: S * =flAX + Es S = I ifS * >0, S = O otherwise L*=JJIX+EL L =1 ifL* > 0, L = 0 otherwise E[eS] = EfEL ] =0 Var[ES] = Var[eL] = 1 COV[CS,EL] = P The error terms have a bivariate normal distribution. The impact of child labor on schooling can be computed as the difference in the two conditional probabilities of schooling: AP = P(S=1 I L=0, X) - P(S=I I L=l, X) The second step consists in estimating the loss in future income when a child leaves school prematurely. For this, we need to know the probability for a child to work after reaching adulthood, as well as the expected wage when working. The standard model in this category is Heckman's sample selection model described in Box 2.2. After estimating this model (again for the various samples: urban boys, urban girls, rural boys, and rural girls), we compute the future stream of income of the child with two levels of education: 6 years of schooling (primary level) and 9 years of schooling (lower secondary level.) In the first case, the child goes to school up to its 12'h birthday, while in the second case, he/she stays in school up to the 15th birthday. We then make the simplifying assumption that if there is substitution between child work and schooling, the child who is working leaves the school after 6 years of schooling (at age 12), while he could have benefited from 9 years of schooling otherwise (until age 15.) Algebraically, if EW6t and EW9, denote the expected labor earnings (taking into account both the probability of working and the expected wage when working) of the child at age t if he/she has completed respectively 6 and 9 years of schooling, and r is the discount rate (assumed to be 5 percent per year), then the loss AEW in life- time future income for a child not completing 9 years of education due to work is: 65 EW9, 65E EW6, AEW=Z6! t=16 (1 + r) t=13 (I + r) Multiplying the discounted loss in future earnings AEW by the substitution effect between child labor and schooling AP, and dividing by the child's life-time earnings if he/she were to remain in school until age 15, provides the percentage cost PC of child labor in terms of forgone income: 65 E W, PC = AP*AEWI L 9, t=16 (1 + r) This computation rests on a number of assumptions (e.g., it assumes zero costs for schooling itself.) Nevertheless, it provides a baseline estimate of the income loss due to child labor. 107 MA.8: MEASURING THE IMPACT OF GROWTH ON POVERTY AND SOCIAL INDICATORS Impact of growth on poverty The World Bank's 1990 and 2001 World Development Reports on poverty recommend broad-based growth as a privileged path for poverty reduction, provided that it is accompanied by policies to promote access to education, health and social services, and by the provision of safety nets. The poverty reduction impact of growth is obvious enough since holding inequality constant, a rise in living standards must lead to lower poverty. However, inequality needs not remain constant. When growth is associated with rising inequality, part of the gains from growth for the poor will be offset by the negative impact of rising inequality. To obtain an estimate of the impact of growth on poverty in Honduras taking into account the impact of growth on inequality, we ran three very simple regressions using EPHPM data. Denoting by Gt the Gini, by W, the mean per capita income, and by Pt the poverty measure in period t, we estimated: ALog P, = m + yALog W, + 5ALog G, + vt ALog Gt = a + fALog Wt + st ALog IPt = + ?ALog W, + T In these regressions, y is the gross elasticity of poverty reduction to growth (we use the term "gross" because we are holding inequality constant); j is the elasticity of inequality to growth; o is the elasticity of poverty to inequality holding growth constant; and X is the net elasticity of poverty to growth (we use the term "net" because inequality is allowed to change). The following relationship holds: - y+38. Impact of growth on social indicators To obtain estimates of the impact of growth on non-monetary indicators, we used a worldwide panel model because there were not enough data points to run regressions on Honduras alone. In the levels modcel, for example, we regressed the social indicators on a spline function of per capita GDP and urbanization rates using data from the World Development Indicators for the indicators and urbanization, and from the World Penn Tables for the per capita GDP. Denoting by SI1, the social indicator for country i at time t, by Y1, the real per capita GDP in constant US dollars of 1985 (Chain index in Penn Tables), by LYI to LY5 the GDP splines in log (1 for less than $2,500; 2 for $2,500 to $5,000; 3 for $5,000 to $10,000; 4 for $10,000 to $15,000; and 5 for more than $15,000), by LUI to LUS the urbanization rate splines in log (I for less than 20%; 2 for 20% to 40%; 3 for 40% to 60%; 4 for 60% to 80%; and 5 for more than 80%), and by a; the set of fixed country effects, we estimate for each indicator a panel model with a log-log specification: Log SI,t = a+ POLYl it +02LY2it + P3LY3it + P4LY4i, + f35LY5it + y1LUl I, +y2LU21, + y3LU3i, + y4LU4i, + y5LU5it + a; + wit. The 3 and y parameters provide the elasticities of the various social indicators to GDP growth and urbanization at different levels of economic development and urbanization. The data available for the regressions varies in quality. The model has also been estimated in differences in logs (without fixed effects). Tests for differences in the growth effects under expansions and recessions were also implemented. In a large majority of cases, there was no statistically significant difference in the coefficient estimates. 108 MA.9: WHO BENEFITS FROM AN IMPROVEMENT IN ACCESS TO BASIC SERVICES? Consider a country with i = 1, ..., N departments. Within each department, the municipalities are ranked by a measure of per capita income. That is, the municipalities are assigned to one of q = 1, ..., Q intervals in their department, and the same number of intervals Q is used in each department. Denote by xqij the value of social indicator x in municipality j belonging to interval q of department i. The mean benefit incidence in interval q for department i is denoted by Xq; and jqi is the number of municipalities in interval q of department i. To assess how various groups (i.e. intervals) of municipalities benefit from an improvement in the social indicator, we run Q regressions: E X,§X!j x4 =q for =q =1, ... , Q E , - J,q q=1 For the poorest interval (q=l), this yields a regression of the level X'i of the indicator in the poorest municipalities in the various departments on the mean level of the indicator in the departments as a whole with one caveat: to avoid the problem of endogeneity (standard department means are obtained over all the municipalities in the department, including those in the first interval), the right hand side variable is computed at the departmental level as the mean on all municipalities except those belonging to interval q. With this setting, it can be shown that the marginal increase for the indicator in interval q is Qfiq/(Q_l+ ,q), where Q is the total number of intervals. It is important to note that the sum of these marginal impacts must be equal to Q. To estimate the parameters 1q at once, one could pool the data and run a single regression where the intercepts and slopes are allowed to differ between intervals: Q Q xX- I,? X q + [q] += q= q=1,j However, there is a restriction in the estimation of this regression in that the sum of all marginal effects E - = 1 must be Q. Writing eiQ in terms of the parameter for other intervals yields: q=1~~~~~~~~~~~~~~= (Q-, 1- Ql+,l Zq=t 1 + 6 q This restriction can be taken into account by estimating the following using non linear least squares: Q 1 l q-l Q-_l _(Q-1_ _ ) _X xq =Eda4 +,/3q Q=1 | - - q=1 q1 Q ) q=l -+q QJ 109 ANNEX 3: AN EXCEL DIALOG BOX FOR SIMULATING THE IMPACT OF POLICIES ON PER CAPITA INCOME AND THE PROBABILITY OF BEING POOR This annex presents an Excel® Dialog Box which can be used for simulating per capita income and the probability of being poor or extremely poor at the household level, given a set of household characteristics. 1. DESCRIPTION AND USE OF THE SIMULATOR To conduct simulations, the user must specify the characteristics of the household for which the simulations or predictions are requested. This is done by choosing the geographic area in which the household lives, as well as other variables such as the household's demographics, the presence of a spouse for the household head, the education of the household head and the spouse, the migration history of the head, and employment variables for both the head and the spouse. For some countries, additional variables are available, such as whether the household belongs to an indigenous population or not. The choices made by the user for the characteristics of the household are used by the Dialog Box to compute a) the expected per capita income of the household; and b) its expected probability of being poor and extremely poor, which depend on the poverty and extreme poverty lines, also defined by the user for both urban and rural areas. In the case of Honduras presented in this chapter, the results displayed in the Dialog Box are obtained as the average of the results obtained for three different time periods or household labor force surveys (the Encuestas Permanentes de Hogares de Prop6sitos Mzltiples - EPHPM - for March 1998, September 1998, and March 1999). Although the results are calculated individually for each period, only the average results are displayed because they are more reliable statistically. T'he Dialog Box can be used to run simple simulations of the impact of selected policies on per capita income and the probability of being poor. This is because when the user changes any one characteristic of the household, the other characteristics are held constant. The new estimates for per capita income and the probabilities of being poor and extremely poor therefore reflect the marginal impact of the change in the specific characteristic for which the simulations are done. For example, one can see what impact raising the education level of the household head (for a household with given characteristics) has on its per capita income and probability of being poor or extremely poor, holding all other variables constant. The results always appear at the bottom of the Dialog Box. Note, however, that for most policies, more sophisticated models than those underlying the Dialog Box should be used to predict outcomes after a policy intervention. While the Dialog Box gives a feeling for the magnitude of the impact of various household characteristics on income and poverty,, it should be used with caution before inferring policy recommendations. In what follows, Section 2 explains the basic structure of the Dialog Box. and provides examples of simulations. 110 2. DESCRIPTION AND USE OF THE SIMULATOR 2.1. Structure of the Simulator When opening the Excel® file, the Dialog Box appears as shown below. At each use, the Dialog Box displays the results of the latest simulation saved by the user (for the first use of the Dialog Box, the simulation shown in the excel file was saved by the authors of the Box). Figure A3.1 shows the dialog box saved by the authors. Figure A3.1: Dialog box saved by the authors Jj¶Xi. g* ~- )wt*U lads 14ta 'dow 4 :0:_ __ _ .- 'i'X -.o;*..'$.:t ,=3 '~~~ ~~ -ff __ Jj4= ;SH The upper part of the Box contains the information on the household, while the bottom part of the Box contains the poverty lines and the results of the simulation. As mentioned earlier, in the case of Honduras, the average results are obtained from results computed for three time periods (Household surveys for March 1998, September 1998, and March 1999) for urban and rural areas separately, using the estimates of separate urban and rural regressions for each time period. Results are shown for three categories: * Per Capita (Household) Income * Headcount Index for Extreme Poverty (i.e., probability of being poor) * Headcount Index for Poverty (i.e., probability of being extreme poor). 2.2. Options for Simulations The Dialog Box can be run using one of the following two methods: i) Use the Tools, Macros, Macros... and then select Run_-Dialog and click the Run command; or ii) Click on the right button of the mouse and choose Run Dialog in the short cut menu (this is faster, but the user choosing this method for running the Dialog Box must first make sure that the cursor is not on any of the control areas; that is, the cursor must be placed outside of the Dialog Box.) The ill Dialog Box is then ready to accept changes for household characteristics. It looks like Figure A3.2. Figure A3.2: Running the Dialog Box _1 lx...- LeuA AZ 1 He"d jiF P^S H-d ZEC*ed SexofHud HOW .i, IPuoayPrP z Spj IEff*ed H.a IS Jr |Ej* oyed isHd |Ay Erboyed AP ofHUd [0ISJ1 1pI~2D< hours of' work <39 l SpoMrCs (=m -orkvK hmor Wha,trf 1u1be ri .....'-'''20S Sh. .Bables(4yr,.) | l A1 ----w*-° kme- -te Aduts(above,)12 ,1 Hud | 4 Head |PrIctue j . (e912]l-r INl^S. .___ smFl 1 j sin10es . . . ; .~ - . H 1 -S; . - lsea sign (to be agreed with) after the choice is made, and when all choices have been made, the user must click again on at the bottom of the Box to obtain the simulation results (or the user can cancel the simulation to keep the previous results in the Box). 2.3. Using the Simulator: Examples When opened with the results saved by the authors, the Dialog Box has the characteristics of a poor family in Honduras. The department of Lempira is where the household lives. The head of the household is a 40 year old male. The household consists of the head, a spouse, one baby and two children under 14 years old. The household head has never migrated, has some primary education, is a self-employed farmer, and is designated as fully employed in the agriculture sector. The farm has less than 10 workers. The spouse has some primary education, works between 20 and 39 hours on the farm, but is not looking to work more due to household obligations. As can be seen in the results area of the Dialog Box, such a family in a rural area would have an expected average per capita income of 170 Lempiras per month. Since 1 US$ is worth about 15 Lempiras, this corresponds to US$11 per person per month. In urban areas, the expected per capita income is 227 Lempiras. The probabilities of being poor and extremely poor would be at respectively 67 percent and 50 percent in urban areas (of the Lempira department), and 72 and 66 percent in rural areas (Figure A3.2). Geographic Location: When any one of the household characteristics is changed, the Dialog Box will return an new estimate of per capita income and the probability of being poor and extremely poor. The difference between the new estimates and the previous ones then provides the likely impact of this change. Consider the effect of geographic location. If one were to switch from the Lempira department to the Cortes department in the Area box, for instance, the Dialog Box would display a per capita income of 335 Lempiras per month in urban areas and 202 Lempiras per month in rural areas. The probability of being poor would drop to 45 percent and 65 percent in urban and rural areas, respectively, and a similar decrease would be observed for the probability of being extremely poor (Figure A3.3). Figure A3.3: Changing the Household's Geographic Area I ~~~~~~~~~~~~~~~~~~tx F T~~~jcaUanLev~~~~~~ - !WpI lEmoyIn d It SexofEkeci jT ; 5p ;JrMPvlyPUd .ij SP-US f ; ] Spouseprae?te Zi3 I [iJ7:*: - AL'htoumkmare? IHSaI I~ j: Hd XH ImyEODyed He: IF |EulyEd rtoibWF -- - Spm120< ho' s o w ow<39 Cinot worknerelf. , J Wa t vw_ _ _ __d i'r 2 Head |N . HHaad rAi4rie Head _______ SL^i^(tdkpa= No I rSpeou i Lag Years ji NO AHed I< to ypd eJLj Heed kiOpeople , I__ __ ___ -,__ Spouse 5dfeo~I ov4bed :1.jSpowe |>* 1 -al :1: Sed.,.ee -.ai . 540hspIPemr wtuic3 ~jJ ] |we rmoeiSf 4 __4 ___ i i S l _ _ _7 _ _ _,: ' . _ . wee I es Ii . Adts bAWU- I DebI eS ) l i I t -l Ol 1 PWCe 1131B64 | 1~~~~~~~~~~~231.09; I=3 r -- -~- I nPovt() l31.76 - 16.s ZD ,< lo . posr6 oes ~Ieedcof fr* f1IW Poveity(%) __ .._ 118 Employment: The Dialog Box can also be used to observe the impact of employment. Not surprisingly, having a head of household searching for employment has a very large negative impact on per capita income in both urban and rural areas. For our baseline household, changing the employment for the head from "Fully Employed" to "Available, Searching" results in per capita income falling by 115 Lempiras per month in urban areas (and corresponding probabilities of being poor and extremely poor of 93 percent and 85 percent). In rural areas, per capita income decreases by 64 Lempiras, resulting in probabilities of being poor and extremely poor of 83 percent and 79 percent (Figure A3.8). These results probably overstate the impact of unemployment on well-being, because households use smoothing strategies in order to cope with unemployment (the volatility of consumption expenditures is lower than the variability of income because households save and borrow). Still, the fact that unemployment can lead to serious consequences for income is clear. By contrast, it can be shown that in Honduras, households with a head not working have higher levels of income, which suggests that those heads who are not in the labor force can afford not to be working. To some extent, the same is true for the spouse, in that in most cases not being in the labor force does not reduce income. However, if the spouse were to become employed full time, and all the relevant boxes were adjusted to reflect this, our baseline household's per capita income would rise to 212 Lempiras per month in rural areas, an increase of 42 Lempiras per month (Figure A3.9). The likelihood of being poor and extremely poor would decrease to 62 percent and 54 percent. The spouse's entry into the labor market would have a similar impact in urban areas, where per capita income would increase to 246 Lempiras per month, 43 Lempiras per month more than if the spouse were to work only part time. The likelihood of being poor and extremely poor would decrease to 63 percent and 46 percent. Figure A3.8: Changing the Household Head's Employment Status Lii s. o u, _ ____,-__;__-_ _ _. sp-- _ _ __-)4. I I I tow tkm _ _ __=_ _ _ _ _ " L 4yi Leaof i f i d eoa jNotrleboFmce ,' AduLs (above 11P __ i i__ Las . YOMtenNots 19043 ,dNa.N In Lao Foc --1 -Mtrlaoroc Z ! l~~~~~~~~~~FEs LTSo ::!:t l t~~~~~~~~~~V --- ow hxo ,XE1l , In 0 T; 119 Figure A3.9: Changing the Spouse's Employment Status .~~~- , 3 fo t , b S u Pnba of"I2 . .F; 44>¢ pd- ll d '' rvwy*i JH.~ . . ~~~~~~~~~~~~~~~~~'. 7 r;; ..atl . H n d orspous sr l unerel. yd (i.w in ____ ___ ~~~~ ) Iz .I ZJX5 ia week) reduces expected per capita income by about 30 percent in both urban and rural areas. If the number of hours worked per week by the spouse were to fall below 20, the household's per capita income would fall to 134 Lempiras in rural areas and 183 Lempiras in urban areas (Figure A3.10). The likelihood of being poor would be 78 percent and 77 percent in rural and urban areas respectively. T'he likelihood of beinig extremely poor would rise to 73 percent in rural areas and 63 percent in urban areas. If this decrease in the number of hours worked by the spouse resulted in her being able and willing to work more hours, in the "Able to work more?" category, the entry "Cannot work more due to health/family reasons" could be changed into "Can work more." Moreover, the "Want to work more?" box could be changed from "No" to "Yes." This would also result in changes in expected per capita income and probabilities of being poor or extremely poor. 120 Figure A3.10: Spouse Wants to Work More : 'eafr~ .dI, Hd 1P Zw i .: H -Eu.oy:d SOXdHa of Now SPOUPBlPIVPartII Spo'kilable &" 7j S- pe-ert? F-e- - hi. -ko-wa,b e7 Age of Hea :| | -; of.l rlXlb=d,,,-.i; : * o w . j .:u P work < 20 j S ICan workmore O*w (S-L4syr.12 H ea d J >==r4et MAd (above 's'r7 s | S-ouse Se |PiVaI . I ;~~SM Rb. .. . 1 ' ; 1.d of '< ' i,,,l,f, ' >~~ ~ ~ ~~~~~ ~~~~ H.W 150n* '-~ I< ' 105g 10b People sf Ltb P- -fT 7' t2a; '0_20 'ff0fS S0 Povarty Lki pAyf ve 36eg w ;; ~~~~~~~~~~~'ExleoPovrY()'S 7A6SEE X,%S >'5LfGeoi Xf 7oeIb'%,I.Wor.6. Sector: The sector in which the household head and/or spouse are employed can also be changed. Using the Dialog Box, one can see that changing the sector for the head to "Services" results in improvements in the household's per capita income and headcount index. In rural areas, the household's per capita income would rise by 31 Lempiras to 182 Lempiras. The probability of being poor would fall to 71 percent, and to 65 percent for extreme poverty (Figure A3.11). In urban areas, the change in sector would have a more modest impact on per capita income, but similar results with regards to poverty. Per capita income would rise by 18 Lempiras, and the headcount index for poverty and extreme poverty would fall to 68 percent and 52 percent respectively. Different sectors can be tested for both the household head and the spouse with varying results. Figure A3.11: Changing the Sector in which the Household Head is Employed w~~~~~~~~~~~~~~~~~~~~~~]rnpyeA Se ofeaead Spus prewst? :F: D- ;___ AP c 4g He d :" ;3 H.W |FFJ Eqb 7i ~: ' I |t dSiff -Number of.S pouse 120< htss of work <39 J1 5 JCerot workmore- heath/fa.m Oabes (0-4 yr.) | 31WarttoWorkmore?- --edm - _ Childirm (ebO S) . 14t ie yr. 2 Hpea IA5vkue ZJ H S Prwate a J =T 5 h [No :- hc ok- j5edFm : Head HSe.erPadSd : 10 people", and the user would see that per capita income rises to 205 Lempiras in rural areas and 267 Lempiras in urban areas (Figure A3.13). The probability of being poor would fall to 66 percent in rural areas and 58 percent in urban areas. The probability of being extremely poor would decrease to 60 percent and 41 percent in rural/urban areas. 122 Figure A3.13: Changing the Size of the Household Head's Firm \ Area ~~ jLera ,I Head |Ptmary PaatW Haed |Byd SOX of Head |Spxe |Prh E J o DyM 'r0~r i. j -2i_ w:r 5pue2<39g tJ jj S Cret wor n;he s Babes (04w4) I * urtoerokmird-5ectWr : _ more? ab e F J sse Rl Spe I~Wrk. be of F:m > 0Pa* Ziij Heardcmetfor jse .j16soO rExtr~~~~ Po"wty Uw$Pvert (% ti Ste :2W s 4:f 0 123 REFERENCES Banco Central de Honduras. 1998. Memoria. Tegucigalpa, Honduras. Banco Interamericano de Desarrollo. 1999. "Problemas y Oportunidades para el Desarrollo de la Economia Rural." Documento de Trabajo. Washington, DC. Bedi, Arjun Singh and Jason Born. 1995. "Wage Determinants in Honduras: Credentials Versus Human Capital." Social and Economic Studies. Vol. 44, No. 1:145-163. Bedi, Arjun Singh and Jeffrey H. Marshall. 1999. "School Attendance and Student Achievement: Evidence from Rural Honduras." Economic Development and Cultural Change. 47:657-684. Behrman, Jere R., Steven G. Craig. 1987. "The Distribution of Public Services: An Exploration of Local Government Preferences." American Economic Review. Vol. 77, Issue 1:37-49. Bergeron, Gilles, Saul Sutkover Morris and Juan Manuel Medina Banegas. 1998. "How Reliable are Group Informant Ratings? A Test of Food Security Ratings in Honduras." World Development. Vol. 26, No. 10:1893-1902. Bradshaw, Sarah. 1995. "Female Headed Households in Honduras," Third World Planning Review. 2:117-131. Del Cid, Jose Rafael, Ian Walker and Helmis Cardenas. 1998. Sociedady Ambiente: Los Desafios para el Desarrollo sostenible de Honduras. Tegucigalpa, Honduras. Delaney, Patricia L. and Elizabeth Sharader. 1999. "Gender and Post-Disaster Reconstruction:The Case of Hurricane Mitch in Honduras and Nicaragua." Mimeo. World Bank. ESA Consultores. 2000. "Estudio de Costo Eficiencia del Fondo Hondurefio de Inversi6n Social (FHIS)." Mimeo. ESA Consultores. 1999. "Honduras en el Siglo XXI: Una Agenda para la Competitividad y Desarrollo Sostenible." Tegucigalpa, Honduras. Fiedler, John, Tina Saghvi, Beatric Rodgers, Jere Behrrnan, Margaret Phillips, Peter Tatian and Gustavo Saenz. 1995. The Unit Costs and Cost-effectiveness of MCH Food and Cash Transfer Programs in Honduras: An Assessment of the Bonos (BMI) and PL-480 Title II MCH Food Distribution Programs. USAID. Washington, DC. Gobierno de Honduras, Municipalidad de Tegucigalpa, and Banco Mundial. 1998. Informe Preliminar Sobre Estimaci6n de Dan os Del Hluracan Mitch a la Infraestructura Pu blica y Costos de Recuperaci6n. Godoy, Ricardo, Kathleen O'Neill, Stephen Groff, Peter Kostishack, Adoni Cubas, Josephiene Demmer, Kendra Mcsweeney, Johannes Overnan, David Wilkie, Nicholas Brokaw, Marques Martinez. 1997. "Household Determinants of Deforestation by Amerindians in Honduras." World Development. Vol. 25, No. 6:977-987. Government of Honduras. 2000. Reuni6n del Grupo Consultivo: Avanzando en la Ruta de la Reconstruccion y la Transformacion NVacional. Tegucigalpa, Honduras. 124 Govemment of Honduras. 1999. Plan Maestro de la Recontrucci6n y Transformacion Nacional: Estrategia para impulsar el desarrollo acelerado, equitativo sosentible y participativo. Tegucigalpa, Honduras. Government of Honduras. 1998. Plan Nacional de Desarrollo Educativo 1998-2001. Tegucigalpa, Honduras. Government of Honduras. 1997. Teorias y Modelos de Economia Politica, Evaluaci6n del Mercado Laboral y Planificaci6n de Recursos Humanos. Secretaria de Educaci6n - Protecto ASED: Esudio Sectoral Grupo 6. Tegucigalpa, Honduras. Green, Duncan. 1998. Hidden Lives: Voices of Children in Latin America and the Caribbean. Cassell, London. Hall, Gillette. 1999. "Honduras: Statistical Information and Systems for Poverty Work Mission." Mimeo. Humphries, Sally. 1998. "Milk Cows, Migrants, and Land Markets: Unraveling the Complexities of Forest-to-Pasture Conversion in Northern Honduras." Economic Development and Cultural Change. 47:95-124 Inter-American Development Bank. 2000. Social Protection for Equity and Growth. Washington, DC. International Finance Corporation and The World Bank. 2000. "Honduras: The Climate for Foreign Direct Investment." Mimeo. Jack, William. 1999. "The Design and Appraisal of Social Investment Funds." Mimeo. University of Maryland, College Park. Jansen, Kees and Ester Roquas. 1998. "Modernizing Insecurity: The Land Titling Project in Honduras." Development and Change. 29:81-106. Lanjouw, Peter and Martin Ravallion. 1999. "Benefit Incidence, Public Spending Reforms and the Timing of Program Capture." World Bank Economic Review. Vol. 13, No. 2:257-273. Larson, Janelle B. Moritaner, Theodosios Palaskas, and Godfrey J. Tyler. 1999. "Land Titling and Technical Efficiency Among Small Coffee Producers in Honduras," Canadian Journal of Development Studies. Vol. XX, No. 2:361-381. Lopez-Pereira, Miguel A. and John H. Sanders. 1992. "Market Factors, Government Policies and Adoption of New Technology by Small Honduran Farmers: A Stochastic Programming Application." Quarterly Journal of International Agriculture. 31:55-73. Martin, Micheal J. and Timothy G. Taylor. 1995. "Evaluation of a Multimedia Extension Program in Honduras." Economic Development and Cultural Change. Vol. 43, No. 4:821-834. Mausolff, Christopher and Stephen Farber. 1995. "An Economic Analysis of Ecological Agricultural Technologies Among Peasant Farmers in Honduras." Ecological Economics. 12:237-248. 125 Morris, Sautl S. and Juan Manuel Medina Banegas. 1999. "Desarrollo Rural, Seguridad alimentaria del hogar y nutrici6n en el Oeste de Honduras." Archivos Latinoamericanos De Nutrici6n. Vol. 49, No. 3:244-252. Newsgroup, Inc. 2000. Ninios Saludables: El Exito de AIN en Honduras. World Bank. Washington, DC. [Video] Pender, John and Sara J. Scherr. 1999. Organizational Development and Natural Resource Management: Evidence from Central Honduras. EPTD Discussion Paper No. 49. International Food Policy Research Institute. Washington, DC. Phillips, Margaret, Gustavo Sanez, John Fiedler, Beatrice Rogers, Peter Tatian, Tina Sanghvi and Jere Behrman. 1995. The Cost Effectiveness of School Feeding and School Bonos Programs in Honduras. USAID. Washington, DC. Presidencia de la Repuiblica. 1999. Memoria del PRAF (Programa de Asignacion Familiar) Memoria de 1998. Tegucigalpa, Honduras. Programa de las Naciones Unidas para el Desarrollo. 1999. Informe sobre el Desarrollo Humano Honduras 1999: El Impacto Humano de un Huracdn. Tegucigalpa, Honduras. Programa de las Naciones Unidas para el Desarrollo. 1998. Informe Sobre Desarrollo Humano Honduras 1998. Tegucigalpa, Honduras. Progresa. 2000. Esta dando buenos resultados Progresa? Informe de los Resultados de una Evaluaci6n Realizada por el IFPRI. Mexico, Secretaria de Desarrollo Social. Republica de Honduras Secretaria de Finanzas. 1999. Memoria 1998. Tegucigalpa, Honduras. Reyes, Joel and Darlyn Meza. 1999. Una estrategia de eficiencia y participaci6n comunitaria. PROHECO-World Bank. Rodger, Beatrice L., Tina G. Sanghvi, Peter Tatian, Jere Berhman, Miguel Calderon, Sally Crelia and Magdalena Garcia. 1995. Food and Income Subsidies and Primary Schooling in Rural Honduras: An Evaluation of the Impact of the Bonos (7BMJF) and PL-480 - Title II School Feeding Programs. USAID. Washington, DC. SECPLANJUNICEF. 1995. Situation Analysis of Children, Women and Youth: Honduras, 1995. Tegucigalpa, Honduras. Secretaria de Educaci6n. 1997. El VII Censo Nacional de Talla en Escolares de Primer Grado 1997. Secretaria de Educaci6n - Proyecto Ased. 1997. "Teorias y Modelos de Economia Politica, Evaluaci6n del Mercado Laboral y Planificaci6n de Recursos Humanos." Tegucigalpa, Honduras. Secretaria de Educacion Escuela Morazanica. 1997. Educaci6n y Desarrollo: Estudio Sectorial Plan Decenal. Tomo I.. Tegucigalpa, Honduras. Secretaria de Educaci6n Escuela Morazanica. 1997. Educaci6n y Desarrollo: Estudio Sectorial Plan Decenal. Tomo II, Anexos. Tegucigalpa, Honduras. 126 Secretaria de Estado En El Despacho Presidential Unidad de Apoyo Tecnico. 1999. "Estudio Sobre el Gasto En Servicios Sociales Basicos Iniciativa 20/20." Tegucigalpa, Honduras. Schiefelbein, E., Helmis Cardenas and Gloria Maria Palacios. 1997. Financiamiento y Costos de la Educaci6n. Mimeo. Secretaria de Planificaci6n and Coordinaci6n y Presupuesto. 1994. Honduras, Libro Q. Tegucigalpa, Honduras. Smith, Katie. 1994. The Human Farm: A Tale of Changing Lives and Changing Lands. Kumarian Press. West Hartford, Connecticut. Stanley, Denise L. 1999. "Labor Market Structure, New Export Crops, and Inequity: The Case of Mariculture in Honduras." Economic Development and Cultural Change. 48:71-89. Stonich, Susan C. 1992. "Struggling with Honduran Poverty: The Environmental Consequences of Natural Resource-Based Development and Rural Transformations." World Development. Vol. 20, No.3:385-99. UNICEF. 1999. El Trabajo Infantil en Honduras. Honduras. Unidad de Analisis de Politicas Econ6micas. 1997. "Construyendo Nuestro Progreso: Una Propuesta tecnica." Tegucigalpa, Honduras. Unidad Regional de Asistencia T6cnica para el Sector Social. 1996. "Honduras: el Gasto Social y su Eficiencia." Tegucigalpa, Honduras. Villalobos, Carlos, Judith McGuire and Magdalene Rosenm6ller. 2000. AIN-C integreated Child Care: Improving Health and Nutrition at Community Level. IESE and the World Bank. Washington, DC. Walker, Ian and Jon Halpem. 1999. Potable water pricing and the poor: Evidence form Central America on the Distribution of Subsidy and on the Demandfor Improved Services. World Bank. Washington, DC. Walker, Ian, Fidel Ordoinez, Peter Nebel, Sonia L6pez, and Helmis Cardenas. 1996. "Estudio de la Pequefia y Micro Empresa en Honduras." Tegucigalpa, Honduras. Walker, Ian and Max Velasquez. 1999. "Regional Analysis of Decentralization of Water and Sanitation Services in Central America and the Dominican Republic." ESA Consultores. Washington, DC. Walker, Ian, Max Velasquez, Fidel Ordofiez and Florencia Rodriguez. 1997. "Regulation, Organization and Incentives: the political economy of potable water services in Honduras." Inter-American Development Bank, Office of the Chief Economist, Research Working Paper R-314. Webb, Anna Kathryn Vandever, Kye Woo Lee, Anna Maria Sant' Anna. 1995. "La Participaci6n de Organizaciones No-Gubernamentales en la Reducci6n de la Pobreza: Estudio de Casa del Proyecto Fondo Hondurefio de Inversi6n Social." World Bank Discussion Papers No. 295S. Washington, DC. Wodon, Quentin. 1997. "Food Energy Intake and Cost of Basic Needs: Measuring Poverty in Bangladesh", Journal of Development Studies . 34: 66-101 127 Wodon, Quentin, and Mohamed Ihsan Ajwad. 2000. "The timing of program capture: An alternative approach." World bank, Washington, DC. Wodon, Quentin, Gilette Hall, and Laura Rawlings. 2000. "Promoting community participation: Consultation, contribution and usage for a social investment fund." World Bank, Washington, DC. Wodon, Quentin, with contributions from R. Ayres, M. Barenstein, N. Hicks, K. Lee, W. Maloney, P. Peeters, C. Siaens, and S. Yitzhaki, 2000, Povertv and Policy in Latin America and The Caribbean, World Bank Technical Paper No. 467, World Bank, Washington, DC Wodon, Quentin, Rodrigo Castro-Fernandez, Gladys Lopez-Acevedo, Corinne Siaens, Carlos Sobrado, and Jean-Philippe Tre. 2001. Poverty in Latin America: Trends (1986-1998) and Determinants. Cuadernos de Economia. Forthcoming. World Bank. 2001. Honduras: Public Expenditure Management for Poverty Reduction and Fiscal Sustainability. Washington, DC. World Bank. 2000. Securing our Future in a Global Economy. Washington, DC. World Bank. 2000. Ayuda Memoria de la Misi6n Conjunta de Supervisi6n de Preparaci6n de una Nueva Operaci6n con el Fondo Hondureno de Inversi6n Social. Mimeo. World Bank. 1999. Memorandum of the President of the International Development Association and the International Finance Corporation to the Executive Directors on a Country Assistance Strategy of the World Bank Group for the Republic of Honduras. Washington, DC. World Bank. 1998. Honduras: Toward Befter Health Care for All. Washington, DC. World Bank. 1995. Honduras: Reforming Public Investmnent and the Infrastructure Sectors. Washington, DC. World Bank. 1994. Honduras Country Economic Memorandum/Poverty Assessment. Report No. 13317- HO. Washington, DC. Wright, James D., Martha Wittig and Donald C. Kaminsky. 1993. "Street Children in North and Latin America: Preliminary Data from Proyecto Altemativos in Tegucigalpa and Some Comparisons with the U.S. Case." Studies in Comparative International Development. Vol. 28, No. 2:81-92. Yuinez-Naude, Antonio, and Raiil Hinojosa-Ojeda, eds. 2000. Cambio estructural y aperlura comercial en America Central, en la Republica Dominicana y en Norteamerica: en enfoque de equilibrio general aplicado. El Colegio de Mexico: Mexico, DF.