Report No. 20208-BR Brazil As Assessment of the Bolsa Escola Programs March 15, 2001 Human Development Sector Management Unit Brazil Country Management Unit Latin America and the Caribbean Regional Office Document of the World Bank CURRENCY EQUIVALENTS Currency Unit = Real (R$) US$1.00 = R$1.87 FISCAL YEAR January I - December 31 ACRONYMS AND ABBREVIATIONS AM - Amazonas AP - Amapa DF - Distrito Federal IBGA - Instituto Brasileiro de Geografla e Estatisticas IDB - Inter-American Development Bank IPEA - Instituto de Pesquisa Econ6mica Aplicada MW - Minimum Wage PETI - Programa de Erradicaqdo do Trabalho Infantil PGRM - Programa de Garantia de Renda Minima PME - Pesquisa Mensal de Emprego PNAD - Pesquisa Nacional por Amostra de Domiilios PRAF-BID - Programa de Asignaci6n Familiar Banco Inter-Americano de Desarrollo PROGRESA - Programa de Educaci6n, Saludy Alimentacion SAEB - Basic Education Evaluation System (Sistema de Avaliacdo da Educaqdo Basica) SM - Salario Minimo SP - Sao Paulo TO - Tocantins UFIR - Unidade Fiscal de Referencia Vice President, LCR: David de Ferranti Country Director, LCC5C: Gobind T. Nankani Sector Director, LCSHD: Xavier Coll Team Leader, LCSHD: Guilherme Sedlacek BRAZIL An Assessment of the Bolsa Escola Programs Table of Contents Executive Summary ............................................................... i 1. Background of the Programs ......................................................, 1 2. The Rationale for Bolsa Escola .................................................................. 2 A. Background .................................................................. 2 B. Why Bolsa Escola ............... 4 3. Selection Criteria and Targeting ................................................................. 7 4. Design and Implementation .................................................................. 12 5. Impacts-Preliminary Evidence ........................ 14 6. Evaluation Instruments .................................................................. 17 7. Monitoring Instruments .................................................................. 18 8. Political and Fiscal Sustainability ............................................................... 19 9. Expansion .................................................................. 21 10. Conclusions and Outstanding Issues ............................................................. 23 References .................................................................. 51 Boxes Box 1.1 Program History ...................................................................1 Box 2.1 Poverty and its Many Faces in Brazil ....................................................... 2 Box 2.2 The Bolsa Escola Programs are part of a Larger Trend of Poverty-Targeted Cash Transfer Programs in Latin America ........................................................ 7 Box 3.1 Targeting the Poor -The Income Means Test and Scoring System ...................... 8 Box 3.2 The "Non-Covered Population" .....................................................1.......... 0 Box 3.3 The Dilemma of the Residency Requirement .............................................. 11 Box 9.1 The Child Labor Eradication (PETI) Program ............................................. 23 Data Tables Table 1. I School Scholarship Program Coverage in Different Areas ..................................... 26 Table 2.1 Selected Indicators -Metropolitan Nuclei and the Federal District 1996 .................... 26 Table 2.2 Enrollment Rates for Working and Non-Working Children in Urban Areas ................ 27 Table 2.3 Lag in Grade for Working and Non-Working Children in Urban Areas ..................... 27 Table 2.4 Incidence of Child Labor in Urban Areas - Percentages ........................................ 27 Table 2.5 Child Labor in the Federal District - 1996 ....................................................... 28 Table 2.6 Child Labor, School Attendance and Overlap By Region ...................................... 28 Table 2.7 Children Out of School, According to the Level of Family Income ........................... 29 Table 2.8 Child Labor in the Ten Metropolitan Areas with Bolsa Escola ................................ 29 Table 3.1 Basic Criteria for Benefit Concession Federal District, Belo Horizonte, & Belm .30 Table 3.2 Number of Families According to the Age of Children ......................................... 30 Table 3.3 The Effects of Selection Criteria on the Number of Families Covered by the Program.. 31 Table 3.4 Reduction in Potential Beneficiary Families Due to Residency Requirements and Estimates of Program's Expenditures as Proportion of Current Receipts ................... 31 Table 4.1 Number of Families with 7 to 14 Year Old Children Whose per Capita Income is Below .5 MW and Estimates of Program-Expenditures as Proportion of Current Receipts According to the Value of the Family Benefit ........................................ 32 Table 4.2 Fiscal Costs of Giving 1/2 Minimum Wage Based on the Number of Children in the Family ........................................................................ 32 Table 5.1 Dropout Rates among Beneficiaries and Non-Beneficiaries (%) ............................. 33 Table 5.2 Promotion Rates Estimates - 1995-1996 ......................................................... 33 Table 6.1 Learning Performance of Beneficiaries and Non-Beneficiaries Selected from the SAEB Sample Fourth Graders - 1997 .................................................................... 33 Table 8.1 The Effects of Selection Criteria on the Fiscal Sustainability of the Program .............. 34 Table 8.2 Program's Cost as % of Current Municipality Revenues by Level of Transfer to Families ........................................................................ 34 Table 1-11 General Program Data ........................................................................ 37 Table 2-II General Program Data for the State of Sao Paulo ............................................... 38 Table 3-II Program Selection Criteria ........................................................................ 39 Table 4-II Program Selection Criteria for the State of Sao Paulo .......................................... 40 Table 5-II Program Requirements of the Children .......................................................... 42 Table 6-II Program Requirements of the Children in the State of Sao Paulo ............................ 43 Table 1-III Time of Residence in Paranoa, D.F ............................................................... 44 Table 2-III Family Composition - Paranoa .................................................................... 45 Table 3-III Labor Status of Potential Beneficiaries and Non-Beneficiaries - Paranoa ....... ........... 46 Table 4-Ill Average Family Income of Beneficiaries ........................................................ 46 Table 5-IlI Income Per Capita ........................................................................ 47 Table 6-Ill Selection Criteria of Program Applicants in the F.D. - 1995 .................................. 47 Table 7-Ill Income, Family Size, and Intensity of Poverty Comparison between the Selected Population (Beneficiaries) and the Parameter Population ..................................... 48 Appendices Appendix I Data Tables ..................................................................... 26 Appendix II Program Criteria, Design and Implementation ............................................ 35 Appendix III Evidence of Program Targeting in the Federal District .................................. 44 Annexes Annex 1 Table 1: Scoring Criteria for the School Scholarship Program Federal District and Belem Selected Topics .................................................................. 49 This report was prepared by a team comprised of Guilberme Sedlacek, (task leader, LCSH1D), Emily Gustahsson-Wright, Nadeem Ilahi (IW) and Margarita Lannon (Program Assistant, LCSHD). The paper benefited frxn background papers by Alexandrina Sobreira de Moura, Lena Lavinas, and Sonia Rocha. While the report is the responsibility of the World Bank alone, we also would like t:o thank the Government of Brazil for its collaboration We also would like to thank Suman K. Bery (LCCWA), Ana-Maria Arriagada, Indermit Gill (LCSHD), Eliana Cardoso (LCSPE), Charles Criffin (LCSHH), P. Zafiris Tzannatos (MNSHD), Chris Patel (LCC5A), Joachim von Amsberg, Mark Thomas, Dorte Verner (LCC5C), Margret Grosh, Bona Kim (HDNSP); Harry Antiony Patinos (1DNED), Harold Alderman, Martin Ravallion (DECRG), Emmanuel Jinemez (SASED), Peter Orazemn (Iowa State University), Emmanuel Skoufias (EFPRI), Nora Lustig, Arianna Legovini, Suzanne Duryea, Ruthanne Deutsh (I)B), Armand Pereira (ILO-Brazil), Reiko Niimi, Arabela Rota, Sizue Imanishi (UNICEF), Cristovam Buarque (Pres. NGO Missio Crian9a), Ricardo Paes de Barros, Lena Lavinas, Sonia Maria Rocha (IPEA), Alexandrina Sobreira Moura (Sectma - PE), and Francisco Ferreira (PUC RJ-Brazil) for their inputs and comments. Finally, we would like to thank the many individuals in Brazil who contributed tine, infomation and insight. EXECUTIVE SUMMARY i) The Bolsa Escola programs are poverty-targeted social assistance programs that give cash grants to poor families with school-age (7-14 year old) children. These grants are given on the condition that the children attend school a minimum number of days per month. These programs began in 1995 at the municipal level in Campinas and the Federal District of Brasilia. By 1999, there were 60 programs in operation in various urban municipalities. Over the same period, and with design similar to the Bolsa Escola programs, two major Federal Programs (FGRM-Fundo de Garantia da Renda Minima and PETI-Programa de Erradicacao do Trabalho Infantil) have been instituted. By the end of 2000, PETI's coverage had reached close to 400 thousand children, while FGRM had a coverage estimated at 2 million households. Recently, the Brazilian Government has moved to scale-up (to reach all municipalities with the worst human development indicators by 2002) and integrate the FGRM and PETI cash-grant initiatives under one coherent program, the Alvorada Program. These programs have four objectives. First, they hope to increase educational attainment among today's children and thus reduce future poverty. Second, by restricting the grants to the current poor, the programs aim to reduce current poverty. Third, by requiring children in beneficiary households to have minimum attendance in school, the programs aim to reduce child labor. Last, an implicit objective is that by providing income support to poor families, they act as a partial safety net-i.e., they prevent these families from falling further into poverty in the event of an adverse shock. ii) The rationale for these programs is best understood by considering the current picture of poverty in Brazil. Poverty is pervasive in Brazil, and one of the strongest correlates of current poverty is low education. Thus, improving the education of children is an effective way of preventing future poverty. Enrollment of children is high, but children enter the school system late and attend infrequently; leading to slow progression through school. The average school-going child in Brazil is behind his counterparts in other Latin American countries. One reason for the slow progression is the poor quality of education services and another is child labor. The majority of workin.g children are enrolled in school. Working children tend to fall behind much faster than those who do not work. As educational attainment and child labor are intimately associated with poverty, there is a strong rationale for programs like the Bolsa Escola. But the focus of this report on the demand size programs does not imply ignoring supply side constraints. Additional improvements to the access and quality of education would complement the necessary demand-side intervention provided by Bolsa Escola. Supply-side interventions in education are especially critical in the poorer areas of Brazil. iii) The Bolsa Escola programs have widespread support and they fit well into Brazil's larger social protection and poverty reduction strategies. Because they explicitly target the poor, the Bolsa Escola programs have an edge over others (such as school feeding programns and other untargeted subsidies) that do not. At the same time, i by focusing on future poverty (children) they complement other programs for poverty reduction that focus more on current poverty (such as provision of basic services). The Bolsa Escola programs have generated strong commitment within and outside the government. The local schools are important stakeholders and various agents of civil society are actively involved in monitoring and enforcement of these programs. This promises better success for these programs than other social protection programs. iv) The programs have sound criteria for selection and targeting. Within municipalities, the programs target the poorest localities first. For the selection of beneficiaries, the programs go beyond a simple income means test. They use a score system which comprises various indicators of the family's living standards. Two issues merit attention. First, there is potential for improvement in the score system, especially in customizing it by municipality. Second, since these programs are implemented at the municipal level, there is no mechanism to ensure that they target the poorest regions in Brazil. Therefore, there is a role for the federal government to encourage the installation of these programs in the poorest municipalities. v) Preliminary evidence (available from the program in Brasilia) suggests that the programs have been reasonably successful in targeting. First, leakage rates (i.e., the non-targeted benefiting from the program) have been low. A comparison of the beneficiary population with the parameter population (those limited by selection criteria) and non-beneficiaries (those that do not meet program criteria) reveals that the program beneficiaries are indeed the poorest. However, under-coverage rates (the targeted population that is missed by the program) have also been high. This is partly due to the weak financial capacity of the poorest municipalities. vi) Preliminary evidence also points to improvements in education and poverty outcomes, but evidence on reduction in child labor is inconclusive. First, school attendance is higher and dropout rates are lower for beneficiaries. More beneficiary children enter school at the right age compared to their non-beneficiary counterparts. Second, as a consequence of good targeting, the program reduces the income gap between beneficiaries and non-beneficiaries, implying short term poverty reduction. However, it has been difficult to ascertain the effects on child labor. This is partly because in Brasilia, where the evaluation was undertaken, there is little child labor. vii) A design-related issue that requires attention is the need to reach the "non- covered" population. Through their design the programs create two types of "non- covered" groups. First, because the programs explicitly focus on poor families with children of school age, they tend to exclude poor families with only pre-school and adolescent children. Second, most of the programs impose a minimum residency requirement on beneficiaries, ostensibly to discourage in-migration. By doing so, however, they can end up excluding the poorest. The exclusion of these groups raises concerns about equity. The report argues that for the benefit of the latter group, the residency requirement should be reduced to one year. For the former groups, this report suggests two options. First, other programs for pre-school children and adolescents could be run in tandem with the Bolsa Escola Programs. Second, the role of the Bolsa Escola Programs could be expanded along the lines of the Progresa program in Mexico to ii include poor families with pre-school and adolescent children and catering to the needs of these groups such as providing nutrition support along with school grants. Any expansion should ensure the programs still stay rooted in current administrative set up (such as pre-schools) so that compliance with the program (minimum attendance) is monitored. viii) The level of the cash transfer needs to be determined carefully. There is a wide variation in the cash transfer level between municipalities which leads to the question of which is the most appropriate level. There are three issues to consider: First, it should be clarified what the cash transfer is meant to do-does it compensate beneficiary families for lost wages, or does it pull them up to the poverty line? Since the objective chosen is likely to result in different levels of the cash transfer, there is a need to go back to program objectives and determine which one is paramount. A conservative approach, and that suggested by this report, would be to link the cash transfer directly with the foregone earnings of the child. The second issue is that the programs currently use a cash transfer that does not vary by recipient. There may be gains from tailoring it to subgroups in the population. For instance, older children are likely to have a higher opportunity cost of going to school and it may be better to have the cash transfer increase with age. This is essentially the design of the Progresa program in Mexico. Third, some of the Bolsa Escola programs give one cash transfer per family rather than per child. This is likely to reduce the incentive of families with many children from participating in the program. We suggest that the cash transfer should be based on the number of children in the family. While attempts to tailor the cash transfer will be financially efficient, it is also likely they will add to the cost of administration, since it is expensive to identify and monitor different subgroups. However, more evaluation is necessary to specify the optimal level. ix) The Bolsa Escola Programs should not be seen by local governments as substitutes for investments in schools. Poor educational attainment in Brazil is a consequence of both demand failure (poverty) and supply constraints (poor school quality). While school quality is not a major issue in Brasilia-where Bolsa Escola was first implemented-it is likely to be important in other urban areas, particularly the poor rural ones. Hence, local governments should not see the Bolsa Escola. programs as replacing other education-related interventions. This report recommends that the Bolsa Escola programs follow the example of other programs in Brazil (PETI) and in the region (Progresa in Mexico) that provide explicit financial support to local schools. x) Perhaps the biggest consideration in the successful expansion of these programs is their fiscal affordability. Brazil today is highly decentralized. Not only are programs such as the Bolsa Escola implemented at the municipal level, but their financing also comes from local sources. This has created a particularly perverse problem in that the poorest municipalities are the ones that need these programs the most (since they have a higher incidence of poverty) but are unable to afford them (since they also have low local revenues). The calculations in this report reveal that except for the richest urban municipalities, the program would amount to a fairly large burden on local finances. There is evidence even today that in some of the poorer municipalities where the programs are being implemented, shortages of funds have resulted in under-coverage of iii the deserving population. There are two ways of resolving the financing problem. One is to alter the selection criteria or the size of the cash transfer to reduce the size of the beneficiary population or the arnount of support respectively. The other is to top-off local deficit through financing from the federal government. This report recommends the latter because changes in benefit size will ultimately compromise the effectiveness of the program. In the extreme, such adjustments would start affecting the effectiveness of these programs in reaching their objectives. Federal transfers are perhaps the best option. However, we recommend that to preserve the incentives of local municipalities to efficiently run the programs, there should be a local share that each municipality must pay. xi) These programs are only partial safety nets-they protect the structurally poor during crises and equip the next generation with risk reducing human capital. There are two ways in which these programs serve the purpose of a safety net. Economic crises ca seriously affect the health, nutrition and educational outcomes of the extreme poor. Thus, by supporting the structurally poor, these programs provide them with an assured source of income and smooth income shocks. Second, higher levels of education reduces vulnerability to future economic shocks. Hence, by encouraging human capital investments, these programs better prepare the poor for future shocks. These programs act as a partial safety net because they do not cover the transitory poor-those who are near poor now but can fall into poverty as a result of an economic shock. There is a need to explore how to design complementary programs to address the needs of the transitory poor. xii) The conclusions of this report have implications for the debate of program expansion. Based on existing evidence, this report finds that the Bolsa Escola programs are appropriately designed and well administered. They have a role to play in the larger social assistance strategy of Brazil. They are likely to remain successful because they enjoy broad support in the government as well as in civil society. However, the concerns raised in this report here should be kept in mind when replication or expansion of these programs is being considered. Some have been discussed above. First, adjusting the size of the cash transfer or the selection criteria in order to make the programs affordable, can compromise the effectiveness of the program. Second, the administering municipalities and the federal and state governments need to agree on cost-sharing for the programs. A related issue is that the federal government needs to make its own calculations to prioritize geographical areas where interventions are needed most so as to remain consistent with the objective of reaching the most needy in Brazil. Third, there is a need to either expand the existing Bolsa Escola programs to include the "non-covered" populations or serve them with other complementary programs. Fourth, the Bolsa Escola programs should complement rather than substitute investments in school quality. Fifth, since the programs act only as a partial safety net, there is a need to examine how to reach the transitory poor through complementary programs. Last, but not least, there is a need to carry out a broad-based and systematic evaluation of these programs. The existing evaluation, which only covers the Federal District of Brasilia, is not enough. A formal evaluation, such as the one for Progresa in Mexico, should include control and experiment groups that are tracked over time so that it is possible to make a scientific iv judgment on the programs' effectiveness.' Only with a proper evaluation can the true effectiveness of these programs be determined. xiii) Programas Alvorada and PETI. Recently, the Brazilian Government- has moved to scale-up and integrate separate cash-grant initiatives financed by the federal government (FGRM-Fundo de Garantia da Renda MInima and PETI-Programa de Erradicagilo do Trabalho Infantil) under the Alvorada program. This program will seek, over the next few years, the gradual implementation of a Bolsa Escola program in poor mun.icipalities with the worse indices of social development in country. This policy initiative represents a major step forward in program design and fiscal commitment, it integrates within one coherent program both preventive initiatives (Bolsa Escola/FGRM with two million beneficiaries in 2000) and the remedial child-labor eradication program (PETI, with four hundred thousand beneficiaries in 2000). Program Alvorada will permit the Government to address design and implementation issues hindering the effectiveness of the FGRM program - very low cash transfer level, complex benefit design and target exclusion of many poor municipalities in the Northeast and North. By building on the complementarity of existing conditional cash transfers to the poor, and improving coordination with educational initiatives aiming at universalizing basic education, it has the potential do significantly increase targeting and effectiveness. 'Progresa (1999) Evaluaci6n de Resultados del Programa de Educaci6n, Salud, y Alimnentacion.-Primeros Avances. v 1. BACKGROUND OF THE PROGRAMS 1.1 The Bolsa Escola programs are poverty-targeted social assistance programs that give cash transfers to poor families with school-age children. In return for the transfer, the beneficiary families agree to keep their children in school. Specifically, the cash grants are given on the condition that children will attend school at least 90 percent of the required number of school days per month. 1.2 The four central objectives of the programs are to: * Increase educational attainment and subsequently reduce poverty in the long term. The main objective of the programs is to increase educational attainment among current school-age children (7-14). The ultimate hope is that higher educational attainment will lead to higher future incomes, and thereby decrease the chance that these children will grow up to be poor * Reduce short-term poverty. By restricting the grants to families that are poor now, the programs aim to reduce current poverty. * Reduce child labor. By requiring children who benefit from the program to have minimum attendance at school of 90 percent, the programs aim to reduce the time children would have available to engage in labor, thus reducing the incidence of child labor among the target group. * Act as a potential safety net. Though the programs are not a full-fledged safety net at this point, they do have the potential of being one. By providing income support to poor families the programs prevent such families from falling further into poverty in the event of a negative shock such as an economic recession. BOX 1.1 Program History * Since the early 1970's, minimum income programs have been proposed as a poverty alleviation tool in Brazil. The Bolsa Escola programs are part of a larger government social assistance strategy. * In 1995, the first of these types of programs was implemented by the local governrient in Campinas, as an additional component of the social assistance network. At about the same time it was also installed in the Federal District (Brasilia), as largely an educational program. * By November, 1998 there were more than 60 programs in operation (four of which were by states: Amapa (AP), Amazonas (AM), Distrito Federal (DF), Tocantins (TO); and the rest by Municipalities. Table 1.1 in Appendix I lists the coverage of the programs in three metropolitan areas). 1 2. THE RATIONALE FOR BOLSA ESCOLA 2.1 This section assesses the current situation of poverty, school attendance and child labor at the national level in Brazil as well as in those metropolitan areas where expansion is being considered. In addition, this section considers alternative types of social protection programs in Brazil. It demonstrates why the Bolsa Escola can easily complement some of these programs and is perhaps more appropriate than some. BOX 2.1 Poverty and its Many Faces in Brazil * Brazil has a poverty headcount ratio of 23%, which is high for a middle-income country. * There is a strong inverse relationship between education level and poverty rates in Brazil. About 90% of household heads have not graduated from high school, and 74% have not completed 4Ih grade. * School attendance of children is negatively associated with their poverty status. Brazil lags behind other countries in Latin America in children's grade-for-age. * Child labor is common-about 14.9% of children in the 7-14 age group work. Child labor is directly associated with poverty-the incidence of child labor is 20% in the bottom quintile of the income distribution, but only 5% in the top. A. Background 2.2 Poverty is pervasive in urban Brazil, especially in the North and Northeast. The overall incidence of poverty in Brazil is approximately 22%. For the purpose of this report, it is better to focus on the incidence of poverty in the urban metropolitan areas where the Bolsa Escola programs are in place or are being proposed. Table 2.1 in Appendix I presents the distribution of the poor in the ten examined cities by income as well as by access to basic services. There are wide disparities in the distribution of urban poverty along regional lines-i.e., between the North/Northeast and the Center South. In the poor Northeast, average monthly per capita family income is as low as R$254 (Fortaleza) and R$285 (Recife). In the more affluent Center South, it is as high as R$1,563 (Porto Alegre) and R$1,589 (Curitiba). At the poverty line of 1/2 of the minimum wage per capita, the incidence of poverty is the highest in the Northeast, such as in Salvador (23%) and Fortaleza (22%), and the lowest in Southern Curitiba (3%). 2.3 School enrollment is high, but children enter the school system late and progress slowly. Most children in urban Brazil today are enrolled in school. Household survey data (PNAD, 1998) indicate that enrollment rates are as high as 2 96%, with little difference between boys and girls (Table 2.2 in Appendix I). Enrollment is not a problem, but low quality of education, late entry into school and progression through the school system are the main reasons for low educational attainment among Brazilian children. Only 62% of the children entering the school system in the Federal District-a relatively mnore wealthy metropolitan area-are of the right age foe their grade, the rest are old for their grade. As many as 8% enter the school system four years late. Partly from late entry and partly due to irregular attendance, Brazilian children tend to fall behind while progressing through the school system. Calculations from the household survey reveal that the educational attainment of children in urban EBrazil is only 60% of what it should be for their age (Table 2.3). In other words, a typical 13 year-old who should have completed 6 grades (assuming he/she entered school at 7 years of age) is currently in 4th grade-behind for the age by about 40%. Boys tend to lag behind more (45%) than do girls (38%). Part of this gender disparity in education lags may be explained by child labor, in which boys are over- represented. 2.4 The vast majority of working children are enrolled in school but child labor contributes to slow progression (due to repetition), and hifh drop out rates. The incidence of child labor in urban Brazil is 9% (Table 2.4). It is twice as high for boys (12%) than for girls (6%). The vast majority (80%) of working children are enrolled in school (Table 2.2) and a large proportion of these children work more than 15 hours per week (Table 2.5). Children who work exhibit a lag in attainment that is 18% greater than that of those who do not workc (Table 2.3). Among other factors, this is due to the direct time that labor takes away from school and school work, the physical and mental exhaustion and the health effects the child may experience from working. Thus, even though child labor is not incompatible with school enrollment, work seems to slow progression through the school system (Table 2.6). 2.5 Both education and child labor are intimately associated with poverty. While child labor and low educational attainment are problems in themselves, they are also intimately associated with poverty in Brazil (Table 2.7). First, there is a strong negative correlation in Brazil between education levels and poverty. About 90% of heads of poor households have not graduated from primary school and 74% have not completed fourth grade. The key missing asset among poor households is human capital-education. Even though attendance differences are small between poor and non-poor children, children in poor households have much lower grade attainment than those in non-poor ones. Child labor is also closely associated with poverty. The incidence of child labor in households in the bottom 20% of income distribution is 22%, compared to only 6% in the top 20%. The problem of poverty and child labor is more stark in rural than in urban Brazil (Table 2.8). 2 Child labor is defined here as work that children do, both for wages as well as on the family farm or enterprise. It should be noted that most of the work girls do consists of household chores, an activity that goes largely unaccounted for in the PNAD survey data. 3 B. Why Bolsa Escola? 2.6 If the quality of the education system is inadequate, then why not just give more money to the schools? It is indisputable that many Brazilian schools, especially in poorer areas, have low quality of education as well as poor infrastructure. It may be argued that the levels of educational attainment in Brazil are low because of this. In part, this is true. Nevertheless, there is another major factor contributing to the low levels of educational attainment in Brazil. As discussed above, both child labor and education are intimately associated with poverty. For many poor families, the only way to achieve a minimum standard of living is to keep their children out of school. Thus, for these families the opportunity cost of a child attending school-i.e., the foregone potential income of the working child-could outweigh the potential benefits received through education. As discussed above, many children both work and attend school (80%) leading to an average lag of 18% in grade attainment. This indicates that it is not necessarily only the quality of education that is the cause of low attainment, but that it may also be a fact that children are both working and attending school. This constraint is best tackled by providing a monetary incentive to attend school. But arguing in favor of demand-side interventions such as the Bolsa Escola does not imply ignoring supply side constraints. Additional improvements to the access and quality of education would complement the necessary demand-side intervention provided by Bolsa Escola. This supply-side intervention is especially critical in the poorer areas of Brazil. 2.7 Bolsa Escola is more appropriately targeted and effective than food subsidy anti-poverty programs. Brazil has a number of government programs in place that provide assistance to poor families by giving food or food-related subsidies. These programs aim to overcome nutritional deficiencies among the poor considering that such nutritional deficiencies may contribute to poverty. Some examples are Milk Distribution ("Milk is Health") and Basic Food Basket Distribution Programs. These programs suffer from a number of drawbacks. First, they do not provide the transfer in cash, but rather give specific goods. There is consensus today that when it comes to poverty alleviation cash transfers are better at improving the welfare of individuals than in-kind ones. Second, subsidies on food products tend to distort relative prices in the economy, which in turn can have negative implications for food production and marketing. Third, these programs tend to be poorly targeted and administered. Poor administration is largely due to the fact that these programs, unlike the Bolsa Escola, are not anchored in the existing education administration system. 2.8 Bolsa Escola can act as a complement to other anti-poverty programs in Brazil which focus on basic services. Since access to basic services, such as sewage facilities and water, is an indicator of general well-being, this access is a crucial component to the poverty reduction equation. Infrastructure investments that provide basic services are essential not just for poverty reduction but also for the welfare of the population-for instance, they may have positive health externalities. However, evidence shows that even when such basic services have 4 been provided, poverty often remains an issue. Anti-poverty programs in Brazil such water and sanitation, lack a stronger link to long-term poverty reduction. The Bolsa Escola programs have the potential to tackle short-termn poverty by providing a cash grant as well as its long-term determinants-low educational attainment. By breaking the poverty cycle through increased educational attainment, Bolsa Escola is able to combat long term poverty. Government policy should therefore think of the provision of basic services as a complement to poverty targeted programs such as the Bolsa Escola. Both types of programs are needed. 2.9 Bolsa Escola has broad support among civil society. Though minimum income programs have only recently begun to be used as a nationwide tool for poverty alleviation in Brazil, the pros and cons of these types of prograrns have been debated since the early 1970's. In fact, political scientists, sociologists and economists have centered many discussions over the years around the idea of these types of programs. Over the last several years, the Brazilian Government has shown an increased commitment to poverty reduction throngh increased social spending and the support of minimum income programs. The increasing demand for these programs over time demonstrates the positive momentum that these actions have achieved. The Bolsa Escola programs are consistent with the government goal of improving protection of poor children through demand-side programs that address inequalities in the educational system. The participation of social assistance organizations and schools emphasizes the pledge of many stakeholders in attempting to attain the objectives set out by the Bolsa Escola programs. In several of the cities in which Bolsa Escola is in place, social workers play a crucial role in the selection of beneficiaries. Schoolteachers and administrators are committed to handling the monitoring component of the programs together with municipalities. 2.10 These programs are only partial safety nets-they help protect the structurally poor during crises and equip the next generation with risk reducing human capital. Safety nets address vulnerability. The Bolsa Escola programs are not a complete safety net, but they do provide protection for some sections of the population. There are two ways in which they do this. 2.11 The Bolsa transfer cushions family income during a crisis. There are two types of vulnerable groups-those who are below the poverty line-the structurally poor-and those who are above it, but can fall below it when a shock hits-the transitory poor. The Bolsa Escola programs target the structurally poor (i.e. if times are normal then these would also be the structural poor). The structurally poor are particularly vulnerable to both economic and idiosyncratic shocks. Existing evidence from the region confirms this. In Brazil, recent episodes of economic downturn and upswings have particularly exposed the poor to even worse outcomes (World Bank, 2000-M. Thomas). Thus, the fact that these programs promise to continue to provide transfers to the current poor means they mitigate the risks the poor face. 5 However, in their current design, these programs do not cover the transitory poor. This is because the transitory poor are the current non-poor and therefore theoretically, invisible to these programs targeting mechanisms. They could potentially cover the transitory poor if they capture this group after it has slumped into poverty, however the targeting mechanism would have to be adjusted (beneficiary selection is based on proxies for household permanent income). 2.12 Education and health are good risk reducing assets. During economic crises in Brazil, the least educated experience the greatest vulnerability (World Bank, 2000-M. Thomas). Further, when a crisis hits, one of the coping strategies of the current poor is to pull their children out of school and put them in work to supplement family income (see Duryea, 1998; Ilahi 1999). While such strategies do reduce some of the short term costs of a crisis, they create long term vulnerability. This is because they reduce the educational attainment of children, who then are more likely to be vulnerable to economic risks in the future. Hence, by providing cash transfers to the poor conditional on school attendance, these programs increase educational attainment among future generations and hence reduce future risks. 6 BOX 2.2 The Bolsa Escola Programs are Part of a Larger Trend of Poverty-Tar,geted Cash Transfer Programs in Latin America The prototype of the Bolsa Escola-the program in Campinas-began in 1995. This program and subsequent ones in Brasilia, D. F. and other cities, as well as the Federal Program for the Eradication of Child Labor (PETI) have become models for the rest of Latin America. Mexico replaced its untargeted tortilla subsidy with Progresa, a variant of the Bolsa Escola Program with the objective of providing poverty-targeted social assistance. Honduras has installed the Programa de Asignaci6n Familiar (PRAF-BID II). Similarly, Ecuador's Beca Escolar is considering programs closely modeled along the same lines and Nicaragua is on the verge of installing its own demand-side intervention in education, health, and nutrition. Since most of these programs are too recent for evaluation, it is premature to pass judgment on their success. One that is fairly advanced in implementation is Progresa. Some preliminary evidence shows its promise. Mexico's Progresa: While Progresa closely resembles its predecessor in Brazil, it also has distinct differences in design. First, since Progresa (unlike Bolsa Escola) is a federally designed and administered program, it uses nationwide geographic targeting to identify the pooresit areas where the intervention is to be made first. The process is based on a geographic poverty map. This allows Progresa to achieve optimal targeting within Mexico (the poorest are served first) and improves administration, something the Brazilian programs cannot claim. Second, while Bolsa Escola requires only one kind of good behavior from the program beneficiaries-that their children attend school regularly-Progresa also requires that mothers in beneficiary families follow prescribed pre- and post- natal care and that all family members visit a health clinic a minimum number of times per year. It also provides food supplements to families in extreme poverty. Third, while Bolsa Escola gives a standard cash transfer to all families, Progresa's cash grant increases with the grade in school- recognizing that the opportunity cost of children's time increases as they grow older. Furthermore, it offers higher transfers for girls than for boys attending secondary school. This is to ensure that gender lags in secondary enrollment are reversed. Progresa has demonstrated success in improving educational attainment. As part of an on-going evaluation being conducted by the Intemational Food Policy Research Institute (IFPRI), T. Paul Schultz (1999) has used a difference-in-difference approach to systematically test the effects of Progresa on education using experiment and control groups. His findings are that primary enrollment has increased in Progresa regions by about 2%, from a high baseline of 92% enrollment. Secondary enrollment has risen in Progresa areas by 5%, from a baseline of 65%. 3. SELECTION CRITERIA AND TARGETING 3.1 This section discusses the criteria established by the program for the selection of beneficiaries based on the objectives of poverty reduction and increased school attaimnent. In the selection of potential beneficiaries two issues are considered. First, the criteria must ensure that the programs reach the intended beneficiaries; i.e., poor families with school-age children. Second, the criteria must ensure that 7 the number of potential beneficiaries is narrow enough to be sustainable for the programs -an issue that will be discussed further in section 8. Appendix II describes in detail the actual selection criteria used in the programs in place today. 3.2 The programs target geographically within municipalities. Even though the choice of the municipalities themselves is not based on any larger (Brazil-wide) notion of geographical targeting, the programs have good targeting within regions where they operate. This is achieved in two ways: geographical as well as income-based targeting. First, the programs start with the poorest neighborhoods within a municipality. Beneficiaries are then chosen in such neighborhoods based on an income means test or the score system (discussed in Box 3.1). This approach is cost-effective and ensures that the poorest localities in the municipalities take higher priority. BOX 3.1 Targeting the Poor -the Income-Means Test and Scoring System Given the programs' objective of reducing poverty through increased educational attainment, the target population includes those deemed relatively most impoverished in each area. Initially, an income-means test was used to identify the beneficiary group. This method took into consideration the characteristics established by the program-that families have children in the 7-14 age group, they earn less than one half of minimum wage and, in some cases, they have a minimum amount of residency in the area. The income measure had difficulties. One was that a significant portion of family income of the poor in Brazil is derived from informal occupations and unregistered sources. This income is extremely hard to measure accurately. For this reason, a scoring system was developed. The score, which is demonstrated in Table I in Annex 1, considers various aspects of the potential beneficiaries' living standards. In addition to income, the characteristics considered include housing status, engagement in the labor market, educational attainment, presence of durable goods in the household and number of children in the household, to name a few. Beneficiaries are selected based on the number of points they receive. In each area or federated unity, as in the case of Brasilia, each category may take on a different weight. In Brasilia, for example, priority is given to families with children below fourteen years of age who rent their homes. A heavy weight is assigned to dependents in the household while less weight is given to the different types of labor market engagement. Some issues remain outstanding. First is the weight assigned to bi-parental versus single parent households. Since the latter group tends to live under more adverse conditions, it makes sense to give more weight to this group. Second, there should be a differentiation between those who have a regular source of income and social guarantees and those who do not and consequently are engaged in activities that do not provide a reliable income flow. Overall, while it is clear that the scoring system provides an appropriate method for targeting given the circumstances surrounding the income of families in these areas, tailoring the scoring system to each area by weighting according to specific characteristics would probably improve the targeting method. 8 3.3 Is half a minimum-wage a reasonable poverty line for each mnunicipality? This is the per-capita income level below which the families are eligible to participate in the program.3 This is also a key parameter in determining the sustainability of the program. Because socioeconomic conditions and local prices can vary substantially between municipalities, it may not be appropriate to use a single poverty line across the board. It may be better to start with a nationally determined poverty line (half of minimum wage) and to then adjust it for spatial variation in prices. 3.4 The selection criteria established by the programs narrovws the target population to the intended beneficiaries, but is this too narrow? The underlying objectives of the prograrns-to provide income support to poor families with school-age children-allow the program to naturally narrow the population of potential beneficiaries. When the program was originally designed, the three selection criteria of the program were: a) poor households; b) children 7-14 years of age; and c) a minimum residency of 5 years (such as in Brasilia but not in other metropolitan areas). Tables 3-II and 4-II in Appendix II show the selection criteria for the other municipalities. The number of families in the areas according to age, shown in Table 3.2 in Appendix I, allows us to obtain the total number of potential beneficiaries if only the age criteria were used. Using household data from the PNAD, Table 3.3 simulates how each one of the three conditions narrows the group of those who are eligible for the program. Recall that economic affordability of the program is the most important reason for having a small target population. Imposing the condition of "children in the 7-14 age bracket" on poor families substantially reduces the number of eligible families. The overall drop in the ten cities would be about 50%. This drop would be greater in the larger metropolitan areas of Rio de Janeiro and Sdo Paulo (more than 50%) than in the cities of the Northeast-Belem and Recife. When the residency requirement (see Box 3.3), is added to the selection criteria-ostensibly to reduce the incentive for in-migration of poor families-the numbers of eligible households can drop substantially. This drop is much greater in the Center South-an area that has been a magnet for in-migration in recent years-than in the North or Northeast. Box 3.2 discusses the issue of whether or not the strict selection criteria create a "non-covered population" in terms of age and the residency requirement. 3 Note that this criteria does not vary by municipality. 9 BOX 3.2 The "Non-Covered Population" the Brasilia Case The three proposed eligibility criteria for selection into the programs leave two main types of "non- covered populations." This box identifies what these two populations are and how large they are likely to be in the ten metropolitan areas. . By targeting poor families with children of school going age the programs overlook poor families without children in that age bracket. This is a potential problem since poor families with very young children-under 6 years of age or over 14-are also deserving of poverty-targeted interventions. However, as is discussed elsewhere in this report there are two reasons why it is best to leave this group out of the Bolsa Escola programs. First, the Bolsa Escola are rooted in the school system and so the compliance (i.e., school attendance) of children of school-going age is easy to monitor. Second, very young children have specific needs such as post-natal care and daycare which are best addressed by a separate program (see Pies de Barros and Mendonqa, 1999). Table 3.2 shows that the non-covered population in this category is not small. Of the families that have children in the 0-6 age bracket, as many as 57% do not have children in the 7-14 age group, implying they will not benefit at all from the program. . By restricting to a subset those who have been residing in the city for a minimum 5 in years or more, the program further "excludes" those families that are poor and have children of school- going age but are recent migrants. Table 3.3 in Appendix I shows that the imposition of the 5-year residency requirement reduces the number of eligible families by as much as 21%. This dilemma is discussed further in Box 3.3. . While this report recognizes that these populations are "excluded" it is important to note that it is not as a result of design failure of the programs. The programs intentionally select the specified groups based on the trade-off between the objectives and fiscal sustainability of the programs. 10 BOX 3.3 The Dilemma of the Residency Requirement In the three metropolitan areas where they operate, the Bolsa Escola programns require that beneficiaries must have been residents in the metropolis for more than the minimum period of residency. This period varies from two to five years. The reason for this requirernent is that the programs wish to avoid creating a rural-urban migration problem. Since a sizable proportion of the in- migrants to urban metropolises are poor families with young children, the program designers fear that if they do not use a time of residence filter, the programs will start attracting the poor from the outlying areas. In the extreme, the programs would ultimately result in benefiting only the recent migrants. In order to minimize the chance that the programs create an incentive for rural-urban migration and to make them more affordable (by trimming the numbers who qualify), the programs choise to install the minimum residence requirement. The discussion above notwithstanding, a 3-5 year residency requirement severely curtails the ranks of the deserving poor in these cities. In fact, this filter is biased against the poorest who live in the cities now (and who did not in-migrate after hearing about the program). In that sense, the time of residence filter may be seen as counter-productive. How severely does the program restrict the qualifying population? Table 3.4 in Appendix I shows the percent reduction in qualifying families when the time of residence filter is used in selecting beneficiaries in the ten cities. The overall reduction is about 20% and there is little difference between the three and five year requirement The filter trims a much larger proportion of the qualifying families in the cities of the Center South (41% drop in Sao Paulo and 81% in Brasilia) than in the Northeast (5- 6% in Recife and 6% in Belem). This is understandable since the rich and fast-growing cities of the Center South are most likely to have attracted migrants in the last 3-5 years. How can the program designers resolve the dilemma? It may be best to use a shorter residence requirement-say one year. This would dent the incentive outsiders have from in-migrating and benefiting from the program immediately. At the same time, the programs would not "over kill" by dropping a large number of current poor who came into the city 1-5 years ago. One problem would still remain. Relaxing the residency requirement would undoubtedly increase the costs of the programs. The answer here is that it would swell the ranks of the qualifying families a lot more in the Center South than in the poor Northeast. Therefore, the biggest increase in costs would result in the richer cities of the Center South and not in the Northeast. As is obvious from Table 3 4, local finances in the latter are adequate to meet the increased demand on program funds. 11 4. DESIGN AND IMPLEMENTATION 4.1 This section discusses issues related to program design and the practical disbursement of the Bolsa (cash grant) to beneficiary families. The most pertinent elements of this discussion are why school investments should complement cash grants; why giving cash is appropriate; how the amount of the Cash transfer ought to be set; and to whom it should be given. 4.2 The Bolsa Escola programs should be seen as complements to, not substitutes for investments in schools. The poor record of educational attainment in Brazil is a consequence of both demand failure (i.e. the poor cannot afford to send their children to school) and supply constraints (parents consider the quality of schools to be inadequate). The Bolsa Escola programs are primarily demand-side programs. This is partly because school quality is not an issue in Brasilia-where Bolsa Escola was first implemented. However, it is likely to be important in other urban areas, particularly the poorer ones. Thus local governments should not see the Bolsa Escola programs as replacing other education-related interventions. Rather, the Bolsa Escola cash grants should be given in tandem with investments in local schools. This is essentially the approach of similar programs (PETI in Brazil and Progresa in Mexico). 4.3 Giving cash is better than giving other forms of goods or services There are two main reasons why programs that want to alleviate poverty are better off giving cash than some form of goods or service. Given that increasing household welfare is the ultimate objective of a poverty alleviation program, a cash transfer, as compared to an in-kind transfer such as a school meal, best increases household welfare because it leaves to the household how to spend the money. One counter argument is that if the welfare of vulnerable groups within the household were an issue-such as the well-being of children-then it may make more sense to give an in-kind transfer specifically to the children rather than give cash to the household head. However, a sizable body of literature on the intra-household allocation of resources demonstrates that earmarked transfers to specific individuals in the household do not always result in achieving their objective.4 This is because intra-household allocations depend on the bargaining power of individuals and until the sharing rules are known, little can be said about the success of such transfers in increasing the welfare of specific individuals in the household. Second, non-cash transfers end up distorting the market for the goods or services in question. For instance, if a school milk program replaces the cash- based Bolsa Escola then the government would either have to take over the production and/or marketing of milk, or use a subsidy to ensure adequate supply for the program. Furthermore, it would become very difficult to exclude the non- 4 Haddad, Hoddinott and Alderman (1997). 12 beneficiary students from the beneficiaries. Thus, non-cash-based transfers come at a higher cost to society than a simple cash-based transfer. 4.4 How much should the cash transfer be? It is important from the standpoint of benefits as well as costs that the level of the cash grant ought to be set at the right level. There are two issues related to the optimal level of the cash grant. First is at what level it should be set, given program objectives. Given the facit that these programs have multiple objectives, with no clear weighting ascribed to each one, it may be difficult to arrive at with a precise level of the cash grant. In other words it is not obvious theoretically whether the program should pay a cash grant that brings the families up to the poverty line, or a grant level that compensates families for the opportunity cost of participating in the program (i.e., the opportunity cost of the child's time, time cost of visiting the health celnter, etc.). If the primary objective of the program is to reduce poverty then the grant should be adequate to bring the family up to the poverty line. If the primary objective is to have children in school then compensating families for the opportunity cost of child time as well as other indirect school costs may be best. However, it should be noted that since households themselves ascribe a positive value to the education services these programs provide, the optimal level of cash grants should necessarily be below whatever is determined as the economic optimum. 4.5 Should the cash transfer vary by age or gender? The theoretical discussion notwithstanding, there are practical issues regarding the timing or the level of cash transfer that need attention. This is because not everybody who participates in the program is identical and therefore, there may be gains in efficiency to be had if the program were to alter the timing or level of the grant to account for this heterogeneity of participants. There are two examples that illustrate this. First, the analogous program in Mexico, Progresa, changes the level of the grant according to the school grade and gender of the child. This is based on the explicit understanding that families with older children face a higher opportunity cost than those with younger children so the grant for children in secondary school would have to be higher than for those in primary school. Also, the female- male differential in enrollment in Mexico widens at the secondary school level so a greater incentive is given to girls to attend. The Brasilia version of the Bolsa Escola program deposits the equivalent of one minimum wage into a savings account in the name of the beneficiary child each year. The child is allowed to withdraw the money once he/she completes 8th grade. The idea here is to alter the timing of the grant to create an incentive to complete secondary school. 4.6 The benefits of customizing the cash transfer to participants ought to be weighed against the costs. While tinkering with the cash grant level, to tailor it to the various sub- groups in the population makes economic sense (from an efficiency standpoint), it also adds to the cost of administering the program, since additional criteria have to be checked when making the cash transfer to the family or when the banking system has to be involved. If these marginal administrative costs are not large then it would make sense to tailor the grant to the different 13 subgroups in the program. However, if these costs are large then it may be better to leave the cash grant scheme simple. 4.7 The cash transfer should be based on the number of eligible children in the household. Currently in many of the programns, a family qualifies for the cash transfer if it has at least one school-going age child and the level of the cash transfer does not increase with the number of eligible children in the household. So a family with three children in the 7-14 age group would receive the same cash transfer as a family with just one child. This structure of the transfer lowers the incentive for the family with many children to participate in the program since the opportunity cost of putting three children in school is much higher than that of keeping one in school. This defeats the purpose of the program. One solution is to base the cash transfer on the number of eligible children in the family. Since the average number of children in eligible families is under two, giving one cash transfer per child rather than one per famnily would improve the incentive for families with many children to keep all of them in school. Further, if such changes were accompanied by a reduction of the amount of the cash transfer (per child) from one minimum wage to one-half the minimum wage, such a structure of the cash transfer would improve fiscal sustainability by reducing the overall costs of the program by about 12%, with sizable reductions in costs in the Northeast (about 10%) (see Table 4.2). 4.8 The cash transfer is (and should be) given to the mother of the children. Evidence from around the world shows that a transfer to the woman rather than the man in the family, will result in more appropriate use of the funds. Additionally, it ensures the participation of the mother in the child's development and education and gives the woman some financial responsibility, which has obvious implications for gender equality. The Bolsa Escola programs require that the cash be transferred to the mother of the beneficiary household, and this is the correct approach. 5. IMPACTS-PRELIMINARY EVIDENCE 5.1 This section evaluates the performance of Bolsa Escola along the lines of its objective-reaching the poorest subset of the population in the chosen areas; increasing education; reducing short and long-term poverty, and reducing child labor. Since a full-fledged evaluation of the impacts is planned but has not yet been conducted, this section provides a preliminary assessment based on evidence available from program data as well as that available from indirect sources. 14 5.2 The programs have been well targeted to the poor families with children of school-going age and to the poorest regions within the metropolitan areas.5 In the absence of data on experiment and control groups, only indirect evidence on targeting is available. The evidence from Brasilia suggests that the program there has been well-targeted (Saboia and Rocha, 1998), meaning that the actual beneficiaries of the program were the intended beneficiaries based on the selection criteria established by the program. Appendix III demonstrates in detail the evidence supporting the success of the program in capturing the intended beneficiaries or "parameter population." By comparing the characteristics of both beneficiaries and non-beneficiaries (see Tables 1-III through 7-111 in Appendix III) it is clear that the program did capture poor families with children of school- going age. Part of the reason for the success in reaching the intended beneficiaries is the program's use of a score system--a quasi-means test that allows the program to select beneficiaries by "scoring" their observed characteristics and then using a cutoff point to choose the most needy (see details in Section III, Box 3.1). Another reason for the targeting success of the program is the manner in which it was expanded. In Brasilia, the program was implemented gradually by starting from the poorest neighborhoods and then expanding to less and less poor neighborhoods. This method of expansion uses a mix of geographical targeting (neighborhoods) and within a geographic area, targeting based on a quasi-means test (household score). 5.3 While many municipalities have been successful in reaching all of the potential beneficiaries, on average the programs have not been successful in reaching all potential beneficiaries. Referring back to Appendix II, it may be noted that although the programs demonstrated appropriate targeting, as described above, the programs on average were not able to serve nearly 43% of the parameter population (the population established by the selection criteria). This is not a question of "non-covered population" due to the selection criteria, however. These families were actually captured in the measurements based on the selection criteria but for some reason they were not served. Nevertheless, it is important to note that 21% of the programs were able to serve 100% of the selected (or potential beneficiaries) which demonstrates that many of the progranis have been successful in reaching all of the intended beneficiaries.6 The reason why some programs served less than 100% of the selected beneficiaries was financial-the program costs with 100% service may just were too high. 5.4 The programs show improved educational outcomes. Four outcorne indicators of education can be used to assess the effectiveness of the programs in the ten metropolitan areas-drop out rates, age at first entry, promotion and learning. Since a minimum school attendance rate of 90% is required for a family to continue to receive the cash grant, the program directly forces high attendance 5 Appendix III discusses in detail the evaluation of the programs' ability to target appropriately. 6 The mean of potential beneficiaries over actual beneficiaries was approximately 57%, the median was 53% and the mode was served 100%(12 out of 55 served 100% -where three values were excluded due to incomplete data). 15 and low dropout. First, a comparison of program data (experiment) and school census data (control group) reveals that drop out rates among beneficiaries are much lower (0.4% in 1996) than among non-beneficiaries (5.6%) (Table 5.1).7 Second, a larger proportion of the children in beneficiary households enter the school system at the right age than do their non-beneficiary counterparts. Third, children in beneficiary households do exhibit a higher promotion rate (80%) than their counterparts in non-beneficiary households (72%) (Table 5.2). Finally, there was little difference between the beneficiary and non-beneficiary groups in learning outcomes. It should be noted, that the recent controversy relating to the impact of the Brasilia Bolsa Escola program suggests the urgent need of a competent third-party evaluation. It should also be pointed out that all the comparisons cited in this paragraph are based on uncontrolled comparisons and that a more systematic evaluation that controls for the fact that beneficiaries are not a random sub-group of the population may give better indications of program success. 5.5 There is evidence of immediate poverty reduction. The Federal District Program had remarkable success with respect to its income distribution component. Because of relatively low poverty incidence coupled with comfortable public finance conditions, it has been possible to transfer a monthly minimum income benefit to 25 thousand families amounting to less than one per cent of the local current receipts.8 Transferring a minimum wage - R$130 in November 1998 - to families whose average household per capita income is R$44.35, significantly improved living standards. Considering half a minimum wage as poverty line, the income gap ratio is decreased from .41 to .04 as the average per capita income is raised to R$72 because of the monetary transfer.9 That is, the aggregate monthly transfer would be sufficient to raise all families who receive the benefit to the poverty line level. It is noteworthy that per capita income among beneficiary families varies from zero to values above the proposed criteria of half minimum wage. As a consequence, after the same monetary transfer to all families, some attain per capita income above the half minimum wage poverty line, while others still remain below this parameter. Since the transfer corresponds to almost 40%, the most obvious result is the immediate increase in family income and private consumption (Rocha, 1999). Evidence is lacking in the other areas and should also be examined for any robust results of positive poverty impacts. 5.6 The evidence on the effects on child labor is inconclusive. The elimination of child labor is another measure of program success. Since school hours are short child labor can be concurrent with school attendance. The programs' direct effect 7 Note that the difference in dropout rates between beneficiaries and non-beneficiaries obtained in this way provides an upper bound on the actual difference since the former group consists of children from poor families who are more likely to drop out compared to a randomly chosen child. In other words the true effect of the program would be obtained when a randomly selected child is given the cash transfer and their drop out rate compared with that of the rest of the population. 8In 1997 the payment of benefits corresponded to 0.7% of current receipts. 9 Estimates based on data from families receiving the benefit in July 1998. 16 on reducing child labor is to be investigated directly from program data, which is difficult since families who receive the benefit will probably deny using child labor. 17 6. EVALUATION INSTRUMENTS 6.1 How to measure program success in increasing educational attainment. In Section III, a number of indicators of educational attainment were discussed. The question is, if given the choice of using an ideal indicator of education attainment in order to evaluate the success of the program, which would it be? Attendance is certainly not appropriate. This is because the program monitors on the basis of attendance and it drops those families in which children fall below a minimum acceptable level of attendance (90%) over a given period. Thus, after the fact, all program participants would show adequate attendance. Proficiency test scores such as the SAEB would be ideal (Table 6.1 in Appendix I). 10 Ultimately we are interested in whether children are learning more as a result of the program, not if they are attending regularly. Given the lack of a well functioning testing system, this is difficult at this point but may be an option in the future. Progression through the school system-as captured by promotion-may be the best available indicator of educational outcomes. The child completing grade eight is a potentially appropriate measure of promotion that is being used in Mexico's Progresa Program.I 6.2 How to measure program's success in reducing current poverty. Poverty reduction is usually measured by per capita-based consumption measures such as the headcount-i.e., the proportion of the population below a poverty line. A comparison of regions where Bolsa Escola interventions were made and regions where they were not would provide the contribution of the programs to poverty reduction. However, since the poorer regions are more likely to attract Bolsa Escola investments, it is important to compare changes in poverty (difference-in- differences) rather then levels between program and non-program regions. Further, economic growth could also contribute to poverty reduction and if this growth differs by program and non-program regions then this would bias the poverty reduction effects ascribed to the program even under a difference-in- difference approach. Controlling for growth and other exogenous factors would provide a cleaner assessment of the program's impact on poverty reduction. 1° Note that the difference in dropout rates between beneficiaries and non-beneficiaries obtained in this way provides a lower bound on the actual difference, since the latter group consists of children from poor families who are more likely to drop out compared to a randomly chosen child. In other words, the true effect of the program would be obtained when a randomly selected child is given the cash transfer and their drop out rate compared with that of the population. "1 However, it suffers from two potential drawbacks. First, automatic promotion has been implemented in some grades in Brazil. Second, the school system may artificially inflate the promotion rate if it stands to fain resources on the basis of how well program participants are doing. 2 Note that the difference in dropout rate between beneficiaries and non-beneficiaries obtained in this way provides a lower bound on the actual difference, since the latter group consists of children from poor families who are more likely to drop out compared to a randomly chosen child. In other words, the true 18 6.3 How to measure programs' success in reducing child labor. While assessing the programs' impact on education by monitoring student performance is relatively easy given the structure of the program, it is more difficult to measure their impact on child labor reductions. Ultimately this has to be done at a broader level by not using control and experiment groups, but rather modifying the existing household survey data (such as the PNAD) to do a baseline and then a post-program evaluation. If the modified household survey included a question on whether the household benefited from the program, then after controlling for endogenous placement a difference-in-difference in the incidence of child labor between participant and non-participant households would demonstrate the effect of the program. 7. MONITORING INSTRUMENTS 7.1 Participation of locals and stakeholders ensures good targeting. Participation of local members of the community, because each of the programs is local, ensures that the programs reach the intended beneficiaries as well as ensuring that monitoring occurs at various independent levels. In Beldm, for example, the programs are supported by the participation of social assistance institutions. Social workers play a significant role in the family selection procedure in that they visit households and help families fill out information forms. Increased participation of civil society would improve the monitoring process. 7.2 How the transfer takes place. The Secretary of Education makes monthly cash payments in each Municipality. The requirements for receipt of the payment are a minimum monthly school attendance of 90 percent. If the child does not meet the attendance requirement for the month, the monthly transfer is not made to the family. The payment is resumed once school attendance is deemed satisfactory. The Accounts Tribunal of the municipality carries out an audit. 7.3 Enforcing school attendance. There are a few weaknesses in the system which should be addressed. The first is the difficulty in ensuring that children actually are attending school. While the school has the responsibility of keeping track of attendance and sending the records to the Municipal Secretariat of Education, there may be some discrepancy between schools in the measure of absenteeism. The results depend on how strict controls are in each school, characteristics of individual children and the type of labor in which children may be involved if they are working and attending school. This may be addressed through greater effect of the program would be obtained when a randomly selected child is given the cash transfer and their drop out rate compared with that of the population. 13 However, it suffers from two potential drawbacks. First, automatic promotion has been implemented in some grades in Brazil. Second, the school system may artificially inflate the promotion rate if it stands to gain resources on the basis of how well program participants are doing. 19 participation of members of civil society and families to ensure greater convergence between program goals and implementation. Additionally, independent spot checks besides the attendance checks carried out by a third party would help to ensure accurate measurement of absenteeism. The second issue is that in those areas where labor is very seasonal, families may pull their children out of school during the high work season and then send them to school for the remaining months. As the system stands, while children may be out of school for a significant portion of the year, during the off-season the families receive the cash transfer since the child is attending school regularly during those months. Since promotion to a higher grade is not necessarily based on ability, the result will clearly be a lower level of performance. Thus, a more appropriate manner to ensure consistent school attendance would be to require an additional minimum number of months of attendance per year. A final suggestion is an annual re- registration to the programs to reinforce program requirements and goals. 7.4 The programs cannot enforce a reduction in child labor. Children work both for wages and as unpaid family workers. Even though the reduction of child labor is an expressed objective of the program, it is difficult for the program to monitor if children in beneficiary households are indeed not working. Good standing in the program does not require that children be pulled out of work, only that they be attending school at a minimum rate which may be compatible with working. Thus, although the program has no mechanism with which to monitor child labor, the hope is that the cash grant and increased education will provide an incentive to not work. 8. POLITICAL AND FISCAL SUSTAINABILITY 8.1 This section discusses issues related to the sustainability of the programs. Political sustainability is ensured through the active support of the government in improving the social protection system in Brazil and its commitment to minimum income guarantee programs such as Bolsa Escola. The fiscal sustainability of the programs is heavily dependent upon program affordability. Data show that in areas where the programs are needed the most, they tend to be the least affordable by local governments. 8.2 The programs promise high political sustainability. Brazilian governments have been increasing their spending on the social sector over the last several years, showing an increasing commitment to poverty reduction and programs such as Bolsa Escola. Primary education programs as well as other types of demand-side programs have received increased financing, demonstrating the willingness of the government to address inequalities in the educational system. The financing of programs addressing basic health and primary education are some examples of this commitment. In addition, the government has made it a 20 priority to: a) improve targeting of its social assistance programs; b) improve quality; and c) reduce the gap in the social protection system. 8.3 Fiscal sustainability is the weakest in poorest municipalities and it may be improved somewhat by establishing more strict selection criteria. Each selection criteria helps to reduce the size of the eligible population (see section 3 for details). If we use all three criteria-poverty, children school age, and a residency requirement-the cost of the programs is about 3% of current receipts (Table 8.1 in Appendix I). However, this average for the ten largest metropolitan areas masks large inter-city differences. For instance, program costs are a much larger share of current receipts in Salvador (19%), Fortaleza (16%) and Belem and Recife (both 12%) than in Brasilia (0.2%) and Sao Paulo (0.90%). There are two reasons this. First, poverty is much higher in the Northeast, so the pool of families that satisfy selection criteria is much larger than in the Center South. Second, the cities of Northeast are poorer than those in the Center South. As a result their fiscal revenues are much smaller than those of their counterparts in the Center South. Not surprisingly then, the programs are the least affcrdable in the areas where they are needed most. 8.4 Should affordability be improved by reducing the size of the cash transfer?. Governments may be tempted to make the programs affordable by cutting the size of the transfer. Table 8.2 shows how program affordability increases as the cash transfer is cut to 1/2 minimum wage per family and then to 1/4 minimum wage. With a transfer of V2 minimum wage, the average program cost is 1.4% of current receipts. With 1/4 minimum wage this is further lowered to less than 1%. Of course vast differences between cities still remain, with the program costs being above 3% in Northeast and minimal in the Center South. But cutting the size of the cash transfer runs the risk of jeopardizing the effectiveness of the programs in inducing child enrollment and reducing poverty. The optimal size of the transfer should be set taking into account either the opportunity cost of the child's time, or the transfer necessary to reduce poverty (see section 4). Any gain from trimming the cash transfer has to be weighed against the cost of not meeting the goals of the programs. 8.5 The federal/state governments should top off the shortfall. Tinkering with the design of the programs (e.g., changing the level of the cash transfer or changing the poverty line), would make them more affordable, but it would also defeat the purpose. Thus keeping the current targeting and transfer design of 1the programs requires that the federal government "top off" with transfers to the poorest municipalities. The question is under what formula should these transfers from the federal government come? The answer is that the federal government can equalize the load of the prograrn on the municipalities. In this way every municipality pays up to a local share of the program and the remainder is taken up by the federal government. This formula would mean that there would be no federal transfer to municipalities where program costs are less than the maximum local share. On the other hand, where program costs are very high ('e.g., 19% in Salvador), the federal government would pay most (for instance 17%) of the 21 costs. Since the poorer municipalities are also likely to need help with their school system, some of the transfer from the federal government can come in the form of "in kind" transfers for improving the school system. A federal financed program would also have the advantage to minimize problems related to migration from covered to uncovered areas. Regardless, it is important to keep in mind that in order to maintain the stake of the municipalities in the program, the federal government should not pay the full cost. 9. EXPANSION-WHERE, WHY AND HOW? 9.1 The success of the Bolsa Escola in targeting and delivery in the cities and states where it exists raises the question whether the programs should be expanded to other areas. Ideally, given the programs' objectives of targeting the poor and considering that the poorest in Brazil live in rural areas, it would make sense to argue that the programs be expanded to rural Brazil. However, there already is another initiative in place that specifically focuses on rural Brazil, and its particular nature of low education and high child labor (see box 9.1). The expansion of the Bolsa Escola into poorer municipalities may require very high subsidies from federal or state governments to ensure fiscal sustainability. Therefore, for now, an expansion of Bolsa Escola to the ten largest metropolitan areas makes sense. 9.2 Why should the program be expanded first in the large urban and metropolitan areas? * These areas have a large proportion of Brazil's poor. According to recent calculations, one-third of Brazil's poor live in these ten cities-in line with the population share of these cities in the national population. These cities have, over time, become less attractive places for rural migrants partly as a result of short-term economic cycles. Nevertheless, they still hold close to 9.5 million poor. * The supply of schools is better in urban than in rural areas. Both the quantity of school supply (number of seats per child) and the quality of education are better in these urban metropolitan areas compared to the rural ones. Thus, pure demand side programs such as the Bolsa Escola make more sense here than in rural Brazil. In rural areas, any demand side intervention in education would have to be accompanied with a strong supply intervention either directly in the school system or through an "extended school day" (Jornada Ampliada) as is being done currently with the PETI program in rural Brazil (see Box 9.1). 22 * These cities are administratively well-equipped to handle the program. Poverty-targeted education grant programs like the Bolsa Escola require efficient administrative machinery to ensure proper implementation. This is where the metro areas, which are also the capital cities of their respective states, have an edge in terms of the administrative machinery already in place. * The government is committed to supporting these types of programs. The support for the Bolsa Escola programs in these metropolitan areas is evident. The results of the government's efforts to increase social programs are due largely to the increased demand and the generation of positive feedback from society that supports the notion of expansion. 9.3 However, any expansion must account for the fact that the lessons from Brasilia are "non-representative". The richest evidence on the Bolsa Escola is from the Federal District. However, when thinking of expanding to other urban metropolitan areas, one has to be careful in recognizing that part of the reason for the program's success in Brasilia has to do with the following factors: • Brasilia has a low incidence of poverty. Thus the proportion of eligible families (according to program criteria) in the population of Brasilia is low. This is not the case elsewhere, especially in the cities of the North and Northeast. - Because of low eligibility and high revenues, the program imposes a lower fiscal burden on Brasilia than on other cities. * The quality of schools in Brasilia is very good. However, this is not always the case in other municipalities. Thus there is a need to supplement the Bolsa Escola with a supply side intervention in these municipalities. 9.4 Program affordability and sustainability must be considered when expanding to other areas. The discussion above, as well as that in Section 8 on sustainability, brought to light some points which are crucial to the expansion phase of the programs to other areas. It is evident that in those areas where the programs are needed the most (the poorest areas and rural areas), they are the least affordable. If the programs were to be expanded to these areas, there are several options which might attempt to reduce program costs. First, the selection criteria for the programs could be altered to narrow the number of potential beneficiaries. Second, the amount of the cash transfer could be reduced to lower program costs proportionate to percentage of current receipts. When expanding to rural areas, which given poverty levels, should be a priority for the ]3olsa Escola programs, the amount of the cash transfer could be set at a level similar to that set in the rural PETI program (Box 9.1) of R$25. Overall, however, it is not suggested that program design be altered drastically to make the program affordable as this risks affecting the program objective of poverty reduction. Therefore, one crucial policy option is federal/state subsidies to balance the lack 23 of resources and funding in the poorer municipalities. This may be done by the government equalizing the load of the program on municipalities through a monetary transfer and/or through in-kind transfers to improve the school system. BOX 9.1 The Child Labor Eradication (PETI) Program The PETI program was initiated in 1996 in rural areas of Brazil and has managed to reach 977 municipalities in 27 states and has provided assistance to 396,546 working children in 2000. The PETI program is scheduled to reach close to 800,000 children by 2002. * The program has objectives similar to those of Bolsa Escola, including increasing educational attainment and reducing poverty, though its goal is to simultaneously eradicate the "worst forms" of child labor. Child labor in rural areas accounts for 60% of overall child labor, which in 1999 was 14.9% of all children ages 10-14 in Brazil. * The program first chooses geographical areas throughout Brazil with the greatest incidence of the worst forms of child labor. Among the worst forms of child labor are work in sisal and cassava processing, charcoal production, mining, quarrying, and the collection of sugar cane, tobacco and cotton. * The program provides stipends of approximately R$25 per child to poor families who have children ages 7-14 which is given to the mother. The stipend is conditional on the promise that the children will attend school, participate in after-school activities and agree to not work. The program also transfers to local govermments R$20 per child to finance after-school activities (Jornada Ampliada) and ensure that children do not both attend school and work. * The advantage of the PETI program over other types of minimum income guarantee programs is that it allows for the reduction of the target population, making it more feasible in areas with less than sufficient funds. Also, this program is advantageous because it is anchored in the school system along with having the support of civil society (NGOs, rural worker unions, etc.) for selection as well as monitoring. * A qualitative assessment of the PETI program in Pernambuco (Moura, 1999) revealed the following. a) There was little evidence of mis-targeting; b) participating children were found complying with the minimum days of school attendance and were not working; c) the after-school program (Jornada Ampliada) seemed to be functioning well. 10. CONCLUSIONS AND OUTSTANDING ISSUES 10.1 According to the preliminary evidence available, the Bolsa Escola programs are effective and appropriately targeted, considering the extreme level of poverty, low educational attainment and the child labor situation in Brazil. In addition, the programs evaluated were shown, through the provision of cash grants, to immediately improve living conditions among the poor. Because the Bolsa Escola programs provide monetary incentives to attend school, the demand for education is expected to rise and the programs are able to combat long-term poverty at its roots. These outcomes must, however, be evaluated in the long-run. Given the successful momentum demonstrated by existing evaluations of the Bolsa Escola type programs, an expansion to other areas of Brazil is recommended and could 24 be complemented by supply-side improvements such as school quality and infrastructure programs. 10.2 There are several outstanding challenges which must be considered in order to most effectively expand the programs to other areas of Brazil. The main issues which we have discussed in this paper are: a) the necessity to solidifying design issues such as the size of the cash grant; b) the need to ensure program affordability and sustainability; c) the need for other complementary programs which reach the "non-covered population"; d) the need for complementary supply-side investments in schools; e) the need to reach consensus on the final program objective ;and f) the inherent critical need for a formal impact evaluation. 10.3 The first issue of solidifying the design of the programs, includes twc elements: a) improving the programs' capacity to select the appropriate beneficiaries; and b) optimally setting the level of the cash transfer. As discussed in this paper, overall the program has been very successful in targeting the intended population. Nevertheless, the selection process using scoring system must be tailored and made flexible to allow for differences between municipalities in which the programs are implemented. Appendix II demonstrates that the programs do in fact differ in their selection criteria, which indicates this flexibility (though the differences between the municipalities are perhaps even more significant). The ability of the programs to reach as many of their potential beneficiaries as possible should also be reflected in the setting of the cash transfer. The optimal level must be one which at a minimum compensates families for the foregone earnings of the child. 10.4 In the expansion of the programs, the two design issues above are absolutely crucial and directly linked to the affordability and sustainability of the programs. Clearly the programs must be affordable fiscally in the municipalities in which they are implemented. Program affordability is affected by the selection of beneficiaries (number), the amount of the cash grant as well as the existing infrastructure in the municipality. As discussed in the sustainability section of this paper, the selection criteria help in reducing the size of the eligible population. For example, using the three basic criteria can reduce the cost of the program to 3% of current receipts in some areas. However, in other areas such Salvador and Fortaleza -the costs are still high.. As pointed out, usiIng Brasilia as a model is hardly appropriate considering that there the costs are only equal to 0.2% of current receipts. Another factor to consider in attempts at reducing program costs is the size of the cash transfer. Again though, there are some cities whose costs are still too high for the program to be affordable. It is clear that in areas where a program is most needed -the poorest and rural areas- it is the least affordable. The trade-offs in terns of selection of the population and setting of the cash transfer are inherent given the multiple objectives of the programs. However, attempting to change the program design in terms of selection criteria and cash transfer size to address this factors, runs the risk of jeopardizing program effectiveness. We therefore would recommend, for rural areas, criteria which resemble that first established by the PETI program. We do not recommend 25 sacrificing program effectiveness in order to reach a larger population, a potential solution is to call for federal/state subsidies to supplement those poorer municipalities. 10.5 The third issue is the need for complementary programs that reach the "non- covered population" such as poor families with very young children. By restricting itself to poor households with children of school-going age, the Bolsa Escola program tends to overlook an equally deserving pool-poor families with children under 6 years of age or over 14. This neglect raises questions about why it does not include the latter group and to what extent the program covers the most needy. There are essentially two reasons why this is so. First, the main reason for the success of Bolsa Escola is that by anchoring itself in the existing public school system, it is able to avoid the higher administrative costs that come with a massive poverty-targeted program. Unfortunately, this means that the program cannot monitor the attendance of children under 6 or over 14. Second, children under 6 (preschool children) have distinctly different educational and developmental needs than do those in the school going age. It may be best to address the needs of the former by installing a separate poverty targeted program for poor families with very young children. Existing programs for pre-school children in Brazil (Pdes de Barros, 1999) may provide the best example of what this special group needs. 10.6 The fourth issue is the need for complementary investments in schools. The Bolsa Escola Programs should not be seen by local governments as replacements for other education-related intervention investments in schools. Poor educational attainment in Brazil is a consequence of both demand failure (poverty) and supply constraints (poor school quality). This report recommends that the Bolsa Escola programs follow the example of other programs in Brazil (PETI) and in the region (Progresa in Mexico) that provide financial support to local schools. 10.7 The fifth issue, of reaching consensus on the final project objectives refers to the question whether programs such as the Bolsa Escola can be used as a safety net. In other words, can Bolsa Escola provide support during a negative shock. Negative shocks such as layoff of family members can have impacts on other members of the household-namely children who are pulled out of school and put into the labor market as a household coping strategy. There is evidence from Brazil today that exactly this may have happened during previous crises (Duryea 1999). The Bolsa Escola programs in place might prevent such negative outcomes. However, this would be accurate only to the extent that the program already covers all the households that suffered from the shock. Since today's poor would be receiving the cash transfer before the shock hits, The cash transfer would smooth the impact of the shocks. It is likely that some of today's non-poor will become tomorrow's poor when such a shock hits. These transient poor are hard to track (because they have permanent characteristics associated with the non-poor and would consequently not qualify according to the targeting mechanisms currently in use). Therefore, complementary programs to address the needs of the transitory poor would be most appropriate alternative. 26 10.8 The final issue is emphasized in the report is that the Bolsa Escola programs need to be formally evaluated more in terms of impacts. The evidence from the evaluations up until now is not robust enough to systematically evaluate the program in light of its multiple objectives. Continued assessments of program success will enhance program design, targeting effectiveness, implementation, and monitoring and evaluation. 10.9 Expansion of the Programas Alvorada and PETI. Recently, in September 2000, the Brazilian Government has moved to scale-up and integrate separate cash-grant initiatives financed by the federal government (FGRM-Fundo de Garantia da Renda Minima and PETI-Programa de Erradica,ao do Trabalho Infantil) under the Alvorada program. Implementation continues the primary responsibility of states and municipalities. This national program will seek, over the next few years, the gradual implementation of a Bolsa Escola program in poor municipalities with the worse indices of social development in country. This policy initiative represents as a major step forward in program design and fiscal commitment, it integrates within one coherent program both preventive initiatives (Bolsa Escola/FGRM with two million beneficiaries in 2000) and the remedial child-labor eradication program (PETI, with four hundred thousand beneficiaries in 2000). Program Alvorada will permit MEC to address design and implementation issues hindering the effectiveness of the FGRM program - very low cash transfer level, complex benefit design and target exclusion of many poor municipalities in the Northeast and North. By building on the complementarity of existing conditional cash transfers to the poor, and improving coordination with educational initiatives aiming at universalizing basic education, it has the potential do significantly increase targeting and effectiveness. 27 APPENDIX I DATA TABLES TABLE 1.1 School Scholarship Program Coverage in the different areas Belem 44,762 4,423 (June 1998) Belo Horizonte 51,730 3,300 (planned Dec./98) Brasilia 59,926 25,312 (Jan. 1999) (*) Estimates obtained from PNAD-1996 (Special Tabulations by Sonia Rocha) (**) Information from the local programs TABLE 2.1 Selected Indicators Metropolitan Nuclei and the Federal District 1996 t:0 dififi-ul At00:: i TAi00; Tot 00 ti Per cai# Nme (%) Garb \f gei W0tr Sewerge North / Northeast Belem (N) 965 308 44,762 19.66 87.84 81.16 8.31 Fortaleza (NE) 814 254 114,726 21.77 91.40 73.80 12.78 Recife (NE) 809 285 76,308 19.84 90.03 85.96 29.05 Salvador (NE) 898 318 148,323 23.11 89.53 89.44 42.64 Center South Belo Horizonte (SE) 1307 452 51,730 8.95 93.50 96.62 87.86 Rio de Janeiro (SE) 1410 558 113,874 6.28 97.50 97.13 69.95 Sao Paulo (SE) 1547 538 168,298 6.11 99.16 98.93 86.42 Curitiba (S) 1589 567 13,791 3.26 98.62 97.86 47.93 Porto Alegre (S) 1563 633 30,099 6.88 98.64 95.58 33.89 Federal District (C) 1435 480 59,926 12.01 94.99 86.96 72.67 Source: IBGE / PNAD 1996 (*) Data refer to 1995 Concepts refer to the number of persons subjected to the following conditions: * Adequate water - public pipeline system and internal plumbing; * Adequate sewerage - public sewerage system; * Adequate garbage collection - directly or undirectly collected. 28 TABLE 2.2 Enrollment rates for working and non-working children in urban areas Boy Girl Totall Working? No 0.94 0.95 0.9'i Yes 0.84 0.83 0.83 Total 0.93 0.94 0.94 Source: Rocha, March, 1999. TABLE 2.3 Lag in grade for working and non-working children in urban areas Boy Girl Total Working? No 43.91 37.82 40.77 Yes 50.34 41.19 47.32 Total 44.65 38.01 41.33 In this table we compute the mean of gfalag (gfalag=100-(attainme/age-6)*100). Source: Rocha, March, 1999. TABLE 2.4 Incidence of Child Labor in urban areas- percentages A_e ROY Girl Total 10 0.05 0.01 0.03 1 1 0.07 0.03 0.05 12 0.10 0.05 0.07 13 0.13 0.07 0.1( 14 0.22 0.11 0.16 Total 0.12 0.06 0.09 Source: Rocha, March, 1999. 29 TABLE 2.5 Child Labor in the Federal District - 1996 Alt 10o 14rs old04 chidldren 208,502 Who work - number 7,463 - percentage (%) 3.58 10 to 14 yes old hildn in flimiies ws peTrpr incomeslow Ai M"W 36,627 Who work - number 1,810 - percentage (%/O) 4.94 Who work at least 15 hours a week - number 1,584 - percentage (%) 4.32 Source: IBGE/PNAD 1996 (Special Tabulations). TABLE 2.6 Child Labor, School Attendance and the Overlap by Region W O } QiS & Wr Ntr ________ North 1.7 % 82.8 % 9.9% 5.5 % 100% NE 3.6 % 72.6 % 17.2 % 6.6 % 100 % MG/ES 2.5% 82.5% 11.0% 4.0% 100% Rio 1.0% 91.9% 2.9% 4.1 % 100% Sao Paulo 0.8 % 91.6 % 4.9 % 2.8 % 100 % Sul 1.9% 81.0% 13.2% 3.9% 100% Centro - Oest 2.4 % 81.3 % 10.8 % 5.5 % 100 % DF 0.6 % 94.1% 3.5 % -1.9 % 100 % Urban 2.1 % 85.1% 8.3% 4.2% 100% Brasil 2.3 % 81.5% 11.4% 4.8% 100% Source: PNAD, 1996. Special Tabulations by Sonia Rocha 30 TABLE 2.7 Children out of school, according to the level of family income 7 to 14 years ohi out of school Nucei Below .5 MW per capita income Below .25 MW per capita inome Number % Number % Belem 3,924 9.71 1,742 11.99 Fortaleza 10,683 9.54 5,127 14.81 Recife 4,992 6.88 1,664 5.60 Salvador 10,359 9.31 3,523 8.22 Belo Horizonte 4,822 8.49 1,521 8.82 Rio de Janeiro 4,467 5.06 1,676 4.77 Sao Paulo 8,057 7.63 2,687 6.13 Curitiba 1,858 14.30 265 8.33 Porto Alegre 1,662 6.51 415 6.67 Distrito Federal 4,746 7.55 1,521 7.82 Total 55,570 8.07 20,141 8.16 Source: IBGE/PNAD 1996, Special Tabulations by Sonia Rocha. TABLE 2.8 Child Labor in the Ten Metropolitan Areas with Bolsa Escola Nuclei Children who work (10 to 14) Number _ Belem 873 3.37 Fortaleza 5,554 8.91 Recife 2,496 5.88 Salvador 1,245 1.73 Belo Horizonte 2,791 8.88 Rio de Janeiro 558 0.99 Sao Paulo 895 1.35 Curitiba 265 2.94 Porto Alegre 624 4.17 Distrito Federal 1,810 4.94 Total 17,111 4.10 Source: IBGE/PNAD 1996, Special Tabulations by Sonia Rocha. 31 TABLE 3.1 Basic Criteria for Benefit Concession Federal District, Belo Horizonte and Belem Crti Fea fliDsuiet &1eMi::: : :.m,te j: Per Capita Family Income * formal criteria (income per capita) 0.5 minimum wage 65.64 UFIR's 0.5 minimum wage * in Nov/98 R$ 65.00 63.09 65.00 Presence of Children (age bracket) 7 to 14 7 to 14 4 to 14 Minimum period of residence 5 years 5 years 3 years Source: Lavinas (1998). Minimum wage at R$130 in November 1998 TABLE 3.2 Number of families according to the age of children (Families living in the area for at least five years) N400i Wt_t hi u Wth 77 total without 7-14 Belem 2,182 24,122 15,405 Fortaleza 6,404 69,227 35,467 Recife 7,278 36,384 21,204 Salvador 12,846 64,210 38,730 Belo Horizonte 3,297 31,447 13,949 Rio de Janeiro 17,862 47,450 24,561 Sao Paulo 25,066 34,913 25,064 Curitiba 1,062 8,483 4,243 Porto Alegre 4,357 18,270 10,589 Brasilia 1,808 13,115 8,368 Total 82,162 347,621 197,580 Source: IBGE/PNAD 1996, Special Tabulations by Sonia Rocha 32 TABLE 3.3 The Effects of Selection Criteria on the Number of Families Covered by the Program AU Families be'w Fuilr belw 11I MW Fa*ii1s b*6w 1/2 MW & 112 MW quaWy & chidren 7-44 quwly eb 7-14 & 5 yr residency No. of Families (000) No. of Families (000) No. of Families (000) Belem 45 22 21 Fortaleza 115 62 :57 Recife 76 40 :38 Salvador 125 62 .56 Belo Horizonte 52 30 :26 Rio de Janeiro 114 54 44 Sao Paulo 139 59 35 Curitiba 14 7 6 Porte Alegre 30 13 13 Brasilia 60 32 6 Total 770 382 302 Based on a transfer of I min wage per family. Source: IBGE/PNAD 1996 TABLE 3.4 Reduction in potential beneficiary families due to the residency requirement and estimates of program's expenditures as proportion of current receipts ~~~~~~~~~~~lamO . _ , at last three yrsofreince Ateastfiveyears ofresidence Nuclei % Of Current % ofCtrrent T () Receipts (*) Receipts Belem -5.8 12.01 -6.46 11.93 Fortaleza -4.5 15.96 -6.92 15.55 Recife -4.6 11.70 -6.18 11.,1 I Salvador -7.7 19.30 -8.42 19.1.6 Belo Horizonte -10.2 3.68 -11.87 3.61 Rio de Janeiro -17.5 2.51 -18.55 2.48 Sao Paulo -40.9 0.85 40.91 0.85 Curitiba -14.8 1.48 -14.80 1.48 Porto Alegre -3.1 3.11 -4.69 3.06 Brasilia -81.1 0.23 -81.12 0.23 Total -19.7 2.73 -20.70 2.70 Source: Special Tabulations by Sonia Rocha*Restricting first by criteria that children are 7 to 14 years olil and that per capita income is below .5MW. 33 TABLE 4.1 Number of families with 7 to 14 year old children whose per capita income is below .5 MW and estimates of program-expenditures as proportion of current receipts according to the value of the family benefit Nudlei Nwn0ber % o Curret Recipts Familie One Mm. Wg W Belem 22,524 12.75 6.38 2.55 Fortaleza 61,743 16.71 8.35 3.34 Recife 40,342 12.27 6.13 2.45 Salvador 61,527 20.92 10.46 4.18 Belo Horizonte 29,921 4.10 2.05 0.82 Rio de Janeiro 54,145 3.04 1.52 0.61 Sao Paulo 59,081 1.44 0.72 0.29 Curitiba 7,160 1.74 0.87 0.35 Porto Alegre 13,285 3.21 1.61 0.64 Brasilia 32,335 1.23 0.62 0.25 Total 382,063 3.40 1.70 0.68 Source: IBGE/PNAD 1996, Special Tabulations by Sonia Rocha. TABLE 4.2 Fiscal costs of giving V/2 minimum wage based on the number of children in the family Nucle t Num000|br of %ofrent U % 000 :;7s Red,ucton00030 Belem 37,485 10.61 11% Fortaleza | 103,823 14.05 10% Recife 67,998 10.34 11% Salvador 101,712 17.30 10% Belo Horizonte 50,459 3.46 4° Rio de Janeiro 71,457 2.01 19% Sao Paulo 56,394 0.69 19% Curitiba 10,877 1.32 11% Porto Alegre 24,500 2.96 3% Brasilia 9,045 0.17 26% Total 533,750 2.38 12% Source: IBGE/PNAD 1996, Special Tabulations by Sonia Rocha. 34 TABLE 5.1 Dropout Rates among Beneficiaries and Non-Beneficiaries (%) Year E3Beficiaries Non-Benerfciaries Total 1994 - - 6.2 1995 0.2 6.5 6.0 1996 0.4 5.6 5.2 Source: School Scholarship Program (1996) and School Census (1995) as cited by Sant'Ana and Moraes (1997). TABLE 5.2 Promotion Rates Estimates - 1995-1996 ________________ _ 1995 1996 Beneficiaries (a) 79.8 87.9 (b) 67.0 80.2 Non-Beneficiaries (a) 79.5 79.9 _______________ _ t b) 70.9 72.2 Total (a) _ 79.6 80.4 -(b) 70.8 73.5 (a) Estimates by SanfAna and Moraes (1997) including the Literacy Cycle where promotion is automatic. (b) Estimates which consider the 3rd. to the 8th. grades only TABLE 6.1 Learning Performance of Beneficiaries and Non-Beneficiaries Selected from the SAEB Sample Fourth Graders - 1997 Subjec Beneficiaries Non-Beneficiaries Number % correct number °/O correct Portuguese 143 38.7 374 37.4 Mathematics 159 37.8 377 37.5 Science 162 42.4 377 43.5 Source: Abramovay, Andrade and Waiselfisz (1998), p.134 35 TABLE 8.1 The Effects of the Selection Criteria on the Fiscal Sustainability of the Program All Famtii# beloW Fautiies bIOMw 1/2 MW od"6 14 */2~ MW qafy Ciru714qwf ChIl"e 744 ya rsdec Costs Costs Costs Belem 25.3 % 12.8 % 11.9 % Fortaleza 31.1% 16.7 % 15.6 % Recife 23.2 % 12.3 % 11.5 % Salvador 42.6 % 20.9 % 19.2 % Belo Horizonte 7.1 % 4.1 % 3.6 % Rio de Janeiro 6.4 % 3.0 % 2.5 % Sao Paulo 3.4 % 1.4 % 0.9 % Curitiba 3.3 % 1.7 % 1.5 % Porte Alegre 7.3 % 3.2% 3.1% Brasilia 2.3 % 1.2 % 0.2 % Total 6.8 % 3.4 % 2.7 % Based on a transfer of I min wage per family. Source:IBGE/PNAD 1996. TABLE 8.2 Program's Cost as % of Current Municipality Revenues by Level of Transfer to Families Belem 11.9 % 6.0 % 3.0 % Fortaleza 15.6 % 7.8 % 3.9 % Recife 11.5 % 5.8 % 2.9 % Salvador 19.2 % 9.6 % 4.8 % Belo Horizonte 3.6 % 1.8 % 0.9 % Rio de Janeiro 2.5 % 1.3 % 0.7 % Sao Paulo 0.9 % 0.5 % 0.3 % Curitiba 1.5 % 0.8 % 0.4 % Porte Alegre 3.1 % 1.6 % 0.8 % Brasilia 0.23 % 0.1 % 0.0 % Total 2.7 % 1.4 % 0.7 % Source: Special Tabulations by Sonia Rocha. 36 APPENDIX II PROGRAM CRITERIA, DESIGN AND IMPLEMENTATION This section describes more explicitly the actual criteria, design and implementation details of the programs that have been put in place beginning in 1995 until today.'4 Tables 1-II through 6-II demonstrate when the programs began in each municipality, the selection criteria used, the number of families benefiting relative to the number of famnilies eligible by municipality, how much and what type of support (income and/or in- kind services) families receive, and requirements of children to remain in the program. Program Initiation and Duration The first Bolsa Escola programs began in Brasilia and Campinas in 1995. However similar programs began as early as 1986 in some municipalities of the state of Sao Paolo (see Tables 1 and 2). In Sao Paolo state, up until March of 1999, forty-five programs had been implemented. Another eleven municipalities across Brazil, from Amapa to Porto Alegre, have received the prograrm since 1995. This makes a total of 58 programs in all of Brazil according to the latest program data by Lavinas and Bittar. Selection Criteria The selection criteria used by the programs are much more varied than the originally proposed selection criteria (see Tables 3-II and 4-II). The basic categories uitilized are: a) the presence of children of a certain age; b) a maximum family incorme; and c) a residency requirement. Within these categories there is a great deal of discretion. Some additional criteria involved a socioeconomic evaluation, having a mother as head of household, and other various characteristics. Children's Age Only one out of the twelve programs in municipalities other than in the state of Sao Paolo require that children are between the ages of 0 and 14, seven require that children are between ages 7 and 14 and the remaining four have requirements which vary to include both younger and/or older children. In Sao Paolo, the majority of municipalities require that children be between the ages of Oand 14 while a great number have other requirements; for example, children between the ages of 0 and 20 or 0 to 14, or even that the child possesses some sort of learning deficiency. Family Income The three possible criteria used for income are a) per capita family income up to R$35, b) up to R$60 or c) the other category which varies to include from 1/2 of minilmum wage to 1/4 of minimum wage, one or even two times of minimum wage. Out of the 58 municipalities, nine require that beneficiaries eam at least R$35, fifteen accept 14 Lavinas, L. and M. Bittar. 1999. 37 beneficiaries who earn up to R$60, and the remaining fall in the other category. Within this third category, income varies greatly. Residency Requirement The criteria used for residency includes 2 years, 5 years or the other category (which indicates either 1 year, 3 years or after a certain date). Forty-three of the municipalities require only two years of residency for eligibility in the program, ten require a minimum residency of 5 years and the remaining municipalities vary between I and 3 years. Mother Head of Household Out of the 58 municipalities, 19 give preference to families in which the mother is the head of the household. The remaining do not use this as a criteria for selection. Other Criteria Some of the municipalities also use other criteria which may include, for example, the requirement that children are deemed to be "at risk" or that they live in a certain area chosen by the program, or criteria based on a socioeconomic evaluation of the family. Number of Beneficiaries Based on the selection criteria discussed above, a number less that or equal to the number of potential beneficiaries are benefiting from the program. A large number of the programs are serving close to the total potential beneficiaries (see Tables 1-II and 2-II). In some cases such as Belem and Belo Horizonte, the number of beneficiaries is much less than the potential beneficiary population (52% and 19% respectively). In Sao Paolo state as well, there are many municipalities which are not reaching all potential beneficiaries. The reasons for the programs not reaching all potential beneficiaries are not clear from this data; however, it may be posited that the inability to reach the entire parameter population has to do with programn affordability in many cases. Program Benefits There are differences between municipalities in the type, amount and periodicity of the distribution of the program benefits. Out of the 58 municipalities, there are four programs which provide only some type of in-kind or service transfer, three programs which provide this and a cash grant, and 51 programs that provide only a cash grant. Some examples of transfers other than cash include gas for use in cooking, construction materials, food baskets and transportation vouchers. The amount of the cash grant as well as the periodicity (monthly, annually or quarterly) varies between municipalities as well. 38 Requirements of Children A family's ability to continue to receive benefits from the program includes five main criteria and also varies by municipality (see Tables 5-11 and 6-11). School attendance, school performance, participation in other programs and another category which includes several different requirements, are those considered. The majority of the programs use school attendance as a criteria to remain in the program. Only three out of the twelve non-Sao Paolo municipalities use school performance as a criteria for children to remain in the program. In Sao Paulo, however, this criteria was much more commonly used (22 out of 46 municipalities). The same is true in all of the municipalities for the requirement of participation in other types of programs. Some of the other requirements include such things as vaccination of children, children not working, and "appropriate" use of the funds. This additional criteria may add substantially to the benefits received by children involved in the program. Table 1.11 General Program Data _ _v_ _ of Benefit No. of Adtul MonWtyear No. of Families Benefieqy Income Food Other Mumicipality Ji~ogram Began Seleted F, milies ______ Bashet ________k 1. Amapa January/96 1200 1200 X 2. Bel6m January/97 9320 4820 X 3. Belo Horizonte August/97 27419 4500 X 4. Boa Vista June/99 423 423 X 5. Brasilia May/95 25680 23000 X 6. Goias January/96 110000 110000 X Kitchen gas 7. Mundo Novo May/98 300 300 X 8. Paracatu May/98 3040 880 X 9. Porto Alegre April/96 1621 1604 X 10. Recife September/97 1000 800 X 11. Sao Luis March/98 29615 29615 X X 12. Tocantins August/96 60 20 X Source: Lavinas, Le Bittar, M., IPEA, 1999 39 Table 2-11 General Program Data for the State of Sao Paulo :Nl n n a edpaliy ; 0 Name ofthe Prosm inp Ff Aet; 0 f ad ; I Aparecida/SP Programa Complementando a Renda Mar/99 60 20 X 2. Areias/SP Programa Complementando a Renda Mar/99 38 20 Family Support 3. Cagapava/SP Programa Complementando a Renda Jan/86 208 25 X 4. Campinas/SP Programa de Garantia de Renda Familiar Minima Mar/95 9180 4530 X 5. Campos do Jordao/SP Programna Complementando a Renda 431 89 X 6. Caraguatatuba/SP Programa Complementando a Renda Jul/93 120 15 X 7 Catanduva/SP Programa Bolsa Escola Jun/97 721 721 X 8 Cruzeiro/SP Programa Complementando a Renda 0 50 X 9 Ferraz de Vasconcelos/SP Programa de Renda Familiar Minima Jul97 310 109 X 10 Franca/SP Programa Bolsa Educagao Jun/96 330 180 X 11 Guaratingueta/SP Programa de Renda Familiar Minima May/95 68 40 X 12 Guariba/SP Programa Complementando a Renda Oct/86 400 35 X 13 llhabela/SP Programa Complementando a Renda Mar/99 42 45 X 14. Jaboticaba/SP Programa de Renda Familiar Minima Apr/98 104 55 X 15. Jambeiro/SP Renda Minima Jan/00 114 0 X 16. Jambeiro2/SP Programa Complementando a Renda May/97 135 44 X 17. Jambeiro3/SP Fortalecendo a Familia Mar/99 95 95 X I8. Jundiai/SP ProduSao Associada com Garantia de Renda May/96 534 243 X 19. Lagoinha/SP Programa Complementando a Renda Mar/99 21 21 X 20. Lavrinhas/SP Programa Complementando a Renda Mar/99 14 14 X 21. Lavrinhas2/SP Fortalecendo a Familia Mar/99 225 225 X X 22 LimeiralSP Programa de Garantia de Renda Familiar Minima Mar/98 578 528 X 23. Monteiro Lobato/SP Programa Complementando a Renda Sep/87 200 42 X 24 Natividade da Serra/SP Programa Complementando a Renda Mar/99 43 43 X 25. Osasco/SP Programa de Combate a Miseria e Garantia de Sep/96 649 370 X 26. Ourinhos/SP Programa de Garantia de Renda Familiar Minima Apr/98 692 133 X 27. Paraibuna/SP Programa Complementando a Renda Jun/87 455 29 X 27. Paraibuna2/SP Fortalecendo a Familia Jan/98 758 560 X Food Stamps 28. Pindamonhangaba/SP Programa Complementando a Renda Jan/97 118 43 X 29. Presidente Prudente/SP Programa de Garantia de Renda Minima Sep/97 3100 387 X 30. QueluzlSP Programa Complementando a Renda Mar/85 400 18 X 31. Reden9io da Serra/SP ProgwramaComplementando aRenda Mar/99 31 31 X 32 Ribeirao/SP Programa de Garantia de Renda Familiar Minima Dec/95 2089 1300 X 33. Santa Branca/SP Programa Complementando a Renda Mar/99 70 25 X 34. Santo Andre/SP Programa Familia Cidadg Aug/98 466 422 X 35. San Antonio de/SP Programa Complementando a Renda Dec/86 210 24 X 36 Santos/SP Programa Nossa Familia Aug/98 301 117 X X Transportation 37 S3o Bento do Sapucai/SP Programa Complementando a Renda Jan/92 300 150 X 38. SAo Jose dos Campos/SP Programa de Garantia de Renda Minima e de Nov/97 4537 450 X 39. S3O Sebastigo/SP Programa Complementando a Renda Mar/99 40 22 X 40 53o Sebastiao2/SP Fortalecendo a Familia Mar/99 60 43 X 41. Silveiras/SP Programa Complementando a Renda 0 24 X 42. Taubate/SP Programa Complementando a Renda Jan/96 108 44 X Construction Materials 43. Tremembe/SP Programa Complementando a Renda Jan/97 62 32 X 44. UbatubW/SP Programa Complementando a Renda Mar/99 24 21 X 45. Ubatuba2/SP Fortalecendo a Familia Mar/99 41 36 X Source: Lavinas, Le Bittar, M., IPEA 1999. 40 Table 3-IL Program Selection Criteria _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ~~~~~~~~Crkteria _______ ______ __ __ Childs Age Pe Capita Family Icome Municipality Cbildrn or From 0 to From I to Other Up to Up toRS6Q,00 Other Dependents 14years 14 years R$354)O _ 1. Amapa X X X X 2. Belem X 4 - 14 years X up to 12 s.m. 3. Belo Horizonte X X X up to R$65,00 4. Boa Vista X X X up to /2 s.m. 5. Brasilia X X X up to ½/2 s.m. 6. Goias X 0 - 6 years X up to I s.m. 7. Mundo Novo X X X X 8. Paracatu X X X X 9. Porto Alegre X 0 - 18 years X up to 14 s.m. 10. Recife X X X X 11. Sao Luis X 7 - 16 years X X 12. Tocantins X X X up to 2 sm. Time off*sideace Time Of Socio- Families wI Municpality Residence 2 years 5 years Other Other Criteria Ekonomic Mother Head ._______________ ._______ ._.._.._...__ Status of Household 1. Amapa X X X X 2. Belem X 3 years Living in are X X selected by program 3. Belo Horizonte X X X X 4. Boa Vista X X 14 years X X 5. Brasilia X X X X 6. Goias X X X 7. Mundo Novo X 2 years in munic. & X 5 years in the state 8. Paracatu X X X X 9. Porto Alegre X In risk situation X X 10. Recife X x X 11. Sao Luis X 3 years Children x X attending school 12. Tocantins I_I Source: L. Lavinas and M. Bittar. IPEA. 1999 41 Table 4-11 Program Selection Criteria for the State of Sao Paulo Mnnidpslity Childr or fromO0 to Pr~u Ito te tC4t ptRVt*SA t 1. Aparecida/S? X X X x 2 Areias/SP X X X X 3. Cagapava/SP X Up to lMW 4. Campinas/SP X X Disabled X X 5 Campos do JoTdao/SP X 0 to 16 years X Up to R$26 6 Caraguatatuba/SP x 0 to 20 years X X 7. Catanduva/SP X X X X 8 Cruzeiro/SP X X X Up to RS20 9. Ferraz de Vasconcelos/SP X X X X 10 Franca/SP X 7 to 16 years X X I1. Guaratingueta/SP X X X Upto¼MW 12. Guariba/SP X 0 to 18 years X Up to RS40 13. Ilhabela/SP X X X X 14. Jaboticaba/SP X X X Up to RX50 15. Jambeiro/SP X X 16. Jambeiro2/SP X X x x 17. Jarnbeiro31SP X X 18. Jundiai/SP X X 19. Lagoinha/SP X X X Up to R$50 20. Lavrinhas/Sp X X X Up to RS50 21 Lavrinhas2/SP X X 22. LimeiratSP X 6 to 14 years Family income < 2 MW 23 Monteiro Lobato/SP X X 24. Natividade da Serra/SP X up to 15 years X Up to R$50 25 OsascolSP X X disability X X 26. OurinhostSP X X X X 27. ParaibuntaSP X X Family income < R$50 27. Paraibunsa2/SP x X Family income < RS50 28. Pindamonhangaba/SP X X disability X Up to R$50 29. Presidente Prudente/SP X X X Up to 37.5% MW 30. QueluzlSP X X 31. Redencao da SerTa /SP X o to 15 years X Up to R$50 32. Ribeiruo Prieto/SP X X Family income < 33. Santa Branca/SP X X Variable 34. Santo Andre/SP X X X Up to ¼/ MW 35. San Antonio de PinhaltSP X X Family income < 36. Santos/SP X 0 to 16 years X Up to R$50 37. Sao Bento do Sapucai/SP X X Up to 18 years X - disability 38 Sao Jose dos Campos/SP X X X Up to / MW 39 Sao SebastiaolSP X X X X 40. Sao Sebastiao2/SP X X X X 41. Silveiras/SP X X X X 42. Taubate/SP X X X Up to R$S50 43. TremernbdlSP X X X Up to RS50 44. Ubatuba/SP X variable X X 45. Ubatuba2/SP X X Source: Lavinas, Le Bittar, M., IPEA 42 _ _ _ _ _ _ _ __ _ ._ _ _ _ _ _ _ Othet tistf,ntnS_,_ rime of __. . Fa2=wIw/ Mot* municip*11y Resience Z yCan 5yars Other OtberCri* EA.a.mk Motlid Of Hieo8d ._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Ie is h atio a _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1. Aparecida/SP X X X X 2 Areias/SP X I year X 3 Cacapava/SP X X X X 4. Campinas/SP x 2 years after law X 5 Campos do Jordao/SP X X X 6. Caraguatatuba/SP x I year X 7. Catanduva/SP X 3 years Application and X home visit 8. Cruzeiro/SP X X X 9. Ferraz de Vasconcelos/SP X X X X 10 Franca/SP X X X X 11. Guaratingueta/SP X X X X 12. Guariba/SP X I year X 13. Ilhabela/SP X X X 14. Jaboticaba/SP X 3 years X 15. Jambeiro/SP X I year X 16 Jambeiro2/SP X X X 17. Jambeiro3/SP X X X 18. Jundiai/SP X X Presence of child & X school attendance 19. Lagoinha/SP X X X 20. Lavrinhas/SP X 21. Lavrinhas2/SP X Unemployed X 22. Limeira/SP X Since 1993 X 23. Monteiro Lobato/SP X I year Application and X home visit 24. Natividade da Serra/SP X X x 25. Osasco/SP X Before July 1994 X 26. Ourinhos/SP X X X X 27. Paraibuna/SP X X X 27. Paraibuna2/SP X X x 28. Pindamonhangaba/SP X X X 29. Presidente Prudente/SP X 3 years X 30. QOueluztSP X 31 Redencao da Serra /SP X X X 32. Ribeirao Prieto/SP X X Housing X X 33. Santa Branca/SP Live in the X Municipality 34. Santo Andre/SP X 3 years X X X 35. San Antonio de/SP X 36. SantosISP X I year Situation of nsk X X 37. Sao Bento do Sapucai/SP X X Number of children X X 38. Sao Jose dos Campos/SP X X x x 39. Sao Sebastiao/SP X X X 40. Sao Sebastiao2/SP X X X 41. Silveiras/SP Live in the X Municipality 42. Taubate/SP X X X 43. Tremembe/SP X X x 44 Ubatuba/SP 45. Ubatuba2/SP _ Source: Lavinas Le Bittar, M., IPEA. 43 Table 5-11 Program Requirements of the Children IffrlSiipa~lty . mpose School Sch~*ln Others 1. Amapa X X X 2. Belem X X Children not working 3. Belo Horizonte X X 4. Boa Vista X X X 5. Brasilia X X Vaccinations 6. Goias X X 7. Mundo X X X Mirim Patrol 8. Paracatu X X X 9. Porto X X X Good use of money 10. Recife X X 11. Sao Luis X X 12. Tocantins X X X 44 Table 6-11 Program Requirements of the Children in the State of Sao Paulo ._r..dp,ti" i n idtPwky Impoap ScboeR Attewn School Perfornue oter Progyms Others I Aparecida/SP X X X X 2 Areias/SP X X X X 3. Cagapava/SP X 4 Campinas/SP X X X Using resources to improve the quality of life 5 Campos do JordAo/SP X X 6. Caraguatatuba/SP X X 7 CatanduvatSP X X X 8. Cruzeiro/SP X X X X 9. Ferraz de Vasconcelos/SP X X X Vaccination 10. Franca/SP X X 11. Guaratingueta/SP X X X 12. Guariba/SP X X 13. Ilhabela/SP X X 14 Jaboticaba/SP X X X X 15. Jambeiro/SP X X X 16. Jambeiro2/SP X X 17. Jambeiro3/SP X 18. Jundiai/SP X X X 19. Lagoinha/SP X X X X 20. Lavrinhas/SP X X 21. Lavrinhas2/SP X X X X 22. Limeira/SP X X 23. Monteiro Lobato/SP X X 24 Natividade da Serra/SP X X X X 25. Osasco/SP X X X 26. Ourinhos/SP X X X 27. Paraibuna/SP X X X 27. Paraibuna2/SP X X X 28. Pindamonhangaba/SP X 29. Presidente Prudente/SP X X X 30 Queluz/SP X X 31. Reden,co da Serra /SP X X X X 32. Ribeirao Prieto/SP X X X 33. Santa Branca/SP X X 34. Santo Andre/SP X X X X 35 San Antonio de Pinhal /SP X X X X 36. Santos/SP X X X X 37. SAo Bento do Sapucai/SP 38. SAo Jos6 dos Campos/SP X X 39. Sao Sebastiao/SP X X X X 40. Slo Sebastiao2/SP X X X 41. Silveiras/SP X X X 42. Taubate/SP X X X 43. TrememWSP X 44 Ubatuba/SP X 45. Ubatuba2/SP X X X Source: Lavinas, Le Mariana, B., IPEA 1999 45 APPENDIX III PROGRAM TARGETING IN THE FEDERAL DISTRICT This appendix discusses in more detail the evaluation of the pioneer program of Bolsa Escola in Paranoa, D.F done by Saboia and Rocha in 1998. This evaluation demonstrates the effective capacity of the program to distinguish the most destitute families and to incorporate them into the program. In the study, first family characteristics of potential beneficiary and non-beneficiary families are observed and second, using the PNAD- 1995 dataset, they calculate the number of families who satisfied the program criteria and compare this with the number of families who actually benefited after a year and a half of implementation. FAMILY CHARACTERISTICS OF POTENTIAL BENEFICIARIES15 Comparing potential beneficiaries to non-beneficiaries they first examine family characteristics and find that 96.8% of potential beneficiaries are female, demonstrating that the program would be successful in making the monetary transfer to the mother or the adult responsible for the child in the household. They also find that 90% of potential beneficiaries were born outside of the Federal District -an indication of high migration to the area. However, since one of the selection criteria established by the program is a minimum residence of five years in the area, they also compare potential beneficiaries and non-beneficiaries in terms of number of years spent living there. They find that only three out of the 2,629 families had lived in the area less than five years. In fact, almost 23 percent of the beneficiaries had lived in the area for more than twenty-five years. This demonstrates that the program is not capturing recent migrants and that it is not providing an apparent incentive for migration. Table 1-IH Time of Residence in Paranoa, D.F. No. of Cases % No. of Cases % Less than 5 years 3 0.1 --- 0.0 5/9 years 511 19.4 88 17.3 10/14 years 497 18.9 83 16.3 15/19 years 575 21.9 125 24.6 20/24 years 443 16.9 96 18.9 25 or more 599 22.8 116 22.8 Total 2,629 100.0 508 100.0 Saboia and Rocha, 1998. 15 Potential beneficiaries refers to program applicants. 46 In addition, Saboia and Rocha examine age distribution among potential beneficiaries and non-beneficiaries as well as their partners, and find that the average age is 37 years which would be consistent with having children between the ages of 7 and 14. Family composition is another important factor in considering whether or not the program attracts the appropriate beneficiaries. The average number of family members was approximately five and 9.3 percent had seven family members. A larger number of family members indicates a smaller family income especially if many of those members are children. Table 2-Ill shows that of applying families, 54.6 percent have children age 14 or below and 13.7 percent have children older than 15, while the potential beneficiary parent makes up 19.8 of the households and their partners 11.9 percent. Sumrning up, the potential beneficiary families correspond with the model of the organization of a Brazilian family -nuclear family- with few members from the extensive family. Table 2-III Family Composition - Paranoa Potential Bfeneicinries Non-Beneficiaries No. of Cases % No. of Cases % Family representative'6 2.629 19,8 508 22,4 Partner 1.589 11,9 339 14,9 Dependents (0/14) 7.267 54,6 1.136 50,0 Child 7.054 53,0 1.104 48,6 Niece of nephew 181 1,4 25 1,1 Other 32 0,2 7 0,3 Dependents (15 or +) 1.823 13,7 289 12,7 Child 1.750 13,1 278 12,2 Niece of nephew 7 0,1 0,0 Mother/Father 34 0,3 8 0,4 Mother/Father In-Law 6 0,0 1 0,0 Other 26 0,2 2 0,1 Total Number of Indiv. 13.308 100,0 2.272 100,0 Saboia and Rocha, 1998. Saboia and Rocha also examine the potential beneficiaries' (parents') educational attainment levels which is an important factor to consider in that often the poverty cycle is repeated due to lack of education. The evidence demonstrates that 26.4 percent are illiterate, 61.9 percent are barely literate, 10.4 percent have completed the first grade and only 1.3 percent have completed second grade. The evidence also demonstrates for potential beneficiaries the slow progression through school and increase by age in being behind in school -at age 7 the average lag in school attainment is 0.3 years while at age 14 it has reached 3.6 years. Again, this demonstrates that the program would reach families who have a pattern of low school attainment, a result of slow progression through school. 16 Family representative is usually the mother as the benefit goes to the mother by design of the prograrn. 47 The labor status of potential beneficiaries, demonstrated in Table 3-Ill, is also examined showing that only 53.2 percent are working; of those only 45.4 for salary and 53.9 percent are self-employed. Of the partners of the potential beneficiary parents that are working, 86.7 percent are working, which emphasizes that men are the breadwinners in most households. Table 3-Im Labor Status of Potential Beneficiaries and Non-Beneficiaries - Paranoa B*Fami703nly R Bei p retah Pai . e No. of % No. of % No. of % No. of % Cases Cases Cases Cases Working 1,398 53.2 293 57.7 521 86.7 29 96.7 Not Working 1,229 46.8 215 42.3 80 13.3 1 3.3 Total 2,627 100.0 508 100.0 601 100.0 30 100.0 Income is another important family characteristic to examine when determining whether or not the program is attracting the intended beneficiaries. Tables 4-III and 5-III demonstrate that, as expected, average family income of beneficiaries was very low. Average family income was equal to R$ 175.40, while average per capita income was equal to R$35.00. Considering minimum wage as a benchmark for per capita income, 29 percent earned less or equal to 1/4 minimum wage, 85.2 percent earned less up to 1/2 of minimum wage and 99.1 percent up to 0.7 of minimum wage. 17 Table 4-II Average Family Income of Beneficiaries - Paranoi Tow ita come _______._________:_______o _ . - d AmountS (1$ No; of ' O720S ;ig0;002a1 offtI.eomff! Am- (,Ri $ Work Up to 100 630 24,0 Family Representative 170.332 36,9 101/200 1.311 49,9 Partners 198.323 42,9 201/300 514 19,6 Other members 25.911 5,6 301/400 134 5,1 Pension 19.566 4,2 401/500 27 1,0 Retirement 7.851 1,7 501/600 8 0,3 Rent 15.498 3,4 601/700 2 0,1 Food voucher 3.293 0,7 701/800 1 0,0 Other 21.194 4,6 Total 2.627 100,0 Total 461.968 100,0 '7 It should be noted that while the selection criteria requires that beneficiaries have an income of less than or equal to 1/2 of minimum wage, in practice the program has actually included those families with income up to 0.7 of minimum wage. This allowed for a larger number of families to be reached and in Paranoa raised the number by 15 percent. 48 Table s-M Income Per Capita Amoutnt (RS) No-o oCases _____ Up to 25,00 740 29,0 25,01/50,00 1.436 56,2 50,01/70,00 341 13,4 70,01/100,00 33 1,3 100,01/125,00 3 (1,I Total 2.553 100,0 Seboia and Rocha, 1998. Evaluation of Program Targeting Saboia and Rochas' examination of the PNAD data demonstrates, based on the criteria, that the number of beneficiaries reached by the end of 1996 was 27,000 equaling an impressive 77 percent of the total number of potential beneficiaries in 1995 which was 35 thousand families. The question is, however, if these families in effect demonstrated the characteristics of the target population. The evidence demonstrates that the program did reach those intended in that the characteristics of the beneficiary families matched the selection criteria. The indicators also demonstrate that the beneficiaries were actually subject to socioeconomic conditions more adverse than the population limited by the selection criteria. Assessing the population based on the criteria of age (7-14), income (less that 0.7 of minimum wage)' , and residency (at least five years), Table 6-III demonstrates the number of families reached for each criteria. Table 6-111 Selection Criteria of Program Applicants in the Federal District - 1995 thousand (¶) thousaud (%) _iddals families Relevant Population 1.7161 100 477.8 100 Income per capita less than 0.7 MW 342.5 17.1 81.9 17.1 Residency of at least 5 years 221.0 12.9 52.8 11.1 With Children ages 7-14 148.8 8.6 35.6 J 7.4 Source: IBGE/PNAD, 1995 (Special Tabulations). Saboia and Rocha, 1998. Table 7-III compares the characteristics between the selected population (beneficiaries), and the parameter population which refers to the population limited by the selection criteria. For average family income, there is only a three percent difference between the two groups. However, for average per capita income the difference is substantially larger 18 Although the original income requirement established by the program was 0.5 of minimum wage, the program in the Federal District broadened the criteria to include those earning less than 0.7 of minimum wage. 49 -18 percent when considering the difference in family size. In other words, the average income of the selected population is lower than for the parameter population. The last row in the table gives the income ratio, which takes as the poverty line 0.66 of minimum wage. This demonstrates that the intensity of poverty is 0.45 for the selected population while for the parameter population it is much less (-0.35). These results are evidence that the program is reaching an even poorer population than it had targeted through the criteria, which is positive given the objectives of the program. 50 Table 7-11I Income, Family Size and Intensity of Poverty Comparison between the Selected Population (Beneficiaries) and the Parameter Population Selected Population Parameiter Population Average Family Income (R$) 175.4 1:30.4 Average Per Capita Income (R$) 36.3 43.2 Average Family Income 4.8 4.2 Income Ratio 0.45 0.35 Source: IBGE/PNAD, 1995 Bolsa-Escola (Paranoa Registry) 51 Annex 1 Table 1 Scoring Criteria for the School Scholarship Program Federal District and Belem Selected Topics Priority for concession Dependents under special custody 160 50 160 Dependents under socio-educacional custody - 50 160 Undernourished children from 0 to 6 years 160 50 Dependents of hadicapped elders 160 50 160 4 to 14 years old handicapped children - - 160 Handicapped 160 50 Number of children (0 to 14 years) I 0 0 0 2 20 70 20 3 50 140 50 4 80 210 80 5 110 280 110 6 140 350 140 7 180 420 180 8 or more 210 490 210 Marital Status Applicants with spouse 0 0 Applicants without spouse 100 110 Single - - 150 Married - - 0 Divorced - - 150 Widow - - 150 Applicant's Schooling Illiterate 50 50 50 Literate 25 25 25 Incomplete Primary School 15 0 15 Complete Primary School 15 0 15 Incomplete High School 5 -25 5 Complete High School 5 -25 5 Spouse's Schooling Illiterate 50 - 50 Literate 25 - 25 Incomplete Primary School 15 - 15 Complete Primary School 15 - 15 Incomplete High School 5 - 5 Complete High School 5 - 5 Applicants Labor Market Engagement "8iscateiro" 80 50 80 Farmer 75 37.5 75 Employee 65 25 65 Retired 60 12.5 60 Inactive 20 0 20 Spouse's Labor Market Engagement "Biscateiro" 80 - 80 Farmer 75 - 75 Employee 65 - 65 Retired 60 - 60 Inactive 20 - 20 52 Housing Status Rented 120 100 120 "Cedido" 80 0 80 Owner occupied dwelling in irregular property 20 0 20 Owner occupied dwelling in regular property 0 0 0 Per Capita Family Income (R$) 0 - 13.00 - 100 - 0 - 16.00 - - 40 14.00 - 28.00 - 75 - 17.00- 32.00 - - 25 29.00 - 33.00 - 50 - 33.00 - 48.00 - - 10 34.00- 56.00 - 25 - 48.00 - 65.00 - - 5 57.00 - 70.00 - 0 - 66.00 or more - - 0 71.00 or more - -100 - up to 25 50 - from 26 to 50 35 51 or more 0 53 REFERENCES Duryea, Suzanne. July, 1998. Children's Advancement Trough School in Brazil: The Role of Transitory Shocks to Household Income. Inter-American Development Bank. Haddad, Lawrence, John Hoddinott and Harold Alderman, 1997. Intra-Household Resource Allocation in Developing Countries, Baltimore:Johns Hopkins Press. Lavinas, Lena. ,Renta Minima o Beca Escolar? Paralelos entre las experiencias internacionales y las iniciativas brasilefias. De Igual a Igual: El Desaflo del Estado Ante Los Nuevos Problemas Sociales 1999 Fondo de Cultura Econ6mica de Argentina, S.A., pp.181-213. Lavinas, Lena Programas de Garantia de Renda Minima: Perpectivas Brasileiras. IPEA, 1998. Lavinas, Lena. Avaliagdo de Dois Programas Bolsa Escola- Belo Horizonte e Recife. 1999. Lavinas, Lena, and M. Bittar. Special Tabulations of the Minimum Income Programs, December, 1999. IPEA Lobato, Ana L. (org.) Garantia de Renda Minima - Ensaios e Propostas, Brasilia, IPEA, 1998. Moura, Alexandrina Sobreira, 1999. Programa de Erradicacao do Trabalho Infantil na Zona da Mata de Pemambuco, Brasil. Recife, Brazil. Paes de Barros, Ricardo and Rosane Mendon,a. November, 1999. Uma Avaliagao dos Custos e dos Beneficios da Educacao Pre-escolar. IPEA Rocha, Sonia. March, 1998. Minimum Income Programs -How Do They Apply to the Brazilian Nuclei? World Bank Report. Rocha, Sonia and Joao Sab6ia. 1999. Programas de Renda Minima Linhas Gerais de uma Metodologia de Avaliacao a partir da Experiencia Pioneira do Paranoa no Distrito Federal. IPEA. Sant'Ana, Silvio Rocha, Moraes, Andrea. Avaliacao do Programa Bolsa Escola do GDF. Fundacao Grupo Esquel Brasil, 1997. Schultz, T.Paul. June,1999. Preliminary Evidence of the Impact of Progresa on School Enrollment from 1997 and 1998. International Food Policy Research Institute. Waiselfisz, Julio J., Miriam Abramovay, and Carla Andrade. Bolsa Escola -Meloria Educacional e Reducao da Pobreza. Projecto Conjunto: UNESCO, UNICEF, POLIS. Brasilia, March, 1998. 54 Waiselfisz, Julio J. Resumen Ejecutivo del informe "Programa de Becas Escolares del Distrito Federal" de Brasil. De Igual a Igual: El Desaflo del Estado Ante Los Nuevos Problemas Sociales 1999 Fondo de Cultura Econ6mica de Argentina, S.A., pp214-227. 55