69307 A WORLD BANK STUDY Early Child Education M A K I N G P R O G R A M S W O R K F O R B R A Z I L’ S M O S T I M P O R TA N T G E N E R AT I O N David K. Evans Katrina Kosec A W O R L D B A N K S T U D Y Early Child Education Making Programs Work for Brazil’s Most Important Generation David K. Evans Katrina Kosec © 2012 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 15 14 13 12 World Bank Studies are published to communicate the results of the Bank’s work to the development community with the least possible delay. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally edited texts. This work is a product of the staff of The World Bank with external contributions. Note that The World Bank does not necessarily own each component of the content included in the work. The World Bank therefore does not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The �ndings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomi- nations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are speci�cally reserved. Rights and Permissions This work is available under the Creative Commons A ribution 3.0 Unported license (CC BY 3.0) h p://creativecommons.org/licenses/by/3.0. Under the Creative Commons A ribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: A ribution—Please cite the work as follows: Evans, David K. and Katrina Kosec. 2012. Early Child Education: Making Programs Work for Brazil’s Most Important Generation. Washington, D.C.: World Bank. DOI: 10.1596/978-0-8213-8931-7 License: Creative Commons A ribution CC BY 3.0 Translations—If you create a translation of this work, please add the following disclaimer along with the a ribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ISBN (paper): 978-0-8213-8931-7 ISBN (electronic): 978-0-8213-9563-9 DOI: 10.1596/978-0-8213-8931-7 Cover photo: Courtesy of Barbara Bruns. The girls in this photo were a ending a creche (daycare center) in Rio de Janeiro, Brazil. Library of Congress Cataloging-in-Publication Data has been requested. Contents Acknowledgments .................................................................................................................. vii Acronyms and Abbreviations .................................................................................................ix Executive Summary ..................................................................................................................xi 1. Early Child Education—A Top Priority for the Coming Years .................................. 1 Why Are Early Child Development and Early Child Education So Important? ....... 1 How Has Brazil Advanced in Early Child Development and Early Child Education in Recent Years?......................................................................................... 7 Key Issues Facing Brazil in Early Child Education ...................................................... 11 2. Ensuring High Quality Early Child Education for Brazil’s Children .................... 14 Current Quality of Early Child Education in Brazil ..................................................... 14 Curricular and Program Structure Improvement......................................................... 25 Monitoring Program Quality ........................................................................................... 25 Improving Quality through Improved Incentives........................................................ 28 Improving Quality through Improved In-Service Training and Supervision.......... 29 Improving Quality through Knowledge Sharing ......................................................... 29 Lessons ................................................................................................................................ 31 3. How to Reach the Very Poorest Children .................................................................... 33 Access to Early Child Education around the World .................................................... 33 How Has Access to Early Child Education Evolved in Brazil? .................................. 34 Reaching the Poorest: How Many to Plan For .............................................................. 44 Building New Centers ...................................................................................................... 47 Alternative Ways to Deliver Early Child Education .................................................... 50 Provision and Financing of Early Child Education ...................................................... 52 Lessons ................................................................................................................................ 57 4. The Next Steps for Brazil’s Children ............................................................................ 59 Cross-Sectoral Collaboration ........................................................................................... 59 Leveraging Private Sector Provision, Funding, and Innovation ................................ 64 Compensating for Differences across Municipalities................................................... 66 Lessons ................................................................................................................................ 69 Appendixes................................................................................................................................ 71 Appendix A: Pre-school Enrollment with Alternative De�nitions ............................ 73 Appendix B: Survey of Evidence for Early Child Education in Brazil ...................... 74 Appendix C: Rio de Janeiro’s Creche Lo ery ............................................................... 75 Appendix D: Quality Rating Systems ............................................................................ 76 iii iv Contents Appendix E: Selected Child Development Instruments .............................................. 77 Appendix F: Curriculum for Primeira Infância Completa .......................................... 78 Appendix G: Selection of MEC Publications for Early Childhood Development ... 79 Appendix H: Jamaica’s National Strategic Plan for Early Childhood Development .............................................................................................................. 80 References.................................................................................................................................. 81 Boxes Box 2.1: Impact evaluation for early child development .....................................................28 Box 2.2: Training and supervision in two municipalities....................................................29 Figures Figure 1: ECE coverage by state, 2009 ....................................................................................xii Figure 2: ECE access by income quintile ............................................................................. xiii Figure 3: Quality of creches and pre-schools in six capital municipalities..................... xiv Figure 4: Additional children enrolled following revenue shock that leads median municipality to enroll 100 more children.......................................................................xv Figure 5: Fraction of creche students in public vs. private institutions by quintile of income (2009) .............................................................................................................. xvii Figure 6: Fraction of pre-school students in public vs. private institutions by quintile of income (2009) .......................................................................................................... xvii Figure 1.1: Cognitive development across income differentials over time.........................2 Figure 1.2: Impact of high-quality, highly targeted early child education programs.......4 Figure 1.3: Sample of large-scale programs ............................................................................7 Figure 1.4: Public child investments per capita, 2006–2009 ................................................10 Figure 1.5: Fraction of children enrolled in early child education (1996–2009) ...............12 Figure 2.1: Average quality index of creches and pre-schools by region (2001 and 2009).....................................................................................................................................16 Figure 2.2: Average number of students per classroom, by year and region ...................17 Figure 2.3: Average infrastructure quality index of creches and pre-schools in Brazil by administrative dependency (2009) .................................................................17 Figure 2.4: Distribution of creche quality by broad quality domain .................................19 Figure 2.5: Distribution of creche quality by selected speci�c areas .................................19 Figure 2.6: Distribution of quality across ECE institutions .................................................20 Figure 2.7: Level of education of teachers in ECE versus primary and lower- secondary school ...............................................................................................................21 Figure 2.8: Fraction of teachers with higher (post-secondary) education, by year and region ...........................................................................................................................22 Figure 2.9: Relationship between observed creche quality and parents’ subjective measure of quality, Rio de Janeiro, 2001 ........................................................................24 Figure 3.1: Creche enrollment around the world, age 0–3 (or as speci�ed) .....................33 Figure 3.2: Pre-school enrollment around the world ...........................................................34 Figure 3.3: Fraction of 0–3 children in creche, 1996–2009 ...................................................35 Figure 3.4: Fraction of 4–6 children in pre-school, 1996–2009 ............................................35 Contents v Figure 3.5: Creche access by state, 2009 .................................................................................36 Figure 3.6: Pre-school access by state, 2009 ...........................................................................36 Figure 3.7: Fraction of 0–3 children in creche by gender (1996–2009) ...............................37 Figure 3.8: Fraction of 4–6 children in pre-school by gender (1996–2009) ........................38 Figure 3.9: ECE institution a endance by income (1996–2009) ..........................................38 Figure 3.10: Fraction of children enrolled in creche in public vs. private institutions by income quintile (2009) ............................................................................39 Figure 3.11: Fraction of children a ending pre-school in public vs. private institutions by income quintile (2009) ............................................................................40 Figure 3.12: Fraction of 4–6 children in pre-school by rural/urban, 1996–2009 ...............41 Figure 3.13: Fraction of 4–6 children in pre-school by income and rural/urban, 1996–2009 ............................................................................................................................41 Figure 3.14: Creche enrollment by mother’s work status, 1996–2009................................43 Figure 3.15: Pre-school enrollment by mother’s work status, 1996–2009 .........................43 Figure 3.16: Creche participation in Rio de Janeiro, by income level ................................46 Figure 3.17: Number of public creches by region, 2001–2009.............................................47 Figure 3.18: Share of creches of different administrative dependencies, by region ........52 Figure 3.19: Share of pre-schools of different administrative dependencies, by region .............................................................................................................................53 Figure 3.20: Fraction of early child education students in unregistered institutions, by year and region.......................................................................................54 Figure 4.1: Additional children enrolled following revenue shock that leads median municipality to enroll 100 more children ........................................................67 Figure C.1: Lo ery process for municipality of Rio de Janeiro, 2007–2010 ......................75 Tables Table 1: Sample of evidence on impacts of early child education in Brazil .......................xi Table 1.1: Sample of Brazilian evidence on impacts of early child education ....................4 Table 1.2: Policy changes in early child education in Brazil .................................................8 Table 1.3: Starting age for compulsory education in Europe, North America, and selected other countries ......................................................................................................8 Table 1.4: Child health outcomes across the world in 2007 ................................................10 Table 2.1: Impact of creche quality on months of child development, measured in months of child development ..........................................................................................15 Table 2.2: Domains in early child education observation instruments .............................18 Table 2.3: Teacher quali�cations required in Brazil and comparator countries ...............23 Table 3.1: Labor force participation across countries ...........................................................42 Table 3.2: Total number of children not enrolled in ECE, by region .................................44 Table 3.3: Number of 4–5 year olds not enrolled in pre-school in 2009, by rural/ urban and region ...............................................................................................................45 Table 3.4: Estimated demand for creche, by region .............................................................46 Table 3.5: Average enrollment size of pre-schools in 2009, by region ...............................47 Table 3.6: ECE institutions, by type of institution (2005, 2009)...........................................49 Table 3.7: Method for identifying locations for new ECE centers ......................................50 vi Contents Table 3.8: Annual municipal early child education expenditures in 2009, by region .....55 Table 4.1: Brazil’s National Plan for Early Childhood .........................................................61 Table 4.2: Services offered under Chile Crece Contigo ........................................................63 Table 4.3: Steps to establishing public-private partnerships for ECD ...............................66 Table A.1: ECE enrollment rate by year and de�nition of enrollment ..............................73 Table B.1: Survey of studies of impact of ECE in Brazil ......................................................74 Table D.1: Quality standards used to rate ECD centers ......................................................76 Table D.2: Rating systems for early child development centers.........................................76 Table E.1: Sample of child assessment instruments used in creches and pre-schools ....77 Table F.1: Curriculum implemented as of November 2009 ................................................78 Table G.1: Selection of MEC publications for ECE ...............................................................79 Table H.1: Jamaica’s National Strategic Plan for Early Childhood Development ...........80 Acknowledgments T his report draws deeply on the extraordinary efforts and innovations demonstrat- ed by early child development policy makers around Brazil. Many of these policy makers have given generously of their time and shared information about programs highlighted in this report, including Maria do Pilar Lacerda Almeida, former Secretary of Basic Education; Rita Coelho, Early Child Education coordinator at the Ministry of Education; Claudia Costin, Secretary of Education for the Municipality of Rio de Janeiro; Maria Correa, former Secretary of Education for the State of Acre; Osmar Terra, former Secretary of Health for the State of Rio Grande do Sul; Célia Gedeon, Elaine Pazello, and Gabriela Moriconi at the National Education Research Institute (INEP); Ricardo Paes de Barros, Secretary of the Office of Strategic Action in the Secretariat for Strategic Affairs of the Presidency of Brazil; and Mirela de Carvalho at the Rio de Janeiro State Secretariat of Education. This report draws on background papers about innovations in early child education in Rio de Janeiro and in caregiver training and supervision in two municipalities within São Paulo state. The authors of the Rio de Janeiro study are Ricardo Paes de Barros, Mire- la de Carvalho, Samuel Franco, Eduardo de Pádua, and Andrezza Rosalém. The authors of the São Paulo studies are Joseane Bom�m, Marisa Vasconcelos Ferreira, and Zilma de Moraes Ramos de Oliveira. The report also draws on a host of early child education research from around Brazil: Maria Malta Campos of Fundação Carlos Chagas, Aimee Verdisco of the Inter-American Development Bank, and Fabiana Felicio all contributed with valuable information on their research. Peer reviewers at both the beginning and completion of the project provided valu- able feedback, speci�cally Harold Alderman, Felipe Barrera, João Batista Oliveira, Ma- ria Thereza Marcilio, Robert Myers, Sophie Naudeau, Daniel Santos, and Mary Eming Young. Steven Barne of Rutgers University supplied helpful feedback at various points. The Maria Cecilia Souto Vidigal Foundation provided excellent comments and support- ed the publication of the Portuguese edition of this report: “Educação Infantil: Program- as para a Geração Mais Importante do Brasil.� Colleagues at the World Bank contributed with readings of the draft and strategic suggestions, including Erica Amorim, Laura Chioda, Tito Cordella, Margaret Grosh, and Emiliana Vegas. In particular, Madalena dos Santos, Michele Gragnolati, Chingboon Lee, and Barbara Bruns provided detailed, crucial inputs. The World Bank Brazil administrative staff—Mariane Brito, Marize San- tos, and Carla Zardo—facilitated countless meetings and communications with Brazil- ian policy makers. The report has especially bene� ed from the support and encouragement of the World Bank country director for Brazil, Deborah We el, and the former World Bank country director for Brazil, Makhtar Diop. The team accepts full responsibility for any errors. vii Acronyms and Abbreviations Asinhas Asas da Florestania Infantil/Children’s Wings of Florestania ChCC Chile Crece Contigo/Chile Grows With You CLASS Classroom Assessment Scoring System CNPq National Council for Scienti�c and Technological Development CONAFE Consejo Nacional de Fomento Educativo ECD Early Child Development ECE Early Child Education ECERS-R Early Childhood Environment Rating Scale—Revised for pre-schools EDIs Espaços de desenvolvimento infantil/Early Child Development Spaces Educa a Su Hijo Educate Your Child ERSs Environmental Rating Scales ESAC Escala de Avaliação de Ambientes para Bebês e Crianças Pequenas/ Escala de Avaliação de Ambientes da Educação Infantil EU European Union FLFP Female Labor Force Participation FPE/ FPM State and Municipal Participation Funds FPG Child Development Institute FUNDEB Fund for the Maintenance and Development of Basic Education FUNDEF Fund for the Maintenance and Development of Elementary Education and Teacher Development GDP Gross Domestic Product IADB Inter-American Development Bank IBGE Instituto Brasileiro de Geogra�a e Estatística/Brazilian Institute of Geography and Statistics ICMS Imposto sobre Operações relativas à Circulação de Mercadorias e Prestação de Serviços de Transporte Interestadual e Intermunicipal e de Comunicação (state tax on goods and services) INEP Instituto Nacional de Estudos e Pesquisas Educacionais Anisio Teixeira IPEA Institute for Applied Economic Research IPIexp The export-proportional excise tax IPVA The motor vehicles tax IRmuni and IRest Income tax bene�ts levied on income paid by the municipalities or the state ITCMD The tax on inheritance and donations ITERS-R Infant and Toddler Environment Rating Scale—Revised for creches ITR The 50% share of the rural territory tax due to municipalities LC 87/96 Transfers under the complementary law ix x Acronyms and Abbreviations LDB Law of Directives and Bases of National Education LFP Labor force participation MEC Ministry of Education MINEDUC Ministerio de Educación NACCRRA National Association of Child Care Resource & Referral Agencies NAEHCY National Association for the Education of Homeless Children and Youth NGA National Governor’s Assocation NSP National Strategic Plan for Early Childhood Development OECD Organization for Economic Co-operation and Development PB Participatory budgeting PIM Primeira Infância Melhor/Be er Early Childhood PNAD Pesquisa Nacional por Amostra de Domicilios QRS Quality Rating Systems Rede Nacional National Network for Early Childhood Primeira Infância SEE State Education Secretariat UNDIME The National Union of Municipal Education Leaders UNESCO United Nations Educational, Scienti�c, and Cultural Organization UNICEF United Nations Children’s Fund WB World Bank WHO World Health Organization Exchange Rate Effective May 28, 2011 Currency Unit = Brazil Real (BRL) BRL$1.00 = US$0.63 US$1.00 = BRL$1.59 Executive Summary The Facts E arly child education (ECE) can have lasting positive impacts on children, with bene�ts far exceeding the costs. But quality is crucial. Evidence from the United States, Argentina, Chile, and elsewhere has shown long-term positive impacts of early child education. That evidence is now complemented by data from Brazil showing posi- tive impacts of early child education—particularly pre-school—on short-run cognitive development, medium-run test scores, and long-run educational a ainment and income (table 1). But simply enrolling children in ECE is no guarantee of success. Evidence from Brazil shows that children who a end low-quality pre-schools perform the same on literacy tests two years later as do children who a end no pre-school at all, whereas children in high-quality pre-schools perform much be er. Good ECE centers are char- acterized by quality in a range of areas: safe and appropriate infrastructure, effective daily program structures, personal care routines, interactions with staff, and activities to promote physical, social, and cognitive development. Table 1: Sample of evidence on impacts of early child education in Brazil Sample Results Southeast & Northeast Almost half an additional year of schooling by adulthood Two to six percent increase in future earnings for men (more for those coming from illiterate households) Campo Grande, Florianópolis, Positive signi�cant impact on Provinha Brasil (second grade literacy) results, Teresina particularly for children who attended high-quality pre-schools Sertãozinho Provinha Brasil results six percent higher for children with ECE Rio de Janeiro Positive impacts of high-quality creche participation on mental and social development while still in creche Nationally representative Positive impact of creche participation on fourth-grade math scores Source: World Bank. ECE ma ers most for the poorest children. ECE is likely to have the greatest posi- tive impacts on children from poorer and less-educated families where there is less cog- nitive stimulation at home. In Ecuador and in the United States, poor children have been shown to start out cognitively disadvantaged due to less stimulation at home, and to grow more disadvantaged over time. Poor children need these programs the most and see the highest returns from them. In Brazil, a study of adults in the Northeast and Southeast regions demonstrated that pre-school had bigger impacts on children with illiterate parents than on children with literate parents. Internationally, the strongest evidence for the value of ECE comes from high-quality programs closely targeting the most vulnerable children. Pro-child government policies have laid the groundwork for past and future ad- vances in ECE. The last 20 years of early child policy in Brazil have signi�cantly im- proved the situation for children and paved the way for future improvements. The Fed- xi xii Executive Summary eral Constitution of 1988 made children citizens and added ECE to the purview of the Ministry of Education, underscoring the educational as well as the social welfare-based motivation for ECE. The starting age in Brazil for formal education has been decreasing over time, from seven prior to 1996 to six in 2007, and �nally—by constitutional amend- ment—to four in 2009. Brazil’s school starting age is now among the youngest in the world. Two school �nance laws have also helped partially equalize access to education �nancing. First, in 1998, FUNDEF provided more education funds to municipalities— particularly those in the poorest states—potentially freeing up funds for ECE. In 2007, FUNDEB speci�cally paid municipalities for enrolling children in ECE. Both laws have helped make an early school starting age a real possibility for Brazil’s children. There are stark disparities in ECE coverage across states, with some requiring massive expansions in the coming years to achieve intended universal pre-school cov- erage by 2016. Achieving universal enrollment in pre-school for children ages 4–5 will require almost 1.6 million new spaces. Expanding creche enrollment to even thirty per- cent of children would require over 1.3 million new spaces. Behind these massive num- bers lies great variation across states (�gure 1). Six states have pre-school enrollment rates under 60 percent, meaning that universal coverage would require almost doubling enrollment over �ve years. Likewise, six states have creche enrollment rates under 10 percent, requiring massive expansion to reach the most vulnerable children. Poor children and rural children are being left behind in early child education. Brazil’s poorest children are by far the least likely to be enrolled in creche or pre-school, and those poor children that are enrolled are much more likely to rely on public schools than are their richer peers. The richest children are three times as likely to be in creche as the poorest children and 24 percent more likely to be in pre-school (�gure 2). This dif- ference exacerbates disparities in opportunity that—without intervention—will follow these children throughout their education. Not only are wealthier families much more Figure 1: ECE coverage by state, 2009 Creche (age 0–3) Pre-school (age 4–5) Source: Data from PNAD (2009). Note: Creche access is the enrollment rate in schools for children aged 0–3. Pre-school access is the enroll- ment rate in schools for children aged 4–5. Executive Summary xiii Figure 2: ECE access by income quintile 0–3 children in creche 4–6 children in pre-schools by income quintile by income quintile 0.34 1.0 Fraction enrolled Fraction enrolled 0.26 0.8 0.18 0.6 0.10 0.4 0.02 1996 2001 2006 1996 2001 2006 Year Year 1st quintile 2nd quintile 3rd quintile 4th quintile 5th quintile Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996-2009. Note: Before 2007, the pre-school a endance rate is the fraction of 4–6 year olds in pre-school. Beginning in 2007, the pre-school a endance rate is the fraction of 4–6 year olds in institutions intended for 4–6 year olds. likely to provide their children with early child education: they are much more likely to use the private system. More than 85 percent of the poorest children who are in ECE use public institutions, compared with only about 20 percent of the wealthiest children. Be- cause private institutions tend to be of higher quality (at least in terms of infrastructure and teacher training), this gap further intensi�es inequality of opportunity. Rural households are much poorer than urban households, so it is unsurprising that rural children have lower enrollment in ECE. However, the poorest urban children participate in pre-school at the same rate as the wealthiest rural children, so income does not capture the full rural-urban disparity. Rural children have less access to early child education, likely because low population density makes center-based care much more expensive in rural areas. Quality in ECE centers in Brazil has improved over time but is still weak, particu- larly in activities to stimulate cognitive development. In certain key indicators, such as physical infrastructure and caregiver-child ratios, ECE centers have improved over the past decade. Creches across the country have become more likely to have a dedicated school building, an indoor bathroom, electricity, a library, a computer, and connections to public sewer networks. Pre-schools likewise have improved across an array of infra- structure measures. Good infrastructure is an essential but far from sufficient condi- tion for a high-quality ECE experience. A ention to children is also essential. Creches in Brazil maintained a constant child-caregiver ratio of 26 over the last decade, while pre-schools improved from 39 to 32. However, this compares unfavorably with the rec- ommendations from the United States’ National Association for the Education of Young Children that even �ve-year-olds have a maximum of ten children per caregiver, and that younger children have much lower ratios. A recent study of creches and pre-schools in six capital municipalities around Brazil found that ECE centers were strongest on xiv Executive Summary interactions between caregivers and children, and weakest on activities and a consistent program structure that lends itself to cognitive, social, and emotional development. Spe- ci�cally, ECE centers lacked a program structure with a schedule, free play, group time, provisions for disabled children, and activities such as block play, musical and movement activities, and science activities. Overall, using a scale that has been applied in several countries, 50 percent of creches and 30 percent of pre-schools rated inadequate, and none rated excellent (�gure 3). Quality outside capital municipalities is likely to be even lower. Figure 3: Quality of creches and pre-schools in six capital municipalities 60 50 Creche 40 Percent Pre-school 30 20 10 0 Inadequate Basic Adequate Good Excellent Quality Source: Fundação Carlos Chagas (2010). The quality of ECE centers varies dramatically across regions by infrastructure, teacher quality, and activities, as does spending per pupil. On an index of infrastruc- ture quality, pre-schools in the Southeast had twice the infrastructure quality of pre- schools in the North (and creches were almost as disparate). Concretely, all pre-schools in the South and Southeast had electricity in 2009, whereas only three-quarters of pre- schools in the North did. Creches in the North had 70 percent more students per class- room than did creches in the South, clearly affecting students’ experience at an age when caregiver-child interaction is very important. Private creches also greatly surpass public centers in infrastructure quality. Not only do the North and Northeast need expansions in coverage; they also need substantive quality improvements (as do even the highest quality regions, the South and Southeast). These quality differences likely reflect spend- ing differences; the Southeast spends �ve times as much per ECE student as does the North, and more than six times as much as the Northeast. Municipalities with higher income and more income inequality are less likely to expand investment in early child education. A study of Brazil’s over 5,000 munici- palities during 1995–2008 demonstrates two key points about municipal characteristics and ECE investment. First, public creches and pre-schools do not bene�t all citizens equally. They disproportionately bene�t poorer citizens, since the poor are most likely to use these public services. Rich families tend to enroll their children in private ECE, Executive Summary xv and therefore are less invested in the public education system. Second, the distribution of income in a municipality hugely affects public ECE investment. Given extra revenue, poor and equal municipalities are more likely to expand public ECE than are richer and more unequal municipalities. For example, if the municipality with median inequality is given enough new revenue to expand ECE and enroll 100 new children, more unequal municipalities given the same amount of money will enroll fewer than 100 children, and more equal municipalities will enroll more than 100 children, as illustrated in �gure 4. Likewise, if the municipality with median income is given enough new revenue to ex- pand ECE and enroll 100 new children, richer municipalities given the same amount of money will enroll fewer than 100 children, and poorer municipalities will enroll more than 100 children. This is an important �nding given the many efforts over the last two decades to transfer money between municipalities in order to partially equalize access to education resources. FUNDEB has explicitly aimed to equalize funding for ECE by guaranteeing all municipalities a minimum amount per child enrolled. Policy makers can be er de- sign future education �nance reforms by taking into account that not all municipalities have the same incentives to invest additional revenue in ECE. Targeting funds at poor and equal municipalities means the funds are more likely to end up in ECE rather than in higher levels of education or in goods other than education, like public infrastructure. These �ndings imply that FUNDEB would create be er incentives for investment in ECE if it were a national rather than a within-state redistribution. Figure 4: Additional children enrolled following revenue shock that leads median municipality to enroll 100 more children 200 Income 180 181 169 Inequality 160 Children enrolled if median enrolls 100 155 140 135 125 120 113 106 103 100 100 94 87 80 81 71 69 60 40 42 20 23 0 0 10th 20th 30th 40th 50th 60th 70th 80th 90th Decile of median income or of inequality (Gini coefficient) Source: Kosec (2011), using data from Censo Escolar (1995–2008), Tesouro Nacional (1995–2008), and IBGE Censo (2000). xvi Executive Summary Much innovation in ECE in Brazil is taking place at the municipal level, provid- ing models within the country to improve both access to and quality of ECE services. Many municipalities are investing in their ECE programs by developing specialized cur- ricula, monitoring systems, improved training for caregivers, and more. With over 5,000 municipalities across Brazil, there is great opportunity for these municipalities to learn from each other to enhance quality and efficiency. Some public programs target speci�c areas of interest; their curricula may be of interest to other municipalities. For example, the Municipality of Santarém (State of Pará) has developed the program Eco-Schools, which provides lessons about the environment beginning in ECE. Likewise, the Mu- nicipality of Rio de Janeiro has developed a curriculum to provide training to parents of creche children that spans education, health, and social assistance. The curriculum in- cludes DVDs, workbooks, and discussion guides. The State of Acre’s program of home- based education visits includes workbooks and educational agent guides which demon- strate—both to the agents and then to the parents—how to use common objects available near the home as tools of cognitive stimulation. Clearly, these programs and others like them represent a potent educational resource for other municipalities. The Policy Implications Brazil will need to be strategic in where it invests in ECE and use creative models to reach more children. A 2009 constitutional amendment lowered the mandatory school starting age to four years old, and the nation has a stated goal of achieving universal cov- erage by 2016. Given the 1.6 million pre-school aged children not in pre-school, roll-out must be strategic to reach that goal. In areas of lower population density, ECE centers may need to be constructed on a smaller-than-typical scale, or home visits and other delivery modalities may need to be used in order to cost-effectively satisfy needs for ECE. In some rural areas, it is hard to justify center-based education since sparse popu- lation densities would mean that very young children would have to travel very far to �ll a creche or pre-school. For example, in the municipality of São Paulo, 0.4 square kilometers of territory would be sufficient to �ll an average-sized São Paulo pre-school with 4–5 year olds. However, in more rural Barra do Turvo municipality, also in São Paulo state, over 320 square kilometers would be required to �ll an average-sized São Paulo pre-school. Programs such as the State of Acre’s Asas de Florestanhia Infantil (for pre-school aged children) or Rio Grande do Sul’s Primeira Infância Melhor (for younger children) serve as models for potential alternative outreach. In addition to providing stimulation for children, they provide training to parents to reinforce the bene�ts from external a ention. Municipalities should be er target new centers and spaces at the poorest children (whose parents are not able to self-�nance ECE) and open new centers in areas that will achieve this purpose. Among children from the richest �fth of families in Brazil who a end creches, almost 20 percent use publicly-provided creches (�gures 5 and 6). If those spaces were instead allocated to the poorest �fth, then their overall enrollment rate in creche would increase by 50 percent. This inefficient targeting of scarce spaces has not improved over time; reliance on public creches by the richest has hovered around 20 per- cent in the decade since 2002. Furthermore, evidence presented earlier demonstrates that the returns to ECE are particularly high for the poorest and most vulnerable households. These households have the fewest resources to fund alternatives to public creche care. Executive Summary xvii Figure 5: Fraction of creche students Figure 6: Fraction of pre-school in public vs. private institutions by students in public vs. private quintile of income (2009) institutions by quintile of income (2009) 0.85 0.81 0.1 0.92 0.8 0.76 0.82 0.8 Fraction enrolled Fraction enrolled 0.62 0.72 0.74 0.6 0.53 0.47 0.6 0.59 0.4 0.38 0.41 0.4 0.24 0.28 0.26 0.2 0.19 0.15 0.2 0.18 0.08 0.0 0.0 Bottom Second Third Fourth Top Bottom Second Third Fourth Top Quintile Quintile Public Private Public Private Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 2009. By taking advantage of existing Brazilian data (the IBGE’s PNAD household survey and demographic census), municipalities can determine which areas have the highest con- centrations of low-income households and the least coverage of ECE services. Another solution is allocating existing creche spaces by progressive methods that target vulner- able children in the face of excess demand. Municipalities should consider the fairest, most transparent methods of allocat- ing public ECE spaces, including a centralized, means-based selection process or a lo ery, rather than leaving the decision to individual creche directors. Most munici- palities currently allocate spaces via a completely decentralized process in which each creche director is responsible for assigning spaces—usually with official guidelines to prioritize vulnerable children. The disadvantage of this system is that directors may fail to be objective, or they may even be biased against children with special needs or other difficulties, whose care may be more difficult or expensive. A second method, less used in Brazil but more efficient from a resource allocation standpoint, is to have a centralized, means-tested selection process at the municipal level. Using eligibility for a cash-transfer program such as Bolsa Familia or data from the social protection register (Cadastro Único) can facilitate placement of children with the greatest need for public assistance. A third method, which the Municipality of Rio de Janeiro adopted for cohorts entering creches from 2008 through 2010, is to use a lo ery to assign scarce spaces among equally eli- gible families. An advantage of the lo ery is that it provides every interested child with the possibility of a space. Because no testing of children’s socioeconomic status is perfect, this means that even children who are judged to be less vulnerable by standard indicators but who perhaps have some particular, less visible needs have the possibility of gaining a place. This also leads to more diversity in socioeconomic background across creche chil- dren. Either a means-tested or a lo ery system leads to greater transparency and fairness than allocation based on director discretion. Both methods require the development of reli- able information systems to track children from their �rst entry into the education system. Teachers need speci�c guidance on the best stimulation activities to use in the classroom, to complement existing ECE curricular guidelines. Teachers at Brazilian ECE centers rate relatively high on interactions, suggesting that these teachers are moti- xviii Executive Summary vated to engage with children. But Brazilian ECE centers rate very poorly on effective ac- tivities for stimulating children’s cognitive and social development. Teachers lack strong guidance on effective activities. A valuable next step will be to build on the Ministry of Education’s 1988 three-volume curricular guide for ECE and the curricular directives laid out by the National Education Council in 2009. Speci�cally, providing practical workbooks or guides with lesson plans containing de�ned activities in pre-reading, pre- mathematics, and other areas will help educators not only understand what activities are effective, but also how to implement them. Some municipalities—such as the Mu- nicipality of Rio de Janeiro—are developing their own curricular guides with speci�c guidance on program structure and activities. These could serve as a model. ECE educa- tors will also need hands-on training (i.e., not merely a seminar held far from their own classroom on how to teach more effectively) and careful supervision by experienced, effective ECE educators. The federal government should encourage strong municipal monitoring systems that hold ECE institutions accountable for their results, introduce a standard obser- vational tool, and provide licensing guidelines for minimum quality standards that municipalities can adapt. The Ministry of Education has set minimum standards of in- frastructure quality for creches and pre-schools. The next step will be to establish addi- tional licensing guidelines regarding child-caregiver ratios and caregiver quali�cations. Municipalities can greatly improve quality by introducing standardized observational tools, such as the ITERS-R and ECERS-R instruments, to allow for regular, systematic monitoring of activity quality and program structure. In some international systems, ECE observational results are published so that parents can use quality data to choose the best centers for their children and policy makers can use the same data to target lower-quality centers for improvement interventions. The Municipality of Rio de Janeiro recently introduced testing of child cognitive development in all creches. Results (con- cealing individual child pro�les) have been used to identify the creches most successful in stimulating development, to promote cross-creche learning. Brazil should facilitate knowledge sharing across ECE providers, so that they can learn from one another’s success stories. Given the great array of experimentation and innovation at the municipal level, the Ministry of Education can encourage knowledge sharing on a number of levels. Several networks for early child development (ECD) exist: the National Network for Early Childhood focuses principally on improving national policy; the National Union of Municipal Education Leaders (UNDIME) likewise dissem- inates some documents on national early education policy. Recently, two initiatives for more direct sharing of information have been launched. One is the Network for Coop- eration and Peace for Children, which is a social networking site (a là Facebook) where policy makers, nonpro�t workers, educators, and other interested citizens actively share information. A second initiative launched by the National Forum for Early Child Devel- opment and the World Bank, called the Network for Cooperation in Early Child Devel- opment, seeks to provide a map of ECD services across Brazil. It has also developed a social networking site focused on sharing exactly the kinds of materials described above: information on innovative programs, experiences, and best practices. An initiative that the federal government formerly sponsored is an annual Prize for Quality in Early Child Education, which ran from 2000 through 2005 and highlighted innovative projects im- plemented by individual municipalities or ECE centers. These kinds of programs can help municipalities learn from one another. Executive Summary xix Integrating health services into creches and pre-schools can provide opportuni- ties to improve child welfare in cost-effective ways, and would be facilitated by estab- lishing a cross-sectoral coordinating agency. Multisectoral programs help parents learn about and access in one place all of the services they need to help their children flourish, rather than having to separately seek out services about which they may be unaware. The most established example of cross-sectoral collaboration in ECD in Brazil is the Rio Grande do Sul program, Primeira Infância Melhor. Although the program is housed in the State Health Secretariat, it is managed by the State Technical Group, which includes technical staff from the Secretariats of Health, Education, Culture, Justice, and Social De- velopment, regional coordinators from health and education from across the state, and others. The education secretariats in the State of Acre, the Municipality of Rio de Janeiro, and elsewhere in Brazil are working to establish partnerships with health secretariats to offer these comprehensive services in a single locale. Brazil has made progress in this regard through the development of the National Plan for Early Childhood, prepared collaboratively by the National Network for Early Childhood. But if the Government of Brazil is serious about cross-sectoral collaboration on ECD, it will need to establish a coordinating agency to oversee its implementation. Partnerships outside the public sector can provide signi�cant resources to expand ECD services, improve their quality, and innovate in reaching the most vulnerable populations. Data from 2009 show that the private sector provides more than one third of creche spaces: 29 percent are fully private and 14 percent are conveniado (or govern- ment contracted). Likewise, a quarter of provision at the pre-school level is through the private sector: 19 percent fully private and 5 percent conveniado. The number of conve- niado centers is growing rapidly, with hundreds of early child institutions contracted each year. This expansion signals signi�cant capacity in the private sector to supplement the efforts of the public sector in providing early child services. Furthermore, the data on quality suggest that contracted institutions have be er infrastructure on average than public institutions and are not distinguishably different in other observable measures from either fully private or fully public ECE centers. Expansion of contractual relation- ships with careful, ongoing monitoring of quality might help the government satisfy immediate demands for ECE. Although FUNDEB reimbursements are equal for public and conveniado spaces, municipalities have the liberty to compensate conveniado spaces at higher or lower levels, providing the freedom to create incentives for private provision for difficult-to-reach populations, as has taken place in Chile. In addition to subsidized provision through the private sector, Brazil may also bene�t from expanding its use of public-private partnerships to offset some of the budget impacts of coverage expan- sion and quality improvements. Many corporations have a philanthropic arm which can contribute to ECD, and a number of corporations across Brazil are already doing this. In many countries, these partnerships are mobilized to provide matching funds to provide and improve ECD services. Using participatory budgeting (PB) to distribute educational resources has the potential to lead to more equitable outcomes and to target resources at the poorest children. Participatory budgeting began in Porto Alegre, Brazil, and has since spread to many municipalities across the country. It allows citizens to vote on how to use a share of municipal revenue designated for their neighborhood, and to elect neighborhood rep- resentatives to make municipality-wide spending decisions. Additionally, policy mak- xx Executive Summary ers must publicize budgets and expenditures to promote transparency. PB increases po- litical participation of marginalized groups and can lead to more pro-poor expenditures, including public investment in ECE. It does so by ensuring that the poor—who have the highest demand for public ECE—are more directly and fully involved in decision- making. Resulting policies are more likely to reflect the will of the people rather than the preferences of the most politically-powerful citizens. The federal government may wish to encourage and create incentives for PB as a part of a broader early child education policy. The government could even create incentives for more limited forms of PB that speci�cally apply only to education spending. CHAPTER 1 Early Child Education—A Top Priority for the Coming Years T he year 2011 marked the beginning of a new administration in Brazil. The Ministry of Education clearly identi�ed early child education (ECE) as one of the top priori- ties of the new administration, along with secondary school and improving the reputa- tion of the teaching profession (Weber 2011).1 Within early child education, the Ministry of Education identi�ed two key priorities: expanding access to creches (centers for chil- dren ages 0–3) and pre-schools (ages 4–5), and improving their quality. Brazil has made great advances on both fronts in the past two decades, but the quality of service delivery must improve, and major gaps remain in reaching the poor. Fifteen years ago, less than half of Brazil’s children aged four to six were enrolled in pre-school, and fewer than one in twelve a ended creche. A scant one in �ve pre- school teachers had training beyond secondary school. Now, the number of children in pre-school has increased by more than �fty percent, to 75 percent of 4–5 year-olds and 81 percent of 4–6 year-olds; moreover, 18 percent of 0–3 year-olds a end creche, and half of pre-school teachers have post-secondary training.2 School infrastructure has also improved steadily, and class sizes have declined. At the same time, these numbers indicate that many children still do not a end pre-school—compulsory since 2009—and most do not a end creche.3 The poorest chil- dren stand to bene�t the most from both early education opportunities and the income gains of having an additional working parent, but only one in nine of these children at- tend a creche. This exacerbates existing income gaps. Recent in-depth studies of creches and pre-schools across Brazil also reveal that their quality lags far behind that found in higher-income countries. Early child development interventions are essential to both increasing the produc- tivity of Brazil as a whole and to providing equitable opportunities for the disadvan- taged. These programs bene�t the poor more than other populations, and the poor are most in need of these bene�ts. Education interventions are crucial. Creches and pre- schools provide opportunities for stimulation and development that can wire children for future success. For these centers to be effective, however, they must satisfy certain quality standards, and expose children to the right types of activities and experiences. Why Are Early Child Development and Early Child Education So Important? The �rst few years of a child’s life hold massive sway over long-term outcomes. From birth to age �ve, children develop “foundational capabilities� on which the rest of their development builds. Just as positive environments and opportunities can wire children for success, failure to provide those opportunities can signi�cantly reduce future oppor- tunities (Shonkoff and Phillips 2000). Children’s health, wealth, and home environment have a major impact on many long-term outcomes, pu ing poor children at high risk 1 2 A World Bank Study for slower development. For example, researchers examined vocabulary test scores for Ecuadorian children between three and six years old (Paxson and Schady 2007). They found that while rich and poor children had very similar vocabulary test scores at three years of age, those scores had diverged dramatically by age six, when the poorest 25 percent had only two-thirds the vocabulary of the richest 25 percent (�gure 1.1). Figure 1.1: Cognitive development across income differentials over time 110 Standardized TVIP (vocabulary) score 100 Fourth decile 90 Third decile 80 Second decile 70 60 First (poorest) decile 50 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 Child age (months) Source: Paxson and Schady (2007). While this evidence demonstrates that children from different backgrounds are on different trajectories, the encouraging news is that public interventions have been shown to mitigate the effects of unfavorable circumstances. Conditional cash transfer programs such as Bolsa Familia have improved children’s schooling and health. Certain home vis- iting programs and parent training programs have led to not only improved health, but also a reduction in crime and other adverse behaviors when children reach adolescence. Several child care and pre-school programs have had long-term impacts on health, edu- cation, employment, income, and other indicators of well-being (Almond and Currie 2011). Likewise, much of the evidence on early child development demonstrates that impacts are particularly effective for poor children, leading Economics Nobel laureate James Heckman to argue that “investing in disadvantaged young children is a rare pub- lic policy with no equity-efficiency tradeoff� (Heckman and Masterov 2007). The importance of early child education No single area of investment—be it education, health, home environment, or child pro- tection—is sufficient to ensure the success of Brazil’s children. Children need a range of investments in different areas. But education and cognitive stimulation interventions— Early Child Education: Making Programs Work for Brazil’s Most Important Generation 3 whether through center-based care or through parent training and home-based stimula- tion—have enormous potential to make a difference for children. Reviews of early child development programs show that interventions targeting one area have large spillover effects in other areas. Nutrition programs have been shown to affect cognitive devel- opment, conditional cash transfer programs have improved health, and education and child care interventions affect health and behavior (Almond and Currie 2011). However, in a review of thirty early child development interventions around the world, research- ers found that those interventions which provided education services or a combination of education and nutrition services tended to have be er impacts on children’s cognitive development than those that provided nutritional or �nancial transfers alone (Nores and Barne 2010). For example, a program in Jamaica combined cognitive stimulation with nutritional supplementation and found that—sixteen years later—the cognitive stimulation cut high school dropout rates in half, whereas the supplementation had no impact (Walker et al. 2005). Likewise, an integrated program coordinating home-based and center-based cognitive stimulation and nutritional supplements in the Philippines had major impacts on cognitive outcomes and on some health outcomes (Armecin et al. 2006). In both cases, early child education was critical. Center-based pre-schools and creches have relatively high start-up costs, but pro- vide clear opportunities to add other child care programs (such as nutrition monitoring and supplementation or parental education) at low additional cost. That is, formal early child education can create opportunities for a multitude of successful interventions. The constantly expanding array of evidence points to enduring impacts of early child care and education programs around the world, although with two important quali�cations. First and most importantly, having a program is no guarantee of success: quality is es- sential, and low quality programs can be worse than no program at all. Second, the strongest evidence comes from programs targeting the most vulnerable. The evidence for this impact comes from around the globe: the United States, Europe, Asia, Latin America, and—increasingly—Brazil itself. Signi�cant and lasting positive impacts from around the world Rigorous evaluations of high-quality early child education programs have demonstrated long-term impacts on educational a ainment, poverty, and participation in crime, as demonstrated in �gure 1.2. The strongest evidence comes from the United States, where evaluations of early child education programs such as the High/Scope Perry Pre-school program, the Abecedarian program, and the Chicago Parent Child program have dem- onstrated long-term impacts on a wide range of outcomes. The Perry Pre-school pro- gram has an estimated social return of 7–10 percent, which is signi�cantly higher than average investments (Heckman et al. 2010). These were relatively small programs: the Perry Pre-school program had 123 original participants, and the Abecedarian program had 56 (Masse and Barne 2002). However, evidence on Head Start, the largest pre- school program for vulnerable children in the United States (currently serving about 800,000 children each year), suggests positive impacts on cognitive development, grade progression, and college participation, and reductions in crime, depression, other be- havioral problems, and even mortality (Almond and Currie 2011). Likewise, large-scale pre-school programs in the United States, Chile, and Argentina have demonstrated posi- tive impacts on children’s cognitive development and—sometimes—positive behavioral impacts (�gure 1.3). 4 A World Bank Study Figure 1.2: Impact of high-quality, highly targeted early child education programs Chicago Parent Perry Preschool Abecedarian Child Age 10 Age 14 Age 27 Age 21 Age 21 Skilled job or higher edu High school graduation Earn $2,000+ monthly Achievement test Juvenile arrests HS graduation Reading Arrests Math IQ IQ 0% 227% 214% 3% 6% 6% 31% 68% 28% -50% -32% Source: Barne (2004). Studies from Brazil—both those using national data and those examining speci�c municipalities—also demonstrate gains from early child education on short-run cogni- tive development, medium-run academic performance, and long-run educational a ain- ment and earnings, as demonstrated in table 1.1. (A more complete review of evidence from Brazil is summarized in Appendix B.) However, most studies of the effectiveness of ECD in Brazil have relied on assumptions that children with similar observable income and family background characteristics would experience similar outcomes in the absence of pre-school. This may not be the case; families that look similar may have different Table 1.1: Sample of Brazilian evidence on impacts of early child education Sample Results Southeast & Northeast Almost half an additional year of schooling by adulthood Two to six percent increase in future earnings for men (more for those coming from illiterate households) Campo Grande, Positive signi�cant impact on Provinha Brasil (second grade literacy) results, particularly Florianópolis, Teresina for children who attended high-quality pre-schools Sertãozinho Provinha Brasil results 6% higher for children who attended ECE Rio de Janeiro Positive impacts of high-quality creche participation on mental and social development while still in creche Nationally representative Positive impact of creche participation on fourth-grade math scores Sources: The Southeast & Northeast study comes from Young (2001). The Campo Grande, Florianópolis, and Teresina study comes from Fundação Carlos Chagas (2010). The Sertãozinho evidence comes from Felício, Menezes, and Zoghbi (2010). The national study is Rodrigues, Pinto, and Santos (2010). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 5 unmeasured qualities, such as their commitment to education, which contribute to child outcomes independently of pre-school. Still, the Brazil data generally points to positive returns to participation in early child education, from the time of the intervention (in a Rio de Janeiro study), to primary school outcomes (in several studies), to outcomes in adulthood (in a study in the Northeast and Southeast). The most credible evidence from the United States, in contrast, comes from studies which use randomized provision of ECE and rigorous evaluation methods. Quality of early child education matters Providing a child with a place in a center or program is insufficient to confer the prom- ised returns of early child education. Recent studies show that the set of cognitive and non-cognitive skills children acquire at school is much more relevant to longer-term life outcomes than is schooling a ainment (Hanushek and Woessman 2008). In Brazil and elsewhere, higher quality ECE centers lead to be er cognitive development. A variety of studies have used observational instruments to rate the quality of creches (for children 0–3 years old) and pre-schools (for children 4–5 years old) accord- ing to infrastructure, activities, interactions between staff and children, and so on. For example, a recent study examined pre-schools in three municipalities (Campo Grande, Florianópolis, and Teresina) across 43 measures of quality and then linked the �ndings to students’ second-grade reading scores. A high-quality pre-school would have—for example—a well-lit main space, good supervision of child safety, books, puzzles, and small blocks for �ne motor skill development, affectionate and respectful caregivers, a daily routine that includes both individual time and small-group activities, and regular information provision to parents. A lower quality pre-school would lack most of these characteristics. After controlling for some observable differences between students, the study showed that children who a ended higher quality pre-schools did much be er on their second grade reading exam. In fact, the difference in reading scores between children a ending a higher quality pre-school and those a ending a lower quality pre- school was even larger than the difference between children with a mother who had �nished secondary school and those without. The test scores of children who a ended a low-quality pre-school were not signi�cantly be er than those of children who did not a end pre-school at all (Fundação Carlos Chagas 2010). Of course, there may be other, unob- servable student and family characteristics that explain both the selection of a high-quality pre-school and student performance. However, these results are strongly suggestive that the quality of a pre-school (and not merely a endance) determines student outcomes. Evidence at the creche level reveals analogous results. In 2001, researchers used similar observational instruments in 100 creches around Rio de Janeiro, and looked at the association between creche quality and children’s cognitive, social, and physical development. Overall development of children in the highest quality creches was sig- ni�cantly higher than in the worst creches, particularly in terms of social and cognitive development. Importantly, the authors also note that improvements in the quality of activities and in program structure are relatively inexpensive ways to enhance the qual- ity of early child education (Barros et al. 2011a). In contrast, other kinds of investments such as infrastructure improvements had no impact on child development for the same group of creches. Thus, investing in quality—and identifying the most needed elements of quality—has large payoffs. In Chapter Three, we use this quality analysis to inform the quantity-quality trade-offs municipal governments must make. 6 A World Bank Study The results of these Brazilian studies complement international studies from the United States, Canada, and elsewhere, demonstrating that high-quality center-based care makes a huge difference, and showing that low-quality center-based care can be worse for development than home care.4 In one study from the United States, children in low-quality programs performed worse on post-tests than on pre-tests of social and behavioral skills, motivation, and self control after one school year, while children in medium- and high-quality programs improved (Thornburg et al. 2009).5 Brazil cannot afford to merely provide spaces in child care centers. It must provide quality. Impacts for vulnerable children While there is evidence for the importance of center-based early child stimulation and education for all children, the impacts seem to be greatest for vulnerable children. Vul- nerable children come from poor families with relatively low levels of parental edu- cation, and receive relatively less cognitive stimulation and exposure to vocabulary at home. In the absence of interventions, these children are at a great disadvantage, and that disadvantage expands over time. As in Ecuador (�gure 1.1), evidence from the United States shows that the gap between the highest and the lowest income quintiles in terms of both social skills and school preparation (reading and math) is signi�cant upon school entry: children in the lowest income quintile have math scores 20 percent lower than children in the highest income quintile (Barne , Brown and Shore 2004). The vul- nerable children are just beginning their school career, and they are already far behind more privileged children. Therefore, early child education can particularly bene�t the poor, helping to close the gap in cognitive development across income groups. A World Bank study compares adults from two regions of Brazil (the Northeast and the Southeast) who a ended pre- school to those who did not and found that pre-school a endance is associated with additional total years of education. However, individuals bene� ed more from pre-school if their parents were illiterate than if their parents had four years of schooling. Pre-school was associated with 0.4 additional years of lifetime educational a ainment for those whose parents had four years of schooling, and with 0.6 additional years for those whose parents were illiterate (Young 2001). Other evidence from programs in the United States, the United Kingdom, Nepal, Vietnam, Guatemala, and elsewhere con�rms that many early child devel- opment programs have particularly high impacts for the most vulnerable (Engle et al. 2007). On the other hand, the more ambiguous evidence on early child education pro- grams comes from universal programs, where vulnerable children are not targeted. A large-scale pre-school program in Canada (Quebec) showed small but negative impacts on children’s health and behavioral development when the program was expanded to middle-class children who would otherwise have been at home with a parent (Baker, Gruber and Milligan 2008; Almond and Currie 2011). Likewise, a universally-available pre-school program in Denmark showed no impacts (Gupta and Simonsen 2010). To be sure, some universal programs have shown positive effects on average (�gure 1.3), but these �ndings are less consistent than are those for the most vulnerable children. These �ndings highlight two key reasons to target the poorest, most vulnerable chil- dren in providing early child education. First, early child education programs show the most consistent returns for vulnerable children. These children are least likely to receive adequate cognitive and non-cognitive development at home, and least likely to be en- rolled in private early child education. Second, these vulnerable children are most likely Early Child Education: Making Programs Work for Brazil’s Most Important Generation 7 Figure 1.3: Sample of large-scale programs Oklahoma Quebec Pre-school Denmark Chile Argentina (concurrent) (concurrent) (age 11) (concurrent) (3rd grade) Socio-emotional Pre-reading Test scores Vocabulary Pre-writing Behavioral Pre-math Cognitive Cognitive Behavior Behavior Health Health + No impact Ambiguous – Sources: The Quebec �ndings come from Baker, Gruber, and Milligan (2008). The Oklahoma �ndings are from Gormley and Gayer (2006). The Denmark �ndings are from Gupta and Simonsen (2010). The Chile �ndings are from Urzúa and Veramendi (2011). The Argentina �ndings are from Berlinski, Galiani, and Gertler (2009). to enter school with a signi�cant de�cit, so there are equity reasons to target them and allow them to catch up. Brazil’s policy for pre-school is that it be universal and com- pulsory, but it will take several years to achieve this (80 percent were enrolled in 2009), and municipalities will have until 2016 to come into compliance. Creche access will not be universal for the foreseeable future (18 percent were enrolled in 2009). Policies that target the most vulnerable are the most likely to raise enrollment and to cost-effectively improve child outcomes. How Has Brazil Advanced in Early Child Development and Early Child Education in Recent Years? The situation of young children in Brazil has improved signi�cantly in recent decades, in terms of government policy, government programs, and—most importantly—child outcomes. These gains can be observed across sectors, since child policy spans educa- tion, health, legal protections, water, sanitation, and other areas. Government policy advances Brazil’s government policy over the last 20 years can broadly be characterized as pro- child.6 The Federal Constitution of 1988 declared children to be citizens and identi�ed the Ministry of Education as responsible for the education of children ages 0–6. The 1990 Statute for Children and Youth clari�ed children’s rights to education, to a fam- ily or guardianship, and to protection from labor and from full penalties for crimes. Other laws in more speci�c areas have improved life for children. The Law for Free Civil Registry (1997) eliminated any fee associated with birth registration, which is essential for helping children take advantage of social programs.7 In health, the National Pact to Reduce Maternal and Neonatal Mortality8 (2004) and the National Policy on Food and Nutrition9 (1999) have also advanced the state of children. 8 A World Bank Study Table 1.2: Policy changes in early child education in Brazil Year Law Policy change 1975 Pre-school Education Coordinating Body Education of 4–6 year olds put under responsibility of MEC (Coordenação de Educação Pré-Escolar) 1988 Federal Constitution (Art 208.IV) De�ned ECE (age 0–6) as government responsibility 1996 Law of Directives and Bases of National Placed responsibility for ECE in municipalities Education De�ned ECE (age 0–6) as part of Basic Education Set starting age for obligatory education at 7 1998 Fundamental Education Fund (FUNDEF) Increased education revenue to municipalities 2005 Law 11.114 Lowered obligatory starting age to 6 2006 National Policy for Early Child Education De�ned strategies and goals for ECE at each level of government 2007 Basic Education Fund (FUNDEB) Provided municipalities with capitation grants for ECE starting from birth 2009 Constitutional Amendment 59 Lowered obligatory starting age to 4 Source: Compiled by authors. Over the last 30 years, Brazil has made policy advances in early child education (table 1.2), and these advances are clearly linked to expansions in programs to bene�t children. The precursor to these was the 1975 establishment of the Pre-School Education Coordinat- ing Body, which put the education of 4–6 year olds under the responsibility of the Ministry of Education, or MEC (Ministério de Educação 2006). Next came the Federal Constitution of 1988, which squarely identi�ed early child education as a responsibility of the gov- ernment—speci�cally, the Ministry of Education. Then, the Law of Directives and Bases of National Education (LDB) of 1996 gave municipal governments the responsibility for ECE and made age seven the starting age for compulsory education. LDB also de�ned ECE as part of basic education, continuing a transition of creches (0–3) from the Ministry of Social Development to the Ministry of Education that stemmed from the 1988 constitu- tion but had not been completed in practice. A 2005 amendment to the LDB reduced the compulsory education starting age to six, and a 2009 constitutional amendment reduced it to four, making Brazil’s starting age among the youngest in the world (table 1.3). The transition of creches to the purview of the Ministry of Education is essential in viewing them not as a holding place while mothers work, but rather as a place of education. The transition has been a slow one, signaled by the fact that an inter-ministry commission was established as late as 2005 to propose the transition of creches and pre- schools still supported through the Ministry of Social Development to the Ministry of Table 1.3: Starting age for compulsory education in Europe, North America, and selected other countries Age Countries 3 Mexico 4 Brazil, Northern Ireland 5 Argentina, Greece, Hungary, Latvia, Malta, the Netherlands, England (UK), Scotland (UK) 6 Austria, Belgium, Canada, Chile, Cuba, Denmark, France, Germany, Hong Kong SAR, China; Iceland, Ireland, Italy, Norway, Poland, Republic of Korea, Romania, Singapore, Slovak Republic, Slovenia, Spain, Turkey, United States* 7 Bulgaria, Estonia, Finland, Lithuania, Sweden Source: Eurydice (n.d.) and UNESCO Institute for Statistics. Mexico: Yoshikawa et al. (2007). Note: *True in two thirds of states in the United States. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 9 Education. This was also one of the priorities highlighted in January of 2011 by Minister of Education Fernando Haddad (Paraguassu 2011). A major legislative effort to ensure sufficient municipal investments in basic educa- tion (primary and lower-secondary school) was the Fund for the Development of El- ementary Education and Teacher Development (FUNDEF), implemented in 1998. The law did not target early child education, but it made additional education revenue avail- able to many municipalities, potentially freeing funds for early child education. The law mandated that all municipal and state governments contribute 15 percent of their rev- enue from four existing intergovernmental transfers to a state-level fund.10 This resulted in 27 education funds: one for each state and one for the Federal District. The money in each fund was redistributed to contributors according to their share of the state’s en- rolled public primary and lower-secondary school students. All receipts from the fund had to be spent on education, and 60 percent had to be spent on teacher compensation. Each municipality established a council to oversee expenditures. If fund receipts alone did not reach a federal minimum per enrolled student, the federal government would top-off the fund, bringing it up to the required level.11 FUNDEF laid the groundwork for further education �nance reforms that had much more direct impacts on early child education. In June 2005, Congress voted on a bill pro- posing the Fund for the Maintenance and Development of Basic Education (FUNDEB). This reform proposal extended the provisions of FUNDEF to also cover pre-school stu- dents, thus ensuring municipal governments a minimum amount of funding per en- rolled pre-school child. Initially, the FUNDEB proposal excluded children enrolled in creche. The Brazilian Congress’ Select Commi ee charged with reviewing FUNDEB then decided to additionally include children enrolled in creche, and a vote on the floor of Congress eventually con�rmed their inclusion (Campos, Fullgraf and Wiggers 2006). The FUNDEB law was implemented in 2007.12 In its �nal form, FUNDEB modi�ed FUNDEF’s redistribution algorithm to take into account total creche, pre-school, pri- mary, and secondary school students (MEC 2008). It also gradually increased the frac- tion of revenue paid into the state-level funds.13 This created clear, new incentives for municipalities to invest in early child education. Improvements in government programs Public investment in programs for children and adolescents in Brazil doubled between 2006 and 2009 (�gure 1.4). Some of the most far-reaching and high-pro�le interventions are the social assistance programs that also bene�t education and health, Bolsa Escola and Bolsa Alimentação, which combined to become Bolsa Familia in 2003. Coverage of children ages 0 to 7 by Bolsa Familia rose from 3.7 million to 6.2 million children between 2005 and 2010 (Ministério da Saúde do Brasil n.d.). Child outcome improvements These policies and programs have truly borne fruit. In the last decade, unregistered births in Brazil dropped by 60 percent, from 30 percent in 1995 to about 10 percent in 2007 (Muzzi 2010). In health, the proportion of stunted children under age �ve (i.e., extremely low height-for-age) has fallen from 14 percent in 1996 to just 7 percent in 2007.14 Likewise, child mortality among under-�ve children and infant mortality (under age one) have both dropped by almost two-thirds over a decade. While cognitive development of young children is not measured on a wide scale, participation in ECE programs has in- 10 A World Bank Study Figure 1.4: Public child investments per capita, 2006–2009 1,000 Child investments per capita 900 800 in reais 700 600 500 400 2006 2007 2008 2009 Year Source: UNICEF (n.d.). creased dramatically, reaching 18 percent of children ages 0–3 and 80 percent of children ages 4–6 in 2009. While Brazil has made a great deal of progress on early child development policy in recent years, in some ways it still lags in early child outcomes. Child health is a major concern, with high mortality and medium stunting rates. Brazil has an under-�ve mor- tality rate that is in the middle of the range for Latin American countries, but the highest for middle income Latin American countries. In 2007, Argentina, Chile, Colombia, Costa Rica, Mexico, Uruguay, and República Bolivariana de Venezuela all had lower under- �ve mortality rates than did Brazil, while most of the other (poorer) Latin American countries had higher mortality rates (table 1.4). Early child education enrollment in Bra- zil is high relative to countries in the region, but still not comparable to that found in the richest nations, as will be discussed in detail in Chapter 3. Table 1.4: Child health outcomes across the world in 2007 Mortality rate per 1,000 Stunting rate (%) Region population (under 5) (under 5) GDP per capita, PPP* Brazil 23 7% $9,181 Colombia 20 — $8,084 Mexico 19 16%† $13,371 Latin America and Caribbean 25 16% $9,712 United States 8 4%‡ $43,662 Sweden 3 — $34,782 OECD 9 — $30,860 Sources: World Bank World Development Indicators; UNICEF State of the World’s Children. Note: * GDP is purchasing power parity (i.e., adjusted for spending power in different environments), in constant 2005 International dollars. † 2006 data (2007 data unavailable) ‡ 2004 data (2007 data unavailable) — Not applicable. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 11 This report, in addition to examining Brazilian data and policy over time, draws on data from a variety of international comparisons. Some of them are comparable to Brazil in income (such as Colombia and Mexico), others in scope of population (such as the United States and Mexico), and some in terms of urbanization and population density (Sweden). All have expanded or improved their early child education systems in recent years and provide insights on which Brazil can draw. The report also highlights innova- tive practices from elsewhere in the world that are relevant to Brazil’s challenges. Key Issues Facing Brazil in Early Child Education Despite many improvements in early child education, including strengthening some ele- ments of quality and expansions in access, Brazil has essential issues to resolve if it seeks to close the opportunity gap and make long-term, high-return investments in Brazil’s children. This report explores each of these issues in detail. How to provide world-class early child education, especially in the most dif�cult locations Research on early child education centers across Brazil demonstrates both that the qual- ity of the center has a massive impact on child development and that quality is sorely lacking in many creches and pre-schools. While infrastructure quality and teacher quali- �cations have steadily improved, the quality of activities is still de�cient. In addition to the broad curricular guides published by MEC in 1998 and the curricular directives laid out by the National Education Council in 2009, teachers need speci�c guidance on the most stimulating activities. To close the quality gap in early child education, other levels of government should support municipalities with tools for monitoring the qual- ity of activities, for developing speci�c programs of intellectual and social stimulation, and for holding centers accountable for the quality of their programs. Fundação Carlos Chagas (in partnership with the Ministry of Education and the Inter-American Develop- ment Bank) as well as the Institute for Applied Economic Research (IPEA) have already developed tools for evaluating quality. The federal government can also provide clear guidelines for monitoring and licensing ECE centers, ensuring that minimum standards are available for municipalities. Chapter 2 of this report details the evolution of quality in ECE centers in Brazil and the variety of ways that Brazil can strengthen them. Reaching the children who need these programs the most High-quality early child education must be extended to children who, to date, have been marginalized. While access has expanded impressively (�gure 1.5 below), more than one million children remain out of pre-school, which is now a constitutionally com- pulsory level of schooling (albeit compliance will not be mandatory until 2016). The poorest children are most in need of these programs, and targeting them should be the government’s �rst priority. However, enrollment for the poorest children is only 67 per- cent for pre-school and 12 percent for creches. Currently, 13 percent of public creche spaces—which are entirely free to the child—are taken by children from the richest �fth of families, while only 12 percent of poor children are in any creche at all. Brazil must be strategic in where it rolls out ECE in the short run, and creative in how it does so, to ensure that new spaces reach the underserved. These income-based disparities in access mirror those between rural and urban ar- eas. In rural areas, alternative delivery mechanisms like home visits—as are done in the state of Acre (through education) and in Rio Grande do Sul (through health)—may be 12 A World Bank Study Figure 1.5: Fraction of children enrolled in early child education (1996–2009) 0.8 0.6 Fraction enrolled 0.4 0.2 0.0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Creche Pre-school Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. the only way to reach the most distant children. Chapter 3 of this report explores ECE coverage in detail and the implications of the coming expansion. The next steps in ECD for Brazil While improving quality and expanding coverage are at the top of Brazil’s stated agenda for early child development, Chapter Four of this report explores three other areas inte- gral to success in delivering effective ECD services on a large scale. How to effectively integrate early child development services? Considering areas in ECD, child health is one in which Brazil is weakest relative to comparator countries, whereas coverage in social protection and in education is stronger. The relative strength of education provides an opportunity to bundle services and take advantage of the stron- ger education infrastructure to provide improved health services for children. Bundling services can improve early child health and cognitive development simultaneously. Evi- dence shows that combined programs can be highly effective (see, e.g., Nores and Bar- ne 2010). Examples from Australia, New Zealand, Sweden, and Chile all demonstrate the effectiveness of cross-sectoral child development programs in improving outcomes across sectors. In Brazil, the municipality of Rio de Janeiro, the state of Acre, and the state of Rio Grande do Sul, among others, are experimenting with these kinds of programs. How to leverage private support? Many other countries have succeeded in incor- porating private sector support effectively to provide large-scale ECD services. This can include direct provision, as Brazil does through creches and pre-schools contracted (con- veniado) by the government. It can also include partnerships to augment public �nanc- ing—likely essential in the face of Brazil’s massive required expansion in coverage. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 13 How to compensate for municipal differences? Although reforms in education spending have created incentives for all municipalities to invest in ECE, poorer children in wealthier and in more unequal municipalities are likely to receive fewer public ECE options because of lesser public commitment to ECE provision. Brazil can take steps, such as encouraging elements of participatory budgeting practices, to protect the poor- est children’s ECE opportunities everywhere. Notes 1. In this report, Early Child Development (ECD) refers to all activities intended to promote cog- nitive, emotional, and behavioral development in children age zero to six. Early Child Education (ECE) refers speci�cally to those ECD activities under the purview of the Ministry of Education. Thus, creche and pre-school a endance is part of ECE, whereas child nutrition and health care provided at a clinic are part of ECD. 2. These enrollment rates are based on parent reports from the 2009 Pesquisa Nacional por Amostra de Domicilios (PNAD). They are the share of children of the given age range that are enrolled in school—whether in early child education or primary school institutions. 3. While Article 6 of Constitutional Amendment 59/09 made four the compulsory starting age for education, municipalities and states have until 2016 to reach full compliance. 4. The �nding of negative outcomes for low-quality center-based care has been among children from higher-income households (see Almond and Currie’s discussion of Baker et al. 2008). How- ever, the Brazil �ndings of no impact of low-quality center-based care applied to a representative sample of pre-school a endees. 5. Another study shows a clear correlation between low quality creches and worse outcomes than no daycare at all (Howes 1990). 6. A more comprehensive summary of these can be found in the National Plan for Early Childhood (Rede Nacional Primeira Infância 2010). 7. Lei da Gratuidade do Registro Civil (No 9.534/1997) 8. Pacto Nacional pela Redução da Mortalidade Materna e Neonatal 9. Política Nacional de Alimentação e Nutrição 10. These four transfers—the state and municipal participation funds (FPE/FPM), the tax on goods and services (ICMS), the export-proportional excise tax (IPIexp), and transfers under the comple- mentary law (LC 87/96)—constituted the majority of revenue for most governments. 11. In 1998, six states received federal top-off: Bahia, Ceará, Maranhão, Pará, Pernambuco, and Piauí. The per-child minimum changed annually. It began to vary with child grade level in 2000 and with school type (rural or urban) in 2005. 12. FUNDEB was created by Constitutional Amendment No. 53/2006 and regulated by Law No. 11.494/2007 and Decree No. 6.253/20071. FUNDEB was established for the period 2007–2020, and effectively replaced FUNDEF. 13. FUNDEB gradually increased state and municipal contributions from their four main transfer revenue sources (FPM/FPE, ICMS, IPIexp, LC 87/96): from 15 percent in 2006 to 16.67 percent in 2007, 18.33 percent in 2008, and 20 percent in 2009 onward. FUNDEB also expanded the list of transfer revenues of which a percentage had to be contributed to the state education fund. These transfer sources included: the tax on inheritance and donations (ITCMD), the motor vehicles tax (IPVA), the 50 percent share of the rural territory tax due to municipalities (ITR), and income tax bene�ts levied on income paid by the municipalities or the state (IRmuni and IRest). The fraction of these transfers that had to be contributed to the state education fund was 6.67 percent in 2007, 13.33 percent in 2008, and 20 percent in 2009 and onward. 14. Stunting is an indicator of child malnutrition. The population of stunted children under age �ve is the population of under-�ve children whose height for age is more than two standard deviations below the median for the international reference population ages 0–59 months. For children up to two years old, height is measured by recumbent (lying down) length. For older children, height is measured while standing. The data are based on the World Health Organization’s new child growth standards released in 2006. CHAPTER 2 Ensuring High Quality Early Child Education for Brazil’s Children A s Brazil expands its access to pre-school and creche education, hopefully it will cap- ture the major returns promised to investments in early child development. How- ever, as shown in Chapter One, placing children in early child education establishments of low quality makes no promise of returns. The average quality of early child education in Brazil is low, but some municipalities are delivering quality care. Brazil can improve ECE quality by providing be er materials to strengthen activities, encouraging effective monitoring systems which keep ECE institutions accountable for their results, and facili- tating knowledge sharing among providers. Current Quality of Early Child Education in Brazil Global evidence supports the importance of quality for ECD outcomes, but measuring quality for young children is elusive. Clearly high-quality early child education is the product of many elements: the teacher, the infrastructure, the activities, the hygiene rou- tines, and other factors. At higher levels of education, quality is often measured through value-added indicators of student improvement. However, while there are many tools available for measuring child development at early ages, they are less precise than mea- sures for older children and few systems universally test young children. Also, linking those measures to creche and pre-school incentives can have the perverse effect of lead- ing centers with discretion over enrollment to exclude developmentally-delayed chil- dren. As a result, creche and pre-school quality is usually measured with multi-dimen- sional observational instruments, in which enumerators observe the creche or pre-school in session and characterize its quality across a number of areas. A 2001 study of 100 creches in Rio de Janeiro demonstrates just how these various el- ements have an impact on different aspects of child development. Researchers observed hundreds of characteristics of creches over several days, which were then summarized across three dimensions: infrastructure, health and sanitation, and activities and pro- gram structure. (Most of the standard observation instruments approximate these di- mensions.) They then implemented a test of children’s social, physical, and mental de- velopment, to see how observed creche quality correlated with children’s development (table 2.1). They observed that the quality of the infrastructure had a strong positive association with social and physical development: A ending one of the best creches in terms of infrastructure would advance a child by 3.8 months in her social develop- ment and 2.4 months in her physical development, relative to a ending one of the worst creches. Alternatively, activities and program structure in the best creches added over 3 months to children’s social and mental development. Thus, high-quality creches can contribute to children’s development across the range of cognitive areas. 14 Early Child Education: Making Programs Work for Brazil’s Most Important Generation 15 Table 2.1: Impact of creche quality on months of child development, measured in months of child development Dimensions of development Statistical Overall Social Physical Mental signi�cance Overall 1.2 2.3 0.3 1.8 99% Dimensions of quality Infrastructure –0.9 3.8 2.4 1.2 95% Health & sanitation –1.7 –2.8 –3.0 –2.9 90% Activities & program structure 2.5 3.2 –0.04 3.1 Insigni�cant Human resources 1.7 –0.7 2.0 1.2 Parents & community relations –0.2 –0.9 –0.6 –0.5 Source: Barros et al. (2011a). Note: Shades of gray refer to statistical signi�cance (white is insigni�cant). Darker shades refer to more highly signi�cant correlations. Infrastructure quality In Brazil, the quality of facilities for early child education is on the rise. High-quality infrastructure can expand learning and interaction opportunities, minimize distractions and wasted time, and make parents feel more comfortable entrusting their children’s care to another. Average infrastructure quality improved between 2001 and 2009 in all regions and for both creches and pre-schools. In �gure 2.1, infrastructure quality is mea- sured by an index that averages seven quality indicators, available in the census of all registered pre-schools and creches. These measures are the fraction of institutions with a dedicated school building (as opposed to being located in a church, teacher’s home, or another place), an indoor bathroom, electricity, a connection to the public water net- work, a connection to the public sewer network, a library, and a computer. In addition to improving on the overall quality index during 2001–2009, creches in each region im- proved on nearly every one of the seven individual quality measures.1 For example, the likelihood of being in a dedicated school building rose from 83 percent to 91 percent, the likelihood of having an indoor bathroom rose from 90 percent to 95 percent, and the likelihood of having a computer rose from 35 percent to 68 percent. The gains, however, have not been evenly distributed across regions of the country. Pre-school infrastructure in the North deteriorated in some respects over 2001– 2009.2 Pre-schools there became more likely to have a dedicated school building, elec- tricity and a computer. But they became less likely to be connected to the public water or sewer networks, or to have a library. The North saw a 51 percent increase in the total number of registered pre-schools between 2001 and 2009; this rapid expansion might ex- plain drops in some infrastructure quality indicators. Existing data make it impossible to determine whether pre-schools with low-quality infrastructure already existed but were not registered in 2001 (but then had registered by 2009), or whether pre-schools built in the North between 2001 and 2009 had lower-quality infrastructure than did pre-schools already in place in 2001. The inability to distinguish between these stories underscores the importance of registering schools. ECE infrastructure quality in the North and in the Northeast has consistently lagged behind that of other regions, both in the overall index and in the individual factors in- cluded. For example, in 2009 the average infrastructure quality index score of pre-schools 16 A World Bank Study Figure 2.1: Average quality index of creches and pre-schools by region (2001 and 2009) 0.90 0.85 0.84 0.79 0.78 0.81 0.82 0.79 0.79 0.80 0.76 0.76 0.70 0.70 0.69 0.70 0.68 0.64 0.63 0.60 0.58 0.60 0.54 0.53 Index 0.50 0.47 0.47 0.40 0.42 0.40 0.30 0.20 0.10 0.00 2001 2009 2001 2009 2001 2009 2001 2009 2001 2009 2001 2009 Brazil North Northeast Southeast South Central West Region Creche average Pre-school average Source: Censo Escolar (2001, 2009). in the North was only half that of pre-schools in the Southeast. And while all registered pre-schools and creches in the Southeast and in the South had electricity in 2009, only 75 percent of pre-schools in the North had electricity. In all regions but the South, creches have higher-quality infrastructure than do pre- schools. Part of this difference may come from the fact that creches are more likely to be private than are pre-schools. In 2009, 43 percent of creches were private, but only 24 percent of pre-schools were private. If parents pay for private schools because they want higher quality infrastructure than is available in public schools, this could explain part of the infrastructure quality differential between creches and pre-schools.3 Changes in the number of students per room shed light on whether new enrollments are accommodated through an expansion in the supply of schools or through more in- tensive use of existing schools. Program quality may decline with class size. While total enrollment has increased, the number of students per classroom has decreased.4 In pre- schools, class size has decreased from 39 students per classroom to 32 students. Creche class size has remained constant at 26 students per classroom (Evans and Kosec 2011). However, this compares unfavorably with recommendations from the United States’ National Asso- ciation for the Education of Young Children that even �ve-year-olds have a maximum of ten children per caregiver, and that younger children have lower ratios (NAEYC 2008). The number of students per classroom remained the same or shrank in every region and for both levels of ECE during 2001–2009 (�gure 2.2). The number of classrooms seems to be growing at least as fast as the student population. Regional disparities are again apparent, with the North and the Northeast having the largest number of students per classroom, and the South and the Southeast having the smallest. But these regional disparities in student-room ratios are shrinking over time. Public creches and pre-schools tend to be of much lower quality than their private counterparts (�gure 2.3). Interestingly, contracted (or conveniada) private schools tend Early Child Education: Making Programs Work for Brazil’s Most Important Generation 17 Figure 2.2: Average number of students per classroom, by year and region Creche Average no. of students 40 40 36 30 34 32 29 26 26 26 20 21 23 23 21 10 0 Brazil Central West North Northeast South Southeast 2001 2009 Pre-school Average no. of students 50 49 40 42 39 40 38 38 30 32 34 34 34 29 30 20 10 0 Brazil Central West North Northeast South Southeast 2001 2009 Source: Censo Escolar (2001, 2009). Figure 2.3: Average infrastructure quality index of creches and pre-schools in Brazil by administrative dependency (2009) 1.0 0.8 0.6 Index Public 0.4 Private 0.2 0 Creche Pre-school Administrative dependency Source: Censo Escolar (2001, 2009). to have infrastructure quality that is higher than that of public schools. This is contrary to the perception in many municipalities, and may reflect a greater concentration of contracted schools in municipalities where the overall creche and pre-school quality is higher. In 2009, the infrastructure quality index of public creches in Brazil was about three-quarters that of private creches, and the quality index of public pre-schools was less than two-thirds that of private pre-schools. 18 A World Bank Study Activities and program quality While quality infrastructure has proven impacts on children’s physical development, it is insufficient to provide children with the cognitive development they need from an early child education center. This is central to Minister of Education Fernando Haddad’s statement that, “It is necessary to change the environment of the creche and the pre- school, to fortify these centers as learning establishments� (Weber 2011). Last year, Fundação Carlos Chagas, the Ministry of Education, and the Inter-Amer- ican Development Bank collaborated on a study of creches and pre-schools in six capi- tal cities in Brazil: Belém, Campo Grande, Florianópolis, Fortaleza, Rio de Janeiro, and Teresina (Fundação Carlos Chagas 2010). This study employed two of the most stan- dard observational instruments for early child education: the Early Childhood Environ- ment Rating Scale—Revised for pre-schools (ECERS-R, Harms, Clifford and Cryer 2005) and the Infant and Toddler Environment Rating Scale—Revised for creches (ITERS-R, Harms, Cryer and Clifford 2006). The domains of quality measured (see table 2.2) are similar to those used in Rio de Janeiro in 2001. Table 2.2: Domains in early child education observation instruments ITERS-R (for centers of children 0–2½ years of age) ECERS-R (for centers of children 2–5 years of age) Space and Furnishings Space and Furnishings Personal Care Routines Personal Care Routines Listening and Talking Language-Reasoning Activities Activities Interactions Interactions Program Structure Program Structure Parents and Staff Parents and Staff Source: FPG Child Development Institute (n.d.). The standard ITERS-R and ECERS-R instruments rate ECE centers—overall and in each domain—in four quality categories: inadequate, basic, good, and excellent. The Fundação Carlos Chagas study adapted the scale into �ve categories: inadequate, basic, adequate, good, and excellent. This adaptation makes comparisons between the Brazil- ian results and international applications of the instruments difficult. However, a recent study of almost 700 American pre-schools placed the average pre-school evenly within the “adequate� category (LoCasale-Crouch et al. 2007), as did a study of pre-schools in Germany and in Portugal (Tie e et al. 1998).5 Quality in Brazilian ECE seems sig- ni�cantly lower: Brazilian pre-schools rated “inadequate� in two cities, “basic� in three cities, and “adequate� in just one city. Brazilian creches in two cities rated “basic� on average, whereas creches in the other four cities were “inadequate.�6 Even this measure of quality is troubling, and quality is likely to be even lower in rural areas with access to fewer resources. Across areas, the study found that—for both creches and pre-schools—the most lacking areas are activities and program structure (�gure 2.4). At both age levels, the low quality ratings are largely driven by a lack of good activities, inadequate program struc- ture, and non-stimulating space and furnishings. Within activities, there are de�ciencies across the board: in every area from playing with blocks, to physical activities, �ne mo- Early Child Education: Making Programs Work for Brazil’s Most Important Generation 19 Figure 2.4: Distribution of creche quality by broad quality domain 10 9 Excellent 8 7 Good Score 6 Adequate 5 4 Basic 3 Inadequate 2 1 es n ct m d nts d nd n d al of io re p m an f ru r a iti af e g ct ta a an are R ent ca on s tiv ur in st og st rs ine ra ui e rs n g e q ac Ac P Pr te de k i pe out Sp In un e a Sp Quality domain Source: Fundação Carlos Chagas (2010). Note: The distribution of scores is very similar for pre-schools. Creches are rated on a scale of 1–10, with 1–3 characterized as inadequate, 3–5 as basic, 5–7 as adequate, 7–8.5 as good, and 8.5–10 as excellent. tor activities (puzzles, games, etc.), nature or science activities, use of books, etc. (�gure 2.5). The creche level, for example, lacks activities involving books, those incorporating music and movement, nature and science activities, and appropriate use of videos and computers (i.e., age-appropriate content, for appropriate periods of time, and only for children over 12 months of age). Figure 2.5: Distribution of creche quality by selected speci�c areas 10 Excellent 9 8 Good 7 Adequate 6 Score 5 Basic 4 3 Inadequate 2 1 0 en & iti e pl n ct ild in of s ks ok tiv nc of isio em ic ra h oc ld n t es ay n t'n te t-c bo ac cie ov us hi io io rv bl -c is in ul M s pe of ith ild v Ad & er Su se w ch p e m Su U ur ay at Pl N Quality domain Source: Fundação Carlos Chagas (2010). 20 A World Bank Study On the other hand, the strongest area for both creches and pre-schools is interac- tions. Within that area, interactions between caregivers and children are rated relative- ly highly. That said, interactions are still rated as “adequate,� failing to achieve either “good� or “excellent� status. Still, this suggests that caregivers are engaged with chil- dren, and mostly need help involving those children in constructive activities. These averages suggest particularly bad conditions for some of the worst creches and pre-schools. In fact, half of creches and almost one-third of pre-schools were found to be “inadequate� whereas only 1 percent of creches and 4 percent of pre-schools achieved a “good� rating (�gure 2.6). Not a single institution achieved an “excellent� rating. Likewise, the previously-discussed 2001 study of 100 creches in Rio de Janeiro found major quality gaps between the best and worst creches. The size of these gaps is striking: The best �fth of creches were rated two and a half times be er in activities and program structure than were the worst �fth of creches (Barros et al. 2011a). Figure 2.6: Distribution of quality across ECE institutions 60 50 40 Percent 30 Creche 20 Pre-school 10 0 Inadequate Basic Adequate Good Excellent Quality Source: Fundação Carlos Chagas (2010). The Fundação Carlos Chagas study did not �nd any systematic difference in ob- served quality across private, public, or conveniado schools. However, Early Child Devel- opment Spaces (EDIs)7—or educational institutions with only creche and/or pre-school— tended to be of signi�cantly higher quality than those connected to a primary school.8 What about teachers? Measuring the quality of teaching is difficult, and the current best measures applied to higher levels of education rely on student test scores. This is impossible at the level of early child education, where universal student testing is nonexistent and measurement of child outcomes may be much less exact. In primary education, the teacher has been identi�ed as the single most important factor for student success (Hanushek and Rivkin 2010). Yet while teachers have an essential influence on children, years of teacher educa- tion does not seem to ma er much for student outcomes; a survey of 170 different stud- ies examining the relationship between teacher education and student outcomes found that 91 percent of them found either no relationship or—in a few cases—even a negative relationship. Among the best studies in terms of methodology, none found a positive relationship between teacher education and student outcomes (Hanushek and Rivkin Early Child Education: Making Programs Work for Brazil’s Most Important Generation 21 2004). Thus, teacher quality is clearly important, but a teacher’s formal educational level is a poor proxy for actual effectiveness. Dozens of studies have explored the relationship between caregiver education and observed center quality at the ECE level (Barne 2003 and Whitebook 2003), and many have found a positive correlation. However, many of these studies fail to control for other characteristics that are likely correlated with teacher education, such as the wealth of the community, which could both a ract more educated teachers and have other char- acteristics that would suggest be er observed quality. A recent study overcomes some of those failings by examining seven major pre- school programs in the United States using the same method and including a broad range of control variables (Early et al. 2007). They examined the relationship between teacher education and four different outcomes: classroom quality, receptive vocabu- lary (i.e., the words one understands well enough to recognize when read or heard), pre-reading skills, and early math skills. When the researchers examined whether the highest degree a ained by the lead teachers could predict these outcomes, there was evidence of a positive relationship in only six out of 27 different analyses.9 When the researchers focused on the highest education levels among teachers with a major in ECE, the results were once again weak: only two out of 19 analyses showed evidence of a posi- tive association. Finally, the researchers looked speci�cally at teachers with a Bachelor’s degree, to see whether a major in ECE had a signi�cant impact in that case: exactly one of 23 analyses showed a signi�cant positive relationship. In other words, a careful, consistent examination across seven different large-scale studies showed no consistent relationship between teacher education and student per- formance. Does that mean that teacher or caregiver education does not ma er? Rather, the lack of a relationship implies that broad measures of teacher education, even with a specialization in ECE, do not predict outcomes. Most likely, this implies that current programs of teacher education in the United States do not—on average—equip teachers with the skills needed to effectively improve children’s cognitive development. Furthermore, teachers of early child education in Brazil have signi�cantly less educa- tion than do primary and lower secondary school teachers (�gure 2.7). There may be some Figure 2.7: Level of education of teachers in ECE versus primary and lower- secondary school 100 80 46 54 70 Percent 60 40 % Higher education 54 50 % High-school or less 20 30 0 Creche Pre-school Primary & lower secondary school Level of education Source: Censo Escolar, “Sinopse do Professor,� 2009. 22 A World Bank Study Figure 2.8: Fraction of teachers with higher (post-secondary) education, by year and region Creche 0.38 Fraction enrolled 0.4 0.35 0.35 0.34 0.3 0.29 0.22 0.2 0.18 0.17 0.13 0.14 0.1 0.06 0.02 0 Brazil Central West North Northeast South Southeast Region 2001 2009 Pre-school 0.64 0.63 Fraction enrolled 0.60 0.6 0.51 0.4 0.34 0.34 0.37 0.30 0.29 0.25 0.2 0.04 0.07 0 Brazil Central West North Northeast South Southeast Region 2001 2009 Source: Censo Escolar (2001, 2009). threshold below which teachers are simply ineffective. This report does not rule out the importance of a minimum standard of teacher education. For example, the recent Bra- zilian requirements that Brazilian teachers have secondary education should ensure lit- eracy and be er capacity for teachers to take advantage of other resources and materials. It may also raise the prestige of the profession, a racting more competent individuals. The quali�cations of ECE teachers have—as in infrastructure—improved dramati- cally over time, with the proportion of teachers with some post-secondary education more than doubling from 2001 to 2009 in both creches and pre-schools (�gure 2.8). Be- tween 2001 and 2009, the fraction of pre-school teachers with higher education increased by 70 percent or more in all regions, and the fraction of creche teachers with higher education increased by 94 percent or more in all regions. Regional differences in teacher education also became less pronounced over this period. While the top region had a pre-school teacher higher education rate nine times that of the bo om region in 2001, it had just over two times the rate of the lowest-performing region in 2009. Creche teachers showed a similar pa ern. If surpassing a minimum level of teacher education is impor- tant for improved child outcomes, then these trends bode well across the country. Most comparator countries have minimum education requirements for early child education providers, usually a university degree for pre-school teachers and at least a secondary school degree for creche providers (table 2.3). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 23 Table 2.3: Teacher quali�cations required in Brazil and comparator countries Country Teacher quali�cations Brazil Teachers in creches and pre-school • Completion of secondary school Chile Teachers and directors: • Five-year university degree in pre-school education Professional technicians (for 2–4 and 4–6 age groups): • Two-year post-secondary pre-school technician degree Educator assistants: • No mandatory professional quali�cations Mexico (creche) Formal center teachers: • University degree in child development Parent education promoters (in CONAFE, the program designed for rural and vulnerable areas): • Secondary school completion and CONAFE training Mexico (pre-school) General and indigenous pre-school teachers: • University degree in child development CONAFE community school instructors: • Secondary school completion and CONAFE training New Zealand Educators in teacher-led services: • 3-year Diploma in Teaching for early childhood education Educators in parent-led services: • No mandatory professional quali�cations Sweden Pre-school teachers: • 3.5-year university degree in teaching Child assistants: • Upper secondary professional diploma Family day care workers: • No mandatory professional quali�cations Source: Brazil: Decree 6.755 (29 January 2009). Other countries: Bruns et al. (2010). The National Network for Early Childhood already calls for the completion of three ambitious goals over the next decade: (a) in ten years, all directors of ECE institutions will have higher education, (b) in six years, all teachers will have ECD training at the higher education level, and (c) in ten years, all teachers will have training in inclusive education and in Brazilian sign language (for the deaf). The goal that each ECD pro- vider has substantial training in ECD speci�cally (beyond short distance courses) is an excellent one. However, achieving it in the next six years is likely to be a true challenge as many ECD caregivers currently lack that training; to date, ECD-speci�c training has been offered at only a very limited level in Brazil. Key areas to improve in quality Improving early child education quality means improving both program structure and program activities. This is consistent with �ndings from other countries: A major study of the quality of pre-schools in the United Kingdom found activities to be one of the weakest areas despite strong interpersonal interactions, echoing the Brazil study (Sylva et al. 2004). A study of creches in the United States found weaknesses in the same kinds of activities and in children’s opportunities to interact with books, as did a study on Brazil (Cryer and Burchinal 1997). Yet the �nding that interactions are strong in Brazilian pre- schools and creches suggests that educators want to provide quality services, and are sim- ply lacking the knowledge and resources to provide high-quality structure and activities. 24 A World Bank Study Monitoring and supervision by the government could directly help centers improve their program structure and activities. An effective system of knowledge-sharing and the provision of information about best practice would equip centers with the tools nec- essary to invest in the right types of quality. The role of parents in monitoring quality Unfortunately, parents have trouble objectively judging the quality of early child educa- tion centers, thus highlighting the importance of the government’s role as overseer. A 2001 study of creches in Rio de Janeiro asked parents about their perceptions of their children’s creches and showed that parents thought almost all creches were of high qual- ity (�gure 2.9), completely independent of the actual quality, which was measured sepa- rately using an objective rating tool (Barros et al. 2011a). This con�rms similar �ndings in studies of parent perceptions of pre-schools in Germany and in the United States (Cryer, Tie e and Wessels 2002), as well as a �nding that parents vastly overestimate the amount of regulation the government provides for child care centers in the United States (NACCRRA 2009). Parents’ perceptions of child care center quality tend to be particularly misguided when it comes to elements that are not easily monitored, such as activities and supervision, as opposed to elements that are more easily observed such as the physical space (Cryer and Burchinal 1997). Because parents do not have sufficient in- formation to evaluate centers well on exactly the areas—such as activities and program structure—where Brazilian creches and pre-schools need major improvements, there is an argument for federal and municipal governments to step in to assist in measuring and evaluating quality. Figure 2.9: Relationship between observed creche quality and parents’ subjective measure of quality, Rio de Janeiro, 2001 1.00 0.90 Parents’ subjective measure of quality y = 0.08x + 0.88 0.80 R² = 0.03 0.70 0.60 0.50 0.40 0.30 0.20 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Objective measure of daycare quality Source: Barros et al. (2011a). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 25 Curricular and Program Structure Improvement Two of the most widely-touted high-return early child education programs are the Perry Pre-school and the Abecedarian programs, and each of those programs had a high-quali- ty curriculum that guided the activities of caregivers and children. A randomized trial in the United States compared three curricular models for three and four-year-olds; the re- sults showed far be er long-term life outcomes for children with curricula in which chil- dren initiated activities and teachers then responded or in which teachers and children planned activities together, as compared to a more scripted curriculum (Schweinhart and Weikart 1998). This is very different from most primary school curricula, so it is not surprising if a standard education training program would be insufficient to effectively prepare ECE providers. In Brazil, the Ministry of Education published a three volume curricular guide in 1988 which has rich information on how to design activities for early child education centers. Further, the National Education Council published a set of guiding directives for ECE curricular proposals in 2009.10 A valuable next step would be to provide practi- cal workbooks or guides with lesson plans containing speci�c activities in pre-reading, pre-mathematics, and other areas, to help educators not only understand what activities are effective, but also how to implement them. In order to shift to more child-directed activities, ECE educators will need improved, speci�c training in this kind of open cur- riculum, as well as sufficiently small groups of children such that each child can initiate activities. Municipalities can of course take the initiative to supplement the Ministry resourc- es. The Municipality of Rio de Janeiro has developed its own curricular guidelines with speci�c suggestions of activities to develop key skills. The Rio guidelines also provide guidance on program structure, giving priority to cognitive activities in the mornings, followed by reading activities in the early afternoon, and fresh air activities later. The guides suggest how a day could be structured, alternating these activities with free play (Barros et al. 2010). The next step will be shifting to a more child-directed program struc- ture. Greater communication between municipalities might help them to leverage one another’s experiences with effective curricular guides and program structures. Monitoring Program Quality Many countries have implemented programs to strengthen monitoring of quality for early child development (see box 2.1). The federal government in the United States has proposed new legislation to evaluate each of the nation’s Head Start programs over the next three years, examining program performance using an established classroom obser- vation rating instrument, as well as �scal integrity and licensing standards. Each year, the bo om 25 percent of Head Start programs will compete with private centers for government funding (Haskins and Barne 2010, US Department of Health & Human Services 2010). This competition will create incentives for program directors to perform and will reduce the persistence of ineffective programs. Likewise in Sweden, all child care centers must be registered and undergo annual reviews, including supervision, in order to continue functioning (OECD 2006). 26 A World Bank Study Box 2.1: Impact evaluation for early child development Much of the global knowledge on the importance of early child education stems from a few rigorous impact evaluations of child development programs. The High/Scope Perry Pre- school Study and the Abecedarian program both took advantage of the fact that they were not universal programs—in fact, they were both small pilots—to randomly assign the program among those children most in need, to allow comparisons over time between those who par- ticipated and those who did not. Evidence of comparable quality in Latin America is very limited. Most of Brazil’s evidence on Early Child Education comes from comparisons in later schooling outcomes between chil- dren who attended creches or pre-schools and those who did not, which yields some insight but also leaves open the possibility that other factors may be affecting these outcomes. For example, parents who value schooling may be more likely to enroll a child in pre-school and their child may also have strong long-term school outcomes, but the latter could be due to the parent spending time helping the child learn, something not captured in the comparison of pre-schoolers and non-pre-schoolers. The Municipality of Rio de Janeiro has—like the Perry Pre-school program—taken advan- tage of non-universal coverage to use a lottery to assign scarce slots among the neediest of children. This allows for comparisons among comparable children who received public creche care and those who did not. The study began in 2007. Early results show signi�cant impacts on mothers’ labor force participation, and results on children’s cognitive development will be collected in 2011. This study and more like it in the future will strengthen the evidence base for the best ways to deliver quality early child education in Brazil. The United States is an interesting comparator for Brazil given the relatively similar size and level of decentralization in the system. At the most basic level of ensuring qual- ity, almost all states in the United States require child care centers and pre-schools (even completely private ones) to be licensed. Licensing usually includes a renewal either once a year or every other year, and most states will shut down child care centers or pre- schools that do not satisfy licensing requirements. However, while licensing generally covers a broad range of areas, including staff age, facilities quality, and training guide- lines, meeting these standards is a necessary but far from sufficient condition to achieve quality early child education (NARA 2006). Likewise, 26 states or localities have implemented Quality Rating Systems (QRS) for child care and ECE centers. These systems rate centers on a wide variety of qual- ity standards, generally including the following: child-staff ratios and group size, staff training and education, and an assessment of the classroom and learning environment (Zellman and Perlman 2008). The full range of quality standards and how ratings tend to be assigned is shown in Appendix D. Most of these ratings systems then provide conse- quences: inducements for centers to improve, support for centers that are weak, and clo- sure for the direst cases. These programs are voluntary in most cases, yet many centers participate because of the variety of incentives provided. State and local governments also hope that, as more centers participate, parents will use the ratings to choose centers. Thus more centers will have an incentive to join, since a non-rating will be interpreted as a poor rating. In addition, two-thirds of these QRS have built in an evaluation of the rating system itself. These evaluations show systematic improvements in the quality of Early Child Education: Making Programs Work for Brazil’s Most Important Generation 27 Early Child Education centers, as measured by observational instruments, after partici- pating in a QRS (Tout et al. 2010).11 Environmental rating scales Almost all of the QRSs in the United States have employed some sort of observational measure (Tout et al. 2010). Most have used the ITERS-R for childcare centers and the ECERS-R for pre-schools. The instrument currently proposed by the United States De- partment of Health and Human Services for use to evaluate Head Start programs is the Classroom Assessment Scoring System (CLASS), which is designed for pre-school and the �rst few years of elementary school (La Paro, Pianta and Stuhlman 2004).12 Versions of several of these instruments have been implemented in Brazil. In 2009, Fundação Carlos Chagas adapted the ECERS-R and ITERS-R for Brazil and applied them in creches and pre-schools in six capital municipalities, with the results earlier in this chapter (Fundação Carlos Chagas 2010).13 In addition, IPEA in 2001 applied an instrument which evaluates similar areas to the ECERS-R in a study of 100 pre-schools in Rio de Janeiro, as presented at the beginning of this chapter (Barros et al. 2011a). The Municipality of Rio de Janeiro is currently designing a quality monitoring system that will use the Brazilian adaptations of ITERS-R and ECERS-R across its creches and pre-schools, with a similar purpose as the QRS programs in the United States (Barros et al. 2010). Child development measurements Testing children’s cognitive and socio-emotional development can play an important role in gauging the overall health of the ECE system as well as identifying children with special needs. Of course, test results at any age must be safeguarded to prevent abuse. As the consequences of an assessment increase in importance, so does the importance of using high-quality instruments implemented by well-trained examiners. For example, a simple child monitoring test can have implications for how a child’s family views the potential of that child. Other tests, used to screen for whether a child requires special services, can directly affect a long-term education trajectory. It is therefore essential to ensure that tests are implemented appropriately and that the results are used with care (Snow and Van Hemel 2008). This is in line with the National Education Council’s 2009 judgment proposing that ECE evaluations not be used to hold back individual students.14 If the goal is to gauge the quality of the system, then testing a sample of students in the system can be sufficient and cost-effective. The test results can be delinked from child-speci�c identi�ers to protect children’s well-being. At the same time, many QRSs in the United States require child care centers to provide screenings to ensure that chil- dren are developing appropriately and to see if they need any additional services. In no case are these assessments linked to the center’s incentives: They are purely intended to raise the quality of services for the child. Most programs that require assessments, how- ever, must provide evidence that information from the assessment is used for individual planning for the children, and must share the information with parents, as in the New Mexico (United States) program (Look for the Stars 2005). A variety of instruments are in use in U.S. QRSs,15 but the recommended instru- ments tend to have two characteristics in common: (1) Most of them are designed as screening tools, rather than broad measures of child development, and (2) all of them rely on ratings and reports from parents or child care providers, rather than direct as- sessment with children. One that is frequently used is the Ages and Stages question- 28 A World Bank Study naire, which has recently been translated and validated for use in the municipality of Rio de Janeiro (Carvalho et al. 2011). This testing can guide municipalities to improve. Researchers in Rio de Janeiro re- cently applied the Ages and Stages questionnaire to all present students in 500 creches, then compared cognitive and motor development for students in the �rst year to that of students in the last year. This provided an estimate of the value-added of a creche.16 The municipality will now visit the creches with highest value added and seek to learn what they are doing right, to share that with the rest of the network (Barros et al. 2011c). More systematic knowledge sharing within and across early child education systems is a criti- cally important way in which Brazil could identify areas for improvement. Interventions to support improvements While most of the QRSs in the United States are voluntary, one of the reasons to partici- pate is that all participating creches and pre-schools receive access to training, most of- ten in the following areas: (1) assessment of the environment, (2) language and literacy, (3) speci�c curriculum training, (4) business practices, (5) safety, (6) social and emotional development, and (7) child assessments (Tout et al. 2010). A number of programs, both in the United States and abroad, offer improvement grants targeted to some element of pre-school quality. Some aim to improve the quality of teachers with general grants to early child educators to earn higher degrees, in Jamai- ca, Australia (Victoria) and the United States (New York City). Given the tenuous link be- tween general teacher quali�cations and student performance, this approach may have limited effectiveness unless it is restricted to high quality trainings. Another approach, seeking to link teacher training more closely with desired outcomes, is the Maine Roads to Quality Registry (United States), in which early child educators join a program that provides career counseling, eligibility for scholarships, and a 180-hour training program closely-aligned with the state’s identi�ed learning goals (Howes et al. 2008). Improving Quality through Improved Incentives At the level of primary education, education systems around the world are exploring various ways to reward performance by monitoring student outcomes and rewarding teachers and schools that demonstrate improvements. There has been promising evi- dence on these programs, showing that incentives for teachers and schools have major impacts—in India, for example (Muralidharan and Sundararaman 2009)—but also evi- dence to the contrary, showing no effects—in a recent program in the United States (Springer et al. 2010). More than �fteen states and municipalities across Brazil have implemented pay-for- performance programs, linking student improvements to monetary bonuses for teach- ers. At the level of early child education, how can one encourage teachers to focus on activities that will be most helpful to their young students in preparing them for the next phase of their education? A number of programs in the United States have established incentive programs for pre-school and child care facilities. These programs are not linked to direct child devel- opment measurements, but rather to systems that monitor quality through observation. The awards can be linked to a particular quality improvement plan, or in some cases the centers are free to use them as they wish (Tout et al. 2010). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 29 Improving Quality through Improved In-Service Training and Supervision As in later years of education, the provider or teacher is likely the most important factor in determining the success or failure of a child’s early education experience. Two essen- tial components of ensuring high quality care and teaching are the training and supervi- sion of early child educators. Many municipalities in Brazil are seeking to improve in these areas. Box 2.2 includes two examples. Box 2.2: Training and supervision in two municipalities The municipality of Araraquara (in São Paulo) is improving the continuing education of its early child educators, and São Bernardo do Campo (also in São Paulo) is strengthening its program of school supervision. Araraquara has experimented with both internal training (of early child educators by the school director) and external training (by a municipal team). While internal training has shown certain strengths in terms of the directors’ familiarity with the school and the teachers, encouraging directors to take responsibility for this has proved challenging (Bom�m, Ferreira, and Oliveira 2011). In São Bernardo do Campo, the program has struggled with de�ning the precise role of the supervisor to maximize effectiveness. This process of experimentation and revision is reflective of what many municipalities across Brazil experience. In both cases, the federal government also plays a role in developing materials to guide these processes, but much is left to the municipalities ( Ferreira, Oliveira, and Bom�m 2011). As discussed earlier, simple levels of training do not predict success in early child education. Teachers need training, but in particular, they need training in hands-on ac- tivities and interactions to foster child cognitive development. A study of the highly ef- fective Perry Pre-school program in the United States indicates that key elements of the program’s success were (1) systematic in-service training in early child development and education (general education training was found to be unhelpful as the form of peda- gogy is distinctive in early childhood), and (2) ongoing supervision by supervisors who are experts in the curriculum (Schweinhart, Barnes, and Weikhart 2005). Brazil currently has several in-service training programs for early child educators. A federal program entitled Proinfantil provides distance training to teachers in creches and pre-schools; its goal is to train 23,000 educators by July 2011. The Proinfantil cur- riculum has six key areas: four content areas—(1) language, (2) identity, society, and culture, (3) mathematics and logic, (4) life and nature—and two methodological areas— (5) foundations of education and (6) organization of teaching. This program is a start to supplying ECE providers with the building blocks they need, but teachers still need more hands-on training. Nonpro�t initiatives have sprouted to �ll the demand, such as the Education Tables (Mesas Educadores) from the Millennium Fund for Early Childhood, which are resource centers providing training for in-service early child development providers. Likewise, �ve nonpro�ts have joined to support Formar em Rede, which pro- vides distance learning to early child educators in close to �fty municipalities around Brazil. More general in-service programs with strong hands-on training in preparing high-quality activities to improve cognitive development are lacking. Improving Quality through Knowledge Sharing Because early child education programs are the responsibility of the municipal educa- tion secretaries, and because Brazil has over 5,500 municipalities, the variation in pro- 30 A World Bank Study grams is extensive. State governments have no formal role in early child education, al- though occasionally they will support municipalities, as in Acre’s Asas da Florestania Infantil program and in Rio Grande do Sul’s Primeira Infância Melhor program, which is health-based but includes child cognitive stimulation. The federal government also has programs to support municipal efforts. Finally, the non-government sector is actively involved in early child education. Federal role Despite federal funding channeled to municipalities for recurrent costs of early child education through FUNDEB, most municipalities have insufficient early child educa- tion centers to meet demand. The federal program Pró-Infância provides funding to municipalities to address this insufficiency; it has so far funded over 2,000 centers, and provided resources to equip several hundred more. A second major federal program, Proinfantil, that provides training, was discussed in the previous section. The federal government consistently sponsors and publishes guidance for municipalities on a range of topics, including how to develop contracts with nonpro�t ECE providers, indicators of quality in ECE, curricular references, and others.17 Creches and pre-schools are also included in some more general programs, such as the National School Library Program, which selects and provides books for educational institutions of all levels (MEC 2009). Municipal programs have the opportunity to inform others Many municipalities are investing in their early child education programs, developing specialized curricula, monitoring systems, improved training for caregivers, and more. With over 5,000 municipalities across Brazil, there is great opportunity for these munici- palities to learn from each other. Some public programs target speci�c areas, and their curricula may be of interest to other municipalities. For example, the Municipality of Santarém (State of Pará) has developed the program Eco-Schools (Santarém 2010), which provides lessons about the environment beginning in ECE. The Municipality of Campinas (State of São Paulo) has developed Prodança-Criança Escola to teach children dance in order to provide a range of development opportunities. Likewise, the Municipality of Rio de Janeiro has devel- oped a curriculum of 40 lessons to provide training to parents of creche children. These span education, health, and social assistance (see the curriculum outline in Appendix F). The curriculum includes DVDs, workbooks, and discussion guides. Likewise, the pro- gram of home-based education visits developed by the state of Acre (described further in the next chapter) includes workbooks and educational agent guides which demon- strate (both to the agents and then to the parents) how to use common objects available near the home as tools of cognitive stimulation. Clearly, these curricula and others like them represent a potential educational resource for other municipalities. In addition to these curricula, many municipalities have invested in additional inno- vative programs from which others can learn and adapt. For example, the municipality of Macapá (in Amapá), introduced the program Revivendo as Tradições, whereby local cultural traditions are woven into play and learning activities. Rio de Janeiro is adapting Bookstart, a U.K. program to expose children ages 0–3 to positive reading experiences by providing reading materials and encouragement to parents. Salvador (in Bahia) es- tablished Pequeno Cidadão, in which pre-schoolers go to a public office to receive their own identity cards, and are guided through all the necessary steps. They learn the na- Early Child Education: Making Programs Work for Brazil’s Most Important Generation 31 tional anthem, hear stories about their rights as citizens, and are invited to share their own relevant experiences. In Lucas do Rio Verde (in Mato Grosso), creche participants take walks to a variety of locations: the supermarket, where they learn about fruits and vegetables, or a local artisan’s shop, where they learn about regional art and painting. In Belém (Pará), pre-school children learn math by collecting and counting different kinds of used telephone calling cards, going on to practice reading and learning about the Pará landmarks depicted on the cards.18 Access to these types of programs, and evidence on which are most successful, will help municipalities to build on the experiences of others. How can municipalities share information? Municipalities currently have few tools that they can use to learn from each other’s pro- grams. Several networks for early child development exist, including the National Net- work for Early Childhood. However, this network focuses less on program-level knowl- edge sharing and more on improving national policy. The National Union of Municipal Education Leaders (UNDIME) likewise disseminates some documents on national early education policy. Recently, two initiatives for more direct information sharing have been launched. One is the Network for Cooperation and Peace for Children, which is a social networking site (a là Facebook) where policy makers, nonpro�t workers, educators, and other inter- ested citizens actively exchange information. The site currently has nearly 500 members. A second initiative launched by the National Forum for Early Child Development and the World Bank, called the Network for Cooperation in Early Child Development, seeks to provide a map of early child development services across Brazil. It also provides a social networking site focused on sharing exactly the kinds of materials described above: information on innovative programs, experiences, and best practices. The federal government can also play a role in helping municipalities learn from one another. An initiative that the federal government formerly sponsored is an annual Prize for Quality in Early Child Education, which ran from 2000 through 2005 and highlighted innovative projects implemented by individual municipalities or ECE centers. Lessons There are a number of ways in which Brazil can contribute to ECE quality improvements in key areas. First, the federal government can develop more speci�c curricular guides that provide local creches and pre-schools with outlines for activities. Second, Brazil can introduce systematic monitoring by providing access to a standard observational tool. Third, the federal government can set licensing guidelines for minimum standards in quality for municipalities to adapt. This is a step beyond the existing guidelines for the quality of creches: it entails both quality minimums and guidelines for the implementa- tion of a licensing program. Finally, Brazil can encourage more systematic monitoring within municipalities and more fluid knowledge sharing across municipalities. Notes 1. The only exception is that creches in the Central West saw a one percentage point decline in access to the public sewer network. All other regions improved in every infrastructure category. 2. This is also the region with the farthest to go in terms of achieving universal pre-school access. 3. It is also possible that a more representative sample of pre-schools are registered and thus in- cluded in the Census of Schools, whereas only creches with relatively good infrastructure are reg- 32 A World Bank Study istered. This might be the case since pre-school is a higher and more prevalent level of education, and may be more likely to take place in schools that offer primary education (and thus are more closely regulated and monitored). 4. The Census of Schools does not indicate how many rooms are dedicated to each level of educa- tion. However, it does indicate how many total rooms and how many total students there are in schools that offer creche-level education, and in schools that offer pre-school-level education. As a result, our indicator of students per room in creches (pre-schools) is total students per room in schools that offer creche- (pre-school-) level education. 5. The Germany and Portugal study uses the ECERS, whereas the United States study uses the ECERS-R (of which the Brazilian study is an adaptation). Comparability of the two instruments has been demonstrated (Sakai et al. 2003). 6. Again, the scales are not exactly the same, so these comparisons should be taken with caution. 7. Espaços de Desenvolvimento Infantil, commonly referred to as EDIs 8. It cannot be ruled out that this is because EDIs are newer than other institutions. 9. The 27 analyses are the following: 4 outcomes x 7 studies = 28 analyses, minus one study that did not carry out the reading test. 10. National Education Council Judgment 20/2009 (Parecer CNE/CEB No 20/2009). 11. Only one study has examined the relationship between QRS participation and changes in child outcomes. A study of the United States (Colorado) program showed no relationship, although one- third of creches had dropped out by the end of the study (Zellman et al. 2008). 12. The CLASS instrument examines teacher-student interactions in pre-schools in three domains: emotional support, classroom organization, and instructional support. Each domain is broken down into dimensions such as positive climate (to promote children’s enjoyment of the classroom environment) and concept development (to promote higher-order thinking and cognition) (Teach- stone 2011). 13. In Portuguese, the scales are entitled Escala de Avaliação de Ambientes para Bebês e Crianças Peque- nas (ESAC) and Escala de Avaliação de Ambientes da Educação Infantil (Barros et al. 2010). 14. Parecer CNE/CEB No 20/2009. 15. Appendix E provides details on selected instruments. 16. This assumes the absence of cohort effects, in which the average cognitive development of children in Rio born—for example—in 2008 is signi�cantly lower than that of those born in 2010. 17. A list of recent publications is provided in Appendix G. 18. Several of these examples are drawn from Prêmio Qualidade na Educação Infantil 2004. CHAPTER 3 How to Reach the Very Poorest Children Access to Early Child Education around the World B razil has increased creche enrollment extensively over the past �fteen years. With 18 percent of children ages 0–3 covered, it now exceeds the level in Mexico (for 0–3 years old) and has a comparable level to that of Chile (�gure 3.1). However, creche en- rollment still lags signi�cantly behind countries with more established programs, such as New Zealand, Sweden, and the OECD as a whole. These gaps reflect policy differ- ences. Chile and Mexico have targeted policies for child care at the earliest ages. Sweden provides full-time care and education in pre-schools from age one (Bruns et al. 2010). An increasing number of countries are seeking universal pre-school coverage for children ages four and �ve. Brazil—with 75 percent enrollment for 4–5 year olds—lags behind the OECD average of 85 percent, but has enrollment comparable to that found in Chile (�gure 3.2). Among Latin American countries, Brazil and Mexico are the only two that have made pre-school mandatory for children ages four and �ve. Mexico’s ex- perience is especially relevant to Brazil since it, like Brazil, seeks to achieve universal pre-school coverage by 2016. Mexico has achieved almost 100 percent coverage for �ve- year-olds, and well over 90 percent for four-year-olds. But one result of this rapid expan- sion in Mexico has been a drop in enrollment for three-year-olds in 2005, as efforts and resources were focused on four- and �ve-year-olds. The resources provided by Brazil’s education fund (FUNDEB) for both creches and pre-schools should help to mitigate that Figure 3.1: Creche enrollment around the world, age 0–3 (or as speci�ed) 50 40 Percent enrolled 30 20 10 0 Brazil Chile Chile Mexico United Sweden OECD (0–2) (2–4) States (3) average Region Source: Brazil data from PNAD 2009. U.S. data from Aud et al. (2010). Other data adapted from (Bruns et al. 2010). 33 34 A World Bank Study Figure 3.2: Pre-school enrollment around the world 100 90 Brazil Percent enrolled 80 Chile Mexico 70 United States 60 Sweden 50 OECD average 40 Age 4 Age 5 Source: Brazil data from PNAD 2009. Chile data calculated from Mineduc 2010. Data from U.S. Census Bureau, Current Population Survey, October 2009. Other data adapted from Bruns et al. (2010). effect, although there is evidence that some of the states with the strongest pre-school coverage also have weak creche coverage. A second result in Mexico is that the expan- sion led to reductions in some quality measures: the proportion of pre-schools with a student-adult ratio exceeding 30 rose signi�cantly in the course of the expansion (Yoshi- kawa et al. 2007). As Brazil seeks to expand enrollment in coming years, it will need to carefully monitor quality to ensure that access does not leave children in low-quality care. How Has Access to Early Child Education Evolved in Brazil? A endance at creches and pre-schools has increased dramatically in recent years. Only 9 percent of under-four children a ended a creche in 1999, but 18 percent a ended a creche in 2009—a doubling of the creche enrollment rate over a single decade (�gure 3.3).1 Similarly, while 52 percent of children ages 4–6 a ended school in 1999, 81 percent a ended school in 2009 (�gure 3.4).2 This rapid expansion is comparable to expansions in other countries over the same period. While pre-school enrollment expanded by 56 percent in Brazil from 1996 to 2009, pre-school in Colombia expanded by 40 percent, and pre-school in Mexico expanded by 61 percent. It is immediately clear that pre-school a endance is far more common than is creche a endance. However, while the pre-school enrollment rate was more than six times the enrollment rate in creches in 1996, it was only 4.4 times the creche enrollment rate in 2009. Thus, while pre-school is the more common type of ECE, creches are gaining ground. Al- though all regions have increased their creche and pre-school enrollment in recent years, there are stark differences across regions. The South and the Southeast had the highest initial creche enrollment rates in 1996 and went on to experience the greatest percent- age point increases in enrollment through 2009. The Southeast additionally experienced the largest percentage increase in creche enrollment of any region. This has increased the regional gap in creche a endance as the regions that started out with lower enrollment are expanding enrollment more slowly than are the South and the Southeast. In contrast, Early Child Education: Making Programs Work for Brazil’s Most Important Generation 35 Figure 3.3: Fraction of 0–3 children in creche, 1996–2009 0.24 0.20 Fraction enrolled 0.16 0.12 0.08 0.04 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Brazil South North Southeast Northeast Central West Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Figure 3.4: Fraction of 4–6 children in pre-school, 1996–2009 0.9 0.8 Fraction enrolled 0.7 0.6 0.5 0.4 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Brazil South North Southeast Northeast Central West Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Note: Before 2007, this is the pre-school enrollment rate for children aged 4–6. Beginning in 2007, this is the enrollment rate in pre-school and/or the 1st year of primary school for children aged 4–6. the regional gap in pre-school a endance has decreased slightly since 1996. The Central West and the South—the regions with the lowest pre-school a endance rates in 1996— saw the greatest percentage increase in a endance during 1996–2009. While regional gaps persist, there is convergence. 36 A World Bank Study Regional trends in ECE expansion do not merely track income. The Northeast—one of Brazil’s poorest regions—has had higher creche and pre-school enrollment rates than the relatively rich Central West during nearly all of 1996–2009, and has enjoyed some of the highest pre-school enrollment rates of any region. Municipality-speci�c priorities and levels of demand seem to be important factors driving these investments. Thus, beyond the need for resources, there must be a will to invest as well. At the state level, there are also stark disparities in access to creches and pre-schools. Figure 3.5 offers a cross-state comparison of access to creches in 2009, and �gure 3.6 shows the same for pre-schools. A number of states stand out when compared to others in the same region. For example, Roraima in the North has a creche enrollment rate of 14 percent, which is higher than the rest of the region, and a pre-school enrollment rate of 81 percent, which is close to that of the Federal District. On the other hand, Alagoas, in the Northeast, lags behind the rest of the region in both creche and pre-school enrollment. Figure 3.5: Creche access by state, Figure 3.6: Pre-school access by state, 2009 2009 Creche (age 0–3) Pre-school (age 4–5) Source: Data from PNAD (2009). Note: Creche access is the enrollment rate in schools for children aged 0–3. Pre-school access is the enroll- ment rate in schools for children aged 4–5. One of the most important state-level �ndings is that six states have pre-school en- rollment rates of under 60 percent: Acre, Amapá, Amazonas, Goiás, Rio Grande do Sul, and Rondônia. That means that, in those states, massive expansions will be necessary in the coming years to achieve universal coverage by 2016. State-level growth rates of creche and pre-school enrollments between 1996 and 2009 are highly correlated. Generally, states that have boosted creche enrollments have achieved the same with pre-school enrollments, and those that are lagging in one area lag in the other as well. Yet it is also likely that some regions with limited ECE bud- gets are forced to trade off investment in one level of ECE for investment in another, as Mexico did in its recent, rapid pre-school expansion. In Brazil, the South is an interesting example; while it was the region with the lowest pre-school enrollment rate for several Early Child Education: Making Programs Work for Brazil’s Most Important Generation 37 years during 1996–2009, it also enjoyed the highest creche enrollment rate for most of that same period. (Pre-school enrollment is still much higher than creche enrollment.) As both levels of ECE are controlled by the same level of government (the municipality), this suggests that the demand for pre-schools compared to creches varies across regions. These �ndings underscore the need to explore how income, urbanization, and fe- male labor force participation have influenced ECE enrollments. These are factors on which regions and states differ markedly, and which likely account for variance in the rate of expansion of ECE. Gender and ECE There is li le evidence of unequal access to ECE for boys versus girls. In general, differ- ences in a endance by gender appear to be temporary and non-systematic (�gure 3.7 and �gure 3.8). For both creches and pre-schools, none of Brazil’s regions has consis- tently had higher a endance by boys than by girls, or vice versa, during the last decade. This is consistent with broader education a endance pa erns in Brazil, which show li le evidence of gender disparities. Poor children are being left behind While children of all income quintiles a end creches and pre-schools, access is most highly-concentrated among the wealthiest (�gure 3.9). This may be true for several rea- sons. First, in areas with a high concentration of high-income families, the tax base from which public education is funded will be relatively larger. This is especially true in Bra- zil. Municipalities legally must spend at least 25 percent of their revenue on education. Higher revenue automatically translates into higher education spending. Even though the 1998 FUNDEF—and the subsequent 2007 FUNDEB—reform partially equalized ed- ucation spending, it did so within states, which meant that municipalities in wealthier states still enjoyed more education resources than did those in poorer states. The federal Figure 3.7: Fraction of 0–3 children in creche by gender (1996–2009) 0.18 0.16 Fraction enrolled 0.14 0.12 0.10 0.08 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Boys Girls Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. 38 A World Bank Study Figure 3.8: Fraction of 4–6 children in pre-school by gender (1996–2009) 0.8 Fraction enrolled 0.7 0.6 0.5 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Boys Girls Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Note: Before 2007, 6 year olds are pre-school-aged. Beginning in 2007, 6 year olds are in �rst grade. Figure 3.9: ECE institution attendance by income (1996–2009) 0–3 Children in creche 4–6 Children in pre-schools 1.0 0.34 Fraction enrolled Fraction enrolled 0.26 0.8 0.18 0.6 0.10 0.4 0.02 1996 2001 2006 1996 2001 2006 Year Year 1st Quintile 2nd Quintile 1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 3rd Quintile 4th Quintile 5th Quintile 5th Quintile Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Note: Before 2007, the pre-school a endance rate is the fraction of 4–6 year olds in pre-school. Beginning in 2007, the pre-school a endance rate is the fraction of 4–6 year olds in institutions intended for 4–6 year olds. government topped off the education funds in states that did not meet a per-student minimum, but this only bene� ed some states, and did not close the education �nancing gap between states. Second, families’ ability to pay rises with family income. High in- come families can afford private early child education where public education is absent or of low quality. Third, because parental income and parental education are positively Early Child Education: Making Programs Work for Brazil’s Most Important Generation 39 correlated, children from higher-income families are also more likely to have parents who place a high intrinsic value on education. Such parents may be more likely to enroll their children in early child education. Finally, children from higher-income families are less likely to need to work to help support their families. While child labor among children ages 0–5 is uncommon in Brazil, it does occur, and could prevent enrollment. Figure 3.9 reveals not only the strength of the relationship between family income and ECE institution a endance, but also how the situation varies across creches and pre-schools, and how it has changed over time. The income gap in age 4–6 a endance at school appears to be narrowing while the income gap in age 0–3 a endance is expand- ing. This is likely due to the fact that, beginning in 2007, six-year-olds were gradually brought into primary school, which is a compulsory level of education. When all chil- dren—rich and poor—must be provided a space in a public school, family income natu- rally has less of an impact on school a endance. Given a 2009 constitutional amendment to initiate mandatory schooling at age four, the income gap in pre-school a endance is likely to fall even further in the coming years. The case of primary and lower-secondary education—which have been mandatory for decades—suggests such an outcome. As of 2009, 99 percent of the wealthiest 6–14 year olds a ended school, but the poorest chil- dren were not far behind, with an a endance rate of 96 percent. Wealthier families are not only more likely to enroll their children in ECE; they are also much more likely to use the private system (�gure 3.10 and �gure 3.11). At the bot- tom quintile of per capita household income, 85 percent of parents use public creches and 92 percent use public pre-schools. In contrast, at the top quintile, 81 percent use pri- vate creches and 74 percent use private pre-schools. Private ECE not only costs money, but—as shown later—it tends to be of higher quality on several dimensions. If public ECE is not available, or if it is of low quality, a poor family may have no other option than to keep a child at home. Figure 3.10: Fraction of children enrolled in creche in public vs. private institutions by income quintile (2009) 0.85 0.81 0.8 0.76 0.62 Fraction enrolled 0.6 0.53 0.47 0.4 0.38 0.24 0.2 0.19 0.15 0 Bottom Second Third Fourth Top Quintile Public Private Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. 40 A World Bank Study Figure 3.11: Fraction of children attending pre-school in public vs. private institutions by income quintile (2009) 1.0 0.92 0.82 0.8 0.74 0.72 Fraction enrolled 0.59 0.6 0.41 0.4 0.28 0.26 0.2 0.18 0.08 0 Bottom Second Third Fourth Top Quintile Public Private Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Rural children are being left behind Creche and pre-school enrollment are also less common in rural areas. Urban households have higher incomes (more than double as of 2009), which likely explains part of the gap. As described above, this has an effect on the education budget in the municipality. It also has an effect on parents’ ability to self-�nance early child education. Rural areas may also have longer or more complicated commutes to centers or schools, which can be dif- �cult for young children. Gaps in access between urban and rural areas provide unequal opportunities to children based on their geographic location. Vega and Barros (2011) demonstrate that living in a rural area is one of the main sources of exclusion in access- ing housing services and amenities like electricity, water, and sanitation. They also show that socioeconomic status is a huge source of exclusion in accessing pre-school and other basic education, housing, and health services, and in progressing successfully through school. Given that rural areas are generally poor, a large share of rural children—for a multitude of reasons—lacks opportunities to succeed. Policies that can creatively expand early child education in rural areas which cannot be served by center-based care or other common modalities can help to overcome these unequal opportunities. The rural-urban pre-school enrollment gap is narrowing (�gure 3.12). However, as �gure 3.13 demonstrates, rural children ages 4–6 in the highest two income quintiles are about as likely to a end school as are urban children the same age in the lowest two in- come quintiles. That is, the wealthiest rural children a end creche at about the same rate as the poorest urban children. Figure 3.13 suggests that the rural-urban gap in pre-school a endance cannot be at- tributed to the relative poverty of rural areas alone. It may also be the case that children in rural areas have access to more reliable forms of informal child care—such as neigh- borhood cooperatives, relatives, and care by trusted friends—than do children in urban areas. Finally, female participation in the labor force is slightly less common in rural Early Child Education: Making Programs Work for Brazil’s Most Important Generation 41 Figure 3.12: Fraction of 4–6 children in pre-school by rural/urban, 1996–2009 0.8 0.7 Fraction enrolled 0.6 0.5 0.4 0.3 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Urban areas Rural areas Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Note: Before 2007, 6 year olds are pre-school-aged. Beginning in 2007, 6 year olds are in �rst grade. Figure 3.13: Fraction of 4–6 children in pre-school by income and rural/urban, 1996–2009 0.75 Fraction enrolled 0.65 0.55 0.45 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Bottom 2 quintiles, urban areas Top 2 quintiles, rural areas Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009 Note: Before 2007, 6 year olds are pre-school-aged. Beginning in 2007, 6 year olds are in �rst grade. areas than in urban areas of Brazil. In 2009, 40 percent of mothers of 0–3 children in rural areas worked, whereas 43 percent of mothers of 0–3 children worked in urban areas (in- cluding both formal and informal work). Similarly, in 2009, 48 percent of mothers of 4–6 children in rural areas worked, but 50 percent of mothers of 4–6 children in urban areas did. Although these differences are small, the low levels of creche participation in rural areas mean that even small increases in female labor force participation that translate into higher creche enrollments could result in signi�cant percentage expansions. 42 A World Bank Study Female labor force participation (FLFP) and early child education At the same time that ECE enrollment has been expanding, more and more women have been joining the labor force. Labor force participation in Brazil tops 60 percent, which is higher than the Latin American average, and is comparable to OECD countries (table 3.1). Table 3.1: Labor force participation across countries Country 1996 2008 Growth: 1996–2008 Brazil 52% 60% 16% Colombia 33% 41% 23% Chile 35% 44% 25% Mexico 38% 43% 16% Latin America & Caribbean 44% 52% 16% United States 59% 59% 0% Sweden 60% 61% 2% Source: World Development Indicators, World Bank. The labor force participation (LFP) for mothers of young children in particular has expanded dramatically, faster than the average for Brazil or the region. LFP among mothers of creche-age children has risen from 33 percent in 1998 to 42 percent in 2007, a 27 percent increase. LFP among mothers of pre-school age children has risen from 41 percent to 48 percent, a 17 percent increase (just above the Brazil and regional average of 16 percent). Publicly-provided child care has been shown internationally to increase female la- bor supply, and evidence from Brazil supports this �nding. A recent study in Rio de Janeiro (described in box 2.2) compares mothers of children who were provided spaces in public creches to mothers of children who were not. A lo ery determined entry, so mothers from both groups of children are comparable. Among mothers not working prior to the study, the employment rate eight months later of mothers receiving public care was double that of mothers not receiving care: 17 percent versus 9 percent (Barros et al. 2011b). Thus, a policy of expanding creches provides some of the resources to pay for itself by signi�cantly expanding the number of women in the labor force. Fewer than half of mothers with young children (0–3) in Brazil work, although the numbers are expanding. Higher rates of FLFP among mothers of pre-school-aged children than among mothers of creche-aged children indicate greater preference for at-home care of very young children, as well as higher costs of safely transporting very young children and labor market difficulties that stem from taking time off for childbirth and early infan- cy. Maternity leave in Brazil is four months, but it is currently being raised to six months. This expansion may allow new mothers to be er retain an existing job, which would then increase the need for creche spaces in the years following maternity leave. Even non-working mothers have become signi�cantly more likely to send their chil- dren to creches over the last decade. The fact that creche enrollments are expanding among non-working mothers suggests that age 0–3 ECE is seen to offer bene�ts beyond maternal labor force participation. However, the gap in creche a endance rates between children with working and with non-working mothers has expanded from eight to �f- teen percentage points between 1996 and 2009, as shown in �gure 3.14.3 Creche enroll- Early Child Education: Making Programs Work for Brazil’s Most Important Generation 43 ment is thus increasingly associated with a mother’s working status. And yet, less than one-third of children with working mothers are in a creche. Where are those children? Presumably, they are in the care of fathers, older relatives at home, or neighbors. For example, many households with a grandparent living at home do not send their children to creche (INEP 2009). Working and non-working mothers have made similar gains in enrolling children in pre-school over the last decade. The enrollment gap between the children of working and non-working mothers has actually narrowed slightly (�gure 3.15). This decrease may be in part a ributable to the gradual incorporation of six-year-olds into primary Figure 3.14: Creche enrollment by mother’s work status, 1996–2009 0.30 0.25 Fraction enrolled 0.20 0.15 0.10 0.05 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Mother works Mother does not work Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Figure 3.15: Pre-school enrollment by mother’s work status, 1996–2009 0.9 0.8 Fraction enrolled 0.7 0.6 0.5 0.4 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Mother works Mother does not work Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 1996–2009. Note: Before 2007, 6 year olds are pre-school-aged. Beginning in 2007, 6 year olds are in �rst grade. 44 A World Bank Study education starting in 2007, for which a endance—being required—is less dependent on mothers’ employment status. However, the gap was already narrowing in the decade between 1996 and 2006, perhaps because schooling for six-year-olds (and even for 4–5 year-olds) became more of a norm. Reaching the Poorest: How Many to Plan For Given that pre-school education is now compulsory, and municipalities must come into compliance by 2016, the Ministry of Education’s goal is universal enrollment by 2016 (Weber 2011). As highlighted earlier, Brazil has a signi�cant gap to close in order to achieve that goal. Six states have less than 60 percent of children enrolled in pre-schools, and an additional 13 states have less than 80 percent enrolled. The total number of children not yet enrolled is striking (table 3.2). Yet that number is expected to fall signi�cantly over the next �fteen years—in part due to Brazil’s fertility decline over the past 25 years. Enroll- ments in elementary school (Ensino Fundamental) are also expected to drop 23 percent over the same period, freeing up resources to achieve larger-scale ECE enrollments. Table 3.2: Total number of children not enrolled in ECE, by region Creche-aged children (0–3) not enrolled Pre-school-aged children (4–5) not enrolled Region 2009 2025 2009 2025 North 1,190,368 964,198 247,577 188,159 Northeast 3,101,091 2,511,884 360,118 273,690 South 3,410,352 2,762,385 316,093 240,231 Southeast 1,141,513 924,626 482,143 366,429 Central West 749,702 607,259 164,581 125,082 Brazil 9,593,026 7,770,351 1,570,512 1,193,589 Sources: PNAD (2009) survey data, IBGE (2009) population data, and IBGE (2007) population-by-age estimates. In order to reach this magnitude of children, Brazil will need to incorporate an array of delivery mechanisms. Three-quarters of the 1.6 million 4–5 year olds not in pre-school and much of the creche access gap lie in urban areas (see table 3.3) where construction of new centers or adaptation of buildings into centers may be the most efficient solution. Yet coverage rates are much lower in rural areas. In some rural areas, population densities are such that �lling an average-sized pre-school or creche may require too much travel for very young children to justify center-based education. Table 3.5 shows the average size of pre-schools, in terms of the number of students, by region. In São Paulo state, 0.4 km2 of territory in São Paulo municipality would be sufficient to �ll an average-sized São Paulo pre-school with 4–5 year olds.4 However, in more rural Barra do Turvo municipality, also in São Paulo state, over 320 km2 would be required to �ll an average-sized pre-school. Similarly, in Pernambuco state, 0.15 km2 in Recife municipali- ty would be sufficient to �ll an average-sized Pernambuco pre-school with 4–5 year olds. However, in more rural Parnamirim municipality, the same pre-school would have to serve over 110 km2 of territory. Due to economies of scale, it is inevitably more expensive to expand early child education in more rural municipalities. In those areas, the chal- lenge is to either create cost-effective center-based care on a smaller-than-typical scale, or to use home visits and other delivery modalities that might be more suited to rural areas. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 45 Table 3.3: Number of 4–5 year olds not enrolled in pre-school in 2009, by rural/urban and region Children not enrolled, Children not enrolled, Fraction of children not enrolled Region urban areas rural areas that are in urban areas (%) North 157,152 90,425 63 Northeast 215,662 144,456 60 South 247,635 68,458 78 Southeast 393,718 88,425 82 Central West 135,348 29,233 82 Brazil 1,149,515 420,997 73 Sources: PNAD (2009) survey data, IBGE (2009) population data, and IBGE (2007) population-by-age estimates. The policy goal for creche enrollments is less clear. According to the 1988 consti- tution, creche care is a right.5 Yet not all families want to enroll their young children in center-based care, especially in the earliest years. Further, there is no evidence that universal creche care would be bene�cial. Almost all the international evidence is for pre-school, and the evidence that does exist for earlier ages has clearer positive results for highly vulnerable households. What is the ideal enrollment rate? The public provision target rate for children ages 0–3 for European Union countries, set out in the 2002 Barcelona EU meeting, is 33 percent. The rationale for the target, how- ever, is focused on providing working opportunities for women rather than on child de- velopment (European Council Conclusions 2002). A more recent UNICEF report, seek- ing to establish benchmarks for early childhood services in OECD countries, proposes a target of 25 percent of children ages 0–3, reducing the EU target so that countries can focus on ensuring that all care is licensed and regulated (Benne 2008). In Brazil, the Na- tional Plan for Education, passed into law in 2001, set a goal of 30 percent creche enroll- ment by 2005 and 50 percent enrollment by 2010. The new National Plan for Education extends the deadline for that 50 percent enrollment target to 2020. The collaborative Na- tional Plan for Early Childhood, coordinated by the National Network for Early Child- hood6 and developed in concert with a broad range of government and non-government institutions, calls for 40 percent enrollment of children ages 0–3 by 2016, and 70 percent enrollment by 2022. These are clearly more ambitious than the EU and OECD targets. To plan their growth and investment strategies, municipalities need to know the unmet demand for creches. One strategy to estimate demand, proposed by researchers at IPEA, examines the take-up rate of creche care among the wealthiest segment of the population, as this population faces a weak budgetary constraint to ECE enrollment (Barros et al. 2010). Creche use in Brazil is more than three times as high for the richest �fth of the population as for the poorest �fth (34 percent versus 11 percent, shown in �gure 3.9). In the highly urban municipality of Rio de Janeiro, about 52 percent of the wealthiest �fth of the population and 20 percent of the poorest �fth are enrolled. A en- dance by the wealthiest 5 percent of the population is similar to that of the wealthiest 20 percent. In other words, the wealthiest households in Rio de Janeiro—whether de�ned narrowly or broadly—send their children to creches about half the time (�gure 3.16). Of course, using the wealthiest families to calculate the creche demand of the poor- est has potential problems. On the one hand, poorer families may have less access to substitute services like nannies; on the other hand, they may have more underemployed 46 A World Bank Study Figure 3.16: Creche participation in Rio de Janeiro, by income level 60 55 Brazil State of Rio de Janeiro 50 City of Rio de Janeiro Crèche participation (%) 45 40 35 City of Rio de Janeiro 30 (median) 25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Income distribution (poorest to wealthiest) Source: Pesquisa Nacional por Amostra de Domicilios (PNAD), 2006–08. relatives to provide child care service. Also, female participation in the workforce is higher among the wealthiest (although this may in part be driven by access to childcare). Nonetheless, one can learn much about potential unmet demand for creche services among poorer families by assuming that their overall demand is similar to the expressed demand for these services by wealthier families. Creche enrollment rates of children ages 0–3 in the richest �fth of households vary across Brazil’s regions, from 25 percent in the Central West to 43 percent in the South, as shown in table 3.4. Using the assumption that these numbers represent the share of parents in each of these regions desiring their age 0–3 children to be enrolled in center- based programs would imply demand for creche care of approximately 4 million chil- dren. At this point, about 2.2 million children ages 0–3 are enrolled in creche, so satisfy- ing remaining demand would require a near doubling of spaces. However, unmet need (as calculated using this method) varies across regions. In the Southeast, 34 percent of demand is currently unmet, whereas this is true for 66 percent of demand in the North.7 Table 3.4: Estimated demand for creche, by region Enrollment rate of Total estimated Children children 0–3, richest Total children demand for currently Unmet Region 20% of households 0–3 creche spaces enrolled demand North 24% 1,297,686 312,772 107,318 205,454 Northeast 34% 3,670,508 1,235,395 569,417 665,978 Southeast 35% 4,417,646 1,534,584 1,007,294 527,290 South 43% 1,506,833 643,875 365,320 278,555 Central West 25% 879,147 224,171 129,445 94,726 Source: Censo Escolar (2009) and IBGE Censo (2000). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 47 These differences have implications for any national policy that aims to expand creche spaces where unmet demand is greatest. If universal access to pre-school is the goal, Brazil cannot make similarly-sized investments in all regions. Investments need to target areas with the greatest unmet need. Building New Centers Municipalities are continually constructing new early child education centers, causing a rapid rise in creche and pre-school enrollments throughout the country: from just over half a million public creche enrollments in 2001 to almost �ve million in 2009. The total number of public creches has nearly doubled, from 13,382 to 24,731 over the same period (�gure 3.17). The number of pre-schools also grew over that time period, but compara- bility is complicated by the change in the eligible age for pre-school, from 4–6 previously to 4–5 in 2007. Policy makers across Brazil face the important questions of how many centers to build and where to build them. In Brazil, the average pre-school has 45 students, although the regional average var- ies from 71 in the Southeast to 32 in the Northeast (table 3.5). If additional pre-schools of Figure 3.17: Number of public creches by region, 2001–2009 12,000 Number of public crèches 10,000 8,000 North Northeast 6,000 Southeast 4,000 South Central West 2,000 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Source: PNAD (2009) survey data. Table 3.5: Average enrollment size of pre-schools in 2009, by region Average size of pre-schools Additional pre-schools required to reach Region (enrollment) universal enrollment of 4–5 year olds North 44 5,400 Northeast 32 11,535 South 40 8,098 Southeast 71 6,887 Central West 63 2,563 Brazil 45 34,812 Sources: PNAD (2009) survey data, Census of Schools (2009) school-level data, IBGE (2009) population data, and IBGE (2007) population-by-age estimates. 48 A World Bank Study these average sizes were built in each region to accommodate all un-enrolled 4–5 year- olds, almost 35,000 new pre-schools would be needed. This again points to the need to not only construct additional centers, but also to explore alternative mechanisms for reaching some of Brazil’s children. Targeting At the pre-school level, the goal is universal enrollment by 2016. At the creche level, where universal access (but not necessarily enrollment) is the goal, municipalities must prioritize given that enrollment was only 18 percent nationwide in 2009. Indeed, be er targeting will alleviate some of the immediate demand for expansion. For example, in the municipality of Rio de Janeiro, the need to expand in order to satisfy demand for spaces among just the poorest half of the population is 90 percent under current tar- geting, but that includes many creche spaces going to wealthy children whose families could afford to self-�nance creche care. Meanwhile, if those public spaces were given to the poorest, then Rio would only need to expand by 69 percent in order to satisfy de- mands for the poor—still a tall order, but more plausible. The same problem holds across Brazil. In public creches, 13 percent of scarce spots are taken by children from the richest families (�gure 3.10). If those same spaces were allocated to the poorest 20 percent, then overall enrollment in creche amongst the poor- est would increase by half. This misallocation of scarce spaces has not improved: it has hovered around 20 percent for almost a decade, from 2002. Furthermore, the evidence presented in Chapter One demonstrated that the returns to early child education are particularly high for the poorest and most vulnerable households; these households have the fewest resources to fund child development-promoting alternatives to public creches. Given the importance of targeting, how should a municipality assign its scarce spac- es? The method most municipalities currently use to allocate spaces is completely decen- tralized, as each creche director is responsible for assigning spaces. A second method, less used in Brazil, but more effective from a resource allocation standpoint, is to have a centralized, means-tested selection process at the municipal level. A third method, which the Municipality of Rio de Janeiro adopted for cohorts entering creches from 2008 through 2010, is to use a lo ery to allot scarce spaces among equally eligible families. These methods are discussed below. Decentralized selection. Many municipalities permit creche directors to make their own decisions, often based on provided (but not enforced) guidelines. This was the ex- perience in Rio de Janeiro prior to 2007. As a result, creche directors used their discretion and the process was not very systematic. An investigation revealed wide variation in the criteria actually employed to admit children into creches, despite the general guidelines. The advantage of this method is that it permits directors to identify those children who may slip through the cracks of a formal selection process. The disadvantage is that di- rectors may fail to be objective when it is efficient to do so, or they may even be biased against children with special needs or developmental delays that require more involved or expensive care. Means-tested selection. A second method is to use a process to identify the children in greatest need, and to select them. One natural approach would be to employ the so- cial protection register (cadastro único). To improve targeting, the government of Rio de Janeiro recently adopted a means-tested selection process. Speci�cally, the government Early Child Education: Making Programs Work for Brazil’s Most Important Generation 49 decided that children in Cartão Carioca (a Rio de Janeiro-speci�c supplement to Bolsa Familia) would be given highest priority, with levels of vulnerability determining access in the case of excess demand within that group. In the �rst year of implementation, all of the neediest children—those in Cartão Carioca—whose parents solicited creche spots received them. The advantage of this method is that it is much more likely to identify the most vulnerable children than is decentralized selection. However, some highly vulner- able children who do not �t the particular criteria may be excluded. This problem might be ameliorated by allowing directors a certain share of vacancies over which they do have discretion. Lo ery-based selection. Some creches in Brazil use a lo ery to assign scarce spaces. In Bahia, one government creche uses a simple lo ery with a certain number of spaces reserved for public servants (Governo da Bahia 2011). The municipality of Rio de Janeiro implemented a lo ery for the entering creche classes of 2008–2010, allowing any child an opportunity to participate in a creche, but providing higher odds for children from the most vulnerable backgrounds. (In 2011 the lo ery was replaced with a simple means- tested formula, giving priority to participants of Cartão Carioca, the Rio de Janeiro sup- plement to Bolsa Familia.) Public creches were built in vulnerable areas of Rio de Janeiro, so they already targeted the most vulnerable children given that households in wealthier areas were unlikely to want to travel far or into lower-income parts of the city for care.8 One of the advantages of the lo ery system is that it provides every interested child with the possibility of a space in a creche. Because no testing of children’s socioeco- nomic status is perfect, this means that even children who are judged less vulnerable by standard indicators but who perhaps have some particular, less visible need, may earn a place. In addition, a lo ery leads to more diversity in socioeconomic background across children in the creches. On the other hand, this method means that a proportion of creche spaces go to children that are not in the most vulnerable groups. What kind of centers? As municipalities seek to construct new centers, a key question is what kind of cen- ters to build? Creches and pre-schools can be expansions of existing primary schools, self-contained early child education centers, and—in some cases—may have no formal school building at all (e.g., they are located in a church or in someone’s home). Over the past �ve years, the vast majority of growth has occurred through adding creches and pre-schools to existing primary schools, as shown in table 3.6. Table 3.6: ECE institutions, by type of institution (2005, 2009) Creches 2005 2009 Growth (%) In school building with only early child education 19,447 22,887 18 In school building with other education levels 8,732 16,338 87 Not in a formal school building 4,117 3,805 –8 Total 32,296 43,030 33 Pre-schools In school building with only early child education 24,099 23,513 –2 In school building with other education levels 71,670 74,630 4 Not in a formal school building 9,847 8,420 –14 Total 105,616 106,563 1 Sources: PNAD (2009) survey data, Census of Schools (2009) school-level data, IBGE (2009) population data, and IBGE (2007) population-by-age estimates. 50 A World Bank Study Adding ECE to existing schools is more cost-effective than constructing entirely new buildings, most strikingly with regards to the challenges of acquiring new land versus expanding construction on land already managed by the municipal government. How- ever, evidence supports the value of dedicated early child development centers, called EDIs (espaços de desenvolvimento infantil). A recent study of ECE center quality in six Brazilian capitals that judges quality using observational tools covering activities, infrastructure, interactions, etc., showed that creches and pre-schools had signi�cantly higher quality when housed in their own building than when joined with an elementary school (Fundação Carlos Chagas 2010).9 Interestingly, the quality differential was three times larger for pre-schools than for creches. The rationale behind this quality improve- ment is that early child education requires different skills and a different approach than teaching children at higher grade levels; separate ECE institutions have a greater oppor- tunity to organize around a culture of early child education, rather adopting a prevailing elementary school pedagogical culture. Thus, municipalities will have to weigh quality against pragmatism as they seek to expand creches and pre-schools to meet the needs of their citizens. Placement Given existing demand, policy makers are left with the question of where to build the creches. A simple combination of the PNAD household survey and the IBGE demo- graphic census allows for neighborhood-by-neighborhood calculations of the ideal loca- tion for future early child education centers, following the steps in table 3.7. The calcu- Table 3.7: Method for identifying locations for new ECE centers A. Calculate number of children 0–3 by neighborhood [Population Census] B. Apply demand-rate using take-up rate among the wealthy (50%, in the case of Rio de Janeiro) [PNAD] C. Apply poverty rate by neighborhood [Population Census] D. Contrast with existing creche availability by neighborhood [Municipal records] E. Calculate excess demand among poor by neighborhood Source: Barros et al. (2010). lation entails combining estimates on (a) the population of children ages 0 to 3, (b) the percentage of those children with interest in a ending creche (extrapolated from the observed enrollment rate of the richest households), and (c) the percentage of the chil- dren who—by their level of income—would have claim to free public care. This analysis, carried out for Rio de Janeiro by local researchers, resulted in identi�cation of key neigh- borhoods in the West Zone of the city (e.g., Campo Grande, Santa Cruz) and in a few other key areas (e.g., Cidade de Deus, Complexo do Alemão). Such analyses can help municipalities target and deliver creche care effectively. Nonetheless, such analyses may be largely limited to metropolitan areas. PNAD data in particular only allow perfor- mance of this analysis for whole states or metropolitan areas, and comparable datasets are generally not available for small and rural areas. Alternative Ways to Deliver Early Child Education Some states and municipalities are exploring alternative delivery mechanisms, espe- cially for the youngest children. These mechanisms are alternatives to standard, center- Early Child Education: Making Programs Work for Brazil’s Most Important Generation 51 based care, and can be used to reach children that cannot be effectively or efficiently reached by standard modalities. The low population density in rural areas, for example, makes alternative methods especially appealing there. Two examples of alternatives are the Primeira Infância Melhor (Be er Early Childhood, or PIM) program in the State of Rio Grande do Sul—which focuses on children from birth—and the Asas da Florestania Infantil (Children’s Wings of Florestania10) or Asinhas program in the state of Acre— which focuses on children of pre-school age. Primeira Infância Melhor The most developed example of combined home-based and center-based care for chil- dren in Brazil is Primeira Infância Melhor, a program based in the Health Secretariat in the state of Rio Grande do Sul. This program, largely modeled on Cuba’s Educa a Su Hijo (Educate Your Child) program, provides two modalities of care: (1) individual care for pregnant women and for children from birth to two years, eleven months of age, and (2) group care for pregnant women and for children ages 3–6. For the individual care, an agent visits the home (once a week for children, twice a month for pregnant women) and conducts cognitive stimulation activities as well as child health and development monitoring. For group care, participants meet in a community center, church hall, or other space, and take part in games and activities to stimulate children or prepare ex- pectant mothers for parenthood. Topics include nursing and childbirth, among others (Schneider et al. 2009; Schneider and Ramires 2007). An evaluation comparing children before and after entry into the program showed signi�cant cognitive, social, and motor development. Likewise, a comparison between children in the same communities who did and did not participate in the program showed strong gains in all of these dimensions (Primeira Infância Melhor 2011). Asas da Florestania Infantil The State Education Secretariat in Acre has partnered with several municipalities to offer home-based early child stimulation. One of the major challenges that municipalities in the North face is that children live far from schools, which is problematic for all children, but especially for the youngest. The State of Acre created the Asinhas program in 2009 to reach those in remote areas. This program is directed at children ages 4–5 living in the most inaccessible rural areas, most of them in the heart of the forest. The program uses a home-based visit strategy to support children’s social, psycho- motor, and cognitive development. Education agents trained by the State Secretariat of Education make visits twice a week to the homes of children in rural areas. Usually they select one house in a community that will accommodate all of the participating children in the neighborhood. These communities are very isolated, very small and have, on av- erage, �ve families and around �ve to seven children ages 4–5. In 2010, 181 education agents reached 1,700 children ages 4–5 and 1,397 families in 15 municipalities in the state (Tribuna 2010). The program is divided into six modules with speci�c booklets for children, and manuals with guidelines for agents to follow in their interactions with the children and families. Education agents, who are secondary school students or newly graduated teachers, visit the children twice a week in the beginning of the program (module one), and three times a week in the subsequent modules. The municipalities pay their agents and supervisors, and the Education Secretariat provides training, didac- tic materials, and technical assistance. This program has not been carefully evaluated, and a rigorous evaluation of this model of support would be very welcome. 52 A World Bank Study Provision and Financing of Early Child Education Provision A large majority of early child education is publicly-provided—almost entirely by mu- nicipalities, but occasionally by states or by the federal government.11 In 2009, 66 percent of children enrolled in creche and 76 percent of children in pre-school a ended public institutions. These numbers are up from their 2001 levels of 62 percent and 75 percent, respectively—likely due to increased funding available through FUNDEB. Of course, these are the proportions of children accessing public institutions among those who are accessing ECE at all. A number of children a end unregistered institutions, as discussed in the next section. Children using informal or other unregistered care are not captured by these statistics, so they are likely overestimates of public provision in the education of under-�ve children. Private early child education institutions are of two types: either entirely private, or publicly-funded and privately-owned and operated according to the stipulations of a contract with the municipal or state government (contracted, or “conveniada�). Examin- ing the number of institutions in 2009, public creches and pre-schools were more com- mon than private ones in every region of the country. Private creches and pre-schools are least common in the North and the Northeast—the two regions with the lowest per capita income—and most common in the Southeast, as shown in �gure 3.18 and in �gure 3.19. Among private institutions, fully-private schools are more common than conve- niada (contracted) private schools in each region and for both levels of ECE. The results of the next chapter show that conveniada creches and pre-schools have higher qual- ity infrastructure than public schools. If public-private partnerships can provide higher quality care and education at equal or lower cost, then municipalities might consider expanding the use of this modality. Figure 3.18: Share of creches of different administrative dependencies, by region 0.79 0.8 0.75 0.57 0.58 0.6 0.55 Share of crèches 0.43 0.39 0.4 0.30 0.29 0.23 0.19 0.18 0.19 0.2 0.14 0.15 0.16 0.05 0.07 0 Brazil Central West North Northeast South Southeast Region Public Private—Conveniada Fully Private Source: Censo Escolar (2009). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 53 Figure 3.19: Share of pre-schools of different administrative dependencies, by region 1.0 0.91 0.85 0.8 0.76 0.73 Share of pre-schools 0.64 0.6 0.57 0.4 0.34 0.28 0.19 0.2 0.13 0.16 0.11 0.10 0.08 0.08 0.05 0.02 0.03 0 Brazil Central West North Northeast South Southeast Region Public Private—Conveniada Fully Private Source: Censo Escolar (2009). Unregistered institutions Municipal governments only receive education funds—such as transfers coming through FUNDEF or FUNDEB—if they officially register their schools and participate in the Ministry of Education’s annual Census of Schools. The Census of Schools gathers detailed data on students, teachers, and infrastructure in public and private educational institutions. However, there may be private institutions that are insufficiently incentiv- ized to register or to respond to the Census of Schools. Unregistered institutions cannot be regulated and overseen by municipal governments. Also, when governments lack basic enrollment and school quality information, it is harder to plan ECE investments. Household survey data reveal a much higher fraction of children enrolled in ECE than is captured in Census of Schools numbers. This discrepancy suggests that many children are involved in activities that their parents report as ECE, but which do not take place in officially registered ECE institutions. These might include “home creches,� in which a neighborhood woman opens her home and provides care to a small group of children. Simple comparisons from the two data sources suggest that over one-third of pre-school students �t this description, as do as many as 12 percent of children enrolled in creche. The fraction of children in unregistered creches fell dramatically during 2001–2009, from 34 percent to 12 percent (�gure 3.20). This is likely due to incentives for municipali- ties to contract private creches to receive FUNDEB resources. Pre-school students, on the contrary, were about as likely to be in unregistered institutions in 2001 as in 2009. Financing Expanding the supply of creches and pre-schools and increasing enrollment in center- based ECE will require large amounts of additional funding over the coming years. This funding responsibility will fall most directly on municipal governments, as they are the 54 A World Bank Study Figure 3.20: Fraction of early child education students in unregistered institutions, by year and region Creche 0.50 Fraction enrolled 0.5 0.48 0.4 0.34 0.3 0.28 0.28 0.28 0.2 0.15 0.17 0.12 0.1 0.03 0.07 0.03 0 Brazil Central West North Northeast South Southeast Region 2001 2009 Pre-school Fraction enrolled 0.5 0.45 0.4 0.37 0.40 0.40 0.33 0.34 0.31 0.30 0.3 0.29 0.28 0.26 0.21 0.2 0.1 0 Brazil Central West North Northeast South Southeast Region 2001 2009 Source: Censo Escolar (2001, 2009) and PNAD (2001, 2009). primary providers of early child education. Given the unmet demand for creches and pre-schools, and the FUNDEB transfers to municipalities for each child they enroll in early child education, most municipalities will have incentives to further expand ECE. The federal government created additional incentives for investments in pre-school by making pre-school education compulsory as of 2009, although municipalities have until 2016 to reach full compliance. Municipal governments have primary responsibility over early child education (ages 0–5) and primary and lower secondary education (ages 6–14).12 Higher levels are handled by state and federal governments. Municipalities must spend at least 25 percent of revenue on education, but otherwise have remarkable policy autonomy and discretion. This has led to huge cross-municipality variation in early child education investment. Many municipal characteristics influence this variation, some of which are explored in Chapter Four. Over 95 percent of public ECE �nancing is disbursed by municipal governments, although much of that expenditure comes from transfer revenues received from the state and federal governments.13 Total municipal expenditure on early child education in 2009 totaled over 9.7 billion reais (US$5.7 billion). After municipalities pay a �xed share of revenue into a state-level FUNDEB fund, that fund is redistributed among mu- nicipalities in the state according to their share of enrolled public school students (with students at different grades weighted differently). In states where this redistribution does not reach an annually-set federal guaranteed minimum, the federal government agrees to top off the fund receipt. As of 2010, this meant that municipalities were guar- anteed to receive at least R$1,558 (US$873) for each child enrolled in a public full-day Early Child Education: Making Programs Work for Brazil’s Most Important Generation 55 creche (R$1,133 for partial day), and R$1,770 (US$912) for each child in a public full-day pre-school (R$1,416 for partial day). Municipalities in relatively rich states receive more, but no municipality receives less from the state fund.14 Importantly, under FUNDEB, the only action municipalities can take to appreciably increase the receipts from the school fund is to increase the number of children enrolled in early child education. Table 3.8: Annual municipal early child education expenditures in 2009, by region Total municipal Municipal expenditure per Total municipal creche and expenditure on early child enrolled early child education Region pre-school students education student (total reais) North 504,089 281 million 558 Northeast 1,674,525 721 million 430 Southeast 2,364,619 6.83 billion 2,889 South 713,868 1.59 billion 2,228 Central West 305,158 297 million 974 Sources: Census of Schools (2009) and Tesouro Nacional Finbra Database (2009). Municipal expenditure per child enrolled in early child education varies greatly across regions, as shown in table 3.8.15 Cross-regional differences in per pupil expen- ditures are likely an important factor explaining differences in infrastructure quality, teacher training, and average class size. Dramatically, the Southeast spends 6.7 times more per enrolled public ECE child than does the relatively poor Northeast. While hir- ing teachers and building school infrastructure may be less costly in poorer regions of the country than in richer ones, this alone can only be a small part of the explanation: Different regions are clearly choosing different quantity-quality tradeoffs. The share of total municipal education spending going to early child education has increased slightly in recent years, from a national average of 8.3 percent in 2004 to 9.3 percent in 2008. However, relative expenditure on early child education in the total edu- cation budget varies greatly across regions. In 2008, it was 3.2 percent in the North, 3.4 percent in the Northeast, 6.7 percent in the Central West, 12.3 percent in the South, and 14.8 percent in the Southeast. Comparing these actual per-child ECE expenditures at the municipal level with FUNDEB transfers is revealing. It suggests that in the North, Northeast, and Central West, the average municipality spends less on each child enrolled in ECE than it obtains for enrolling that child. In those areas, FUNDEB should effectively remove the �nancial impediments poor municipalities might otherwise face to expanding early child educa- tion in response to demand, since FUNDEB provides more than the current expenditure per child for creche or pre-school. However, municipalities must also remove additional barriers to building and �lling schools, such as by engaging in dialogue with parents about their needs. The 2009 expenditures by region (table 3.8), combined with the number of children not enrolled in pre-school (table 3.2), provides a back-of-the-envelope calculation of the additional operating cost if Brazil were to enroll its 1,570,512 pre-school aged children that are currently not enrolled. The bill would come to about 2.6 billion reais, or US$1.6 billion. Brazil’s total 2009 expenditures on ECE came to US$5.75 billion, so this would 56 A World Bank Study involve almost a 30 percent increase in the ECE bill, only including municipal expendi- tures (which are the great majority), and without any expansion in coverage for children age 0–3. Some of this expense will be off-set by the reduction in enrollments due to falling fertility, as well as by increased tax revenues from expanded female labor force participation. But achieving universal pre-school enrollment will still involve signi�cant trade-offs. Making quantity-quality trade-offs Expanding enrollment in early child education and increasing its quality are twin goals, but they are also potentially competing spending priorities. Municipal governments must prioritize investments in order to grant access to previously-marginalized children (especially the poor and disadvantaged) without sacri�cing the quality needed to make early child education worthwhile. Because pre-school has become a compulsory level of education, there is no longer a substantive quantity-quality trade-off. That is, sacri�cing quantity for the sake of qual- ity is no longer an option. The goal of municipalities is to maintain quality in the face of enrollment expansion, and to make new investments in those aspects of quality that de- liver the highest returns. The trade-off being made is between various different types of potential investment in quality. The evidence presented in Chapter Two suggests some particularly high-return investments; speci�cally, investment in the quality of activities and in program structure. Exchanging investments in other aspects of quality—such as in the general education level of teachers—for these high-return investments is likely a good strategy. Creche education is not compulsory, which introduces a stark quantity-quality trade-off. For example, the fact that the Northeast region spends less per child on early child education than does the North or the Central West, and yet has a higher creche enrollment rate than does either, suggests this trade-off. Local priorities determine the choices made by each municipality. It is instructive to examine what it would cost to meet estimated unmet demand for creche spaces (table 3.4), assuming that new creche spaces have the same “average quality� as existing creche spaces. Alternatively, one can compute how many fewer children would be granted access if the same amount of mon- ey were spent, but if the quality of new creche spaces were increased. The 2009 expenditures by region (table 3.8), combined with the estimated unmet de- mand for creche spaces (table 3.4) provides a back-of-the-envelope calculation of the ad- ditional operating cost if Brazil were to enroll the almost 1.8 million children nationwide who are not enrolled but whose parents would want them enrolled in a creche. The bill would come to about 2.6 billion reais (US$1.6 billion). This is almost the same amount of money that would be required to enroll all not enrolled 4–5 year olds in pre-school. Brazil’s total 2009 expenditures on ECE came to US$5.75 billion, so this expansion in creche enrollment would also involve almost a 30 percent increase in the ECE bill, only including municipal expenditures (which are the great majority). Alternatively, Barros et al. (2011a) imply that a 6 percent increase in creche expen- ditures dedicated to improvements in the quality of activities and in program structure would increase a child’s development age by a sizeable 3.5 months, which is a full 0.5 standard deviation improvement in overall development. This implies that for the same 2.6 billion Reais that could satisfy all unmet demand for creche at the current average quality level of creches, one could enroll over 1.6 million children (94 percent of unmet Early Child Education: Making Programs Work for Brazil’s Most Important Generation 57 demand) at a quality level that implies an additional 3.5 months of development for each of these children. Of course, that increase in quality would mean that some unmet demand (about 6 percent) for creche would remain unmet, and these children would not get any of the bene�ts of creche a endance. Given the returns in child development to improvements in quality, it is highly likely that those quality improvements outweigh the bene�ts of offering lower quality creche care to all children who desire creche care.16 Thus, meeting all unmet demand—at the expense of creche quality—may be a poor investment. Lessons Brazil has made great strides in the provision of early child education in recent years. However, there is a signi�cant gap to �ll, both in rural and urban areas, to achieve uni- versal pre-school education by 2016 and to provide creche services to needy children throughout the country. To achieve this, municipalities must be strategic in the con- struction of new pre-schools and creches, and must use poverty and population data to identify the areas of greatest need. In extremely rural areas, the government will also need to explore alternative delivery mechanisms whenever the transportation of young children is sufficiently complicated to make center-based care unrealistic. While there are alternatives to center-based early child cognitive stimulation in place in a few locales, Brazil can facilitate further exploration by se ing clear minimum standards and per- mi ing municipalities to investigate the possibilities. Finally, given the signi�cant costs associated with expanding access to early child education, municipalities must set clear priorities and make quantity-quality trade-offs that will efficiently achieve those priori- ties. At the level of creche education—which need not be universal—municipalities may �nd that improvements in the quality of activities and in program structure for creches that service the most disadvantaged children yield more bene�ts than does expanding the number of available spaces in existing public creches. Notes 1. Before 2004, the Pesquisa Nacional por Amostra de Domicilios (PNAD) did not survey ur- ban parts of six of the seven states in the North (Rondônia, Acre, Amazonas, Roraima, Pará, and Amapá). As a result, data for the North is not representative of the entire region until 2004, and the North is not shown in this or other �gures until 2004. 2. The trends are similar if one examines the pre-school enrollment rate of only children ages 4–5. For consistency over time, enrollment is reported for children ages 4–6 unless speci�ed otherwise. 3. Regressions at the state aggregate and child levels suggest that this gap and its divergence over time are statistically signi�cant. Speci�cally, regressing creche a endance on mother’s working status, year dummy variables, and their interactions suggests that a mother’s participation in the labor force has a signi�cant, positive effect on creche a endance that is increasing over time. The same thing can be shown by using a time trend in place of year dummy variables; once again, the coefficient on mother’s working status is signi�cant and positive, as is the interaction term. 4. This report uses the São Paulo state average pre-school size of 87 students per school. 5. Federal Constitution of 1988, Title 8, Chapter 3, Section 1, Art 208, Proposition 4. 6. The National Network for Early Childhood is a voluntary network made up of government agencies (the Ministries of Education, Health, and Social Development), international nonpro�ts working in Brazil (e.g., Plan International, Promundo), and a host of local nonpro�ts and private sector representatives, with 86 member organizations altogether. These organizations are joined in dialogue to improve national policies related to children age zero to six years. 58 A World Bank Study 7. Of course, income may be more highly correlated with urbanization in some regions compared to others. In a region where rich families are more likely to live in urban areas than are poor fami- lies, and where demand for creche care is lower in rural areas, then these estimates of demand may be too large. 8. The lo ery is further detailed in Appendix C. 9. The study cannot rule out that EDIs are also concentrated in municipalities with particular in- terest in early child education. However, the study did control for various characteristics of the educational institution, reducing some of that potential bias. 10. “Florestania� is a new word that has been de�ned as “citizenship of the forest,� used to describe a pragmatic philosophy of sustainable environmentally friendly development, coined by the State of Acre’s Jorge Viana administration. 11. As of 2009, only about 1.7 percent of all public creche and pre-school students were in state and federal schools. 12. Municipal governments provide most public primary education (83 percent of students in 2007), but the state government has primary schools in some municipalities. 13. The exact amount is impossible to calculate given that FUNDEB transfers are not broken down by levels of education. However, 88.7 percent of total municipal revenues come from transfers (2008), and as education is one of the sectors most targeted by inter-government transfer programs (i.e., FUNDEB), the number for education is probably even higher. 14. The federal government ensures per-student minima by ‘topping off’ those state funds whose per students amounts fall below the federally-chosen minima. 15. These include both operating and construction costs. 16. In the quality improvement scenario, the quality improvements would lead to 3.5 months of additional development per child times more than 1.6 million children, which implies almost 6 mil- lion collective months of additional development. This is a substantial increase. For it to be a be er choice to enroll all unmet demand at existing creche quality levels than to enroll only 94 percent of unmet demand at a higher level of creche quality, one would have to believe that the 100,000 chil- dren excluded by the la er policy would each earn over 58 months (almost 5 years) of additional development by a ending a creche, as opposed to staying at home. This seems highly unlikely. CHAPTER 4 The Next Steps for Brazil’s Children B razil has made important strides in extending access to Early Child Education. The road ahead is long but if Brazil is pro-active and prudent in its targeting, it can en- sure ECE opportunities for its most vulnerable children in the medium-run and—over time—for all its children. Quality of early child education is also improving along some dimensions, albeit with signi�cant room for improvement. Beyond these two fundamen- tal missions of expanding coverage and improving quality, what are the next steps to ensure healthy early child development for Brazil’s youngest? This chapter explores three innovative areas with the potential to improve early child development in Brazil: cross-sectoral collaboration, leveraging private sector re- sources in a way that improves provision, quality, and innovation in early child educa- tion, and safeguarding opportunities for children in municipalities with less political will for service delivery. Cross-Sectoral Collaboration Coordinated early child development programs, both at the policy level and at the pro- gram or service-delivery level, have the potential to deliver greater bene�ts and to re- duce costs relative to efforts in individual sectors (such as education, health, or social assistance). Several countries have made great strides in coordinating policies as well as combining service delivery. Brazil has recently made strong steps in the direction of coordinated policy, through the National Plan for Early Childhood, but there is much room for expanding and improving coordinated service delivery. Complementarity As discussed in Chapter One, interventions that combine education and nutrition tend to have more positive impacts on children’s cognitive development than those that pro- vide nutritional or �nancial transfers alone. In addition, programs often have strong effects outside the expected mechanisms. Studies have shown that school-based health programs also have large education impacts, as evidenced by a program to treat school- children for intestinal worms in Kenya (Miguel and Kremer 2004) and in an iron sup- plementation program in Indian pre-schools (Bobonis, Miguel and Puri-Sharma 2006). Moreover, cash transfers can boost health outcomes in the same way that nutrition pro- grams can (Nores and Barne 2010). Improved access to services Multisectoral programs also allow parents to �nd information on and access all the ser- vices they need to help their children flourish in one place, rather than needing to sepa- rately seek out services about which they may be unaware. A study of comprehensive 59 60 A World Bank Study social assistance and care for families with young children, including health promotion, employment training for parents, and recreation activities for children, showed signi�- cant effects relative to non-comprehensive services that parents had to search for indi- vidually (Browne et al. 2001). Cost savings These programs also have the potential to generate signi�cant cost savings. Delivering services for two sectors—e.g., education and health—at the same center or during the same home visit is much less costly than making separate visits. Yet these cross-sectoral programs do come at a cost: Government ministries are built to deliver services within sectors, and coordination can be challenging, both logistically and politically. Leaders in various states of Brazil have cited the challenges of seeking to coordinate across sectors for exactly these reasons. In several countries, and in some parts of Brazil, such cross-sec- toral cooperation around early child development has been initiated. This cooperation has occurred at both the level of coordinating efforts (e.g., having a cross-sectoral plan for early child development) and at the level of coordinating programs (e.g., having joint programs that offer services in multiple sectors). Best practices from a number of other countries, as well as efforts within Brazil, can maximize the efficiency of future efforts. Cross-sectoral collaboration in Brazil: The National Plan for Early Childhood Brazil, through the National Network for Early Childhood, has developed a twelve-year National Plan for Early Childhood (Rede Nacional Primeira Infância 2010) (table 4.1). The document was prepared over the course of two years of deliberations and was launched in December 2010. This ambitious plan outlines thirteen goal areas, ranging from educa- tion, health, and social protection to confronting violence against children and protecting children from consumer pressure. In each area, a series of desired actions is presented. The plan then establishes �ve key instruments to achieve these goals, and actions therein. The creation of this plan is an excellent step toward a cross-sectoral policy, complete with goals for different time periods over the next twelve years. In order to achieve even some of the gains proposed here, the Brazilian government will need to study, re�ne, and approve the plan. It will then need to create a coordinating group, like Jamaica’s Early Childhood Commission, with key stakeholders and with a budget line, to oversee implementation in coordination with the many partners in early childhood. The �rst step is underway, as the National Council of Children and Adolescents’ Rights (CONANDA) approved the plan in December 2010. However, in order for the plan to have a true impact, the Government of Brazil will need to establish a coordinat- ing agency to oversee its implementation. The experience of other countries in this area can guide Brazil’s efforts. In several countries, overall ECD policy is housed in the Min- istry of Education: this is the case in the United Kingdom, as well as in Sweden and in several developing countries, such as Nepal and Kenya. Jamaica has adopted a related but distinct approach by creating an explicitly cross-sectoral Early Childhood Commis- sion: although it is housed within the Ministry of Education, the executive director of the commission is a physician. In Chile, the government adopted an alternative approach to coordinating its early childhood program, Chile Crece Contigo, by placing it in the Ministry of Planning to avoid sector identi�cation (V. Silva 2010). Without an executing body with funding, the Brazil plan is likely to have the same fate as the ECE goals of the National Plan for Education for 2001 to 2010: the great majority of states and municipali- ties made no moves to provide resources to achieve the goals, the federal government did Early Child Education: Making Programs Work for Brazil’s Most Important Generation 61 Table 4.1: Brazil’s National Plan for Early Childhood Goal area Example of associated action Healthy children Achieve six months of maternity leave in the public sector and encourage the same in the private sector Early child education In three years, ensure that all Early Child Education centers have formulated their pedagogical improvement plans and are in the process of implementation Family and community of the child Validate, by way of public policies of support, the family unit as the main locus of production of the child’s basic social identity Social assistance to children and their Achieve universal follow-up with families in Bolsa Familia that are not ful�lling families established conditions, prioritizing families with children up to six years old Children in vulnerable situations: foster Guarantee, within two years, the satisfaction of basic human resource norms system, foster families, adoption in the Ministry of Social Development: 1 psychologist and 1 social worker for every 20 children Right to play Annual information campaigns about the importance of play City and the environment Legally mandate that new loteamentosa reserve dedicated spaces for social apparatus attending to children’s rights to health, education, social assistance, and leisure Diversity: Black children, indigenous Create new pre-service training courses for early childhood educators, children, and quilombolab children speci�cally dealing with nuances of indigenous culture Confronting violence against children Campaigns to confront violence in early childhood Assuring documentation of citizenship for Take action so that in three years, every municipality has at least one Cartório all children (Registry Of�ce for Naturalized Persons) Protecting children from consumer Promote the prohibition or limitation of sale of unhealthy food in school pressure canteens as well as advertising in schools Controlling early exposure of children to Prohibit the presence of TVs in creches and regulate their use in pre-schools media (only for teaching purposes) Avoiding accidents in early childhood Promote the creation and ful�llment of legislation intended to avoid poisonings caused by accidental ingestion of medication and cleaning products Instruments to achieve these goals Training professionals for Early Childhood Stimulate through incentive programs, create post-graduate courses relating to early child development The role of social media Mobilize states, the federal district, and the municipalities so that they develop their own plans for early child development Legislative power Provide ongoing and detailed support to the conversion of the National Plan for Child Development into law Research on early childhood Create incentives for research on early child development—for example, a committee and a budget line (in state organizations such as INEP, CNPq, etc.) State & municipal plans for early Construct plans with ample social participation, subject to analysis and childhood approval by the Executive Power; then submit them to the legislative body for analysis, re�nement, and approval Source: Rede Nacional Primeira Infância (2010). Note: a. Loteamentos are subdivisions of land created for residential housing. b. Quilombolas are communities principally populated by the descendants of slaves. not maintain this as a focus, and the ECE coverage goal—particularly at the 0–3 level—was missed completely (Moço 2010).1 Other efforts to promote inter-sectoral coordination in Brazil include the ongoing development of a Uni�ed Program of Integrated Care in Early Childhood by the federal- level Office of Strategic Action in the Secretariat for Strategic Affairs of the Presidency of Brazil, as well as efforts that take place at a more local level through information shar- ing. For example, various organizations—including the World Bank and the non-govern- mental organization Avante—have carried out mappings of early child services in different municipalities to, at the very least, inform each sector of the activities of other sectors. This information is a �rst and necessary step to coordination. 62 A World Bank Study International experience in strategic ECD planning: Jamaica’s National Strategic Plan for Early Childhood Development As with Brazil and its National Plan for Early Childhood, Jamaica in 2008 launched a �ve-year National Strategic Plan (NSP) for Early Childhood Development. The NSP was developed by the Early Childhood Commission and established by the Jamaican parlia- ment in 2003 to ensure that children under age eight have access to high-quality social services. While housed within the Ministry of Education, the NSP coordinates services across the Ministries of Health, Labor, Finance, and the government agencies for technical and vo- cational education and for training, as well as a number of non-governmental partners. The NSP lays out �ve processes to improve early child development, as well as two processes to improve the effectiveness of the NSP itself. Each process has a set of related activities.2 In its �rst two years, the plan has had marked successes, demonstrating how a coordi- nated plan can improve ECD policy. These successes include approval of a Standards and Accreditation System for early childhood parenting education and support programs, de- velopment of a curriculum and delivery model for child development therapy, complete initial inspection of 35 percent of Early Child Institutions, and the design, development, and implementation of a management information system for ECD (JIS News 2010). While the Jamaican National Strategic Plan has led to advances, it has also faced challenges. Because the plan was largely developed by a small group, buy-in from some service providers has been limited, leading to bo lenecks in the implementation of some of the goal activities. Furthermore, communication and coordination among members of the executive commi ee have been hampered by inter-personal conflicts, highlight- ing the importance of selecting leadership for an executing body with a cross-sectoral background and strong management and diplomatic skills. Brazil’s national plan was developed in an extensively collaborative process over a long period of time, but if— upon legalization—the government goes on to create a coordinating body, appointing the right staff for this will be critical to avoiding impediments in implementation. International experience in cross-sectoral service delivery: Chile Crece Contigo In addition to comprehensive ECD plans, several countries have had success with ser- vice delivery that cuts across sectors, providing education, health, and other services from one point of contact. International examples include Chile’s program, Chile Crece Contigo (Chile Grows With You, or ChCC), New Zealand’s Family Start and Well Child Tamariki Ora Service, and England’s Sure Start program. ChCC’s coordinated program can inform similar goals in Brazil. In 2005, the Chilean government organized an inter-ministerial working group, coordinated by the Minis- try of Planning, which included the Ministries of Health and Education as well as the Budget Office. They spent one year carrying out a range of studies to (a) categorize the existing programs related to early childhood, including small-scale programs, (b) ex- plore key issues not previously studied, (c) ascertain Chilean perceptions of early child development, and (d) update statistics, indicators, and results related to early childhood in Chile. In 2006, they prepared a proposal of government measures needed to create a system for integrated protection of early childhood. In 2007 they created the program, and in 2009 the system was integrated into law. ChCC provides comprehensive care to children from pre-natal care for mothers until the children enter pre-school (at four or �ve years old) (table 4.2). Early Child Education: Making Programs Work for Brazil’s Most Important Generation 63 Table 4.2: Services offered under Chile Crece Contigo Target Group Services All children 1. Mass Education Program 2. Interactive information spaces 3. Proposals for improved legislation Children who participate in the public health system 4. Biopsychological development support program Vulnerable children 5. Home visits by health team 6. Automatic access to family subsidy 7. Access to free, high quality nurseries and childcare 8. Preferential access to public programs 9. Integrated attention to children with delays 10. Technical assistance for children with disabilities Source: Adapted from Silva (2010). One valuable lesson from ChCC is that coordinating the policy in a non-sectoral ministry (e.g., planning rather than health or education), can be advantageous both for more efficient management of loans and grants, and to minimize cross-sectoral territo- rial issues. A second lesson is that the sum of cross-sectoral early child development co- ordination has the potential to be much more than the sum of its component sectors. The coordination in ChCC had led not only to increased investment in each area, but also to investments across sectors. For example, children with a developmental delay identi�ed in the �rst few months are referred to a “stimulation room� (sala de estimulación), where they and their parents receive one-on-one or group a ention in cognitive and motor stimulation, employing many of the same skills and exercises used in childcare centers and pre-schools. Rather than merely having larger education and health investments for early childhood, children are now accessing educational interventions within the health sector, providing cognitive stimulation to more young children in need. ChCC also re- lies on an integrated monitoring system, wherein a subset of the data from health and education programs feeds into a single system that sends noti�cations for intervention when the data predict a concern for the child. Such a system is another advantage of an integrated program like ChCC. Center-based cross-sectoral programs Integrated child development programs exist around the world, in both the public and nonpro�t sectors. Some of these programs provide center-based services, some provide home-based services, and many provide a mix, wherein home-based services take place after a referral from the health or education sector (as is the case with Chile Crece Con- tigo). New Zealand’s “Well Child Tamariki Ora Services� program is an example of a well-developed, center-based program that provides comprehensive health care for chil- dren ages 0–5, in addition to parent training and information. Likewise, the Swedish pre-school system provides free pre-school for children from age one to �ve. The health sector works closely with the pre-schools so that screenings for a ention de�cits, hear- ing problems, and motor delays are all available at the center (Ma hews 2010). Movement toward service delivery coordination in Brazil The most established example of cross-sectoral collaboration in early child develop- ment in Brazil is the Rio Grande do Sul program Primeira Infância Melhor, described in some detail in Chapter Two. Although the program is housed in the State Health 64 A World Bank Study Secretariat, it is managed by the State Technical Group, which includes technical staff from the Secretariats of Health, Education, Culture, Justice, and Social Development, regional coordinators from health and education from across the state, and others. At the municipal level, the program is run by the Municipal Technical Group, which also includes representatives from all participating secretariats. This local group runs the program completely, from organizing and monitoring activities to providing training to the �eld workers (Schneider and Ramires 2007). The Education Secretariat of the state of Acre sends educational agents into the homes of rural children twice weekly to provide cognitive stimulation. It is working with the Health Secretariat to coordinate both educational visits and diagnostic infor- mation collection so that children receive the highest possible quality of services. Mu- nicipalities are also seeking to coordinate services: the Education Secretariat of Rio de Janeiro municipality is currently endeavoring to bring health professionals into each creche to ensure that children have access to health services while at an education center. These programs are breaking ground, but increased coordination and cooperation across sectors to deliver services to children will make it possible for more children to receive the range of services they need, at lower cost and to greater effect. Leveraging Private Sector Provision, Funding, and Innovation Given the massive expansion in ECE required over the coming years and the likely bud- get implications of achieving coverage and quality goals, the private sector may be a critical resource for providing care and education. Private investment and innovation in early child education could effectively complement public funding—especially if public construction and expansion is slow to materialize and meet growing demand. Provision The most recent data show that more than one third of creche spaces are provided by the private sector: 29 percent are fully private and 14 percent are conveniado (or government contracted). Likewise, at the pre-school level, a quarter of provision is through the pri- vate sector: 19 percent fully private and 5 percent conveniado. Private provision is grow- ing rapidly, with hundreds of early child education institutions contracted each year (M. d. Silva 2010). The expansion signals signi�cant capacity in the private sector to supple- ment the efforts of the public sector in providing early child services. Furthermore, data on quality suggest that contracted institutions have be er infrastructure, on average, than do public institutions. Brazil’s Ministry of Education has published a guide to assist municipalities in the development of conveniado relationships which may help munici- palities expand the use of this modality to provide ECE (MEC 2009). Given the potential capacity of the private sector and the desired expansion, Brazil may need to leverage non-public providers. Other countries that have expanded rapidly have had to do so. The Head Start pre-school program in the United States, which has expanded by as much as 50,000 spaces in a single year and now serves 900,000 pre-school children, relies exclusively on contracted centers similar to conveniados (OHS 2010). In contrast, during 2001–2005, Mexico expanded ECE for 3–5 year-olds by over one mil- lion spaces, almost entirely through the public system. However, this included major increases in class size—in many cases to over 30 students per caregiver (Yoshikawa et al. 2007). Private investment might be useful in avoiding such deteriorations in quality. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 65 Although FUNDEB reimbursements are equal for public and conveniado spaces, munici- palities have the liberty to compensate conveniado spaces at either higher or lower levels, pro- viding the freedom to create incentives for private provision for difficult-to-reach popula- tions (MEC 2009). Two international examples of this strategy to reach difficult populations come from the United States and Chile. A recent revision of the U.S. Head Start policy cre- ated a range of special incentives for programs to reach homeless children (NAEHCY 2008), who are likely the hardest-to-reach and hardest-to-serve population in the United States. Similarly, the Chilean government included pre-kindergarten and kindergarten in its Preferential School Subsidy program in 2008, which provides additional resources to early child and primary education institutions that cater to the most vulnerable children. In Chile, the vulnerable children are identi�ed either because they are enrolled in Chile’s main social transfer program (Chile Solidario), because they are in the bo om third of vulnerability according to the social protection register (comparable to Brazil’s cadastro único), because they are in the public health program, or—�nally—because they are not in those programs but are still in low-income households (MINED 2011). Institu- tions receive additional resources per vulnerable child reached as well as for enrolling higher concentrations of vulnerable children. The program also includes incentives to encourage schools to use the additional resources to improve opportunities for the most vulnerable children: schools receive some of the monetary incentives upon enrollment and the rest upon approval of their plan to improve education for vulnerable popula- tions. Likewise in Brazil, municipalities can experiment with ways to provide additional incentives to reach the most difficult-to-reach children, and MEC can develop guidelines to reach these children, leveraging the innovation of the private sector. Private sector funding In addition to subsidized provision through the private sector, expanding public-pri- vate partnerships may offset some of the budget impacts of expanding coverage and improving the quality of Brazil’s ECE. Businesses can realize short-term bene�ts from investments in early child development in terms of reduced employee absenteeism and turnover due to child care problems. Indeed, in the United States and elsewhere, pro- ductivity and reduced absenteeism have both been clearly linked to quality child care (Shellenback 2004, Anderson 2009). In addition, the private sector—from a longer-term viewpoint—bene�ts from a well-educated and well-developed workforce, for which early child development is an essential ingredient. Many corporations have a philanthropic arm which can contribute to ECD, and a number of corporations across Brazil are doing this. In many countries, these corporate donations are mobilized as matching funds to provide and improve ECD services. A recent review of ECD partnerships in the United States revealed some common charac- teristics for success. They tend to pool funds from public, private, and nonpro�t coffers and channel them through a state or local central nonpro�t with a particular early child mandate. The focus of these partnerships has been on expanding pre-school, developing both in-home and out-of-home early child education opportunities, and—most com- monly—creating integrated, cross-sectoral early child development systems (as dis- cussed earlier in this chapter). Several partnerships also promote best practices, facilitat- ing the cross-municipal learning discussed in the previous chapter (NGA 2008). These partnerships can be formed wherever there is the will, and a series of steps (outlined in table 4.3) has been successful in leading to their creation. 66 A World Bank Study Table 4.3: Steps to establishing public-private partnerships for ECD 1. Convene leaders from the public, private, and nonpro�t sectors to highlight the importance of ECD and make a call to action 2. Identity resources from each group (�nancial, human, and in-kind) that can form the basis of a partnership 3. Formalize government commitment through legislation of an executing body 4. Determine the executing body’s governance structure 5. Develop an evaluation strategy Source: Adapted from NGA (2008). Steps are derived from partnerships in the United States. In Brazil, an example of public-private cooperation is the Millenium Fund for Early Childhood Education, which promotes in-service training and capacity building for staff and administrators of early child education centers in 15 municipalities of southern Bra- zil (Paiva et al. 2009). One speci�c initiative is the creation of “Education Boards,� a phys- ical space in which early child educators receive training through hands-on participation in the same kinds of play-based activities that they should share with children. In this partnership, municipal education secretariats provide instruction to trainers, UNESCO provides office space for coordinating the program, and private sector donors—princi- pally the Gerdau Institute, among others—provide funding for the program’s operating costs and materials. As Brazil seeks to a ain lofty coverage goals for both creche and pre-school, these kinds of partnerships may prove critical to expanding capacity. Public-private partnerships also provide opportunities to innovate, as private funds can be leveraged to experiment with new modalities and delivery mechanisms for which public funds may incite controversy. An example from primary education is the intro- duction of pay-for-performance for students in New York City, which met with criti- cism but was politically defensible since privately-donated funds were used for the pi- lot transfers. Many existing public-private partnerships outside Brazil provide grants to localities to improve ECD, allowing creativity in the exact application of resources (NGA 2008). In New Zealand, the Centers of Innovation program provides grants to ECE centers to innovate and evaluate, experimenting with new pedagogies, integration of families with center care, and other areas (Gibbs and Poski 2009). Compensating for Differences across Municipalities The Constitution affords extensive municipal government autonomy over early child edu- cation policy, which has led to huge variation in investment. Many municipal characteris- tics influence this variation, including the wealth of the average citizen and the degree of inequality in the income distribution (demonstrated by Kosec 2011). In planning national early child education policy, the federal government should take into account how differ- ent municipalities are differently motivated to expand ECE. This will help the govern- ment develop policies that create incentives for investment where it is needed most. Of course, investment in early child education may be limited by the availability of public funds. When public funds are low, early child education—and especially creche education, still not compulsory—may be viewed as a dispensable luxury good when compared with compulsory pre-school and primary education. However, following the FUNDEF and FUNDEB school �nance equalization reforms, the amount of revenue available for education has become less tied to a municipality’s economic level. By re- distributing revenue within states, from rich to poor municipalities, these reforms have Early Child Education: Making Programs Work for Brazil’s Most Important Generation 67 facilitated ECE investment in precisely those areas where research suggests it will have the greatest impact: poor municipalities. Income, inequality, and ECE investment Shifting funds between different municipalities is important since not all municipali- ties have the same incentives to invest additional revenue in early child education. In a study of Brazil’s over 5,000 municipalities during 1995–2008 (the period during which FUNDEF and FUNDEB were enacted), Kosec (2011) demonstrates two key points about municipal characteristics and ECE investment. First, public creches and pre-schools do not bene�t all citizens equally. They disproportionately bene�t poorer citizens, since the poor are the most likely to use these public services. Rich families tend to enroll their children in private school, and therefore are less invested in the public education system. Second, the distribution of income in a municipality hugely affects public early child education investment. Given extra revenue (such as that provided by FUNDEB), poor and equal municipalities are more likely to expand early child education than are richer and more unequal municipalities. They are especially likely to expand ECE if they have participatory budgeting policies in place, which allow citizens in each neighborhood to directly vote on how their neighborhood will spend a share of municipal funds. Kosec’s results are easily illustrated by comparing a group of pre-school age chil- dren not enrolled in school that live in a municipality with median inequality (Gini coef- �cient of 0.39) with similar groups of children in municipalities with lower and higher inequality. If the municipality with median inequality is given enough new revenue to expand ECE and enroll 100 new children, more unequal municipalities given the same amount of money will enroll fewer than 100 children, and more equal municipalities will enroll more than 100 children, as illustrated in �gure 4.1. Likewise, if the munici- pality with median income is given enough new revenue to expand ECE and enroll 100 Figure 4.1: Additional children enrolled following revenue shock that leads median municipality to enroll 100 more children 200 Children enrolled if median enrolls 100 Income 180 181 169 Inequality 160 155 140 135 120 125 113 106 103 100 100 94 87 80 81 71 69 60 40 42 20 23 0 0 10th 20th 30th 40th 50th 60th 70th 80th 90th Decile of median income or of inequality (Gini coefficient) Source: Kosec (2011), using data from Censo Escolar (1995–2008), Tesouro Nacional (1995–2008), and IBGE Censo (2000). 68 A World Bank Study new children, richer municipalities given the same amount of money will enroll fewer than 100 children, and poorer municipalities will enroll more than 100 children. Poorer municipalities are more likely to enroll new children given any amount of additional revenue. Income appears to ma er more than inequality in determining how likely a municipality is to put extra revenue into expanding early child education. Policy implications These �ndings have at least two important policy implications. First, revenue transfers that target the poorest and least unequal municipalities are likely to be the most effective in expanding early child education. Policy makers in those municipalities already have a relatively strong desire to spend their next Reais of revenue on early child education. Targeting funds at those municipalities means the money is more likely to end up in early child education. In richer and more unequal municipalities, a transfer of funds is more likely to end up in public primary or lower-secondary education, or in goods other than education, like public infrastructure. If the goal is to expand public early child edu- cation, then federally-imposed minimum spending or universal provision requirements may be necessary in municipalities with high income inequality. A corollary of this point is that FUNDEB would create be er incentives for invest- ment in early child education if it involved national redistribution rather than within- state redistribution. A national redistribution would ensure that Brazil’s poorest mu- nicipalities receive the largest influxes of money under FUNDEB, and that its richest municipalities receive the smallest. At present, a poor municipality located in an even poorer state seems rich relative to other municipalities in that state, and therefore loses money through FUNDEB. At the same time, some rich municipalities (in the nation-wide sense) that are poor relative to other municipalities in their state are gaining money. A nation-wide redistribution program would eliminate these imbalances and ensure that money intended to expand public education goes to those municipalities most likely to make these investments. There may be political stumbling blocks to such a policy change, given the strength of state governors and representatives in federal policymak- ing. However, such a policy would be a substantial step in the direction of evening the playing �eld across states and not only within them. Second, these results suggest that improving governance can signi�cantly boost investment in early child education. Increasing the transparency and accountability of local government and enacting participatory budgeting are viable solutions to low in- vestment that do not require externally imposing minimum expenditures or increas- ing the budget. These changes work by ensuring that the poor—who have the highest demand for public ECE but are often socially and politically marginalized—are more directly and fully involved in decision-making. The federal and state governments could increase public ECE investment by encouraging municipalities to make their budgets and expenditures transparent, and by encouraging them to explore ways to give poor citizens a voice in local government. Incentives that encourage participatory budgeting in particular may be effective in expanding early child education. The potential power of participatory budgeting Participatory budgeting (PB) has been shown to make a rich or unequal municipality more likely to convert extra revenue into early child education spending. But what ex- actly is PB, and how can governments bring it about? Following Brazil’s return to de- Early Child Education: Making Programs Work for Brazil’s Most Important Generation 69 mocracy in 1985, many municipalities implemented PB. Under PB, citizens vote on how to use a share of municipal revenue designated for their neighborhood, and they elect neighborhood representatives to make municipality-wide spending decisions. Citizens deliberate and negotiate over the distribution of public resources. Policy makers publi- cize budgets and expenditures to promote transparency. Case studies from Brazil sug- gest that PB increases participation of marginalized groups, leads to more pro-poor ex- penditures, and increases government accountability (Souza 2001). Empirical evidence suggests that Brazilian municipalities with PB spend a slightly larger share of their bud- gets on health and education programs (Boulding and Wampler 2010). By dividing the budget into tracts assigned to each neighborhood, PB mechanically ensures a relatively equal distribution of political power. Thus, resulting policies are more likely to reflect the will of the people and not only that of the most politically powerful citizens. In Brazil, PB has typically been implemented under a situation of divided govern- ment, where the municipal mayor and the municipal legislative council are pursuing two separate political agendas (Melo 2009). The legislative council can veto any budget, but it is politically difficult to do so when the budget comes from the people through PB. When divided government paralyzes policymaking, mayors implement PB to get the policy process moving again. More often than not, PB has remained in place once initially introduced. If PB simultaneously promises to politically empower otherwise marginalized citizens while also increasing spending on early child education, then the federal government may wish to create incentives for and encourage PB as part of a broader ECE development policy. The government could even incentivize more limited forms of PB that speci�cally apply only to spending on education. Lessons Cross-sectoral collaboration Coordination and multisectoral provision of services provide an opportunity to make good programs even be er, to ensure that parents receive the full array of services that their children need, and to reduce costs by offering services at a single location. Of course, integrating early child development services is not a panacea. An important example in the United States is the Comprehensive Child Development Program, which employed case management and home visiting to ensure that low-income children received health, education, and social assistance services. A large evaluation of the program found no average impact whatsoever (St. Pierre and Layzer 1999).3 Coordinated services cannot compensate for poor quality education, health, or social assistance programs. However, as the government of Brazil seeks to improve quality in early child education and in other services, cross-sectoral coordination will be key to ensuring that quality. In Brazil, integrating health services into creches to provide more opportunities to identify developmental delays and health problems and to provide services for those children early on will lead to signi�cant improvements in child welfare. However, in every country where inter-sectoral coordination has been effective, it has required both strong national and local leadership. Because this coordination requires moving outside standard secretariats of health and education, establishing a cross-sectoral coordinating agency and encouraging integration of services can go a long way toward ensuring that children receive the help they need to contribute to Brazil‘s future growth. 70 A World Bank Study Public-private partnerships Partnerships outside the public sector can provide signi�cant resources to expand ECD services, improve the quality thereof, and innovate in reaching the most vulnerable pop- ulations. Municipalities should engage the private sector to innovate on the quality of services and explore the possibility of using private sector providers to reach the most difficult-to-reach populations in innovative ways. MEC can identify and highlight best- practice examples of public-private partnerships. Compensating for municipal differences Even as educational �nancing reforms, especially FUNDEB, have created new incentives to invest in early child education, these incentives are distributed unequally, with poor children in richer and more unequal municipalities receiving fewer services. Elements of participatory budgeting processes can protect the poorest children even in these mu- nicipalities and have received positive feedback in the past. MEC can provide guidance to municipalities and states as to how to implement elements of PB policies to protect the poorest children. Notes 1. The pre-school goal was almost a ained. 2. These are laid out in detail in Appendix H. 3. Later research revealed that the program was effective for certain sub-groups (Ryan et al. 2002). Appendixes 71 Appendix A: Pre-school Enrollment with Alternative De�nitions Given changes in the structure and starting ages of pre-school and primary school edu- cation in Brazil during the period explored (1996–2009), there are several possible ways to compute pre-school enrollment rates. Except where otherwise indicated, we always compute the share of children aged 4–6 that are enrolled in institutions intended for children in that age group. Before 2007, this is pre-school; beginning in 2007, this is either pre-school or primary schools because starting in 2007, six-year-olds were (gradually) included in Year 1 of Primary and Lower Secondary. Table A.1 compares our pre-school enrollment rate in each year during 1996–2009 with the share of 4–6 year-olds in school (whether in pre-school or primary school). This measure equals our measure in 2007 and later. Prior to 2007, the measures differ due to children ages 4–6 being enrolled in schools intended for older students. Table A.1 also shows the share of 4–5 year-olds in school in each year (an age group for which pre-school is always the intended level of education). Table A.1: ECE enrollment rate by year and de�nition of enrollment Pre-school enrollment rate (% share of age 4–6 population enrolled in institutions intended Share of age 4–6 population Share of age 4–5 population Year for children that age) enrolled in any school (%) enrolled in any school (%) 1996 48 54 43 1997 51 56 47 1998 51 58 47 1999 52 60 50 2001 57 66 55 2002 58 67 57 2003 60 68 59 2004 61 71 61 2005 63 72 63 2006 65 76 68 2007 78 78 70 2008 80 80 73 2009 81 81 75 Source: MEC. 73 74 A World Bank Study Appendix B: Survey of Evidence for Early Child Education in Brazil In Brazil, the evidence on early child education is expanding. Table B.1 presents a collec- tion of studies examining the issue. Table B.1: Survey of studies of impact of ECE in Brazil Study Location & Data Design Results Barros et al. Rio de Janeiro Multivariate Impact of high quality creche (2011a) Child development test, regression • Higher quality creches across various observational instrument 2001 dimensions led to better cognitive, social, and physical development Barros et al. Rio de Janeiro Randomized Impact of creche (2011b) Administrative data, assignment • Double labor force participation of initially questionnaire out-of-work mothers Felício, Menezes Sertãozinho (São Paulo) Multivariate Impact of ECE and Zoghbi (2010) Provinha Brasil, student regression, • Students who entered school at 3, 4, or 5 questionnaire propensity score had 6% higher literacy in 2nd grade matching Fundação Carlos Six state capitals Multivariate Impact of high quality pre-school Chagas (2010) Provinha Brasil, ECERS-R regression • Signi�cant impact of higher quality pre- school on Portuguese test score in 2nd grade Pazello and National Multivariate Impact of pre-school Almeida (2010) SAEB, PNAD (1993–2007) regression • No signi�cant impact on primary school completion Rodrigues, Pinto National Instrumental Impact of creche and Santos (2010) SAEB, Uni�ed Health System variables • Improvement in math test scores in 4th (SUS) grade of 0.1 standard deviations Calderini and National Instrumental Impact of pre-school Souza (2009) Prova Brasil 2005 variables • Improvement in math and Portuguese test scores in 4th grade Curi and Menezes- National Multivariate Impact of pre-school Filho (2009) PPV & SAEB regression cross- • Increased primary, lower secondary, sectional analysis secondary, and college completion • Increased standardized test scores in 4th, 8th, and 11th grades • Increased total years of schooling and wages • Impact of creche • Increased secondary and college completion Natenzon (2003) National Multivariate Impact of pre-school SAEB, PNAD (1993–2001) regression • No impact on math or Portuguese test scores in 4th grade Felício and National Propensity score Impact of ECE Vasconcellos SAEB 2003, Prova Brasil matching, �xed • Students from ECE had higher math (2007) 2005 effects scores in 4th grade Young (2001) SE & NE regions Multivariate Impact of pre-school Pesquisa sobre Padrões de regression cross- • Increased years of education (particularly Vida (PPV) 1997 sectional analysis for low-income children) • Reduced grade repetition • Improved earnings (for men) Source: Authors. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 75 Appendix C: Rio de Janeiro’s Creche Lottery Any child could register for the lo ery for a given creche, after which households were split into three groups: Most Vulnerable, Middle Vulnerable, and Least Vulnerable (see �gure C.1). A lo ery was then held for one third of the creche vacancies among only the most vulnerable households. Children who won that lo ery received spots in the creche. A second lo ery was held for the next third of creche vacancies among the Middle Vul- nerable children and the Most Vulnerable children who had lost the �rst lo ery (i.e., they had a second chance to enter). Finally, a third lo ery was held for the �nal third of creche vacancies, among the Least Vulnerable children, as well as the Most Vulnerable and Middle Vulnerable who had lost the previous lo eries. In this way, any child had the possibility of receiving a creche vacancy, but the Most Vulnerable children had three dif- ferent chances to win, whereas the Least Vulnerable children had only one. Figure C.1: Lottery process for municipality of Rio de Janeiro, 2007–2010 Interested children register for the creche lottery they are divided into 3 groups Most Vulnerable Middle Vulnerable Least Vulnerable 1st lottery, for ⅓ 2nd lottery, for ⅓ 3rd lottery, for ⅓ of openings of openings of openings winners others winners winners Spot in creche Waiting list Source: Compiled by authors based on Barros et al. (2010). 76 A World Bank Study Appendix D: Quality Rating Systems Quality rating systems for Early Child Education in the United States generally evaluate the standards outlined in table D.1. Table D.1: Quality standards used to rate ECD centers Licensing Compliance Family Partnerships Ratio and Group Size Administration and Management Health and Safety Cultural and Linguistic Diversity Curriculum Accreditation Environment Provisions for Special Needs Child Assessment Community Involvement Staff Quali�cations Source: Adapted from Tout et al. (2010). Note: Each QRS employs only a subset of these standards. The programs all use a building blocks rating structure, a points rating structure, or some combination of the two, then assigning the programs a number of stars as a quality rating. For example, the Oklahoma QRS employs a building blocks system: centers must satisfy minimum licensing requirements to achieve Level 1, possess weekly lesson plans, a daily reading program, and other factors, to reach Level 2, and so on (see table D.2 for details). Without satisfying all of the requirements for Level 2, a center cannot prog- ress to Level 3. On the other hand, the Colorado system employs points: ECD centers may choose from a range of factors that improve the learning environment such as staff training and child-staff ratio, earning points for each factor and then achieving different levels for points earned. Table D.2: Rating systems for early child development centers Level Requirements Components rated Oklahoma’s Reaching for the Stars system—block-based system 1 1 star is automatic with license Minimum licensing requirements 2 Apply and meet criteria Above plus teacher and director training, weekly lesson plans, activity interest areas, daily reading program, parent involvement (Centers must proceed to Level 3 within two years or they revert to Level 1) 3 Apply and meet criteria or national Above plus teacher credentials, salary compensation, program evaluation accreditation including Environmental Rating Scales (ERSs) 4 Apply and meet criteria and national Same as above accreditation Colorado’s (United States) Qualistar Early Learning system—point-based system 0 0–9 points Learning environment; family partnership; staff training and education; group size and child-staff ratio; accreditation 1 10–17 points Same as above 2 18–25 points Same as above 3 26–33 points Same as above 4 34–42 points Same as above Source: Adapted from Zellman and Perlman (2008), tables 3.1 and 3.2 . Early Child Education: Making Programs Work for Brazil’s Most Important Generation 77 Appendix E: Selected Child Development Instruments Table E.1 presents a list of some of the instruments used in measuring child cognitive and social development in creches and pre-schools. A more complete list can be found in Fernald et al. (2009). Table E.1: Sample of child assessment instruments used in creches and pre-schools Screening tool vs. assessment Ratings & reports vs. direct Instrument of abilities assessment Ages & Stages Screening Ratings & reports Ages & Stages—Socio-emotional Screening Ratings & reports Parents’ Evaluation of Developmental Status Screening Ratings & reports Brief Infant Toddler Social Emotional Assessment Screening Ratings & reports Ounce Scale Assessment of abilities Ratings & reports Pre-school and Kindergarten Behavior Scales Both Ratings & reports High/Scope Child Observation Record for Infants Assessment of abilities Ratings & reports and Toddlers Source: Fernald et al. (2009). 78 A World Bank Study Appendix F: Curriculum for Primeira Infância Completa Table F.1 shows the curriculum implemented as of November 2009. Table F.1: Curriculum implemented as of November 2009 Week Topic Responsible secretariat 1 Within the PIC Education/Health/Social Assistance 2 Health at the childcare center Health 3 Baby on the development path Education 4 And the family, how is it? Social Assistance 5 Good treatment: Your child deserves it Health 6 In my time, children were like this Education 7 Family planning Health 8 Present parent! Social Assistance 9 Congratulations: Child Development Education 10 Growing and developing healthily Health 11 Food, fun, and art Health 12 I want to see you smile Health 13 Affectionate habits Education 14 Diarrhea: When there is a risk of dehydration Health 15 Dedicating time to your child Education 16 Being a child isn’t easy Education 17 Who tells a story? Education 18 Pirlimpimpim: The world of make believe Education 19 Getting current on vaccinations Health 20 Illnesses that color and mark Health 21 Affection: Beyond words and gestures Education 22 Limits: Yes, yes, no, no Education 23 Mental health Health 24 Breathe Rio without allergies Health 25 The environment = my environment Education 26 Rights of the child Social assistance 27 Games are a serious thing Education 28 Prevention of accidents Health 29 Domestic violence Social assistance 30 Community councils Social assistance 31 Uni�ed Health System Health 32 The essence of childhood Education 33 Diversity Education 34 Children and the media Education 35 Leisure for all Education 36 Gender issues Education 37 Sexuality in early childhood Social assistance/Health 38 Strategy for family health Health 39 Family and community living Social assistance 40 The importance of civil registration Social assistance Source: Secretariat of Education, Municipality of Rio de Janeiro. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 79 Appendix G: Selection of MEC Publications for Early Childhood Development Table G.1 shows a selection of MEC publications for ECE. A full list of available MEC publications for ECE is available at the website of the Secretary for Basic Education in the Ministry of Education (MEC n.d.). Table G.1: Selection of MEC publications for ECE Publication title Date Orientações sobre convênios entre secretarias municipais de educação e instituições comunitárias, 2009 confessionais ou �lantrópicas sem �ns lucrativos para a oferta de educação infantil Indicadores da Qualidade na Educação Infantil 2009 Critérios para um Atendimento em Creches que Respeite os Direitos Fundamentais das Crianças, 2ª. ed. 2009 Educação Infantil: Saberes e práticas da inclusão 2006 Parâmetros Nacionais de Qualidade para a Educação Infantil 2006 Política Nacional de Educação Infantil 2006 Parâmetros Básicos de Infra-estrutura para Instituições de Educação Infantil 2006 Referencial curricular nacional para a educação infantil 1998 Source: MEC. 80 A World Bank Study Appendix H: Jamaica’s National Strategic Plan for Early Childhood Development Table H.1 outlines Jamaica’s National Strategic Plan for ECE. Table H.1: Jamaica’s National Strategic Plan for Early Childhood Development Processes to improve early child development Sample associated activity Effective parenting education and support Ensure that antenatal, child health clinics and ECIs provide parenting advice, education, and support for parents Effective preventive health care Develop a certi�cation system to ensure that child health clinics provide the highest quality services Early and effective screening, diagnosis and intervention for Develop and put in place a national policy on screening and “at risk� children and households early identi�cation for children and households that are “at risk� Safe, learner-centered, well-maintained early childhood Inspect ECIs regularly to ensure they maintain high institutions facilities standards Effective curriculum delivery by trained ECD Practitioners Provide all ECIs with an approved early childhood curriculum Working environment processes All the persons and organizations who work with children or Put laws in place to govern the early childhood sector provide programs and services for them must work together to achieve the targets set Decisions on how to improve the quality of early childhood Put in place a computerized system to collect, store and development in Jamaica must be made based on timely, manage important information about young children and clear, current, and accurate information their families Source: Early Childhood Comission (2009). References Almond, Douglas, and Janet Currie. “Human capital development before age �ve.� Handbook of Labor Economics, 2011, 1315–1486. Armecin, Graeme, Jere Behrman, Paulita Duazo, Sharon Ghuman, Socorro Gultiano, Elizabeth King, and Nanne e Lee. Early Childhood Development through Integrat- ed Programs: Evidence from the Philippines. Policy Research Working Paper 3922, Washington, DC: World Bank, 2006. Arne , J. “Caregivers in Day-care Centers: Does Training Ma er?� Journal of Applied De- velopmental Psychology, 1989: 541–552. Aud, Susan, Michael Planty, Thomas Snyder, Kevin Bianco, Mary Ann Fox, Lauren Frohlich, Jana Kemp, and Lauren Drake. The Condition of Education 2010. NCES 2010–028, Washington DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education, 2010. Baker, Michael, Jonathan Gruber, and Kevin Milligan. “Universal Childcare, Maternal Labor Supply, and Family Well-Being.� Journal of Political Economy, 2008. Barnard, Wendy, Wendy Etheridge Smith, Richard Fiene, and Kelly Swanson. Evaluation of Pennsylvania’s Keystone STARS Quality Rating System in Child Care Se ings. Pi s- burgh, PA: University of Pi sburgh Office of Child Development, 2006. Barne , W. Steven. “Be er Teachers, Be er Preschools: Student Achievement Linked to Teacher Quali�cations.� Preschool Policy Ma ers (NIEER), 2003. ———. “Bene�ts of Preschool Education� (presentation), 2004. www.nieer.org/resources/ �les/Barne Bene�ts.ppt (accessed May 3, 2012). Barne , W. Steven, J Tarr, C Lamy, and E Frede. Children’s Educational Needs and Com- munity Capacity in the Abbo Districts. New Brunswick, NJ: Center for Early Educa- tion, Rutgers University, 1999. ———. Fragile Lives, Sha ered Dreams: A Report on Implementation of Preschool Education in New Jersey’s Abbo Districts. New Brunswick, NJ: National Institute for Early Education Research, Rutgers University, 2001. Barne , W. Steven, K Brown, and R Shore. “The Universal vs. Targeted Debate: Should the United States Have Preschool for All?� Preschool Policy Ma ers 6, 2004. Barros, Ricardo Paes, Mirela Carvalho, Samuel Franco, Eduardo Pádua, and Andrezza Rosalém. “The Experience of the Municipality of Rio de Janeiro in Providing Day Care Services for Children Aged 0 to 3.� Background paper for World Bank Brazil ECD study, Rio de Janeiro, 2010. Barros, Ricardo Paes, Mirela Carvalho, Samuel Franco, Rosane Mendonça, and Andrezza Rosalém. “A Short-Term Cost-Effectiveness Evaluation of Be er Quality Daycare Centers.� Inter-American Development Bank Working Paper 239, 2011a. Barros, Ricardo Paes, Mirela Carvalho, Samuel Franco, Andrezza Rosalém, and Daniel Santos. Componentes para um sistema de monitoramento utilizando o ASQ3. Power- 81 82 A World Bank Study point presentation given on February 25, 2011, at the Ministry of Planning in San- tiago, Chile, Rio de Janeiro: IPEA, 2011b. Barros, Ricardo Paes, Pedro Olinto, Trine Lunde, and Mirela Carvalho. The Impact of Free Childcare on Women’s Labor Force Participation: Evidence from Low-Income Neighbor- hoods of Rio de Janeiro. Conference Paper, World Bank Economists’ Forum, 2011c. Bedard, Kelly, and Elizabeth Dhuey. “The Persistence of Early Childhood Maturity: International Evidence of Long-Run Age Effects.� Quarterly Journal of Economics, 2006. Benne , John. Benchmarks for Early Childhood Services in OECD Countries. Innocenti Work- ing Paper 2008–02, Florence: UNICEF Innocenti Research Centre, 2008. ———. “Early Childhood Education and Care Systems in the OECD Countries: The Is- sue of Tradition and Governance.� Encyclopedia on Early Childhood Development. July 8, 2008. h p://www.child-encyclopedia.com/documents/Benne ANGxp.pdf (accessed February 15, 2011). Berk, L. “Relationship of Caregiver Education to Child-Oriented A itudes, Job Satisfac- tion, and Behaviors toward Children.� Child Care Quarterly, 1985: 103–109. Berlinski, S, S Galiani, and Paul Gertler. “The Effect of Pre-Primary Education on Pri- mary School Performance.� Journal of Public Economics, 2009: 219–234. Black, S, P Devereux, and K Salvanes. “Too Young to Leave the Nest? The Effects of School Starting Age.� NBER Working Paper 13969, 2008. Blau, DM. “The Production of Quality in Child Care Centers: Another Look.� Applied Developmental Science, 2000: 136–148. Bobonis, Gustavo J, Edward Miguel, and Charu Puri-Sharma. “Anemia and School Par- ticipation.� Journal of Human Resources, 2006. Boulding, Carew, and Brian Wampler. “Voices, Votes, and Resources: Evaluating the Ef- fect of Participatory Democracy on Well-being.� World Development, 2010: 125–135. Boyd, Brenda J, and Mary R Wandschneider. Washington State Child Care Career and Wage Ladder Pilot Project. Phase 2 Final Evaluation Report, Pullman, WA: Department of Human Development, Washington State University, 2004. Browne, Gina, Carolyn Byrne, Jacqueline Roberts, Amiram Gafni, and Susan Whi aker. “When the Bough Breaks: Provider-Initiated Comprehensive Care Is More Effec- tive and Less Expensive for Sole-Support Parents on Social Assistance.� Social Sci- ence & Medicine, 2001: 1697–1710. Bruder, Mary Beth, Cristina Mogro-Wilson, Vicki D Stayton, and Sylvia L Dietrich. “The National Status of In-Service Professional Development Systems for Early Inter- vention and Early Childhood Special Education Practitioners.� Infants & Young Children, 2009: 13–20. Bruns, Barbara, Andrea Tokman, Amy Walter, and Carla Cutolo. Institutional Frame- work for Early Childhood Development in Chile. Policy Note, Washington, DC: World Bank, 2010. Calderini, SR, and AP Souza. “Pré-escola no Brasil: Seu Impacto na Qualidade da Edu- cação Fundamental.� Anais do XXXVII Encontro da Anpec. Foz do Iguaçu, Paraná, 2009. 1–21. Campos, Maria Malta, Jode e Fullgraf, and Verena Wiggers. “Brazilian Early Childhood Education Quality: Some Research Results.� Cadernos de Pesquisa, 2006: 87–128. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 83 Carvalho, Mirela, and others. “Validating a Brazilian version of the ASQ-3.� Draft report, 2011. City University of New York. The Educational Incentive Program (EIP). h p://www.early- childhoodnyc.org/education/EIP.cfm (accessed February 9, 2011). Cochran, Moncrieff. “International Perspectives on Early Childhood Education.� Educa- tional Policy, 2011: 65–91. Cryer, Debby, Wolfgang Tie e, and Holger Wessels. “Parents’ Perceptions of Their Chil- dren’s Child Care: A Cross-National Comparison.� Early Childhood Research Quar- terly, 2002: 259–277. Cryer, Debby, and Margaret Burchinal. “Parents as Child Care Consumers.� Early Child- hood Research Quarterly, 1997: 35–58. Curi, AZ, and NA Menezes-Filho. “A relação entre educação pré-primária, salários, es- colaridade e pro�ciência escolar no Brasil.� Estudos Econômicos (São Paulo), 2009: 39(4). Department of Education & Early Childhood Development. Scholarships and Incentives. September 21, 2010. h p://www.education.vic.gov.au/careers/earlychildhood/ scholarships/default.htm (accessed February 9, 2011). Dunn, LS. “Proximal and distal features of day care quality and children’s develop- ment.� Early Childhood Research Quarterly, 1993: 167–192. Early Childhood Commission. A Reader Friendly Guide to The National Strategic Plan for Early Childhood Development in Jamaica: 2008–2013. Kingston: The Early Childhood Commission, 2009. Early, Diane, Donna Bryant, Robert Pianta, Richard Clifford, Margaret Burchinal, Sha- ron Ritchie, Carollee Howes, and Oscar Barbarin. “Are teachers’ education, major, and credentials related to classroom quality and children’s academic gains in pre- kindergarten?� Early Childhood Research Quarterly, 2006: 174–195. Early, Diane, Kelly Maxwell, Margaret Burchinal, Soumya Alva, Randall Bender, Donna Bryant, Karen Cai, Richard Clifford, Caroline Ebanks, James Griffin, Gary Henry, Carollee Howes, Jeniffer Iriondo-Perez, Hyun-Joo Jeon, Andrew Mashburn, Ellen Peisner-Feinberg, Robert Pianta, Nathan Vandergrift, and Nicholas Zill. “Teach- ers’ Education, Classroom Quality, and Young Children’s Academic Skills: Re- sults From Seven Studies of Preschool Programs.� Child Development 78, no. 2 (2007): 558–580. Engle, Patrice, Maureen Black, Jere Behrman, Meena Cabral de Melo, Paul Gertler, Lydia Kapiriri, Reynaldo Martorell, Mary Eming Young, and the International Child Development Steering Group. “Strategies to avoid the loss of developmental po- tential in more than 200 million children in the developing world.� Lancet, 2007: 229–242. European Council Conclusions. “Barcelona European Council 15 and 16 March 2002.� SN 100/1/02/02 h p://www.consilium.europa.eu/ueDocs/cms_Data/docs/press- Data/en/ec/71025.pdf, Barcelona, 2002. Eurydice. “Compulsory Education in Europe 2010/2011.� h p://eacea.ec.europa.eu/edu- cation/eurydice/documents/compulsory_education/106EN.pdf (accessed January 1, 2011). Evans, David, and Katrina Kosec. Access to Early Child Education in Brazil. Background Paper, Washington DC: World Bank, 2011. 84 A World Bank Study Felício, F, and L Vasconcellos. “O Efeito da Educação Infantil sobre o Desempenho Esco- lar Medido em Exames Padronizados.� Anais do XXXV Encontro da Anpec. Recife, Pernambuco, 2007. 1–22. Felício, F, RT Menezes, and AC Zoghbi. “The effects of early child education on literacy scores using data from a new Brazilian assessment tool.� Working Paper, 2010. Fernald, Lia, Patricia Kariger, Patrice Engle, and Abbie Raikes. Examining Early Child Development in Low-Income Countries: A Toolkit for the Assessment of Children in the First Five Years of Life. Washington DC: The World Bank, 2009. FPG Child Development Institute. Environment Rating Scales: Assessment Instruments for early childhood and child care program quality. h p://ers.fpg.unc.edu/ (accessed Feb- ruary 12, 2011). Fundação Carlos Chagas. Educação Infantil no Brasil: Avaliação Qualitativa e Quantitativa. Final Report, Inter-American Development Bank, 2010. Gibbs, Robyn, and Jenny Poski . Report on the Evaluation of the Early Childhood Centres of Innovation Programme. Evaluation Associates Ltd, 2009. Gomes, JB. “Política Nacional de Educação Infantil No Brasil: Uma Luta Contínua, Uma Política Descontínua.� 10. Gormley, William, Jr, and Ted Gayer. “Promoting school readiness in Oklahoma: An evaluation of Tulsa’s pre-k program.� The Journal of Human Resources, 2006: 533–558. Governo da Bahia. Creche Nossa Senhora das Graças. 2011. h p://www.pm.ba.gov.br/ creche2011/ (accessed February 23, 2011). Grantham-McGregor, SM, SP Walker, SM Chang, and Powell CA. “Effects of early child- hood supplementation with and without stimulation on later development in stunted Jamaican children.� American Journal of Clinical Nutrition 66, no. 2 (1997): 247–253. Gupta, N, and M Simonsen. “Effects of Universal Child Care Participation on Pre-teen Skills and Risky Behaviours.� European Association of Labor Economists Working Pa- per, 2010. Hanushek, Eric A, and L Woessman. “The Role of Cognitive Skills in Economic Develop- ment.� Journal of Economic Literature, 2008: 607–668. Hanushek, Eric A, and Steven G Rivkin. “Generalizations about Using Value-Added Measures of Teacher Quality.� American Economic Review, 2010: 267–271. Hanushek, Eric A, and Steven G Rivkin. “How to Improve the Supply of High Quality Teachers.� In Brookings Papers on Education Policy 2004, by Diane (ed.) Ravitch, 7–25. Washington DC: Brookings Institution Press, 2004. Harbison, R, and EA Hanushek. Educational Performance of the Poor: Lessons from Rural Brazil. New York: Oxford University Press, 1992. Harms, T, D Cryer, and RM Clifford. Infant/Toddler Environment Rating Scale (Rev. Ed.). New York: Teachers College Press, 2006. Harms, T, RM Clifford, and D Cryer. Early Childhood Environment Rating Scale (Rev. Ed.). New York: Teachers College Press, 2005. Haskins, Ron, and W. Steven Barne . “Finally, Obama administration is pu ing Head Start to the Test.� Washington Post, October 11, 2010. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 85 Heckman, James J, and Dimitriy V Masterov. “The Productivity Argument for Invest- ing in Young Children.� Applied Economic Perspectives and Policy 29, no. 3 (2007): 446–493. Heckman, James J, SH Moon, R Pinto, P Savelyev, and A Yavi . “The rate of return to the High/Scope Perry Preschool Program.� Journal of Public Economics, 2010: 114–128. Helburn, SW (ed.). Cost, quality and child outcomes in child care centers. Technical report, Denver: University of Colorado at Denver, Department of Economics, Center for Research in Economic and Social Policy, 1995. Honig, AS, and A Hirallal. “Which counts more for excellence in childcare staff-years in service, education level or ECE coursework?� Early Child Development & Care, 1998: 31–46. Howes, C. “Can the age of entry into child care and the quality of child care predict ad- justment in kindergarten?� Developmental Psychology, 1990: 292–303. ———. “Children’s experiences in center-based child care as a function of teacher back- ground and adult-child ratio.� Merrill-Palmer Quarterly, 1997: 404–425. ———. “Reconceptualizing the early childhood work force.� In Cost, quality, and child outcomes in child care centers. Technical report., by SW Helburn, 159–170. Denver: Department of Economics, Center for Research in Economic and Social Policy, 1995. Howes, C, DC Phillips, and M Whitebook. “Teacher characteristics and effective teach- ing in child care: Findings from the National Child Care Staffing Study.� Child & Youth Care Forum, 1992: 399–414. Howes, C, R Pianta, D Bryant, B Hamre, J Downer, and S Soliday-Hong. Ensuring Effec- tive Teaching in Early Childhood Education through Linked Professional Development Systems, Quality Rating Systems and State Competencies: The Role of Research in an Evidence-Driven System. White paper, Arlington VA: National Center for Research in Early Childhood Education, 2008. INEP. “Ensino infantil: per�l das famílias com crianças de zero a cinco anos que frequen- tam a escola.� Na Medida, 2009: 1(4): 4–9. JIS News. “Jamaica’s Early Childhood Strategic Plan Gaining Global Recogni- tion .� Jamaica Information Service, August 12, 2010: h p://www.jis.gov.jm/ news/100/24993?mode=redirect. Johns, S. “Early childhood service development and intersectoral collaboration in rural Australia.� Australian Journal of Primary Health, 2010: 40–46. Kosec, Katrina. Politics and Preschool: The Political Economy of Investment in Pre-Primary Education. World Bank Policy Research Working Paper 5647. h p://go.worldbank. org/H6GMC7AWH0, 2011. La Paro, KM, RC Pianta, and M Stuhlman. “Classroom Assessment Scoring System™ (CLASS™): Findings from the pre-k year.� Elementary School Journal, 2004: 409–426. LoCasale-Crouch, J, et al. “Observed classroom quality pro�les in state-funded pre- kindergarten programs and associations with teacher, program, and classroom characteristics.� Early Childhood Research Quarterly, 2007: 3–17. Look for the Stars. “List of Required Documents for 2-, 3-, and 4-STAR Status.� New Mexico Kids: Look for the STARS Library. September 22, 2005. h p://www.newmexi- cokids.org/#.pages.library.stars.index.htm (accessed February 13, 2011). 86 A World Bank Study Marshall, NL, CL Creps, NR Burstein, FB Glan , WW Robeson, and WS Barne . The cost and quality of full day, year-round early care and education in Massachuse s preschool classrooms. Cambridge: Wellesley Center for Women and Abt Associates, 2001. Mashburn, Andrew J, et al. “Measures of Classroom Quality in Prekindergarten and Children’s Development of Academic, Language, and Social Skills.� Child Devel- opment, 2008: 732–749. Masse, LN, and WS Barne . “A bene�t-cost analysis of the Abecedarian Early Childhood Intervention.� In Cost effectiveness and education policy, by HM Levin and McEwan PJ. 2002. Ma hews, Clark. Benchmarking Policies and Interventions in Early Childhood Development: Case Studies from Australia, New Zealand, England and Sweden. Background paper, Washington DC: World Bank, 2010. MEC. Educação Infantil, Secretaria de Educação Básica, Ministério de Educação. h p://portal. mec.gov.br/index.php?option=com_content&view=article&id=12579%3Aeducac ao-infantil&Itemid=859 (accessed March 10, 2011). ———. Orientações sobre convênios entre secretarias municipais de educação e instituições co- munitárias, confessionais ou �lantrópicas sem �ns lucrativos para a oferta de educação infantil. Brasilia: MEC, SEB, 2009. ———. Orientações sobre convênios entre secretarias municipais de educação e instituições co- munitárias, confessionais ou �lantrópicas sem �ns lucrativos para a oferta de educação infantil. Brasilia: MEC, SEB, 2009. ———. Programa Nacional Biblioteca da Escola | Ministério de Educação. 2009. h p://portal. mec.gov.br/index.php?option=com_content@view=article&id=12368@Itemid=574 (accessed February 15, 2011). Melo, Marcus Andre. “Democratizing Budgeting Decisions and Execution in Brazil: More Participatation or Redesign of Formal Institutions?� In Participatory Innova- tion and Representative Democracy in Latin America, by Andrew Selee and Enrique Peruzzo i. Baltimore: The Johns Hopkins University Press, 2009. Miguel, Edward, and Michael Kremer. “Worms: Identifying Impacts on Education and Health in the Presence of Treatment Externalities.� Econometrica, 2004: 159–217. MINED. “Subvención Escolar Preferencial.� Gobierno de Chile | Ministerio de Educación. 2011. h p://www.mineduc.cl/index.php?id_portal=29 (accessed March 5, 2011). Ministério da Saúde do Brasil. Bolsa Família na Saúde (MS-DATASUS). h p://bolsafamil- ia.datasus.gov.br/w3c/bfa_relconsol.asp (accessed February 7, 2011). Ministério de Educação. Política Nacional da Educação Infantil: pelo direito das crianças de zero a seis anos à Educação. Secretaria de Educaçào Básica, Ministério de Educação, 2006. Moço, Anderson. “Balanço do Plano Nacional de Educação (PNE) 2001–2010.� Nova Es- cola, March 2010: h p://revistaescola.abril.com.br/politicas-publicas/legislacao/ pne-plano-nacional-de-educacao-537431.shtml. Muralidharan, Karthik, and Venkatesh Sundararaman. Teacher Performance Pay: Experi- mental Evidence from India. Working Paper No. 15323, NBER, 2009. Muzzi, Mariana. UNICEF Good Practices in Integrating Birth Registration into Health Sys- tems (2000–2009); Case Studies: Bangladesh, Brazil, the Gambia and Delhi, India. Work- ing Paper, New York: UNICEF, 2010. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 87 NACCRRA. Parents’ Perceptions of Child Care in the United States—NACCRRA’s National Parent Poll: November 2008. National Association of Child Care Resource & Refer- ral Agencies, 2009. NAEHCY. “SUMMARY OF HEAD START PROVISIONS ON HOMELESSNESS AND FOSTER CARE.� National Association for the Education of Homeless Children and Youth. January 2008. h p://www.naehcy.org/dl/headstartsum.pdf (accessed March 4, 2011). NAEYC. Child-Teacher Ratio Chart. National Association for the Education of Young Chil- dren, 2008. Natenzon, PE. Efeitos da Educação Pré-Escolar no Brasil. Monograph, Departamento de Economia da Faculdade de Economia, Administração e Contabilidade da Univer- sidade de São Paulo, 2003. National Association for Regulatory Administration & National Child Care Information and Technical Assistance Center. “The 2005 Child Care Licensing Study.� Final Report, 2006. NGA. Partnering with the Private and Philanthropic Sectors: A Governor’s Guide to Investing in Early Childhood. h p://www.nga.org/Files/pdf/0806PARTNERINGEDUCATION. pdf, Washington, DC: National Governor’s Assocation Center for Best Practices, 2008. Nores, Milagros, and W. Steven Barne . “Bene�ts of early childhood interventions across the world: (Under) Investing in the very young.� Economics of Education Review, April 2010: 271–282. OECD. Starting Strong II: Early Childhood Education and Care. Annex E, OECD Publishing, 2006. OHS. “Head Start Program Fact Sheet, Fiscal Year 2010.� U.S. Department of Health & Hu- man Services, Administration for Children & Families, Office of Head Start. February 2010. h p://www.acf.hhs.gov/programs/ohs/about/fy2010.html (accessed March 5, 2011). Paiva, Maria da Graça Gomes, Alessandra Schneider, Maria Lúcia Salle Machado, and Pollyana Vilela Duarte Perinazzo. “A New Look on Early Child Care and Educa- tion (ECCE) as Joint Responsibility.� Current Issues in Comparative Education, 2009: 33–41. Paraguassu, Lisandra. “Precisamos investir em qualidade.� O Estado de São Paulo, Janu- ary 5, 2011. Parks, Casey. “Program launched to improve early childhood education.� The Oregonian, May 11, 2009. Paxson, Christina, and Norbert Schady. “Cognitive development among young children in Ecuador: The roles of wealth, health, and parenting.� Journal of Human Resourc- es 42, no. 1 (2007). Pazello, ET, and RB Almeida. “O efeito da pré-escola sobre o desempenho escolar futuro dos indivíduos.� ANPEC. Salvador, 2010. Primeira Infância Melhor. PIM-Programa Primeira Infância Melhor-Noticiário. February 3, 2011. h p://www.pim.saude.rs.gov.br/a_PIM/php/pagina-Noticiario.php?n=63 (accessed February 25, 2011). Rede Nacional Primeira Infância. “Plano Nacional pela Primeira Infância.� 2010. 88 A World Bank Study Rodrigues, CG, Christine XC Pinto, and Daniel D Santos. “The impact of daycare a en- dance on Math test scores for a cohort of 4th graders in Brazil.� Working paper, 2010. Ruopp, R, T Travers, F Glan , and C Coelen. Children at the Center: Final report of the Na- tional Day Care Study (vol 1). Cambridge: Abt Associates, 1979. Ryan, Carey S, Robert McCall, Debbie Robinson, Christina Groark, Laurie Mulvey, and Bradford Plemons. “Bene�ts of the Comprehensive Child Development Program as a Function of AFDC Receipt and SES.� Child Development, 2002: 315–328. Sakai, LM, M Whitebook, A Wishard, and C Howes. “Evaluating the Early Childhood Environment Rating Scale (ECERS): Assessing Differences between the First and Revised Edition.� Early Childhood Research Quarterly , 2003: 427–445. Santarém. Cámara de Santarém Lançou Programa Eco-Escolas. June 14, 2010. h p://www. cm-santarem.pt/pracapublica/noticias/Paginas/C%C3%A2maradeSantar%C3%A 9mlan%C3%A7ouProgramaEco-Escolas.aspx (accessed February 15, 2011). Saracho, ON, and B Spodek. “Early childhood teachers’ preparation and the quality of program outcomes.� Early Child Development and Care, 2007: 71–91. Scarr, S, M Eisenberg, and K Deater-Deckard. “Measurement of quality in child care centers.� Early Childhood Research Quarterly, 1994: 131–151. Schneider, Alessandra, and Vera Regina Ramires. Primeira Infância Melhor: uma inovação em política pública. Brasilia: UNESCO, 2007. Schneider, Alessandra, Vera Regina Ramires, Maria da Graça Gomes Paiva, and Leila Almeida. “The Be er Early Childhood Development Program: An Innovative Brazilian Public Policy.� Current Issues in Comparative Education (Teachers College, Columbia University), 2009: 24–32. Schweinhart, LJ, and DP Weikart. “Why curriculum ma ers in early childhood educa- tion.� Educational Leadership, 1998: 57–60. Schweinhart, LJ, HV Barnes, and DP Weikhart. “Signi�cant Bene�ts: The High/Scope Perry Pre-school Study Through Age 27.� In Child Welfare: Major Themes in Health and Social Welfare, by Nick Frost, 9–29. Abingdon, UK: Routledge, 2005. Sharp, C. “School Starting Age: European Policy and Recent Research.� 2002. Shonkoff, Jack P, and Deborah Phillips. From neurons to neighborhoods: the science of early child development. Washington, DC: National Academy Press, 2000. Silva, Maria do Pilar Lacerda Almeida. “Políticas Sociais para a primeira infância: Em- preendedores Capeões em Investimento na Primeira Infância.� Brazilian Business Champions for Early Childhood Investment. São Paulo, Brazil: Commi ee for Eco- nomic Development, 2010. Silva, Verónica. Chile Crece Contigo: Política Nacional de Desarrollo Infantil Temprano—Lec- ciones de política pública para la Primera Infancia. PowerPoint Presentation, Washing- ton DC: World Bank, 2010. Snow, Catherine E, and Susan B (editors) Van Hemel. Early Childhood Assessment: Why, What, and How. Washington, DC: The National Academies Press, 2008. Souza, Celina. “Participatory Budgeting in Brazilian Cities: Limits and Possibilities in Building Democratic Institutions.� Environment and Urbanization, 2001: 159–184. Springer, Ma hew, Laura Hamilton, Daniel McCaffrey, Dale Ballou, Vi-Nhuan Le, Mat- thew Pepper, J.R. Lockwood, and Brian Stecher. Teacher Pay for Performance: Experi- mental Evidence from the Project on Incentives in Teaching. Nashville, TN: National Center for Performance Incentives at Vanderbilt University, 2010. Early Child Education: Making Programs Work for Brazil’s Most Important Generation 89 St. Pierre, Robert G, and Jean I Layzer. “Using Home Visits for Multiple Purposes: The Comprehensive Child Development Program.� The Future of Children, 1999: 134–151. Sylva, Kathy, Edward Melhuish, Pam Sammons, Siraj-Blatchford, and Brenda Taggart. The Effective Provision of Pre-School Education (EPPE) Project. Final Report, London: The Institute of Education, 2004. Teachstone. The Class Tool. 2011. h p://www.teachstone.org/about-the-class/ (accessed February 12, 2011). Thornburg, Kathy R, Wayne A May�eld, Jacqueline S Hawks, and Kathryn L Fuger. The Missouri Quality Rating System School Readiness Study. Columbia, MD: Center for Family Policy & Research, 2009. Tie e, W, J Bairrão, TB Leal, and H Rossbach. “Assessing quality characteristics of center- based early childhood environments in Germany and Portugal: A cross-national study.� European Journal of Psychology of Education, 1998: 283–298. Tout, Kathryn, Rebecca Starr, Margaret Soli, Shannon Moodie, Gretchen Kirby, and Kimberly Boller. The Child Care Quality Rating System (QRS) Assessment: Compen- dium of Quality Rating Systems and Evaluations. Washington DC: Office of Planning, Research and Evaluation, Administration for Children and Families, Department of Health and Human Services, 2010. Tribuna. “Asas da Florestania Infantil é sucesso.� A Tribuna: O Jornal de Todos os Acrianos, December 15, 2010: h p://www.jornalatribuna.com.br/MostrarNoticia.do?id=118 28&ano=2010&mes=12&dia=15. UNICEF. Investimento Criança no Brasil de 2006 a 2009. h p://investimentocrianca.org.br/ SimIC/InvestimentoCrianca.aspx (accessed February 5, 2011). ———. Plano Nacional pela Primeira Infância será lançado no dia 7 de dezembro, em Brasília. December 3, 2010. h p://www.unicef.org/brazil/pt/media_19413.htm (accessed February 15, 2011). Urzúa, Sergio, and Gregory Veramendi. “The Impact of Out-of-Home Childcare Centers on Early Childhood Development.� Inter-American Development Bank Working Paper 240, 2011. U.S. Department of Health & Human Services. “HHS Announces National Effort to Reshape and Raise Quality in Head Start Programs.� U.S. Department of Health & Human Services—HHS.gov. September 21, 2010. h p://www.hhs.gov/news/ press/2010pres/09/20100921a.html (accessed February 8, 2011). Vega, José Molinas, and Ricardo Paes de Barros. Human Opportunities for Children in Bra- zil: An Assessment with the Human Opportunity Index. Washington DC: World Bank, 2011. Walker, Susan P, Susan M Chang, Christine A Powell, and Sally M Grantham-McGregor. “Effects of early childhood psychosocial stimulation and nutritional supplemen- tation on cognition and education in growth-retarded Jamaican children: prospec- tive cohort study.� Lancet, 2005: Webappendix, 1–8. Weber, Demétrio. “Ministro quer ensino médio em horário integral.� O Globo, January 4, 2011. Whitebook, Marcy. Early Education Quality: Higher Teacher Quali�cations for Be er Living Environments—A Review of the Literature. h P//eric.ed.gov/PDFS/ED481219.pdf, Center for the Study of Child Care Employment, Institute of Industrial Relations, University of California at Berkeley, 2003. 90 A World Bank Study Whitebook, Marcy, and Laura Sakai. “Turnover begets turnover: an examination of job and occupational instability among child care center staff.� Early Childhood Re- search Quarterly, 2003: 273–293. Whitebook, Marcy, C Howes, and D Phillips. The National Child Care Staffing Study. Final Report: Who cares? Child care teaches and the quality of care in America. Washington DC: Center for the Child Care Workforce, 1990. Whitebook, Marcy, L Sakai, E Gerber, and C Howes. Then & now: Changes in child care staffing, 1994–2000. Technical report., Washington DC: Center for the Child Care Workforce, 2001. YMCA of Greater Sea le. STARS Training. 2011. h p://www.sea lemca.org/page. cfm?id=youthSTARS (accessed February 9, 2011). Yoshikawa, Hirokazu, Kathleen McCartney, Robert Myers, Kristen Bub, Julieta Lugo- Gil, Maria Ramos, Felicia Knaul. Early Childhood Education in Mexico: Expansion, Quality Improvement, and Curricular Reform. Innocenti Working Paper—IWP-2007– 03, UNICEF, 2007. Young, Mary Eming. Brazil Early Child Development: A Focus on the Impact of Preschools. No. 22841-BR, Washington, D.C.: World Bank, 2001. Zellman, Gail L, and Michael Perlman. Child-Care Quality Rating and Improvement Systems in Five Pioneer States. Santa Monica, CA: RAND Corporation, 2008. Zellman, Gail L, Michal Perlman, Vi-Nhuan Le, and Claude Messan Setodji. Assessing the Validity of the Qualistar Early Learning Quality Rating and Improvement System as a Tool for Improving Child-Care Quality. RAND report, Santa Monica, CA: RAND, 2008. Zill, Nicholas, Gary Resnick, Kwang Kim, Ruth McKey, Cheryl Clark, Shefali Pai-Sa- mant, David Connell, Michael Vaden-Kiernan, Robert O’Brien, Mary Ann D’Elio. Head Start FACES: Longitudinal �ndings on program performance. Third progress re- port, Washington DC: Research, Demonstration, and Evaluation Branch & Head Start Bureau, Administration on Children, Youth and Families, US Department of Health & Human Services, 2001. ECO-AUDIT Environmental Bene�ts Statement The World Bank is commi ed to preserving In 2010, the printing of endangered forests and natural resources. this book on recycled paper The Office of the Publisher has chosen to saved the following: print World Bank Studies and Working • 11 trees* Papers on recycled paper with 30 percent • 3 million Btu of total postconsumer �ber in accordance with the energy recommended standards for paper usage • 1,045 lb. of net greenhouse set by the Green Press Initiative, a non- gases pro�t program supporting publishers in • 5,035 gal. of waste water using �ber that is not sourced from endan- • 306 lb. of solid waste gered forests. For more information, visit www.greenpressinitiative.org. * 40 feet in height and 6–8 inches in diameter E arly Child Education: Making Programs Work for Brazil’s Most Important Generation is part of the World Bank Studies series. These papers are published to communicate the results of the Bank’s ongoing research and to stimulate public discussion. In the past 15 years, Brazil has made great strides in improving access to early child education, with an enrollment increase of more than 50 percent for both pre-school and creche (day care). Education programs for young children in Brazil have consistently been shown to have long-term positive effects on the life outcomes of participants. These programs have demonstrated positive impacts on, for example, income, length of schooling, and test scores. However, to have such impacts, pre-schools and creches must be high-quality, and few centers are rated as such, even in capital cities. Representation of the poorest and most vulnerable children among those attending pre-school and creche still lags considerably behind that of more privileged children, although poorer children stand to gain the most from early child education programs. Additionally, large disparities exist both regionally and between rural and urban populations. This book details the literature on the long-term effects of early child education and provides a comprehensive view of the regional variation, the importance of program quality, and the socioeconomic gaps in early child education in Brazil. It further examines existing public and private initiatives in Brazil, and it discusses how they can be leveraged to effectively and ef�ciently provide quality pre-school and creche care. A central aim of the book is to provide decision makers with speci�c policy recommendations to improve the quality and equity of the early child education experience in Brazil. Given the dif�culty of reaching children in remote areas and the need to expand coverage to the poorest segments of the population, Brazil must invest strategically in new centers and allocate existing spaces to the poorest people and areas. Municipal policy makers should designate public creche spaces in a transparent manner, provide guidelines to institutions, and monitor adherence to those guidelines. Teachers need guidance on the best activities to use to improve child outcomes. The use of participatory budgeting could potentially improve access and equity by involving the poor directly in the budgeting process. Increased cross-sectoral coordination could improve child welfare in cost-effective ways, and public-private partnerships could stretch existing resources further and expand coverage more quickly. World Bank Studies are available individually or on standing order. This World Bank Studies series is also available online through the World Bank e-library (www.worldbank.org/elibrary). ISBN 978-0-8213-8931-7 SKU 18931