34349 A W O R L D B A N K C O U N T R Y S T U D Y Central America Education Strategy An Agenda for Action THE WORLD BANK A W O R L D B A N K C O U N T R Y S T U D Y Central American Education Strategy An Agenda for Action THE WORLD BANK Washington, D.C. Copyright © 2005 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First Printing: June 2005 printed on recycled paper 1 2 3 4 5 07 06 05 World Bank Country Studies are among the many reports originally prepared for internal use as part of the continuing analysis by the Bank of the economic and related conditions of its developing member countries and to facilitate its dialogs with the governments. Some of the reports are published in this series with the least possible delay for the use of governments, and the academic, business, financial, and development communities. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally-edited texts. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply and judgment on the part of The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, Tel: 978-750-8400, Fax: 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-10: 0-8213-6258-5 ISBN-13: 978-0-8213-6258-7 eISBN: 0-8213-6259-3 ISSN: 0253-2123 DOI: 10.1596/978-0-8213-6258-7 Library of Congress Cataloging-in-Publication Data has been requested. Contents Preface vii Abbreviations and Acronyms ix Executive Summary xiii 1. Introduction 1 2. Three Key Regional Priorities 7 Substantial Improvements 7 Still Three Key Regional Challenges 8 The Three Challenges are Interrelated 20 The Challenge of Equity is Integral to Two Regional Priorities 23 Fiscal and Expenditure Constraints to the Achievement of the Three Regional Priorities 24 3. Constraints and Policy Actions to Achieve each Regional Priority 31 Regional Priority #1: Improving Learning through National Assessment Systems, Teacher Performance, and Teaching Methodologies 32 Regional Priority #2: Reaching Universal Primary Completion by Improving and Scaling-Up Interventions that Address the Needs of the Poor 38 Regional Priority #3: Expanding Secondary Coverage By Adopting a Comprehensive Policy Package to Address Supply and Demand- Side Constraints 48 4. The Regional Matrix 61 APPENDIXES A: Comparative Tables 69 Educational Outcomes 70 Education Expenditure 75 Teachers 79 Community-based school management 80 Factors associated to non-attendance in secondary 82 B: Secondary targets simulation results 83 C: Boxes 87 References 91 iii iv Contents LIST OF TABLES 1. Some Basic Development Indicators for the Central American Countries 2 2. Primary Gross Completion Rates 15 3. Central America Spends Most of Its Education Budgets on Teacher Salaries 28 4. Highest Level of Education Attained by Teachers in Central American Countries 33 5. Innovations in Education Service Delivery Demonstrate Great Potential in Central America 41 6. Total Education Expenditure in Secondary Education 51 7. Public and Private Enrollment Share in Secondary Education 51 8. Regional Agenda for Central American Countries 62 LIST OF FIGURES 1. Logical Framework for the Five Body Chapters of the Report 6 2. Primary Gross Enrollment Rates 1970­2000 8 3. Secondary Gross Enrollment Rates 1970­2000 8 4. Average Years of Education Attained in Central America across Age Cohorts 9 5. Income Per-capita and Repetition Rate in Primary 11 6. Private Rates of Return in Central America and Other Countries 12 7. Repetition Rate by Income Quintile 13 8. Efficiency Indicators in Urban and Rural Areas 13 9. Rates of Return by Earning Quintile 14 10. Income Per-capita and Gross Completion Rate in Primary 15 11. GER, NER and Gross Completion Rate across Income Quintiles in Primary 16 12. Rates of Return by Level of Education 17 13. The "Secondary Education Gap" 19 14. Comparative Survivor Plots 20 15. GER, NER and Gross Completion Rate across Income Quintiles in Secondary 21 16. Comparative Survivor Function Plots by Age 22 17. Average Secondary GER and GER by Quintile in El Salvador and Nicaragua 24 18. Public Education Expenditure per Student in percent of GDP per Capita, 2000 26 19. Evolution of Public Expenditure in Proportion of GDP 27 20. Public Education Spending as a Proportion of GNP, 1999/2000 27 21. Salary, Non Salary Recurrent and Capital Expenditures, 2002 28 22. Public Education Expenditure/GNP and Primary Completion Rate, 2001/2002 29 Contents v 23. Functional Disagregation of Education Expenditure, 2002 29 24. Notional and Effective Class Hours per Year 35 25. Private Spending per Student per Year 49 26. Public and Private Enrollment by Income Quintiles in Secondary 53 27. Public and Private Expenditure Per School-Age Population in Public Secondary Education 54 Preface T he main objective of this regional report is to provide an indepth diagnostic of where Central American countries stand along several education dimensions, underscoring the most urgent and serious challenges and suggesting policy options to address them. This is the first attempt of providing a comprehensive quantification of educational outcomes in four comparable Central American countries: El Salvador, Honduras, Guatemala, and Nicaragua. These countries face common challenges but can also learn a lot from each other insofar as they differ in their education performance, constraints and interventions. A comparative analysis of country interventions, such as school-based management and multigrade models, will be particularly enriching because countries can learn from each other in the design and implementation aspects of these interventions. This regional report is divided into this summary report, called Central America Edu- cation Strategy: An Agenda for Action, and a main report, called "Central America Educa- tion Strategy Paper."This summary report summarizes the main findings, conclusions, and policy options presented in more detail in the second one. The main report is available in grey cover format and on the website: http://www.worldbank.org/laceducation. The main results of this regional report were presented in a four-day workshop held in Dominican Republic on March 14­17, 2005. The workshop, sponsored by the World Bank Institute, in close collaboration with the Human Development Department of the Latin American Region, and USAID-Dominican Republic, was attended by high level technical delegations from six countries (El Salvador,Guatemala,Honduras,Nicaragua,Panama,and Dominican Republic), led by Vice-Ministers or Ministers of Education. The workshop had three main objectives: disseminate the results of the strategy paper to a regional audience (and complete the set of formal comments to the document), promote a technical dialogue on key issues for the region (secondary education, quality of education, education financ- ing,school-based management) and promote active knowledge sharing between each of the participating countries. The workshop gave rise to an agreement, signed by the Vice-Ministers and Ministers of Education, which established the creation of a regional technical network to discuss and disseminate country and regional studies on education sector issues and compare and monitor key education outcomes indicators (primary completion, secondary education coverage, test scores results and efficiency of expenditure). This report was prepared by Emanuela di Gropello under the general supervision of Helena Ribe (previous Sector Leader, LCHSD), Eduardo Velez-Bustillo (Sector Manager, LCHSD),and Laura Rawlings (new Sector leader,LCHSD).The task team also included Ilana Umansky (LCHSD) and Paula Savanti (LCHSD). Julie B. Nannucci was the administrative assistant. Peer reviewers were Manny Jimenez (EASHD) and Alec Gershberg (HDNED). The report incorporates the comments of the Ministries of Education of El Salvador, Guatemala, Honduras and Nicaragua (its four focus countries). Special thanks go to Nina Heinsohn (PRMPR),Jeff Marshall (Consultant,Stanford Uni- versity), Renan Rapalo (Director of UMCE, Honduras), and Darlyn Meza (Minister of Edu- cation, El Salvador) for their help in putting together Chapter V (of the main report) on vii viii Preface School-based Management in CentralAmerica.MikeWalton (RegionalAdvisor,LCRCE) and Ariel Fiszbein (Lead Economist, LCHSD) also provided very useful comments on that same chapter.ChapterV will also be disseminated separately to provide timely input to the current Central American debate on school-based management. Thanks also to the members of the Central America HD team, Suhas Parandekar, Car- los Rojas, Joel Reyes, Angela Demas, Andy Mason, and Manuel Salazar for support, advice and useful comments on this report draft. Finally,thanks to a special group of advisors who provided invaluable comments on edu- cation sector issues in CentralAmerica.The group includes IanWalker (ESA Consultant,Hon- duras), Cecilia De Cano (Ex-Minister of Education, El Salvador), Regina Cáffaro de Morenos (Director of PRONADE, Guatemala), Abigaíl Castro de Peréz (Director of the OEI for Cen- tralAmerica),Marvin HerreraArraya (General Secretary of the CECC),Arvelio García (Direc- tor of UNESCO for Central America) and Donald Winkler (Consultant, The World Bank). Vice President: Pamela Cox Country Manager/Director: Jane Armitage Sector Director: Evangeline Javier Sector Manager: Eduardo Velez Bustillo Country Sector Leader: Laura Rawlings Task Team Leader: Emanuela di Gropello Abbreviations and Acronyms CAS Country Assistance Strategy CCT Conditional Cash transfer programs CEB Basic Education Centers CECC Coordinación Educativa y Cultural Centro-Americana EBI Educación Bilingüe Intercultural ECAP Evaluación de Competencias Académicas y Pedagógicas EDUCO Educación con Participación de la Comunidad FDI Foreign Direct Investment IFPRI International Food Policy Research Institute LLECE Laboratorio Latinoamericano de Evaluación de la Calidad de la Educación LSMS Living Standards Measurement Survey MDG Millennium Development Goals MECD Ministry of Education (Ministerio de Educación, Cultura y Deportes) MINED Ministry of Education NEU New Unitary Schools OEI Organización de Estados Iberoamericanos PACES Programa de Asignaciones para Educación Secundaria PAES Prueba de Aptitud para Educación Secundaria PRAF Programa de Asignaciones Familiares PREAL Partnership for Educational Revitalization in the America PROHECO Proyecto Hondureño de Educación Comunitaria PRONADE Programa Nacional de Autogestión para el Desarrollo Educativo SAT Sistema de Aprendizaje Tutorial SBM School-based Management SWAP Sector Wide Approach ix Currency Equivalents (Exchange Rate Effective 09/27/2004) COSTA RICA Currency Unit = Costa Rica Colones 447.130 Costa Rica Colones = US$1 US$0.023 = 1 Costa Rica Colone EL SALVADOR Currency Unit = US Dollar GUATEMALA Currency Unit = Guatemalan Quetzales 7.9125 Guatemalan Quetzales = US$1 US$ 0.1264 = 1 Guatemalan Quetzale HONDURAS Currency Unit = Honduran Lempiras 18.420 Honduran Lempiras = US$1 US$ 0.0543 = 1 Honduran Lempira NICARAGUA Currency Unit = Nicaraguan Cordobas 16.0100 Nicaraguan Cordobas Oro = US$1 US$ 0.0625 = 1 Nicaraguan Cordobas PANAMA Currency Unit = Panamanian Balboas 1.0000 Panamanian Balboas = US$1 US$ 1.00 = 1 Panamanian Balboas xi Executive Summary Context and Objectives of the Report The six countries that comprise Central America face many similar development challenges because of their relatively small size, high vulnerability to external events, and, in most cases, political polarization and instability. Among the Central American countries, Nicaragua, Honduras, Guatemala and El Salvador face the most acute development challenges. In addi- tion to the problems mentioned above these countries face the challenges of high external debt (Nicaragua and Honduras), post-crisis reconciliation and democratic consolidation. The main objective of this regional paper is to provide an indepth diagnostic of where Central American countries stand along several education dimensions, underscoring the most urgent and serious challenges and suggesting policy options to address them.The report focuses on El Salvador, Honduras, Guatemala, and Nicaragua because these countries have common geographic, political, and economic features which make them broadly compara- ble also from the standpoint of their education sectors. They face common challenges but they can also learn from each other.A comparison with other countries is also provided to be able to put in perspective the education performance and the development path and inter- ventions applied in these four countries. Three Key Regional Priorities Much progress has occurred over the past several decades in Central American education, particularly in terms of educational access.Primary gross enrollment rates are over 100 per- cent in all countries and net enrollment rates are around 80 percent. The urban-rural gap has decreased as the result of the substantial investment that took place in the rural areas. In most countries, there has been a convergent pattern across income quintiles in terms of both primary enrollment and completion. At the same time there continue to be urgent and severe weaknesses in Central Amer- ican education systems. The report identifies three urgent priorities for improving Central American education systems. These are: (a) improving learning; (b) reaching universal pri- mary completion; and (c) expanding secondary coverage. Improving Learning. The analysis undertaken in the report confirms that the single most important educational challenge facing Central America is improving learning at all lev- els. It is imperative to improve learning for reasons of individual as well as national growth, competitiveness, and quality of life. Improving learning can directly improve cognitive skills and abilities, and help increase educational attainment, by lowering cases of dropout and repetition. The skills learned in schools also have important consequences for wages and productivity. In all four Central American countries, learning, as measured by results on national standardized exams, is low or low-intermediate. There is also no evidence of improvement over time. Other more indirect measures of learning, such as repetition and private rates xiii xiv Executive Summary of return, also indicate insufficient levels of learning. In particular, repetition in Central America is generally higher than it is in other countries of similar per capita income. Rep- etition rates are acutely high in the first grade where they reach "unofficial" levels of about 30 percent. While low levels of learning are a problem across-the-board in Central America, the problem is particularly acute for disadvantaged sectors of the population. Reaching Universal Primary Completion. The second key (and interrelated) challenge is ensuring that all children complete primary school. Completing primary is a necessary requirement for having even minimum prospects on the labor market. Universal primary completion is linked to reduced poverty and disease, stronger democracies, and economic growth. Universal primary completion is a minimum goal for of all nations and as such is one of the Millennium Development Goals (MDGs). While there has been substantial improvement in primary coverage and completion over the last several decades in Central America, primary gross completion rates are lower than 80 percent in all countries. The target 100 percent completion rate by 2015 established under the MDG goals will require major efforts by Guatemala, Honduras and Nicaragua. With the exception of Costa Rica, gross completion rates are substantially lower in Central American nations than the Latin American average as well as countries with similar per capita income. Those children who do not complete primary as well as those who never enter school in the first place come predominantly from severely underprivileged sectors of society. Expanding Secondary Education Coverage. Recently several studies have posited that a pop- ulation with full secondary education is key to boosting national competitiveness and GDP in an increasingly technological world. It is shown that secondary enrollment in develop- ing countries has been positively related to GDP levels over the past three decades and that mass secondary schooling is critical to the ability of countries to attract and benefit from trade and foreign direct investment. Supporting this argument, recent evidence on Central America shows that upper secondary education provides high private rates of return and boosts national growth. The demand for educated workers is likely to only increase in Cen- tral America with the CAFTA (Central American Free Trade) agreement because reforms to liberalize trade regimes, encourage FDI and facilitate licensing of technologies may fur- ther unleash the demand for schooling. According to recent research, the most successful examples of nations increasing their populations'education level, notably in East Asia in recent years, expanded their education systems one level at a time--first ensuring that there was broad enrollment in primary,then in secondary, and finally in tertiary. Central America, on the other hand, has not followed this approach, as most countries have as many or more individuals with tertiary than with secondary education only. Central America has a profound skills gap that corresponds with low enrollment rates in secondary education (as illustrated by gross enrollment rates of at most 60 percent and completion rates below 40 percent). These rates are well below those of comparison regions with similar per capita income. Both the transition to lower secondary, and survival rates once in the cycle, are critical issues for most countries. Finally, upper income and urban youth dominate secondary school enrollment, as illustrated by enrollment and completion rates which increase dramatically with income level. Executive Summary xv Interrelation among Challenges. While we address these three challenges--learning, pri- mary completion, and access to secondary--in separate sections of the report, we in no way wish to infer that they are unrelated challenges. In fact, they are tightly linked. In particu- lar, improvements made to education quality will help children get through primary school (reducing dropouts and repetition) and motivate them to attend secondary school. Equity. The problem of stark inequities in the education that poor and marginalized pop- ulations receive is central to two of the three regional priorities. Universal primary comple- tion will not be achieved without focusing on improving the completion of the most disadvantaged groups, which in turn, is inextricably linked to improving the quality of their learning. Additionally, expansion of lower secondary should occur prior to upper second- ary in order to ensure that large proportions of children from all socio-economic and socio- cultural groups are benefiting from the expansion and broad-based secondary schooling can be achieved in Central America. Fiscal and Expenditure Constraints to the Achievement of the Three Regional Priorities. With low income levels and high debt levels,all four CentralAmerican countries face strong budg- etary constraints in the funding of all government services and spending areas. Because of these fiscal constraints,there will necessarily be some tradeoffs among the three regional pri- orities laid out in this report. The nature of the interrelation between these priorities sug- gests an order of intervention: first, the quality of learning should be targeted; second, universal primary completion should be assured; and third, secondary education expanded in a broad-based manner.While education quality should be the first and foremost priority for all four countries, the later two priorities could be addressed simultaneously or even in reverse order in the short to medium term if there is substantial room for improvement of the grade six to grade seven transition rate.At a minimum,Guatemala and Honduras should aim at encouraging broad-based expansion in lower secondary by expanding equitable access to secondary education for the children who are now completing primary.Nicaragua, that has expanded secondary without emphasizing a broad primary completion basis and equitable transition, should now also follow this route; while El Salvador, where lower sec- ondary is more equitably distributed, should aim at expanding upper secondary. At the same time, it will be essential to envisage strategies which overcome the tradeoffs by giving increased budgetary priority in all (or at least three) of the four countries to the education sector,promote higher resource mobilization also from other sources (private sec- tor, households), and introducing measures which encourage higher efficiency in the use of resources (more effective teacher management, higher shares of non-salary recurrent costs, re-allocation across education levels within the education budget, cost-efficient delivery sys- tems, effective demand-side interventions, lower repetition, and so forth). There are existing constraints in terms of budgetary priority for education and effi- ciency in the use of public funds which will need to be addressed upfront. The main con- straints include: (a) low levels of public expenditure in education, compared to the Latin American average and to countries with similar per capita income (with the exception of Honduras); (b) insufficient budget shares for non-salary expenditure; (c) teacher com- pensation rarely or indirectly linked to teacher performance; and (d) inefficient allocation across education levels, with excessive priority given to tertiary education and insufficient to secondary. Both a resource mobilization and an efficiency effort will therefore be needed to address the three regional challenges. xvi Executive Summary Constraints and Policy Actions to Achieve Each Regional Priority Regional Priority #1: Improving Learning through National Assessment Systems, Teacher Performance and Teaching Methodologies Constraints to Learning. The report identifies a number of constraints to learning in Cen- tral America. Key constraints include: poverty, little or no parental education, social and cultural barriers; lack of a culture of assessment and evaluation; weak teacher performance and accountability; overdependence on traditional pedagogical practices; insufficient instructional time (largely due to student and teacher absences and school closings); and insufficient non-salary investment in items such as material resources and teacher preparation programs. Suggested Policy Options. While Central American countries have implemented several wide-reaching curricular, pedagogical, and management reforms, to date these reforms have not resulted in adequate gains in education quality. In response to these identified bar- riers to learning and education quality, the report details four main areas of policy inter- ventions: (a) improvement of national assessment systems; (b) enhancement of teacher performance; (c) diversification of teaching methodologies; and (d) a few demand-side interventions that could also have an impact on learning and education quality. Institutionalize assessment systems more: This includes starting evaluation at grade 1 and establishing effective dissemination mechanisms. Several innovative assessment dissemi- nation policies are currently used internationally. One example is school report cards pub- lished in local newspapers or posted as laminated posters at school entrances. Improve teacher education and preparation: This includes developing selection criteria and diagnostic tools such as accreditation exams, which may require full teacher certification systems as in El Salvador and Costa Rica; and reducing demand/supply gaps in teacher specializations. Improve teacher effort: This includes analyzing the feasibility of incentive mechanisms such as: (a) carefully designed team-based merit-pay schemes; (b) salary scales which pro- mote higher effective hours of work; (c) decentralized systems of teacher monitoring; and (d) fixed-term contracts or local authority to hire and fire teachers. Setting up effective teacher monitoring systems may be the most effective incentive to increase teacher effort and/or instructional time, possibly combined with more flexible hiring and firing mech- anisms. An analysis of school-based management with community participation in the four Central American countries suggests that empowering parents to hire, fire and mon- itor teachers results in greater teacher effort as measured by teaching hours. Diversify teaching strategies and deepen constructivist teaching methods: This includes empha- sizing classroom practices more in pre-service and in-service training and develop innova- tive ways of organizing in-service teacher support. Executive Summary xvii Design effective demand-side interventions. Improvedqualityof teachingandlearningthrough these interventions could be strengthened by complementary demand-side interventions, which include publicity campaigns addressing learning quality and policies that directly impact children's background characteristics, such as their health, nutrition, and access to clean water and safe homes. Regional Priority #2: Reaching Universal Primary Completion by Improving and Scaling-Up Interventions that Address the Needs of the Poor Constraints to Completion. Primary completion rate is still too low and those children who are not finishing primary come almost exclusively from poor, rural, and indigenous popu- lations. Low completion of the poor can be related to poor learning. Poor learning is in turn related to supply-side constraints which are shown to be particularly acute for the poor (the poor have less access to skilled teachers, textbooks and good infrastructure) and to the same characteristics of poverty which make a good learning experience more difficult. Poverty also has a direct impact on completion (that is not only through poor learn- ing) by discouraging enrollment, timely entrance and permanency in school. Poor families have to spend a substantial share of their annual income to keep several children in school. Existing Interventions That Target Improving the Educational Chances of the Poor. A key con- clusion of the report is that there is no magic bullet in terms of improving schooling and com- pletion for the poor. Many interventions have resulted in improvements in certain contexts, but no single intervention can promote all the changes in the teaching-learning environment that are currently creating barriers to primary completion for underprivileged children. The report reviewed experiences with multigrade schooling,community-based management and bilingual schooling, reporting their positive outcomes but also constraints and limitations. The main results and constraints of these programs are outlined below. Multigrade schooling: Multigrade schools are schools in which one teacher teaches more than one grade level simultaneously.All four Central American countries have multigrade programs, the most well known being the New Unitary School (NEU) in Guatemala.There is evidence that when successful,students in multigrade schools can outperform traditional rural students and even urban students. However, the model has shown to be constrained by the continued dependence on traditional pedagogical methods and insufficient ad hoc teacher training and materials. Community-based school management: School-based management (SBM) is a decen- tralization mechanism that shifts certain decisionmaking powers to the school level, emphasizing the role of community and parental management of school affairs. Like multigrade schooling,school-based management mechanisms exist in all four Central American countries.1 These reforms have led, in some cases, to greater community empowerment and teacher effort, resulting in: (a) a better use of the existing limited capacity of teachers and schools;2 (b) higher coverage in rural areas;3 (c) somewhat 1. EDUCO in El Salvador, Centros Autonomos in Nicaragua, PRONADE in Guatemala and PROHECO in Honduras. 2. More so in El Salvador, Honduras and Guatemala than in Nicaragua. 3. Particularly in Nicaragua and El Salvador. xviii Executive Summary better student flows; and (d) learning outcomes at least as high as in traditional schools.4 However, the SBM model ranks poorly in terms of teacher education and experience, on the adoption of innovative teaching methodologies and on the sub- stantive teacher involvement in the schools. These "missing pieces" limit its impact on educational achievement and other education outputs and outcomes. Bilingual schooling: All countries, except El Salvador, have by now established this type of program although bilingual education is most advanced in Guatemala. Bilingual education can have a positive impact on student performance, as illus- trated by the case of Guatemala. However, it is still in an incipient stage in Central America and is often constrained by lack of teacher preparation and insufficient or inappropriate classroom materials. Additionally, the experience with demand-side interventions has been somewhat limited in Central America.Conditional Cash Transfer (CCT) programs have been applied in Hon- duras and Nicaragua, with varying levels of success. In general, they had a positive impact on school attendance and early dropout in both countries, but their effect on enrollment was more ambiguous. Suggested Policy Options. The main policy recommendations for improving completion are to strengthen and improve the existing supply-side interventions, while also applying effective demand-side policies and broader interventions. Improving community-based school management: The impact of community-based school- ing on student flows and learning outcomes could be greatly enhanced by a set of specific actions, which largely aim at setting up the conditions for pedagogical improvement, improved management and empowerment at the local level, and sustainability of the models. Two possible suggested directions for pedagogical improvement include expand- ing community empowerment, by involving parents in pedagogical decisions, with the adequate support; or enhancing teachers' empowerment, by giving them higher peda- gogical autonomy and support, and attracting more skilled teachers through a better alignment of teachers' benefits and professional development opportunities with those of the traditional system. Improving management and empowerment will include improving parental support, disseminating information on school performance to all education actors,and improving transfer formulas.Finally,ensuring sustainability requires the urgent implementation of strategies to include teachers and teacher's unions in the reform process. Improving multigrade schooling: The potential of multigrade schooling could be more fully realized by making sure that teachers are specifically and adequately trained and experi- enced in multigrade teaching techniques and that they obtain the on-going support nec- essary to successfully apply these methods. Improving bilingual education: Expansion of these programs may require a significant increase in state financing, as implementation requires teacher preparation as well as cur- riculum and material development for new ethno-linguistic groups. 4. Again, this is less true for Nicaragua. Executive Summary xix Other measures: This would include analyzing further the scope for conditional cash trans- fers, abolishing fees in primary and increase public spending in education, while ensuring sufficient learning materials by improving budget allocation towards higher shares of non- salary recurrent costs. Sectorwide approaches in the use of external funds will also be key to ensuring higher primary completion by setting up a common framework for donor inter- vention with common outcome indicators and expenditure targets. It is hoped that Educa- tion For All Fast Track Initiative will maximize the impact of external funds in the subsector by helping design and implement a consistent and coordinated framework for intervention in the primary sector. Regional Priority #3: Expanding Secondary Coverage by Adopting a Comprehensive Policy Package to Address Supply and Demand-Side Constraints Constraints to Secondary Enrollment. An analysis of the reasons for non-attendance in the 13­15 and 15­18 age range undertaken from household surveys reveals four main types of constraints that prevent prospective students from enrolling in secondary schools. These are: (a) work-related reasons (need to work); (b) economic-related reasons; (c) learning- related reasons; and, to a lesser extent, (d) lack of supply (of schools and/or teachers). Eco- nomic-related reasons can be related to the private costs of schooling which are much higher at the secondary than at the primary level. The Case for Public Intervention. East Asian countries offer insight into the important role of governments in ensuring that the evolution of education systems in developing countries sup- ports national growth and competitiveness.Independent of the important impact on national growth and competitiveness, other Latin American, Asian, and OECD countries enjoy wide support for mass schooling for reasons of equality of opportunities and social cohesion. In Central America, like elsewhere, public intervention will be necessary to promote the neces- sary broad-based growth in lower secondary for at least two main reasons: (a) social rates of return can be assumed higher than private ones and, therefore, without public interven- tion, a socially suboptimal equilibrium level of lower secondary education will be reached; and (b) on pure equity grounds, public intervention is necessary to encourage higher partic- ipation of low-income groups in secondary education. Such an approach has not generally been adopted in CentralAmerica,where public inter- vention is still too limited and selective. Our limited information suggests that private financ- ing represents between 50 and 80 percent of total secondary education spending in Central American countries, which indicates low priority for secondary education in the national budget.Additionally,public enrollment in secondary is composed primarily of students from the middle-upper income quintiles (with the exception of El Salvador) largely because of the high opportunity and privates costs faced by middle-lower income groups. Greater private financing of upper secondary may be an appropriate option in Guatemala, Honduras and Nicaragua. In Guatemala, Honduras, and Nicaragua ministries of education are not likely to have enough funds for the expensive tasks of expanding both lower and upper secondary in the short term. Because of this these countries may wish to explore more private financing optionsinuppersecondary.5Asalargerandmorediversepool 5. Higher private rates of return in upper secondary education may make up for the higher opportu- nity costs at that level. xx Executive Summary of students graduate from lower secondary schools in El Salvador, there is in contrast more scope for larger public intervention in upper secondary in that country. Suggested Policy Options. When analyzing specific policy options to increase secondary coverage, the report concludes that only a comprehensive policy package will successfully overcome the range of constraints facing secondary enrollment. The main elements of the policy package are outlined below. Distance education and other flexible delivery mechanisms can be very promising, although they need to be more fully assessed. Flexible delivery mechanisms demonstrate good potential to minimize work,economic,and supply-side constraints.Typically,these programs are offered outside the traditional school day, require fewer hours to complete, and are geared more toward skill rather than academic training.They would be particularly advisable in Honduras (due to the existing fiscal constraints), as well as in Nicaragua and Guatemala, where they would constitute a valuable tool to improve coverage with equity. Programs such as Edu- catodos and SAT have had some success in improving secondary enrollment, particularly in marginalized areas, although they have not yet received sufficient assessment. Demand-side subsidies and other interventions to expand coverage may be needed. At the sec- ondary level demand-side subsidies could resolve some of the economic barriers to sec- ondary, and also compensate for the opportunity costs of schooling for low-income students, easing the work constraint. There are few examples of conditional cash transfers in Central America. There is similarly little use of scholarships, with the exception of two scholarship programs set up in Costa Rica and El Salvador. The lessons extracted from the scholarship program of El Salvador should be used for the design and implementation of similar programs in other Central American countries. Mixed schemes of public financing/private delivery have good potential if there is a cost- efficient private sector. In urban areas, an analysis undertaken on El Salvador shows that there is substantial potential for promoting mixed schemes of public financing-private delivery, taking advantage of idle capacity and higher cost-efficiency of the private sector. This option has the potential of addressing the supply constraint in urban areas and eas- ing the learning constraint for middle or low-income students. The sizable role of the pri- vate sector in secondary education in all countries makes this option a valid one. Basic education centers have good potential to improve the transition to secondary. Basic edu- cation centers that integrate primary with lower secondary in the same school, have good potential to help support the transition to lower secondary.This innovation has been piloted in El Salvador and Honduras with mixed results so far (more favorable in El Salvador than in Honduras). Basic education centers are typically introduced along with legislation estab- lishing a compulsory basic education cycle as well as a more integrated curriculum. They could therefore address both supply-side constraints, in urban and rural areas, and quality constraints of lower secondary education by improving the relevance of that subcycle. Basic education centers should be further encouraged in Honduras and developed in Nicaragua and Guatemala, using the lessons of the El Salvador case. Learning improvements are essential. Improving both access to secondary and the quality of secondary are fundamental to expanding enrollment in this cycle. Learning/quality improve- Executive Summary xxi ments specific to the secondary level such as ensuring that teachers are trained in the second- ary level subject that they teach will be essential.Additionally,urgent interventions are needed to drastically improve the quality and relevance of lower secondary, including more innova- tive pedagogical approaches and movement towards a basic education cycle. Budget shares for secondary education and public financing for education should be increased in El Salvador, Nicaragua and Guatemala. A simulation analysis for El Salvador suggests that this country will need to spend between 0.9 and 1 percent more of its GDP by 2015 on secondary education to achieve universal coverage in lower secondary (that is, also includ- ing full transition of the poor) and wide coverage in upper secondary (gross enrollment rate of 70 percent). To be able to address the multiple imperative challenges of its educa- tions sector, that is also including across the border quality improvements and universal completion, El Salvador would need to increase public education expenditure in terms of GDP to levels close to the Latin American average (about 4.5 percent) by 2015. No specific simulations have been done for Nicaragua and Guatemala. Both will, how- ever, need to increase their secondary education budget share and their overall education budgets as a percentage of GDP to achieve their education targets. In both countries, there may be scope for decreasing the public subsidy to tertiary education. Budget shares for secondary education should be increased and efficiency of public financing in education drastically improved in Honduras. A budget simulation analysis for Honduras suggests that this country will need to spend between 0.3 and 0.5 percent more of its GDP by 2015 on secondary education to achieve broad-based coverage in lower secondary (gross enrollment rate of 70 percent). A substantial expansion in upper secondary (gross enroll- ment rate of 50 percent) would require another 0.5 percent of the GDP. Overall, Honduras will probably need an additional 1 percent of its GDP to finance its education reform by 2015. Drastic efficiency improvements will therefore be needed at all levels to address these education challenges within the already substantial education budget. CHAPTER 1 Introduction Context. The six countries that comprise Central America face many similar development challenges because of their relatively small size, high vulnerability to external events, and, in most cases, political polarization and instability. Despite growing GDP and social ser- vice provision, poverty remains an urgent problem in most of Central America, as do high levels of social and economic inequality. Among the Central American countries, Nicaragua, Honduras, Guatemala, and El Sal- vador face the most acute development challenges. In addition to the problems mentioned above these countries face the challenges of high external debt (Nicaragua and Honduras), post-crisis reconciliation, and democratic consolidation. Costa Rica and Panama tend to have considerably higher per capita income, better social indicators, and greater political stability (see Table 1). These two countries will only be dealt with marginally in this report. Timeliness of a Regional Education Agenda. In most Central American countries, a broad national consensus has developed around the importance of education. Most countries increased education spending during the 1990s, a result of both governments and civil soci- ety identifying education as a key priority. Gross enrollment rates have been increasing substantially at the primary level, and are now above 100 percent in all countries. Inequal- ity in educational access has, in turn, been declining as more rural children enroll in school. Most countries have introduced innovations to address education weaknesses including curricular reforms, national assessment systems, management decentralization, and multi- grade schooling. While the results of these interventions have often been positive, many challenges are still ahead: many children never complete primary school; there is a serious secondary edu- cation enrollment gap in all countries; repetition rates remain high in primary schools; the quality of schooling is poor at all levels; inequities between urban and rural, wealthy and 1 2 A World Bank Country Study Table 1. Some Basic Development Indicators for the Central American Countries Latin America Costa El and the Panama Rica Guatemala Nicaragua Honduras Salvador Carribean World Per Capita GNI 4020 4070 1750 720 920 2080 3580 5120 Life Expectancy 75 78 65 69 66 70 71 67 Literacy Rate (pop. above 15) 92 96 69 77 76 80 89 Under 5 Infant Mortality (per 1000 live births) 25 11 49 41 42 39 34 81 Gini Coefficient 56.8 47 59.6 50.3 53.7 49.9 Notes: Most of this data is from 2002, the Gini coefficients are slightly older. Source: WDI, World Bank. impoverished, and ethnic and linguistic groups are prevalent; educational spending and spending efficiency are generally low; and national assessment systems and community- based management models still need to be consolidated in all countries. Cognizant of the importance of education for national growth and competitiveness on the one hand, and poverty reduction on the other, the Central America countries are cur- rently in the process of defining a forward-looking agenda to address education challenges over the next decade. Additionally, Central American countries have also organized them- selves regionally through a number of Latin American forums, in particular the Inter- American Conference on Education, the Ibero-American Conference on Education, and a permanent Secretary, the Central American Education and Culture Coordination (CECC), based in Costa Rica. Institutionalizing regional cooperation in education has allowed the countries to define a common agenda for the sector. This working agenda led to the estab- lishment of 14 main working areas, which include primary completion, teacher education and professional development, secondary education, standards and evaluation, and over- arching priorities, such as the achievement of improved educational quality and equity in the region. In this context, a Central American education strategy paper is timely for Cen- tral American countries. A Central America agenda is also timely for the World Bank itself. After a decade of strong investments in the region, the Bank is moving towards sectorwide approaches to its operations in order to better address the sectorwide nature of many development chal- lenges. This new approach, called SWAP (Sectorwide Approach), has five main focuses: a focus on policy, a focus on results, a focus on sustainability, a focus on donor coordi- nation and a focus on in-country dialogue. The educational challenges that the region confronts--primary completion rates, secondary coverage, education quality, and educa- tion financing--are sectorwide challenges and we hope to offer policy suggestions that sim- ilarly address them in this light. By helping define a broad sectoral regional education agenda, which takes into account these five "pillars," we hope this report will be a timely contribution to the Bank's interventions in the region. Central American Education Strategy 3 Why are we comparing El Salvador, Guatemala, Honduras and Nicaragua? We have seen that these countries share common geographic, political, and economic features. These common characteristics make these countries broadly comparable and, therefore, partic- ularly meaningful to also compare them across characteristics of their education sectors. First, these countries also share common challenges in their education sectors, in terms of learning, primary completion and secondary coverage, public expenditure constraints and reforms/interventions that they have applied. In particular, we will see that--due to the various mechanisms of regional and international cooperation and policy dialogue mentioned above along with the influence of international organizations--many educa- tion interventions in one Central American country are closely related to interventions in the other Central American countries (such as school-based management and multigrade models). From that perspective, a country comparison will be an effective way of defining a common regional agenda for these countries, suggesting policy options to address remaining challenges and constraints. Second, however, we will see that these countries also differ in their education perfor- mance, constraints and interventions. A comparison of the differences will then be par- ticularly enriching as it will help us define a well grounded regional agenda, through a comparison of good and bad practices, and, when necessary, fine tune the agenda to coun- tries' specificities. Additionally, establishing differences will also make it clear how a coun- try can learn from the other(s). We illustrate further these points below. We will, for instance, find that El Salvador is performing better than the other countries in most key areas, that Nicaragua has higher secondary coverage than Honduras starting from a similar primary completion rate, or, still, that Nicaragua has the worse equity out- comes together with Guatemala. It will then be very interesting to assess possible differences in: (a) the education development paths followed by each of the countries (secondary cov- erage versus primary completion, rapid secondary expansion versus broad-based one, and so forth); (b) the main constraints that they face (type of public expenditure constraints, magnitude of cultural barriers, and so forth); and (c) key interventions in the areas of learn- ing, coverage and internal efficiency (quality-enhancing interventions such as teachers' incentives programs and pedagogical reforms, flexible delivery mechanisms to increase cov- erage in rural areas, and so forth), which can explain these differing outcomes. This will help establishing a well grounded regional and country agenda, while also pro- viding useful insights on what a country can learn from the other. For example, it will turn out that all countries can learn a lot from El Salvador on how to achieve broad-based expan- sion in lower secondary education and manage teachers more efficiently; or, still, that Hon- duras can learn a lot from El Salvador in spending its education budget efficiently; or, finally, that all countries can learn a lot from Guatemala on how to address the needs of indigenous populations and from Nicaragua on how to deal with primary fees. Similarly, and related to this, there will also be value-added in analyzing common interventions (such as school- based management and multigrade models) in a comparative perspective. We will, for instance, see that countries can learn a lot from each other in the design and implementation aspects of the school-based management models; and, as we undertake these comparisons of promising and less promising practices, we will also be able to provide more precision and concreteness to the regional agenda. Comparison with Other Countries and Regions. In any case, whatever the value of this Cen- tral American comparison, it will also be necessary to look outside these four countries to be 4 A World Bank Country Study able to put in perspective their education performance, the constraints that they face and the development path and interventions that they have applied. This will provide further guid- ance on the policy options that can help Central American countries tackle their challenges. Comparisons with countries with similar economic development levels will be particularly useful, but broader comparisons with Latin American and countries in other continents will also provide some guidance on future education challenges and strategies to address them. Objectives and Audience. The main objective of this regional report is to provide an indepth diagnostic of where Central American countries stand along several education dimensions, underscoring the most urgent and serious challenges and suggesting policy options to address them. While the report attempts to take a broad view of the education sector, it was not pos- sible to give equal attention to all education levels or challenges. In particular, it should be mentioned that the report does not focus on tertiary education and training and innovation systems. It also deals only marginally with pre-primary education and cultural barriers to education. These are all important subjects and we hope that other reports will deal with them in depth. The primary audiences of this report are Central American Ministers of Education and their staffs and World Bank managers and project leaders. For Central American education specialists and policymakers we hope that this regional report will: (a) provide Central American countries with a broad set of comparable educa- tion outcome indicators to help more thoroughly assess their relative strengths and weak- nesses; (b) further the evaluation of some of the regional and national priorities that have been established; (c) open up regional debate on a number of issues and topics related to the achievement of these priorities, and policy options to achieve them; (d) help identify new regional and country specific priorities by learning from common problems and suc- cesses and/or from the countries' individual experiences; and (e) provide an indepth analy- sis of school-based management reforms. On the Bank's side, it is hoped that the report will contribute to define future education priorities of the region, providing inputs for the Country Assistance Strategies (CASs) and support the design and implementation of future investment operations and SWAPs. New CASs will soon be under preparation for Guatemala and El Salvador and existing CASs will be updated for Honduras and Nicaragua. Bank projects and operations are active through- out Central America. Investment operations are currently under way in six Central American countries and two more are in the pipeline. At the same time, a SWAP, combining a Poverty Reduction Strategy Credit and an Education For All-Fast Track Initiative, is under prepa- ration in Honduras and Nicaragua. Although all these interventions are prepared on a sin- gle country basis, it is expected that they will benefit from the lessons and strategic guidelines developed in this regional paper. For instance, it is expected that the design of the upcom- ing secondary education project in El Salvador will benefit from the strategic guidelines to improve coverage at the secondary level that are developed in this report. It is also expected that the strategic guidelines established to improve quality will be beneficial to the design of the current and future Education For All-Fast Track Initiatives. The summary report is designed to give an overall picture and summary of findings for high level Ministry of Education and World Bank staff. The main report is intended more for policy-makers, technical staff, and those with interest and expertise in the vari- ous areas of study that inform our final conclusions. Central American Education Strategy 5 Methodology. This regional report builds largely from recent World Bank education sector diagnostics on the Central American countries. These sector works can be grouped into three main categories: (a) country specific education strategy papers; (b) other country spe- cific notes part of broader sector works (poverty assessments, country economic memo- randums, and so forth); and (c) a set of papers looking at the impact of different teacher incentives on teacher and student performance, produced as part of a recent regional pro- ject.6 Recent consultant reports (prepared for the World Bank) on different areas of educa- tion delivery are also used, as well as studies undertaken by other international organizations (PREAL, UNESCO, OEI), work done by the Ministries of Education of Governments of the Central American countries themselves, and peer-reviewed academic scholarship. A special effort was made to obtain similar indicators across countries by making use as much as possible of recent household surveys (including LSMS when available) and avail- able databases of the Ministries of Education. The use of the most recent household surveys is particularly important in estimating similar equity outcomes, spending indicators and teachers' characteristics. A special effort is also made to use countries' existing educational achievement data to identify factors associated with educational quality, achievement and school effectiveness, as well as to assess the impact of school-based management. Chapter V of the main report will go particularly deep in the analysis of school-based management, by gathering and analyzing comprehensive comparable information on the origin, charac- teristics, impact and weaknesses of four different Central America models Structure of the Report. The report is divided into two volumes (this summary and the main report available on the web; World Bank 2004e), which, together, should make it pos- sible to establish a "roadmap" for the education sector in Central America or, more exactly, the four poorest countries of Central America (El Salvador, Guatemala, Honduras, and Nicaragua). Information on Costa Rica is, at times, presented for comparative purposes. This summary report, provides, beyond this introduction, an organized and detailed summary of the main findings, conclusions, and policy options developed in the report, presenting, therefore, an action agenda for Central America. The discussion, and accom- panying matrix, addresses the region as a whole. The report attempts to highlight common trends across the four countries and suggest regional policy options, at the same time also showing how countries can learn from each other, based on commonalities between coun- tries as well as comparisons of differences across countries. The main report, "Central America Education Strategy Paper" (World Bank 2004e) includes the five body chapters of the report (Figure 1 illustrates the structure and logic of the main report). The first chapter provides much-needed basic comparative information on educational attainment, coverage, internal efficiency, quality and equity (by quintile, gender, population group, and so forth) across Central American countries, highlighting differences among the countries but also identifying where Central America stands as a region. The sec- ond chapter concentrates on education quality for both primary and secondary levels, review- ing the main determinants of learning in the Central American countries, with special emphasis on the educational achievement of the poor and disadvantaged. The chapter also presents policy options to improve learning. Chapter three puts special emphasis on 6. This project is the Latin American Teacher Quality and Incentives Study whose main report and complementary volume of background studies is forthcoming. 6 A World Bank Country Study Figure 1. Logical Framework for the Five Body Chapters of the Main Report Quality: Expenditure: V) Constraints and Options (chap. (chap. II) [Inputs for Quality] magnit impact and Educational [Cross-sector ude, Performance: Coverage, allocation Inputs] Efficiency, Quality, Equity (chap. I) Inputs] characteristics [Outcomes] and [Cross-sector incidence Secondary management: Education Coverage: (chap. Constraints and Options IV) (chap. III) [Inputs for School-based Coverage] secondary education in Central American countries presenting information on constraints to coverage, discussing why expansion is important, and suggesting policy options to increase coverage. The fourth and fifth chapters concentrate on two cross-cutting interventions, edu- cation expenditure and school-based management, which, if well used, can potentially have a positive impact on several areas of education delivery, including quality, efficiency, and cov- erage. Specifically, chapter four reviews where the Central American countries stand in rela- tion to indicators of magnitude, efficiency and equity of education spending, including, when relevant, analyses of external public funds and private (household) expenditure by country, and provides policy options to improve expenditure in relation to these dimensions. Chapter five reviews Central American reforms in decentralized management, and, in par- ticular, the experiences of community-based school management programs implemented in El Salvador, Nicaragua, Honduras, and Guatemala. It analyzes the main characteristics of these models, their impact on quality, coverage, and efficiency outcomes, and suggests spe- cific policy options on how to move forward. The remaining sections of this summary report introduce the three key regional pri- orities and then go through each regional priority individually discussing the challenges present and policy options ahead. The report ends with a matrix that depicts these priori- ties, constraints and policy options; and with an annex which includes a set of compara- tive tables and boxes. CHAPTER 2 Three Key Regional Priorities Substantial Improvements Much progress has occurred over the past several decades in Central American educa- tion, particularly in terms of educational access. Figures 2 and 3 show steady and steep improvements in primary and secondary gross enrollment rates in all four countries. Educational attainment has also increased dramatically (see Figure 4). These gains in enrollment and attainment are primarily the result of expansion of education systems that, prior to expansion, served primarily the children of the elite. These gains have meant, therefore, that the gap between rich and poor and urban and rural students has declined somewhat. Primary gross enrollment rates are over 100 percent in all countries and net enroll- ment rates are around 80 percent.7 These access gains are important (although the GER rates being over one 100 percent and the difference between the GER and NER tell us that significant numbers of over-age youth are in primary schools hinting at repetition, dropout and re-entry habits, and late enrollment problems in the systems). The urban-rural gap has decreased as the result of the substantial investment that took place in the rural areas. Over the past decade Central American countries have expanded rural education, especially through autonomous, community-managed schools. For the most part, and with the notable exception of Guatemala, there is satisfactory access to primary schools in rural areas. Similarly, in most countries, there has been a convergent pattern across income quintiles in terms of both primary enrollment and completion. 7. Using both official enrollment data and household survey data (see Tables A.1 and A.2). 7 8 A World Bank Country Study Figure 2. Primary Gross Enrollment Rates 1970­2000 120 110 100 El Salvador 90 Guatemala 80 Honduras Nicaragua 70 60 50 1970 1980 1990 2000 Source: EdStats. Figure 3. Secondary Gross Enrollment Rates 1970­2000 60 50 40 El Salvador Guatemala 30 Honduras 20 Nicaragua 10 0 1970 1980 1990 1995 2000 Source: EdStats. Still Three Key Regional Challenges The accomplishments of the past several decades in education need to be acknowledged. At the same time there continue to be urgent and severe weaknesses in Central Ameri- can education systems. Educational inefficiency is high, children are repeating grades, dropping-out, and learning little all of which hurt children's future prospects and increase the costs of schooling for the government; and few youth are making it through secondary schools and university, particularly few poor children. This report identifies three urgent priorities for improving Central American education systems. These are: Central American Education Strategy 9 Figure 4. Average Years of Education Attained in Central America across Age Cohorts 9.00 8.00 7.00 6.00 5.00 Schooling of 4.00 3.00 Years 2.00 1.00 0.00 El Salvador Guatemala Honduras Nicaragua Countries 25 year cohort 40 year cohort 55 year cohort Source: El Salvador, EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2003; Nicaragua, EMNV 2001. (a) improving learning; (b) reaching universal primary completion; and (c) expanding secondary coverage. Improving Learning The analysis undertaken in the report confirms that the single most important educational challenge facing Central America is improving learning at all levels.8 It is imperative to improve learning for reasons of individual as well as national growth, competitiveness, and quality of life. As is, most children are completing their educational experience without gain- ing sufficient skills to allow them to earn a good living and have an adequate standard of liv- ing. Poor learning outcomes are also a root cause of school abandonment, grade failure, and repetition; and snowball into larger and larger problems as students attempt to move to sub- sequent grades without the necessary skills and knowledge, resulting in more repetition, dropout, and absenteeism at higher levels regardless of the quality of the later teachers. Improving learning can, then, directly improve cognitive skills and abilities, and help increase educational attainment, by lowering cases of dropout and repetition. The skills learned in schools also have important consequences for wages and productivity. Higher quality will make workers more productive, increase the rate of return to education and provide an incentive for further private investments in education. 8. Education quality is central to the Millennium Development Goals as well as several other regional and global development priorities. 10 A World Bank Country Study Learning Results on Standardized Exams are Low. In all four Central American countries, learning, as measured by results on national standardized exams,9 is low or low-interme- diate. There is also no evidence of improvement over time (Table A.3).10 Most third grade Hondurans, for example, cannot identify characters or ideas in what they read nor can they make simple grammar elaborations. As predictable, results are sig- nificantly worse for schools serving disadvantaged populations. In El Salvador, one in four third graders can not identify the main message of a basic text. Results in grade nine are generally even lower (Schiefelbein 2004). Of the Central American countries only Honduras participated and was included in the results of the 1997 Latin American Laboratory Assessment.11 This international com- parison shows that even taking into account Honduras' low per capita income, Honduran students perform lower than almost all other Latin American countries. Other interna- tional exams have, in turn, shown that Latin American countries tend to perform poorly compared with wealthier nations.12 Other More Indirect Measures of Learning Also Indicate Insufficient Levels of Learning. Age-by-grade distortion and repetition. For every year comprising primary and secondary schooling between 50 and 80 percent of children are not the appropriate age. This trans- lates into an average age between 5 and 20 percent higher than the official age for the grade. The average age at grade 1 varies between 7.3 years and 8.4 years and the average age at grade 6 is higher than 12.8 years. Age-by-grade distortion is caused by both late (or, much less so, early) entrance to first grade (a problem particularly acute in Guatemala and Honduras) and high repetition rates in primary. Repetition rates are acutely high in the first grade where they reach "unofficial" levels of about 30 percent (see Table A.4). Repetition in Central America is generally higher than it is in other countries of similar per-capita income. It should be noted that through- out Latin America repetition rates tend to be higher than in other parts of the world. Several hypotheses attempt to explain Latin America's high repetition rates.13 These include: (a) overly-demanding curricula and standards; (b) overly-full schools that use lower grades as filters; (c) low teaching quality; and (d) student background characteris- tics, such as a language barriers or lack of academic support at home, that make it difficult for students to pass. While each hypothesis is certainly partially true, the last two hypothe- ses appear to explain the phenomenon best. At the root of all the hypotheses is the funda- mental fact that students are not learning enough in school, either due to supply or to demand-side constraints.14 9. Which is a somewhat reductive way of measuring acquired skills. First, there are cases, such as Honduras, where the exams themselves are not fully aligned with the curricula taught. Second, broader measures of the quality of life would be a better proxy of the benefits brought by improved skills. Alter- natively, earnings potential may also be a good indicator of learning. 10. National assessment systems are new in Central America, the oldest being in Costa Rica (since 1988) and El Salvador (since 1994) and the newest in Nicaragua (since 2002). 11. The UNESCO/LLECE first Latin American exam undertaken in 1997. 12. We reviewed, for example, the results of the 2000/2001 PISA exam where Chile, one of the best performers at the Latin American level, did quite poorly. Obviously care needs to be taken in comparing the Latin American regional exam with the PISA exam. 13. This is based on Edwards (2004). 14. As also shown by Edwards (2004) for Guatemala and AED (2004) for El Salvador. Central American Education Strategy 11 Figure 5. Income Per-capita and Repetition Rate in Primary, 2000­2002 16 GUA 14 12 10 Rate NICA HON 8 ES CR 6 Repetition 4 2 0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Income Per-Capita Source: Table A.4 and WDI. Rates of return. As mentioned above, the skills learned in schools also have important con- sequences for wages and productivity. Students who graduate with greater and more appro- priate skills and knowledge will be more productive workers and will earn higher wages on the labormarket.Ratesofreturnprovideaforward-lookingmeasurementofthequalityofschool- ing (although they are also subject to the contextual demand for skilled labor as well as the overall skill level of the population). A comparable calculation of private rates of return to edu- cation in our Central American countries,15 indicates that an additional year of education gen- erates an approximate 10 percent increase in private earnings (Figure 6 and Table A.5). While this is not an insubstantial figure, comparing it to calculated rates of return in other areas raises some concerns about the quality of schooling (as well as the availability of jobs/demand for skilled labor). The average private rate of return for an additional year of school in Latin America is higher than for any of the four Central American countries, even though average educational attainment is higher in Latin America than it is in Central America alone. Pri- vate rates of return should be lower when larger proportions of the population are educated because education is less of a restricted commodity. The lower rate of return in the Central American countries demonstrates that an additional year of education provides less added benefit to would-be employers, probably because of lower quality of schooling (it may also mean--either alternatively or in addition--that there are not enough skilled jobs available for people as they complete their schooling16). 15. The methodological aspects of the calculation are developed in the main report (World Bank 2004e), Chapter I. 16. Although there is no evidence that demand for skilled labour is flat in Central America (see AED 2004 for El Salvador and World Bank 2004d for Honduras on this subject). 12 A World Bank Country Study Figure 6. Private Rates of Return in Central America and Other Countries, 2000­2002 14 12 10 8 6 4 2 0 El Salvador Guatemala Honduras Nicaragua Costa Rica LAC Asia OECD Source: Table A.5 Strong Inequities in Learning. While low levels of learning are a problem across-the-board in Central America, the problem is particularly acute for disadvantaged sectors of the pop- ulation. In all four countries educational achievement is higher for urban students than it is for rural students;17 it is higher for wealthier students than it is for poorer students, and it is higher for ethnic majority students than it is for indigenous or other ethnic-minority students. Repetition rates are 30­40 percent higher in rural than in urban areas in El Sal- vador and Guatemala. Similarly, repetition rates are nearly twice as high in these same countries for the poorest quintile compared with the top quintile (Figures 7 and 8). In addition, private rates of return are lower for lower income people in all countries, suggesting, among other factors, lower learning levels (this may also represent barriers to well-paying jobs that the poor face due to nepotism and other reasons). This is the result of two types of observations. First, marginal returns to education are higher for workers in the upper quintile of the conditional earning distribution than they are for the lower quin- tile in all countries (Figure 9).18 Second, there are a number of possible contributing fac- tors that explain this difference (quality of education, labour market connections, and so forth) and it is likely that their distribution is biased against low income households.19 In particular, it is likely that the education received by the upper income quintiles is of higher quality than that received by the lower income quintiles, that cultural capital passed through families, communities, and schools favours upper income groups in the labour market, and, finally, that upper income groups have more access to high paying jobs through personal and professional connections. 17. For instance, on average, rural schools score about 10 percent lower than urban schools on the grade 3 standardized exams in Honduras (Honduras--UMCE 2003a) and Guatemala (World Bank 2004a). They score, on average, about 5 percent less than urban schools in El Salvador on the grade 3 stan- dardized exam (World Bank 2004b). 18. This reflects generally higher rates of return per education level for the upper quintile. 19. See, for instance, Arias (2004). Central American Education Strategy 13 Figure 7. Repetition Rate by Income Quintile 16 14 12 10 8 6 4 2 0 El salvador Guatemala Nicaragua I II III IV V Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Nicaragua, EMNV 2001. Figure 8. Efficiency Indicators in Urban and Rural Areas 16 14 12 10 8 6 4 2 0 age for repetition age for repetition age for age for repetition grade rate grade rate grade grade rate distortion distortion distortion distortion (grade 1) (grade 1) (grade 1) (grade 1) El Salvador Guatemala Honduras Nicaragua urban rural Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002, Nicaragua, EMNV 2001. 14 A World Bank Country Study Figure 9. Rates of Return by Earning Quintile 14 12 10 8 6 4 2 0 El Salvador Guatemala Honduras Nicaragua 0.1 0.25 0.5 0.75 0.9 Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002, Nicaragua EMNV 2001. Reaching Universal Primary Completion The second key (and interrelated) challenge is ensuring that all children complete pri- mary school. Completing primary is a necessary requirement for having even minimum prospects on the labor market. Universal primary completion is linked to reduced poverty and disease, stronger democracies, and economic growth. Universal primary completion is a minimum goal for of all nations and as such is one of the Millennium Development Goals (MDGs). Primary Completion Still Below 80 Percent. While there has been substantial improvement in primary coverage and completion over the last several decades in Central America, all four countries are still far below 100 percent primary completion. Primary gross comple- tion rates20 are lower than 80 percent in all countries (see Tables A.6 and A.7, where both official and household survey data are used). Thus, one in five children in our Central American countries will never complete primary school. The target 100 percent completion rate by 2015 established under the MDG goals will require major efforts by Guatemala, Honduras and Nicaragua. El Salvador has slightly bet- ter indicators. With the exception of Costa Rica, gross completion rates are substantially lower in Central American nations than the Latin American average as well as countries with similar per capita income (see Table 2 and Figure 10 for a comparison with countries with similar per capita income). 20. Calculated as the ratio of enrollees in the last grade of primary (minus the repeaters) and the 12- year school age population. Central American Education Strategy 15 High Inequality in Completion. Those chil- Table 2. Primary Gross Completion Rates dren who do not complete primary as well Country Primary GCR as those who never enter school in the first Guatemala (2001) 57 place come predominantly from severely Honduras (2001) 69 underprivileged sectors of society. Nicaragua (2002) 69 This gap can be categorized broadly as Salvador (2002) 75 a gap between the poor and the non-poor Colombia 85 (or very poor and poor) although indige- Philippines 92 nous populations, other ethnic minori- Ecuador 96 ties, and populations living in rural areas Peru 97 have unique needs that need to be ad- Romania 98 dressed independently of the "poor" label. Jordan 104 While they have been declining, the urban/ rural, socio-economic and indigenous/ Source: Table A.6 in Appendix A. non-indigenous gaps in primary comple- tion are still acutely problematic and underscore a persistent educational equity challenge. In Nicaragua, while four out of every five upper-quintile students complete primary school only one in five complete primary from the lowest quintile. A similar situation is visible in Guatemala (Figure 11).21 Figure 10. Income Per-capita and Gross Completion Rate in Primary, 2001/2002 120 100 CR Rate 80 ES NICA HON 60 Completion GUA 40 Gross 20 0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Income Per-Capita Source: Table A.6 and WDI. 21. Completion rates, as reported in Figure 11, are calculated as the proportion of 12­21 year chil- dren having completed the primary cycle. As such they underestimate the current completion rate (prox- ied by the gross completion rate calculated above). 16 A World Bank Country Study Figure 11. GER, NER and Gross Completion Rate across Income Quintiles in Primary 140.0 120.0 100.0 80.0 60.0 40.0 20.0 0.0 GE GER NER GCR GER NER GCR GER NER GCR GER NER GCR El Salvador Guatemala Honduras Nicaragua poorer I II III IV V richer Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001. There is still a substantial gap in primary completion between urban and rural areas in all countries--between 40 and 60 percent less in rural areas depending on the country.22 While for most countries inequities are relatively small in terms of the current gross enroll- ment rates at the primary level, they tend to increase slightly in terms of the net enrollment rates (indicating higher incidence of late entrance and repetition in poorest quintiles) and are then magnified in terms of completion rates (the lowest income quintile has comple- tion rates which are between 55 percent and 75 percent lower than the ones of the top one; Figure 11). Finally, primary completion rates are substantially lower for indigenous than non-indigenous populations in all countries. Expanding Secondary Education Coverage Recently several studies have posited that a population with full secondary education is key to boosting national competitiveness and GDP in an increasingly technological world (Fuller and Holsinger 1993; World Bank 2003). These papers show that secondary enrollment in developing countries has been positively related to GDP levels over the past three decades and that mass secondary schooling is critical to the ability of countries to attract and benefit from trade and foreign direct investment. As good jobs increasingly rely on skilled labor, an undereducated population will fail to attract new investment. Employment opportunities in countries with low or highly skewed education levels are likely to only exacerbate existing 22. As from household survey information. Central American Education Strategy 17 inequalities and face severe skill bottlenecks. On the other hand, skilled workers are much more productive than unskilled workers in today's technologically advanced firms. Skilled labor generates a beneficial cycle in which higher-skilled workers, by being more productive, generate greater profits for firms, who reinvest a portion of those profits in higher salaries, thereby increasing the assets of workers who, in turn, can invest more in developing more skills. There is, therefore, both a direct link between the relationship of education and wages at the individual level (as captured by private rate of return analysis) and the relationship of education and aggregate income at the country level. What is more, this skill upgrading should have a larger effect on growth rates now than before because of the increasing need for skilled, educated labor. Thus, countries with low levels of education remain in a trap of technological stagnation, low growth and low demand for education. Supporting this argument, recent evidence on Central America shows that upper sec- ondary education provides high private rates of return23 (Figure 12) and boosts national growth.24 High private rates of return in upper secondary suggest a combination of high demand for secondary graduates and low supply of secondary graduates. The demand for educated workers is likely to only increase in Central America with the CAFTA (Central Figure 12. Rates of Return by Level of Education, 2000­2002 30 25 20 15 10 5 0 primar lower secondary upper secondary tertiary El Salvador Guatemala Honduras Nicaragua Source: Table A.5. 23. The pattern of generally rising returns with the education cycle (also visible in Figure 12) is con- sistent with those found in other Latin American countries, although the returns to secondary education tend to be slightly higher in Central America. This pattern is also confirmed for Nicaragua in the recent studies of Laguna (2003) and Porta and Laguna (2004). 24. A recent cross-country study of Loayza, Fajnzylber, and Calderón (2002), by regressing the change in the average per capita growth rate between the 1980s and the 1990s on a set of structural and cyclical variables, finds that changes in educational attainment, measured by the gross enrollment rate in sec- ondary education, significantly contributed to the differential growth rate in Central America (about 15 percent of the projected per capita growth rate between the 1980s and 1990s is explained by education). 18 A World Bank Country Study American Free Trade) agreement because reforms to liberalize trade regimes, encourage FDI and facilitate licensing of technologies may further unleash the demand for schooling.25 Secondary Enrollment Gap. Latin America has a profound skills gap that corresponds with low enrollment rates in secondary education. These rates are well below those of compari- son regions with similar per-capita income. According to recent research (World Bank 2003), the most successful examples of nations increasing their populations' education level, notably in East Asia in recent years and the United States some decades ago, expanded their education systems one level at a time--first ensuring that there was broad enrollment in pri- mary, then in secondary, and finally in tertiary. Latin America, on the other hand, has for the most part not followed this approach. Instead, many Latin American countries have expanded and/or invested heavily in the higher levels of education without first ensuring that there was broad access to and completion of lower levels. A balanced transition is nec- essary because it minimizes inequalities,26 ensures that the most gifted students go to uni- versity, and takes advantage of the complementarities between the phases in an educational transition and those in the technological transition of a country. The research concludes that Latin America, as a region, needs to close its skill gap quickly in order to be competitive in the global marketplace. Furthermore, it asserts that expanding upper and lower secondary education coverage is the first priority for most countries to close this gap. Central American countries represent no exception to this pattern of limited secondary enrollment and limited skills. Though expanding, secondary education coverage is still an enormous challenge for all countries under analysis, some more than others, as illustrated by gross enrollment rates of at most 60 percent and completion rates below 40 percent (Tables A.1 and A.6). Figure 13 illustrates that the Central American countries have con- siderably lower secondary gross enrollment rates than countries with equivalent per capita income. Low enrollment is even more evident when one compares net enrollment rates across these same countries. Costa Rica and El Salvador have education distributions that resemble a "lopsided anvil" with the bulk of the population with only some primary schooling, but more individuals with tertiary education than with secondary schooling only (see Box 1, and World Bank 2003). In Nicaragua, most people have only reached primary school, very few have reached lower sec- ondary, and more have reached upper secondary and tertiary. Finally, in Honduras and Guatemala, the largest share of the population has only reached primary and similarly fewer people have reached lower secondary, upper secondary and tertiary education. East Asian countries, on the other hand, have a distribution resembling a diamond in which more adults have secondary schooling than primary and tertiary one. This diamond shape is more desir- 25. Private rates of return, while they can provide a reason for expanding secondary education, do not provide a reason for public expansion. What will count when deciding for public intervention will be the social rates of return (which also include externalities such as the impact of education on societal devel- opment) and their comparison with the private ones. This topic will be further developed when dealing with secondary education coverage. 26. Latin America is the most unequal region in the world and unbalanced transitions between edu- cation levels can perpetuate inequality. An unbalanced transition, with low secondary education levels, would relegate the bulk of workers to low-skill, low-wage jobs while a privileged few earn the high wages that come from a university education. Because of a close relationship between schooling and earnings, an unequal distribution of education tends to perpetuate an unequal distribution of earnings. Central American Education Strategy 19 Figure 13. The "Secondary Education Gap," 2001­2003 100 JOR 90 BOL ROM PERU 80 PHI ALG COL CR 70 L ES 60 PGY R NICA EC 50 HON GE 40 GUA 30 20 10 0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Income per-capita Source: Table A.1 and WDI. able for a developing country in that it represents less educational inequality and indicates that most of the population has, at the least, attended secondary. Behind the Enrollment Rates: Low Transition to Seventh Grade and Low Survival Rates in Sec- ondary. Both the transition to lower secondary and survival rates once in the cycle are criti- cal issues for most countries. Survival functions, which show a projected survival rate of the current seven year cohort, illustrate clearly that in all four countries there are high drop out rates throughout the primary and secondary cycles. Survival rates from first to sixth grade are generally below 80 percent, then another 5 to 40 percent (depending on the country) of sixth grade graduates never enroll in secondary (see Figure 14). The transition rate between sixth and seventh grade is particularly low in Honduras. In El Salvador, by contrast, the transition from sixth to seventh is approximately the same as the transition rate in any of Box 1. Central American Education the previous years. This has been largely Distributions credited to a successful basic education reform. It is striking to see that, with In El Salvador and Costa Rica more people a higher survival rate to the last year have primary schooling and tertiary schooling of primary than Nicaragua, Honduras than have secondary schooling. is achieving a survival rate to grade 11 which is 10 percent lower. Tertiary Secondary Strong Inequalities in Secondary Coverage Primary and Completion. Finally, upper income and urban youth dominate secondary 20 A World Bank Country Study Figure 14. Comparative Survivor Plots SURVIVOR FUNCTION PLOT BY YEARS OF EDUCATION 1.0 El Salvador 0.9 estimate 0.8 Honduras function 0.7 ution ib Nicaragua distr 0.6 al viv sur 0.5 Guatemala 0.4 0 1 2 3 4 5 6 7 8 9 10 11 yedu Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001. school enrollment, as illustrated by enrollment and completion rates which increase dra- matically with income level (Figure 15). Even in the relatively most equitable country, El Salvador, about five times as many children in the richest quintile complete the sec- ondary cycle compared with the poorest one (versus about 2.5 times for primary). Inequity in education indicators is also very substantial between urban and rural areas and indigenous/non-indigenous populations. The Three Challenges are Interrelated While we address these three challenges--learning, primary completion, and access to secondary--in separate sections of this report, we in no way wish to infer that they are unrelated challenges. In fact, they are tightly linked. Improvements made to education qual- ity will help children get through primary school and motivate them to attend secondary school. Likewise limited enrollment in secondary contributes to low education quality by having direct consequences on the number and quality of applicants to the teaching profes- sion. This next section briefly outlines some of the ways in which the three challenges are interrelated. Poor Learning Leads to Low Educational Attainment. Learning impacts educational attainment. This occurs in three primary ways: (a) poor learning leads to grade failure and repetition because of the lack of necessary skills/knowledge to advance to the next Central American Education Strategy 21 Figure 15. GER, NER and Gross Completion Rate across Income Quintiles in Secondary, 2000­2002 120.0 100.0 80.0 60.0 40.0 20.0 0.0 GER NER GCR GER NER GCR GER NER GCR GER NER GCR El Salvador Guatemala Honduras Nicaragua poorer I II III IV V richer Source: Household Surveys. grade;27 (b) poor learning directly affects student interest and enjoyment of school which contributes to their decisions of leaving school; and (c) learning affects the private rates of return of education, that is the potential for future earnings. The impact of learning on rep- etition is important because repetition can lead to low educational attainment. Survivor functions by age for the countries under analysis indicate that when students reach their teenage years they are much more likely to drop out of school (Figure 16; higher dropout in the teenage years is probably due to a combination of factors including the need to work, pregnancy, or a sense of shame or lack of identification with school materials and younger classmates). This finding implies that the lower the number of grades achieved by teenage years, because of repetition (and/or late entrance), the less schooling will have the children when they drop out. In fact, lower transition rates to secondary (seventh grade) in Honduras and Guatemala are associated with higher age-by-grade distortion in sixth grade in both countries. Supporting this fact, a comprehensive analysis undertaken by Edwards, Fuller and Parandekar (1997) on Honduras determined that 13-year-old and older children are more likely to drop out at some point of the cycle, even if they have not completed pri- mary, and, therefore, that school attainment is substantially determined by the age of initial 27. Evidence on Central America from Edwards (2004) and AED (2004) shows that learning failures are the key reason for grade failure and repetition in Guatemala and El Salvador. 22 A World Bank Country Study Figure 16. Comparative Survivor Function Plots by Age El Salvador SURVIVOR FUNCTION PLOT BY AGE 1.0 0.9 Honduras 0.8 estimate 0.7 Nicaragua 0.6 function 0.5 distribution 0.4 Guatemala 0.3 survival 0.2 0.1 7 8 9 10 11 12 13 14 15 16 17 edad Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001. enrollment28 and repetition. As mentioned in points (b) and (c), poor learning can also sim- ply discourage attendance by making school less interesting and limiting future earning potential. Evidence from household surveys suggests that "lack of interest" is an important reason for being out-of-school in Central America. To support further this evidence, Mar- shall (2003b) finds that gains in test scores are a significant determinant of school attendance in Guatemala. Groups with the lowest primary completion rates also have lower educational achieve- ment, lower returns to education and repeat more. It hardly needs to be said that these stu- dents come predominantly from low socio-economic, ethnic minority, and rural populations. Low Learning Outcomes Produce Little Motivation to Attend or Enroll in Lower Secondary. Low learning outcomes in lower secondary schooling, as also indicated by poor perfor- mance in standardized exams and particularly low rates of return29 (see Figure 12), serve as a disincentive for children, especially poor children, to attend and enroll. In particular, when the low quality and little foreseeable return for lower secondary combine with the significant opportunity cost to poor families of having their teenage children attend sec- ondary, families and students themselves have little or no reason to invest several years in this cycle. This, then, is a major block to increasing secondary education coverage. 28. The study finds that every additional year of age at enrollment increases the probability of drop- ping out of school by 25 percent! 29. Which indicate that there is little added benefit of lower secondary schooling probably due to low quality and low relevance of the cycle. Central American Education Strategy 23 Secondary Coverage and Primary Completion Are Related. It hardly needs to be said that high primary school dropout rates affect secondary enrollment rates. In all four countries, at least one in five students who enters primary does not complete primary education. Although there is substantial margin for improving secondary enrollment through an improvement in the transition from primary to secondary (seventh grade) in three of the four countries (as shown in Figure 14), primary completion will have to be addressed if the countries are to achieve universal secondary enrollment. The Challenge of Equity is Integral to Two Regional Priorities As is evident from this introduction to the central themes of this report, the problem of stark inequities in the education that poor and marginalized populations receive is central to two of the three regional priorities. In this report we do not separate equity as its own regional priority but rather we show that in addressing the challenges of primary completion and sec- ondary coverage we are largely talking about overcoming inequalities in which disadvantaged groups face enormous blocks to learning, and access to and success in primary and secondary. Improving Primary Completion Rates Means Focusing on the Poor and Disadvantaged. Low primary completion massively disfavors poor, indigenous, and rural students. Universal primary completion, will not be achieved without focusing on improving the completion of the most disadvantaged groups, which in turn, is inextricably linked to improving the quality of their learning. Countries with Low Enrollment in Lower Secondary and Limited Budgets Should Focus on Expanding This Level Prior to Expanding Upper Secondary. As discussed above, broad- based secondary expansion is conducive to national growth and social cohesion. Expan- sion of lower secondary should occur prior to upper secondary in order to ensure that large proportions of children from all socio-economic and socio-cultural groups are benefiting from the expansion. Upper secondary expansion is also key, but should not take place prior to ensuring equity at the lower secondary level. Key steps towards this objective include increasing the primary completion rate and ensuring a more equitable access to secondary for the children that are completing primary. El Salvador has managed to do a bit of both in recent years. They are probably ready to move on to expand upper secondary. Nicaragua, on the other hand, also made efforts to expand secondary education in recent years, but did so without emphasizing a broad pri- mary completion basis and equitable transition from grade 6 to 7. This resulted in much less equitably distributed educational enrollment compared with El Salvador (see Figure 17). Survival functions by income quintile in the two countries confirm this result. It may be tempting for Honduras and Guatemala to take a direction similar to that of Nicaragua, which was "cheaper" in terms of public expenditure per child than the direc- tion followed by El Salvador. Yet, the social and private benefits to El Salvador's model appear higher; economic growth is higher than in Nicaragua30 and social unrest lower. 30. As shown by the analysis of Loayza, Fajnzylber and Calderón (2002). 24 A World Bank Country Study Figure 17. Average Secondary GER and GER by Quintile in El Salvador and Nicaragua 120.0 100.0 80.0 60.0 40.0 20.0 0.0 I II III IV V I II III IV V El Salvador Nicaragua GER Aver GER Source:El Salvador: EPHM 2002; Nicaragua, EMNV 2001. Ataminimum,GuatemalaandHondurasshouldaimatencouragingbroad-basedexpan- sion in lower secondary by expanding equitable access to secondary education for the children who are now completing primary (once the most disadvantaged groups of children succeed and complete primary school, secondary expansion will become even more balanced). Nicaragua should now also follow this route, while El Salvador should aim at expanding upper secondary (while also pushing for full transition of the poor and very poor to seventh grade). Fiscal and Expenditure Constraints to the Achievement of the Three Regional Priorities The bulk of the remainder of this volume tackles each of the three regional education priorities--identifying the main weaknesses in the area and proposing policy options to improve those weaknesses. First, however, one constraint affects all three of the areas of intervention, and as such is dealt with prior to looking at each priority separately. This con- straint is public money. Tradeoffs With low income levels and high debt levels, all four Central American countries face strong budgetary constraints in the funding of all government services and spending areas. These governments must make decisions about how much money should go into improving edu- cation, offering health services, repairing and building infrastructure, maintaining a military, and other important areas of government spending. Once money is allocated to education, ministries of education will have to make difficult decisions about where to focus their efforts Central American Education Strategy 25 and how to distribute their limited resources. Because of these fiscal constraints, there will necessarily be some tradeoffs among the three regional priorities laid out in this report. In particular, learning improvements, which are necessary at all levels, may require additional resources for teacher education and training, textbooks, and so forth--which may leave little room for other investments. Although learning improvements will have a beneficial impact on the other two regional priorities (as explained above)--limiting somewhat these tradeoffs--reaching the other two objectives will require other specific interventions. Reach- ing universal primary completion, for example, will require additional resources spent on the most disadvantaged groups to make sure that they are provided with the skills and incentives to complete. Expanding secondary coverage, in particular if in a somewhat broad-based way, will also require substantial additional investment to create schools or programs in areas of low access, modify curricula to improve quality and relevancy, and perhaps offer conditional cash transfers to poor students.31 The nature of the interrelation between these priorities suggests an order of interven- tion: first, the quality of learning should be targeted; second, universal primary completion should be assured; and third, secondary education expanded in a broad-based manner. While education quality should be the first and foremost priority for all four countries, the later two priorities could be addressed simultaneously or even in reverse order in the short to medium term if there is substantial room for improvement of the grade six to grade seven transition rate. At the least, countries could expand lower secondary to accommo- date all children finishing primary. Additionally, all countries should make an effort to expand upper secondary education, although, with the exception of El Salvador, these efforts may need to be more selective to save on scarce resources. At the same time, it will be essential to envisage strategies which overcome the trade- offs by giving increased budgetary priority in all (or at least three) of the four countries to the education sector, promote higher resource mobilization also from other sources (private sector, households), and introducing measures which encourage higher efficiency in the use of resources (more effective teacher management, higher shares of non-salary recurrent costs, re-allocation across education levels within the education budget, cost-efficient delivery sys- tems, effective demand-side interventions, lower repetition, and so forth). We illustrate below the existing constraints in terms of budgetary priority for education and efficiency in the use of public funds.32 Insufficient Public Expenditure on Education Public Expenditure Per Student is Below Average in Central America. With the exception of Honduras, all the countries analyzed here spend less on education than other countries with similar per capita income. This is illustrated by a comparison between public educa- tion expenditure per student (in both primary and secondary) and GDP per capita across 31. The case for public financing in promoting broad-based secondary expansion is made in sub- section II.2.C. 32. Other efficiency constraints which are not the focus of the report but could be very important are institutional constraints such as the poor planning and implementation capacity of the Ministries of Edu- cation. In particular, it will be very difficult to achieve an efficient resource allocation without better devel- oped planning and monitoring systems within the ministries. 26 A World Bank Country Study Figure 18. Public Education Expenditure per Student in percent of GDP per Capita, 2000 40 35 30 25 20 15 10 5 0 a r a s y ntina zil dan ico gua ne CD Bra Chileta Ric Mex ara OE Arge Cos IndiaJor Salvado Malaysi Paraguaylippi Uruguae El Guatemala Honduras Nic Phi primary Averag secondary Notes: Central America 2002 or 2001 data. Source: OECD (2002) and sources mentioned in Table A.8. a sample of 16 countries (including the OECD average; Figure 18). Nicaragua spends par- ticularly little in per student terms in secondary education. In spite of noticeable increases in public education expenditure as a percentage of GDP in all countries in the last decade (also illustrated in Figure 19 and Table A.8), the expenditure ratios of these three countries remain well below the Latin America average (see Figure 20)-- and below a sample of countries with similar income per capita.33 These results suggest an urgent need for mobilizing public funds for education in these countries.34 Honduras is the exception to this rule in that they are already spending substantial public resources on education. In Honduras, the main budgetary challenge is increasing the efficiency of resources use. In particular, there was a substantial wage bill increase due to increasing wage rate and inflated payrolls in the country, which may suggest the need for salary and teacher management reforms.35 Low Efficiency of Public Education Expenditure Too Little Spent On Non-Salary Expenditures. With the notable exception of Nicaragua, the vast majority of Central American education budgets go directly to teachers' salaries (see Table 3). In the past decade, the driving force behind increases in public education expen- diture has also been teacher salaries.36 The lack of funds "leftover" for non-salary recurrent spending is very small--these shares represent less than 15 percent of total expenditure in 33. An average expenditure ratio of about 4.2 percent was reported for a sample including Philippines, Jordan, Bolivia, Algeria, Peru, Colombia, Paraguay, and Ecuador (based on UIS 2000 data). 34. Which will be further justified in the case of secondary education in Chapter 3. 35. An analysis of the Honduras case is undertaken in World Bank (2004e), Chapter IV. 36. While pupil-teacher ratios have been kept generally stable and are in line with international standards. Central American Education Strategy 27 Figure 19. Evolution of Public Expenditure in Proportion of GDP 8 7 6 5 P GD 4 of % 3 2 1 0 1995 1996 1997 1998 1999 2000 2001 2002 El Salvador Guatemala Honduras Nicaragua Costa Rica Source: Table A.8. all four countries as well as in Costa Rica37--meaning that countries have very little money to spend on training teachers, providing classroom and school materials, and other basic needs (see Figure 21 and Table A.9). Without dismissing the importance of fair and com- petitive teacher salaries, these non-salary expenditures are also important to ensure educa- tional achievement and progression of disadvantaged students. Teacher Compensation is Rarely or Indirectly Linked to Teacher Performance. What is more, increasing teacher salaries does not automatically or cost-effectively raise teacher or teaching Figure 20. Public Education Spending as a Proportion of GNP, 1999/2000 7.0 6.0 5.0 4.0 3.0 percentage 2.0 1.0 0.0 il C lize Rica LA yana Chile Peru Bolivia Be Braz Mexico ruguay cuador temala Jamaica Panama Barbados Grenada Gu U E salvador Honduras Paraguay Argentina Costa Nicaragua El Gua Source: WB, WDI database, 2003. 37. This is below the OECD average of 20 percent (see OECD 2002). 28 A World Bank Country Study quality. This is particularly the casewhen,as Table 3. Central America Spends Most in Central America, teacher compensation is of Its Education Budgets on Teacher Salaries not directly related to actual indicators of teaching quality and performance (see sec- Percentage of Total tion on learning constraints). MINED Budget that pays Teacher Salaries Central American Education is Not Cost- El Salvador 70% effective. As a result, we find that Central Guatemala 88% American countries do not rank very well Honduras 93% when compared with countries with similar income per capita, such as Peru, Ecuador, Nicaragua 58% and Bolivia (Figure 22). These South Ameri- Costa Rica 92% can countries tend to get better results for Source: Table A.9. their investment (in particular when consid- ering a combination of educational outcomes).38 Figure 19 compares primary completion rates with public expenditure in terms of GDP across countries of similar per capita income.39 Honduras is particularly cost-ineffective (compared to Peru, Bolivia, and Ecuador but also to El Salvador). Inefficient Allocation Across Education Levels. Another type of inefficiency is related to the allocation of public education expenditure across education levels within each country's education budget. Figure 23 (and Table A.10) below shows the relative expenditure shares of each of the education levels in each of the countries under analysis. Figure 21. Salary, Non Salary Recurrent and Capital Expenditures, 2002 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% El Salvador Guatemala Honduras Nicaragua recurrent (salaries) recurrent (non-salaries) capital Source: Table A.9. 38. Private expenditure has not been considered in the expenditure calculation but its inclusion should not alter these conclusions due to its limited magnitude in primary in all Latin American countries. 39. Nicaragua is more cost-effective when comparing completion and per student expenditure. Central American Education Strategy 29 Figure 22. Public Education Expenditure/GNP and Primary Completion Rate, 2001/2002 120 PE 100 ECU CR 80 ELS PAR BO HON Completion 60 NIC Rate GUA 40 Primary 20 0 0 1 2 3 4 5 6 7 Education Expenditure/GNP Source: Table A.6 and WB/WDI for other countries and expenditures. Figure 23. Functional Disagregation of Education Expenditure, 2002 100% 90% 80% 70% other administration 60% tertiary 50% secondary primary 40% pre-primary 30% 20% Notes: Primary and Secondary 10% Shares have been estimated in El Salvador; Primary includes Pre- primary in Costa Rica; Tertiary 0% includes University own budget in El Salvador Guatemala Honduras Nicaragua Costa Rica Nicaragua. Source: Table A.10 30 A World Bank Country Study Low secondary education share. While, not surprisingly, primary education represents the high- est share of the budget in all countries, the share directed to secondary education is much lower, fluctuating between 8­9 percent and 20­25 percent, reflecting low coverage and quality of the cycle. All Central American countries still spend a significantly lower share on secondary than the Latin American average, which is about 30 percent (ECLAC 1994 and 1996). High tertiary share. In contrast, it is interesting to note that the three poorest coun-tries spend as much or even more on tertiary than they do on secondary. While there are compelling rea- sons to invest in universities, it is also socially regressive in education systems where the bulk of students do not have the opportunity to complete (or even enter) secondary.40 High administration costs in Honduras. Finally, our figures also provide us with some esti- mates of the share of central administration costs. Although these costs are probably underes- timated, we can still say, however, that administration costs seem to be relatively contained in all countries, with the exception of Honduras which spends more on administration than on secondary education! Honduras should learn from El Salvador on how to improve the effi- ciency of its education spending. Both a resource mobilization and an efficiency effort will therefore be needed to address these three regional challenges. The relative role of public and private spending is discussed below, in the section on the challenge of secondary education coverage. Simi- larly, specific measures to improve the efficiency in the use of resources will be discussed below under each of the regional priorities. 40. Argentina, Chile, Uruguay, and Colombia spent approximately 20 percent of their education bud- get on tertiary education in the time-period 1994­2000 (see various World Bank Country Public Expen- diture Reviews). This is in the range of what these three, much poorer countries, spend. CHAPTER 3 Constraints and Policy Actions to Achieve each Regional Priority Supply and Demand-side Constraints. Before turning to the analysis of constraints, interven- tions and policy options for each of the regional priorities, a short introduction is needed on the role of supply and demand-side constraints and the relative focus of the report. Both supply and demand-side factors impact the three main priorities just identified in Central America. Demand-side factors can be broadly understood as any characteristic of a child's background that effects his education, or, more specifically, can be understood as the actual demand students and families have for schooling. Supply-side factors refer to characteristics of schools and education systems. The parents' decision to have their child in school is the result of an unconscious equa- tion which weighs the multiple perceived benefits of having them in school against the mul- tiple perceived costs. Policies can effect this decision by weighing in on either side of the equation. Parental demand for schooling, in this sense, is effected by both supply and demand factors. More parents may choose to send their children to school if classes are taught in family's primary language (supply side policy) or if they receive monthly subsidies for keep- ing their children in school (demand side policy). Any policy that changes the perceived costs or benefits of schooling will impact parental demand. Additionally, children and family char- acteristics, such as the child's nutrition status and family education, directly effect learning, as do school characteristics, such as teachers' education and availability of textbooks. Ideally, one would need to undertake a detailed analysis to determine if the education sectors are more supply or demand-side constrained. In Central America, both types of constraints will apply. While the report develops both demand and supply-side constraints, it explores more deeply the supply-side. Some of the main reasons behind this focus are: (a) the strategic priority given to the quality of schooling and education delivery for the poor and marginalized populations in Central America; (b) the very poor quality of schooling in the region, clearly an important supply-side constraint; and (c) evidence, from household 31 32 A World Bank Country Study surveys, that poor quality of schooling contributes to low demand for schooling. There is also a wide span of supply-side interventions that have been applied and can be applied to address the learning, primary completion and secondary education coverage challenges in the region, such as quality improvements (which will have an impact on all three chal- lenges41) and alternative secondary education options (which will address both the supply and demand-side constraints by decreasing the opportunity cost of staying in school and direct costs of schooling). Recent literature from developing and Central American coun- tries confirms that supply-side interventions that improve education quality have substan- tial impacts on family and community demand for schooling as well as improving learning.42 This is not saying that demand-side interventions will not be effective. By reducing direct and opportunity costs to schooling these interventions will be key in increasing enrollment and completion, in particular of low income children and youth. Additionally, by strengthening family and community awareness and role in the delivery of education, demand-side interventions can also trigger changes in school quality and relevancy. We therefore also address some demand-side interventions.43 Regional Priority #1: Improving Learning through National Assessment Systems, Teacher Performance and Teaching Methodologies Constraints to Learning This report identifies a number of constraints to learning in Central America.44 Key con- straints include poverty, little or no parental education, social and cultural barriers, lack of a culture of evaluation, weak teacher performance and accountability, overdependence on traditional pedagogical practices, insufficient instructional time and insufficient non-salary investment in items such as material resources and teacher preparation programs.45 Importance of Demand-side Constraints. At the family, community, and even state level, the greater the wealth the higher the test scores. Poor families are more likely to lack basic resources that help children learn. Low levels of parental literacy and education are also negatively related to student outcomes, as are certain social and cultural characteristics of students, families, and communities. Linguistic differences also pose barriers to learning when high-quality bilingual education is not available.46 This is particularly the case in 41. The link between quality of schooling, repetition, over-age and dropout will also come into play. 42. Recent literature from developing countries has found evidence that school quality does impact demand for schooling. See, for example Paxson and Schady (2002) on Peru; Younger (2000, 2003) also on Peru; Lavy (1996) on Ghana; Bommier and Lambert (2000) on Tanzania; Lloyd and others (2000) and Lloyd and others (2003) on Egypt and Kenya, respectively; and Marshall (2003a) on Guatemala. 43. Including an intervention very much at the border between a supply and demand-side interven- tion such as school-based management (as implemented in Central America, this intervention increases school and teacher accountability to families and local communities by empowering them). 44. The foundation from which this section is built is a review of recent education production func- tion studies in Central American countries (presented in Chapter II of World Bank 2004e). 45. Other constraints on which the report does not focus but which can also be important are policy and institutional framework constraints such as the lack of continuity in education policies and priorities which hampers the achievement of long term learning improvements. 46. In particular, a lack of contextualized textbooks can significantly hamper learning. Central American Education Strategy 33 Guatemala. In addition, the children of families that do not themselves place a high value on education, or who feel culturally or socially alienated by schools, are also less likely to succeed and advance in school (Honduras--UMCE 2003b; World Bank 2004a). Little Evaluation and Assessment. There is very little culture of assessment and evaluation in Central America. Evaluation is an effective means not only for those working within Ministries of Education to assess the national education system but also for teachers to see how they are doing in the classroom, for parents to compare how the school their children attend compares with other schools in the country, and for students, themselves, to learn to judge their learning. Unfortunately, the nascent evaluation systems in Central America offer little of these benefits. In most countries, the results of assessments are not fully and systematically dissemi- nated even to participating schools. Education ministries typically produce a national report but these have limited availability (OEI 2004). Furthermore, testing is conducted on small samples of the overall student populations and, as such, also lacks institutionalization. Finally, the Central American countries rarely, if ever, participate in international comparative eval- uations, and there is a lack of alignment between the national standardized tests and the newly established regional standards This makes comparison across the countries as well as com- parison with other Latin American or developing countries very difficult. Poor Teacher Performance. Weak teacher performance and accountability is closely related to both weak teacher education and professional development opportunities and a general lack of incentives that promote teacher effort and performance. Insufficient educational attainment and low-quality teacher preparation programs. Table 4, below, shows the highest level of education attained by teachers in Central America. With the exception of El Salvador, the share of teachers with tertiary education is low (52 percent in Honduras, 29 percent in Guatemala, and 13.5 percent in Nicaragua). This is the result of low formal teacher education requirements and lack of enforcement of (or high flexibility in) those requirements. In Honduras and Nicaragua, due to limited access to teacher training schools and insufficient supply of teachers in rural areas, high numbers of underqualified teachers are employed in the classroom. Table 4. Highest Level of Education Attained by Teachers in Central American Countries El Salvador, Honduras, Guatemala, Nicaragua, 2002 2002 2000 2001 % % % % Primary 1 3 18 Secondary 12 46 (a) 67 (a) 37 Teacher School 26 Tertiary, Technical 61 (a) 2 3 Tertiary, University 27 50 27 (b) 11 Other 1 3 6 Total 100 100 100 100 Notes: (a) Including teacher school; (b) Includes technical and university tertiary. Source: The World Bank (2004b,c); Arcia (2003); Guatemala, 2000 ENCOVI Household Survey. 34 A World Bank Country Study There is also some evidence from El Salvador47 and Guatemala48 that the quality and rele- vance of teacher education may still be too low. This is true for both pre-service and in-service professional development. In-service professional development seems to have improved in Central American countries over the last decade but there are still a number of challenges ahead to improve the relevance and quality of these training programs.49 In particular, the sys- tems have become more structured, while at the same time also being more flexible and decen- tralized to increase their responsiveness to different training needs, but there is scope for putting more focus on changing/updating classroom practices; paying more attention to non- standard education contexts; and improving follow-up with trained teachers. Despite pay increases, teacher performance remains poor. Teachers in Honduras (1997), El Salvador50 (1995), and Guatemala (1996) have all benefited from recent increases in the level of real salaries. The current hourly wage levels in these countries are comparable to (or higher than) those of other professionals with similar qualifications and should, therefore attract reasonably talented applicants (see Table A.12). As of yet, however, indicators of teaching quality such as test scores, teacher absenteeism rates51 and weekly hours worked52 do not sug- gest any obvious improvements in teacher quality generated from these salary improvements. This could be because the salary increases are too recent to have elicited large changes in teacher characteristics53 or teaching quality or it could be because these salary improvements do not effectively or sufficiently generate incentives to improve teaching. Lack of incentives to perform. In fact, the current structure of teacher salaries in the region, as well as in Latin America as a whole, has been criticized for undermining teacher quality by rewarding seniority and education level without regard to teachers' actual performance in the classroom. Of the countries in this report, only El Salvador has introduced, and only recently, a merit-pay mechanism,54 which by linking salaries to school efficiency indicators in several areas is designed to provide a performance incentive.55 Additionally, none of the Central American countries penalize teacher absenteeism. In part, this is because teacher hours and attendance are difficult to accurately monitor. On the positive side, several Central American countries are moving towards increasingly involving parents and principals in teacher monitoring in an effort to compensate for the lack of central information.56 Passive, Traditional Teaching Methods Continue to Predominate in Classrooms. In spite of comprehensive curricular reforms in many Central American countries that promote con- 47. Only about 40 percent of the eligible students passed the ECAP exam (teacher accreditation exam) in 2001. 48. Results of the achievement tests of teacher-training students in teacher training institutions, car- ried out in 2000 by the PRONERE program, show that the mean test scores obtained by students were only 28.5 for math and 54.3 for reading, out of a possible 100 points. 49. See, for instance, OEI (2003) and PREAL (2003). 50. The increase compensated for eroded purchasing power in the previous years in Honduras and El Salvador (see World Bank 2004b and 2004c). 51. In Honduras, for instance, the 2000 school census shows that as much as 40% of teachers lose one month or more of school a year; and teachers report working about 30 hours a week (Honduras--UMCE 2003a) while they have 37 notional weekly hours. 52. As exemplified in the cases of El Salvador and, particularly, Honduras (see Chapter II of World Bank 2004e). 53. Although there is some evidence that higher salaries have attracted more professionals with ter- tiary education in Honduras. 54. Bono al buen desempeño institucional. 55. The effectiveness of this incentive still needs to be assessed, but some evidence indicates that school-level performance incentives can boost student achievement, especially on indicators evaluated for the bonus. 56. This is the "asymmetric information" rationale for decentralization. Central American Education Strategy 35 structivist pedagogies, teachers remain largely dependent on traditional pedagogical meth- ods (see Box 1 in Appendix C). Among the most commonly mentioned reasons for this persistent dependence are: (a) a lack of teacher expertise in innovative and dynamic ped- agogies; (b) lack of textbooks available on a wide basis; (c) large class sizes; and (d) teacher and teacher union opposition to the reforms. It is worth repeating that not enough empha- sis has been put on providing the teaching and classroom materials and training needs that should accompany these reforms. Insufficient Instructional Time. All countries analyzed in the report have schooling calen- dars of 180­200 days and a school week of 25­30 class hours in primary. This results in roughly 900­1,200 class hours a year, similar to the rest of Latin America.57 However, there is a widespread perception in Central America that actual instructional time is much lower because of teacher and student absences and school closings. Some simplified estimations58 indicate that there would be between 500 and 800 effective class hours, at most, which rep- resent between 58 and 80 percent of the notional class time (see Figure 24), in Honduras, Guatemala and El Salvador. It is obvious that even 700 class hours per year compare very unfavorably with the more than 1,200 notional class hours of Europe or other countries. This lack of instructional time also surely constitutes a major barrier to educational achieve- ment in Central America. While Central American countries have implemented several wide-reaching curricu- lar, pedagogical, and management reforms, to date these reforms have not resulted in ade- quate gains in education quality. Serious challenges, such as exorbitantly high first grade Figure 24. Notional and Effective Class Hours per Year 1400 1200 1000 800 600 400 200 0 Honduras El Salvador Guatemala Notional Class Hours per Year Effective Class Hours per Year Source: World Bank 2004e. 57. A range of 900­1,200 hours per year is significantly less than the averages for Europe or Southeast Asia (1,200­1,750 hours per year) as cited in PREAL (2002). 58. Details on the methodology adopted are reported in Chapter II of World Bank (2004e). 36 A World Bank Country Study repetition rates, continue to plague Central American schools and indicate that much more needs to be done. Suggested Policy Actions In response to these identified barriers to learning and education quality, this report details three main areas of policy interventions: (a) improvement of national assessment systems; (b) enhancement of teacher performance; and (c) diversification of teaching methodolo- gies. All three sets of interventions will be key to improving the quality of teaching and learning. Additionally, we also discuss (d) a few demand-side interventions that could also have an impact on learning and education quality. Improving National Assessment Systems. Setting up a good national assessment system takes time and learning by doing. Institutionalizing testing for diagnostic purposes can help identify weaknesses in the teaching-learning process. Currently standardized tests are con- ducted in the third and sixth grades in several Central American countries. Some thought should be given to conducting some form of evaluation or test as early as the first grade in order to identify teaching and learning problems early. Perhaps most importantly, coun- tries need to improve their methods of assessment results dissemination. Having reliable means of sharing evaluation results with policymakers, principals, teachers, communities, and parents can broaden accountability as well as make educators themselves more aware of the areas in which they need to improve or change their practice. There are several innovative assessment dissemination policies currently in use internationally. One example is school report cards published in local newspapers or posted as laminated posters at school entrances. Improving Teacher Performance. Different interventions have been suggested to improve teacher performance, but two can be highlighted as having particularly high potential: (a) teacher selection and accreditation mechanisms; and (b) implementation of specific incentives to increase teacher effort (such as the application of teacher monitoring systems, salary scales promoting higher effective hours of work and, potentially, carefully designed, team-based merit-pay schemes). Selecting and accrediting teachers. Teacher selection and accreditation mechanisms have been introduced in El Salvador since 2001. Upon completion of the secondary cycle, teachers must take and score seven or higher on the PAES exam59 in order to enter the teaching career. Salvadoran teachers must also pass a special accreditation exam called the ECAP60 at the end of the third year of teacher training school. Teacher selection and accreditation seem to have pro- duced good results in this country (and elsewhere). When coupled with high quality teacher education, teacher selection and accreditation norms can help obtain a pool of high-quality teachers. The other Central American countries would therefore be well advised to follow the example of El Salvador and introduce mechanisms for teacher selection and accreditation. Setting up an effective teacher monitoring system. An analysis of the evidence presented in the report suggests that of all possible incentives to increase teacher effort and/or instructional time, setting-up an effective teacher monitoring system may be the most effective, possibly 59. The standardized national exam which concludes the secondary cycle. 60. Evaluación de Competencias Académicas y Pedagógicas. Central American Education Strategy 37 combined with the use of more flexible hiring and firing mechanisms.61 From our analysis of community-based schooling in the four Central American countries it appears that empow- ering parents to hire, fire and monitor teachers results in greater teacher effort as measured by teacher hours (see below). There are a few limitations to this finding, however. The first limitation is that this com- bination of incentives occurs predominantly in community-schools that are new and located in poor rural areas, factors that may create a particularly favorable environment for establish- ing a new participative school management model. It is not clear that this model can be repli- cated in different contexts such as pre-existing schools in urban areas. An effective teacher monitoring system for urban areas may require additional or alternative measures to strengthen parental participation and involve local offices of the Ministry of Education more systematically. The second limitation is that while parents can monitor teachers in terms of punctuality, absenteeism, and discipline methods, they may not be the ideal stakeholders to evaluate teacher methods or classroom practices. Pedagogical supervisors, school principals, or Ministry of Education officials may be better positioned to effectively evaluate actual teaching. Finally, we should not forget that the expansion of community-schools is con- strained by teacher unions in most of the countries (as will also be discussed below), making "replicability" and scaling up difficult in the current context. As is clear from this discussion, improving teacher practice through more effective teacher evaluation and management is a difficult and complicated endeavor. Nonetheless, it is critical to ensure both higher education quality and increased efficiency in the education system. Finally, skilled and committed teachers are a necessary requirement for a pedagogical shift to a more diverse, relevant and constructivist pedagogy. Diversifying Teaching Strategies and Deepening Constructivist Teaching Methods. The last set of policy recommendations on the supply-side proposes to improve education quality in Central America by dealing with classroom teaching methods. Teachers in Central America are overly dependent on passive and traditional teaching methods such as dictation, repetition, lecturing, and individual work. Research shows that students learn in multi- ple and varied ways and that teaching methods should be multiple and varied to accom- modate all learners. Furthermore, current teaching methods are criticized for failing to promote critical and creative thinking, group collaboration and interpersonal skills, as well as creating a dull and passive learning experience that does not excite or engage students. Yet, changing teaching methods requires extensive teacher training as well as ongoing support, sufficient and appropriate resources for teachers and classrooms, har- monization with texts, curricula, and exams, and committed and capable teachers. In Central America there is ample evidence that pedagogical reforms, which are complex reforms, have not been fully or effectively implemented due to one or more missing requirements for meaningful implementation. Designing Effective Demand-side Interventions to Improve Learning. Improvements in the quality of teaching and learning through the three areas of intervention discussed above could be further strengthened by complementary demand-side interventions. Some of these demand-side interventions include: (a) Publicity campaigns of various natures can increase demand for quality schooling. Radio announcements, billboards, television commercials, and town meetings are 61. Flexible contracts can take multiple forms. Renewable yearly contracts and easier means of firing ineffective or delinquent teachers are two examples of more flexible contracts. 38 A World Bank Country Study a low cost means of raising awareness about the importance of education. These campaigns can address demand constraints that negatively impact enrollment but they can also address the quality of learning by giving families important informa- tion about exams, timely enrollment, or providing tips on how to support chil- dren's academic success in school. (b) Policiesthatinvolvefamiliesandcommunitiesinthemanagementofschoolscanalso be conducive to higher educational achievement, as will be shown in the next section. (c) Demand-side financing systems, such as vouchers, typically give families a voucher for a certain amount of money that they can then use to enroll their child in a school of their choice. These systems should result in school and learning improvements if, as a consequence of consumer choice, poor-quality or badly-managed schools are weeded out or have to improve in order to attract students. In practice, there is less clarity over the effect of vouchers and they may only be of marginal applicability to the countries in Central America because of the lack of multiple neighboring schools in rural areas, where much of the Central American population lives. (d) Finally, policies that have a direct impact on children's background characteristics, such as their health, nutrition, and access to clean water and safe homes, will be key to ensuring higher learning outcomes. Regional Priority #2: Reaching Universal Primary Completion by Improving and Scaling-Up Interventions that Address the Needs of the Poor This section examines the second of the three regional education priorities discussed in this report: reaching universal primary completion. Because the first priority, improving learn- ing, is so integrally tied into the second, many of the constraints and policy options are the same. Poor teaching, families with little education themselves, these types of constraints impact learning but they also impact children's likelihood of making it through primary school. Rather than repeat ourselves in this section, we focus instead on those constraints and interventions that target specific groups of disadvantaged children. We do this also because those children who are not finishing primary come almost exclusively from poor, rural, and indigenous populations. Whereas the last priority looked at broad interventions to improve learning generally, this section will focus on improving the learning and reten- tion of these populations in order to ensure that they complete primary. The structure of this section is slightly different. It examines first some constraints to primary completion, it then examines four main types of education interventions for poor, rural, and indige- nous populations, and it concludes with policy recommendations that focus on improv- ing and strengthening these existing innovations and introducing other broader measures. Constraints to Primary Completion Poor Learning and Completion. Poor learning has both a direct influence on primary com- pletion through dropout (due to grade failure and/or lack of interest) and indirect through repetition (higher repetition leads to higher age-by-grade distortion and to dropouts related to age before primary is completed). Recent household surveys results show that Central American Education Strategy 39 "lack of interest" is one of the key reasons for not being in school in the 6 to 12 age range in all countries, suggesting that poor learning and quality of schooling are key in the deci- sion to enroll and stay in school. Grade failure and repetition are also mentioned as rea- sons for desertion62 (although these reasons are bound to be underestimated and combined with "lack of interest" to avoid social stigmas). Poor learning is particularly acute for the poor and underprivileged who have lower educational achievement, repeat more, and have generally lower private rates of return to schooling. It can be related to supply-side constraints which are shown to be particularly acute for the poor (rural and low-income areas and schools tend to have less skilled teach- ers, fewer textbooks and poor infrastructure63) and to the same characteristics of poverty which make a good learning experience more difficult. Lack of Economic Means and Enrollment in School. Poverty also has a direct impact on com- pletion (that is not only through poor learning) by discouraging enrollment, timely entrance and permanency in school. Not being interested in school may also suggest that family fac- tors such as lack of a conducive environment to learning come into play. In El Salvador and Guatemala many families identify the private costs of schooling as a key reason of why their children do not complete primary. The evidence presented in Table A.11 suggests that a family in the lowest income quintile would have to spend over a third of its annual income to keep two primary level and one secondary level children in school in El Salvador and, in Guatemala, the same family would have to spend 85 percent of its annual income. Private costs at the secondary level are substantially higher than those at the primary level. It is only logical that the private costs of schooling for the even poorer, say for example the lowest decile, are much steeper than what is seen in Table A.11. It should not be overlooked, too, that the opportunity costs of schooling are much higher for the poor than for the non-poor as the limited amount of income a child can bring into the family constitutes a much larger proportion of total income in a poor household than in a middle class or wealthy household. Existing Interventions that Target Improving the Educational Chances of the Poor As mentioned before, many of the policy options discussed that address low quality school- ing are appropriate as well for increasing primary completion rates. Rather than restate these options this section focuses on particular interventions that target alternative service- delivery mechanisms to support those students and communities most at risk of dropping out of primary. The interventions include multigrade schooling, community-based school management and bilingual education programs. To help us assess these interventions, we also make use of some useful results on the characteristics of high-performing low-income schools. At the end of the section, we also present the experience with conditional cash transfers, a demand-side intervention, in two countries of the region. 62. In Honduras, "low performance" and "did not like it" were the first reason for non-attendance in the 8- to 10-year-old bracket (World Bank, 2004c). 63. This is, for instance, revealed, by an analysis of low and upper income schools in Nicaragua and Honduras. 40 A World Bank Country Study Effective schools and supply-side interventions How do high-performing low-income schools differ from low-performing low-income schools? As part of this report an analysis was carried out to identify how high-performing low-income schools (called Effective Schools) differ systematically from low-performing low-income schools. This analysis on Effective Schools64 revealed that there are some clear and intuitively meaningful differences. These differences indicate that high-performing low- income schools tend to differ from their lower-performing low-income schools in four key areas. Effective Schools have: (a) higher teacher and principal education and/or experience; (b) less use of traditional passive teaching methodologies; (c) substantive and supportive involvement of both teachers and parents in school management and a potentially instruc- tional leadership role for principals; and (d) less student absenteeism.65 These differences should be kept in mind in planning reforms to improve quality and completion for the poor. There is no single magical intervention. A key conclusion of this report is that there is no magic bullet in terms of improving schools and learning. Many interventions have resulted in improvements in certain contexts, but no single intervention can promote all the changes in the teaching-learning environment that are currently creating barriers to primary completion for underprivileged children. Our analy- sis of high-achieving, low-income schools discussed above found that these effective Box 2. Qualities of High-Achieving, schools differed from their less successful Low-Income Schools in counterparts in numerous ways (Box 2 Honduras and Nicaragua reviews again these major differences). EFFECTIVE SCHOOLS TEND TO: We review below the Central American experiences with multigrade schooling, · HAVE TEACHERS AND PRINCIPALS WITH MORE EDUCA- TION AND TEACHING EXPERIENCE; community-based management and bil- · USE MORE INNOVATIVE PEDAGOGY, AND GIVE MORE ingual schooling, reporting their posi- HOMEWORK; tive outcomes (see Table 5) but also · GIVE PARENTS AND TEACHERS GREATER RESPONSIBILITY IN SCHOOL MANAGEMENT; AND constraints and limitations.66 Where pos- · HAVE STUDENTS THAT COME TO SCHOOL MORE REG- sibleweattempttorelatetheseconstraints ULARLY AND ARE LESS LIKELY TO REPEAT GRADES. andlimitationstothefindingsoneffective Source: World Bank 2004e. low-income schools. 64. A group of "effective schools" and control group schools were identified in Nicaragua and Hon- duras. For Honduras, effective schools are defined as those schools that score at least three quarters of a standard deviation below the mean on a constructed socio-economic status (SES) composite variable and at least three quarters of a standard deviation above the mean on a minimum of three of the six exams. The control group of schools are simply the remainder of schools whose SES index rank was three-quar- ters of a standard deviation or more below the mean. In Nicaragua, effective schools were defined as those schools scoring at least one quarter of a standard deviation above the mean on grades 3 and 6 exams and at least one quarter of a standard deviation below the mean on a constructed SES index. The control group was again defined as the remainder of the schools with SES index rankings one quarter of a standard devi- ation below the mean. By and large the background characteristics of the control group of schools matched the background characteristics of the effective schools in both countries. The full analysis is reported in Chapter II of World Bank (2004e). 65. This is an input to effective schools insofar as lower levels of student absenteeism lead to higher student educational achievement. 66. The reader will need to keep in mind throughout the assessment of these different reforms that it is often difficult to associate reforms with outcomes, even if the most sophisticated technical tools are available. In this sense, all results need to be interpreted with caution. Central American Education Strategy 41 Table 5. Innovations in Education Service Delivery Demonstrate Great Potential in Central America Education Innovation Program Impact (a) Multigrade Schooling · In Nicaragua, Honduras and Guatemala rural children in multi- grades schools outperform rural children in traditional schools in a variety of subjects. · In Guatemala, student retention and progression is higher in multigrade schools than it is in traditional rural schools. Community-based Schooling · Community managed schools in Central America have resulted in increased enrollment, generally improved student flows, and, in some cases, higher student achievement. Bilingual Schooling · In Guatemala, bilingual schools have had positive effects on stu- dent achievement, progression through school, and on parental attitudes about school. (a) The reader will need to keep in mind throughout the assessment of these different reforms that it is often difficult to associate reforms with outcomes, even if the most sophisticated technical tools are available. In this sense, all results need to be interpreted with caution. Multigrade schooling. Multigrade schools are schools in which one teacher teaches more than one grade level simultaneously. It is an adaptation of traditional primary education for rural areas where there are not enough students (or teachers) to have one teacher per grade level. All four Central American countries have multigrade programs, the most well known being the New Unitary School (NEU) in Guatemala. The experience with multigrade schooling has shown that this cost-effective methodology has the potential to improve educational access as well as performance in rural areas. Multigrade schooling can improve educational performance . . . In particular, there is evi- dence that when successful, students in multigrade schools can outperform traditional rural students and even urban students. For instance, the Guatemalan NEU program eval- uation found that NEU schools students have better reading skills than traditional rural school students (Kraft 1998). Similarly, in Honduras multigrade schools score significantly higher in math and science than multiple-teacher schools (Honduras--UMCE 2003a). Additionally, measures of student retention and completion also appear to be improving in some of the programs. In Guatemala, for instance, fewer NEU school students repeat and more students reach the fifth grade than in traditional rural schools (Kraft 1998); while in El Salvador, between 1992 and 2000, the progress in increasing the proportion of age- appropriate enrollment was three times faster in rural schools than in urban ones, also thanks to multigrade teaching. . . . but significant challenges remain. However, significant challenges remain even in coun- tries with broad support for multigrade schools. In particular, in Guatemala, monolingual Mayan students are still extremely disadvantaged in terms of educational completion; and, in all countries, students in rural areas continue to be more likely to be over-age and have lower average educational attainment and academic achievement than urban students. Essentially, by itself, multigrade schooling does not ensure the use of more active peda- gogical methods or better-prepared teachers. The model has shown to be constrained by 42 A World Bank Country Study the continued dependence on traditional pedagogical methods and insufficient ad hoc teacher training and materials. Community-based school management. School-based management (SBM) is a decentraliza- tion mechanism that shifts certain decision-making powers to the school level, emphasizing the role of community and parental management of school affairs. Like multigrade school- ing, school-based management mechanisms exist in all four Central American countries (EDUCO in El Salvador, Centros Autonomos in Nicaragua, PRONADE in Guatemala and PROHECO in Honduras).67 These reforms are some of the most interesting in the region and have led, in some cases, to greater community empowerment and teacher effort, resulting in: (a) a better use of the existing limited capacity of teachers and schools;68 (b) higher coverage in rural areas; (c) somewhat better student flows; and (d) learning outcomes at least as high as in traditional schools (while community-managed schools are generally established in the poorest and most isolated rural areas69). Communities have been empowered . . . In particular, there is some encouraging evidence that communities have been empowered in Central American SBM models. In El Salvador, parents of EDUCO students participate more in school affairs and EDUCO parents are also more likely to engage in day-to-day classroom activities than parents in traditional schools (Jimenez and Sawada 1999; Sawada and Ragatz 2004). There is also positive evidence of community empowerment in Guatemala from recently collected information on PRON- ADE schools,70 which shows higher frequency of meetings between teachers (or directors) and parents. In general, empowerment has largely occurred in teacher management issues (hiring, paying, supervising, firing). It should be noted that the Nicaraguan model showed less clear indication of greater community empowerment. In the case of Nicaragua the reform has successfully expanded the role of the school, but, in contrast to the other coun- tries, this came more from an increasing role of the school director.71 . . . with positive impact on teacher effort. As a result, teacher effort, as measured by teacher attendance, hours worked, and frequency of homework assigned, seems to be generally higher in community-managed schools. Table A.13 summarizes some recent evidence from El Salvador, Honduras, Nicaragua and Guatemala. Evaluations of the different Cen- tral American SBM models report findings such as fewer teacher absences, more weekly hours worked, and less school closings than traditional schools. Existing capacity is there- fore generally better used in community-managed schools. 67. About 39 percent of all rural students from grades 1 to 9 were studying in an EDUCO school in 2001; 21 percent of primary school enrollment in rural areas was attended through PRONADE schools in 2002; about 11 percent of pre-primary and primary school enrollment in rural areas was attended through PROHECO schools in 2004; 63 percent of all primary and secondary students were attended through an Autonomous School in Nicaragua in 2002. 68. More so in El Salvador, Honduras and Guatemala than in Nicaragua. 69. Again, this is less true for Nicaragua. 70. PRONADE database 2001. 71. Porta, Arcia, and Laguna (2004a); Nicaragua Reform Evaluation Team (1996); Fuller and Rivarola (1998); Parker (2004); Gershberg (2003). Central American Education Strategy 43 SBM shows some positive impacts on coverage and learning outcomes. More flexible man- agement, including local teacher hiring, was also conducive to increased enrollment in rural areas,72 and student repetition rates and educational attainment are somewhat better in community-managed schools.73 Finally, the evidence indicates that community-managed schools may have enhanced learning in El Salvador and Honduras (where results are at least as high or higher than in tra- ditional rural schools, in spite of a disadvantage in terms of parental assets; see Table A.14) and that this improvement can be related to capacity maximization, including increased teacher effort.74 Evidence of positive impact on learning is less clear-cut in Nicaragua (where autonomous schools do not do any better than traditional schools although they have higher assets) and Guatemala (where PRONADE schools do worse than traditional rural schools, although they do as well when their lower parental assets are taken into account). Limitations of the model. The SBM model provides three of the elements identified in Effective Schools (discussed above): higher parental involvement in school management; a potentially instructional leadership role for directors; and less student absenteeism. Nonetheless, it ranks poorly in terms of teacher education and experience75 (Table A.15), on the adoption of active/innovative teaching methodologies (the existing evidence shows no signs that teaching pedagogy changed in community-managed schools76) and on the substantive and supportive teacher involvement in the schools (there is evidence that teachers feel little empowered in community-managed schools77). These "missing pieces" limit its impact on educational achievement and other education outputs and outcomes. In particular, lower teachers' skills and empowerment are themselves a constraint on the application of innovative pedagogical practices. Bilingual education. All countries, except El Salvador, have by now established bilingual education programs, although this type of program is most advanced in Guatemala (also the country with the largest proportion on non-native Spanish speakers). Bilingual pro- grams in the region not only attempt to improve the quality of schooling for non-native Spanish speakers by teaching them Spanish as a second language as well as ensuring 72. In particular this is the case in El Salvador and Guatemala. 73. For instance, Jimenez and Sawada (1999) find that the EDUCO program contributes significantly to decisions to remain in school beyond grade three. Although results are less clear-cut for Guatemala, Marshall's (2004a) data from rural Guatemala show that PRONADE schools have done a better job than traditional schools at keeping students in school. Finally, for Honduras, Di Gropello and Marshall (2004) find some evidence that community schools have lower rates of dropout and grade failure than traditional schools in rural Honduras. Autonomous schools of Nicaragua are found to have a positive impact on dropout and repetition in the (1997­2004) time period in Porta, Arcia and Laguna (2004), although this contrasts with evidence collected only for 2003 by the MECD. 74. Rigorous analysis of this link is provided in Sawada and Ragatz (2004), Di Gropello and Marshall (2004) and Chapter V of World Bank (2004e). 75. In particular in Honduras and Guatemala. 76. See, for instance, Parker (2004) on Nicaragua, Di Gropello and Marshall (2004) on Honduras and Marshall (2004b) on Guatemala. 77. See King, Ozler and Rawlings (2001) on Nicaragua and Di Gropello and Marshall (2004) on Honduras. 44 A World Bank Country Study dominance in their native language, but these programs also aim to close the cultural gap that commonly exists between ethnic minority populations and mainstream schools. Towards this end bilingual schools have their own curricula that supplement national stan- dards with subjects, contents, and materials that teach about the history, culture, and lan- guage of diverse ethnic groups in the region. Bilingual education can have a positive impact on students and communities. In Guatemala, substantial investment in bilingual intercultural education (EBI) has gone to the develop- ment of a national infrastructure to support the program. This infrastructure includes experts in linguistics, curriculum development, and educational planning as well as trained bilingual teachers and bilingual intercultural education materials (Dutcher 1997; Patrinos and Velez 2004); Vawda and Patrinos 1999; (Cummings and Tamayo 1994). Studies of the Guatemalan EBI program indicate that indigenous children in bilingual schools outper- form indigenous students in traditional schools in nearly all subjects (Dutcher 1997). They also have higher attendance, lower repetition, and lower dropout rates. Importantly, chil- dren in EBI schools scored just as high as indigenous children in traditional schools in tests of knowledge of the Spanish language. Bilingual education is still little developed in Central America. Even in Guatemala, where bilin- gual education is most advanced, only 18 percent of indigenous children attend bilingual schools. In the other countries this proportion is even lower. Because of the need to train experts and teachers, develop new curricula and teaching and learning materials, bilingual education requires a significant initial investment. Bilingual schools are also often constrained by lack of teacher preparation and insufficient or inappropriate classroom materials. Conditional cash transfer programs CCT programs such as those in Mexico and Brazil, have proven effective at increasing attendance in school, educational attainment, and boosting enrollment.78 These programs give money to families--usually targeted to poor families only--on certain conditions such as a target attendance rate in school or a minimum number of doctor visits per child per year. They can address both direct and opportunity costs of schooling. Positive impact of CCT programs on school attendance and early dropout. CCT programs have been applied in Honduras and Nicaragua, with varying levels of success (Box 3). In general, they had a positive impact on school attendance and early dropout in both coun- tries, while their effect on enrollment was more ambiguous. There is evidence that the CCT program of Honduras was poorly designed and implemented, suggesting scope for com- parison with the better developed program in Nicaragua. Poor quality of the supply can also constrain these programs, requiring coordination with supply-side interventions. Suggested Policy Actions Existing Supply-side Interventions Need to be Strengthened and Improved. Multigrade, SBM, and bilingual education models are promising innovations in Central America that attempt 78. See, for example Skoufias and Parker (2001) on Progresa (Opportunidades) in Mexico, and Bour- guignon and others (2002) on Bolsa Escola (Bolsa Familia) in Brazil. Central American Education Strategy 45 Box 3. Conditional Cash Transfers Can Help Increase Human Capital Nicaragua's Social Protection Network (RPS) gives poor families cash transfers in exchange for attending health workshops, bringing their young children to regular healthcare appointments, and keeping their 7­13 year-old children enrolled in and attending school. In its target popula- tion, RPS has resulted in: · 18 percent increase in primary school enrollment; · 23 percent increase in school attendance; · 7 percent increase in progression through grades; · 5 percent decrease in child labor (age 7­13); · Improved child health and nutrition; · Greater improvements for the extreme poor and families in areas going through economic downturn. Source: Maluccio 2004. In Honduras, the PRAF (Programa de Asignaciones Familiares) implements a program to improve Honduran population's human capital through encouraging poor families to use education, health and nutrition services, as well as improving the quality of such services. The program con- sists of several interventions, including demand-side interventions which comprise a set of vouch- ers for nutritional status and school attendance (for children attending school until fourth grade--that is, 6­12 years of age) (85,000 families covered). A recent impact evaluation shows that, in its target population, PRAF has resulted in: · A growth of 4.5 percent in assistance rates; · A reduction in early dropouts; · No significant impact on enrolment rates. Source: IPFRI 2003. to directly confront the constraints poor and marginalized children face and improve their educational opportunities. However, higher benefits could be extracted from these models. Reform processes are complex and take time to be fine tuned. The main strategy suggested here is to improve these existing interventions: address their weaknesses, consolidate their strengths, and propose ways in which they could tackle a wider range of constraints. Community-based School Management: Setting up the conditions for pedagogical improve- ment, improved management and empowerment at the local level and sustainability of the model. The impact of community-based schooling on student flows and learning out- comes could be greatly enhanced by a set of specific actions, which largely aim at setting up the conditions for pedagogical improvement, improved management and empower- ment at the local level, and sustainability of the models. Improving pedagogical processes. Two possible suggested directions for pedagogical improve- ment include: (a) expanding community empowerment, by involving communities in peda- gogical decisions and providing them with enhanced support and information to be able to do this effectively; or (b) enhancing teacher empowerment, by giving them higher pedagogi- cal autonomy and support (while also establishing mechanisms that hold them accountable for their performance), and attracting more skilled teachers through a better alignment of teacher benefits and professional development opportunities with those of the traditional 46 A World Bank Country Study system79--keeping, however, a separation to maintain flexibility. It is clear that the second route will require deeper reforms in the model, towards a more balanced community-teacher based school. On the other hand, the second direction capitalizes on teachers' existing inti- mate knowledge of teaching and learning and may, in the long run, be a more effective means of improving these SBM schools (additionally, there is also no clear evidence that community involvement in pedagogical decisions generates pedagogical innovation). Improving management and empowerment at the local level. This will require improving com- munity support (already fairly well ensured by external trainers, social workers or NGOs, in most of the countries, although sustainability is at stake), disseminating information on school performance to all education actors to facilitate "monitorability,"80 improving transfer formulas,81 training directors, and strengthening institutional capacity of coordinating units. Additionally, at least for primary schools, Nicaragua would be well advised to give a more predominant role to parents in school councils as the other countries do. Ensuring sustainability. Finally, ensuring sustainability requires the urgent implementation of strategies to include teachers and teacher's unions in the reform process, by disseminating information on the reform, negotiating more favourable employment conditions for teachers in the non-traditional sector (see above),82 or granting higher teacher pedagogical autonomy. Teachers, who have not developed "stake-ownership," remain the main opposition and can potentially derail the reform, especially in Honduras and Guatemala. Additionally, it should be gradually envisaged to substitute the coordinating units by departments or offices that are fully part of the structure of the ministries of education (as is already the case in El Salvador and Nicaragua). Finally, it is imperative to undertake analyses of the budgetary implications of the absorption, improvement and expansion of the models. Community-based schooling could also be potentially applied in poor urban areas, although further measures may be necessary to adapt the model to this environment. The model would probably end up being more of a teacher-based schooling model with community participa- tion because of the possible difficulties for mobilizing families as effectively as in rural areas and the existence of a larger pool of skilled teachers. Multigrade Schooling: Providing more pedagogical support to teachers. Similarly, the poten- tial of multigrade schooling could be more fully realized by making sure that teachers are 79. Comparing teacher salaries and benefits across SBM and traditional schools, we see that, for a sim- ilar level of qualifications, in El Salvador and Nicaragua salaries are identical across the two systems (the possibility of raising fees would in fact allow salaries to be higher in the autonomous schools of Nicaragua), while in Guatemala and, particularly, Honduras, salaries are higher in the traditional system. All other monetary benefits are generally higher in the traditional system (where, in particular, there are more fringe benefits), with the exception of Nicaragua. Finally, contract tenure is typically fixed-term, yearly renew- able, for SBM models, which constitutes one of the most salient features of the models, and was designed to create an incentive for teachers to perform well. 80. In particular, it is important that communities count with good information on school and stu- dent performance to monitor teachers more effectively. 81. Transfer formulas should be modified, by basing the allocation of funds on the real demands that may come from the school councils themselves (through their School Education Plans), ensuring there- fore higher shares for non-salary recurrent expenditure for materials and the like, and introducing allo- cation criteria that take into account equity and performance (provision for small schools in Nicaragua, measures of student attendance or retention in Honduras, Guatemala, and El Salvador). 82. While maintaining fixed-term contracts, which is at the very core of the incentive structure of autonomous models, a merit system (escalafon), similar to the one applied in the private sector, should be designed and implemented for all countries in the non-traditional sector to retain teachers (including regular salary increases based on seniority and qualifications, increased responsibilities, training options, and so forth). Central American Education Strategy 47 specifically and adequately trained and experienced in multigrade teaching techniques and that they obtain the ongoing support necessary to successfully apply these methods. Overall, improving community-based and multigrade schooling will help extract higher benefits from fairly cost-efficient delivery models. Bilingual Education: Developing teaching skills and materials for ethno-linguistic groups. Serious challenges remain in bilingual education in all countries, as indicated by the per- sistent indigenous/non-indigenous gap. Most strikingly access to these programs needs to be expanded. Expansion may require a significant increase in state financing for these pro- grams, as implementation requires teacher preparation as well as curriculum and material development for new ethno-linguistic groups. It would also be timely to start evaluating the programs in Nicaragua and Honduras and make a more effective use of the lessons that can be drawn from the experience of Guatemala. Demand-side Initiatives and Other Broader Interventions Are Also Required to Improve Learning and Completion. Improving primary completion, particularly of the poor, will also require applying some effective demand-side policies and implementing some broader interventions. Applying effective demand-side policies that address learning, enrollment, and timely entrance. Poli- cies to improve children's nutrition status and national publicity campaigns will not only be use- ful to learning itself but also to encourage enrollment and timely entrance to the school system. Analyzing the scope for conditional cash transfers. We have seen that conditional cash transfer programs have been applied in Honduras and Nicaragua, with varying levels of success. These programs may be of limited applicability for the primary level in Central America because they could easily be prohibitively expensive. However, further analysis of their costs and benefits in poor countries should be undertaken. If well designed and implemented in coordination with supply-side interventions these programs (as more traditional scholarship programs) could have a positive impact on the schooling of the very poor. These programs may have a wider application at the secondary level and are discussed further in the next section. Abolishing fees in primary. Another way of dealing with private costs to schooling could be simply through a broad supply-side intervention which eliminates school fees in primary and provides free textbooks and uniforms for poor primary school students. These policies may be worth attempting in Guatemala and El Salvador, following the example of Nicaragua which is attempting to enforce the voluntary nature of primary school fees. For comparison purposes, several African countries also recently abolished school fees at the primary level and the response from families in terms of enrolling their children in school was massive.83 Increasing public education spending and improving budget allocation. Increasing the amount of public education spending84 and ensuring sufficient learning materials by improving budget allo- cation towards higher shares of non-salary recurrent costs (which will also follow from better 83. See, for example, World Bank (2004f). 84. Education For All simulations in Honduras, Nicaragua, and El Salvador show that public funds for primary would not need to increase in terms of GDP, or would even decrease, if we assume substan- tial efficiency improvements (decreased repetition). While it is likely that repetition will decrease (other- wise we would not reach universal completion in the first place) and that this will produce substantial savings, it is also clear that addressing the special needs of the most disadvantaged groups will require addi- tional resources which have not been considered in the different scenarios. 48 A World Bank Country Study teacher management and scaling-up of cost-efficient delivery models) would both help address the completion challenge. These two areas for improvement were discussed above. Improving aid effectiveness. Finally, sector-wide approaches in the use of external funds will also be key to ensuring higher primary completion by setting up a common framework for donor intervention with common outcome indicators and expenditure targets. The new Edu- cation For All-Fast Track Initiatives that are developing in both Honduras and Nicaragua at the primary level represent a shift from traditional project-based approach to a programmatic sector-wide approach, with five main focuses: (a) policy dialogue; (b) results; (c) sustainability; (d) donor harmonization; and (e) in-country dialogue. It is hoped that by helping design and implement a consistent and coordinated framework for intervention in the primary sector, these initiatives will maximize the impact of external funds in the subsector. Regional Priority #3: Expanding Secondary Coverage By Adopting a Comprehensive Policy Package to Address Supply and Demand-Side Constraints The final Central American education priority proposed in this report is that of expand- ing enrollment in secondary schools. This section will first address the primary constraints to secondary coverage. It will then lay out a case for why public monies are important to invest in this goal. Finally, it suggests various policy options to overcome these constraints. Constraints to Secondary Enrollment An analysis of the reasons for non-attendance in the 13­15 and 15­18 age range under- taken from household surveys (see Table A.16) reveals four main types of constraints that prevent prospective students from enrolling in secondary schools. These are: (a) work- related reasons; (b) economic-related reasons; (c) learning-related reasons; and, to a lesser extent, (d) lack of supply (of schools and/or teachers). Need to Work. The need to work, either in the home or outside, appears to be the first or sec- ond most significant factor of non-attendance in the 13­15 age range in all four countries and the most important factor in the 16­18 age range. This finding confirms the importance of age/work-related issues that emerged from the survival rate analysis discussed earlier in this volume. Economic Constraints. As mentioned earlier, the private costs of schooling are much higher at the secondary level than at the primary level in all countries (see Figure 25). Household surveys show economic constraints to be very relevant in El Salvador and Guatemala, where private costs are also particularly high (in Guatemala sending one child to secondary requires an investment of almost 50 percent of the annual income of a poor household and 15 percent of the annual income of a middle-lower income household; see Table A.11). However, in all four countries, families with multiple school-age children, and poor and extremely poor families, may face prohibitively high private costs. Poor Quality. A substantial share of the stated reason for non-attendance in the 13­18 age range lies, again, in "lack of interest" in schooling. This explanation suggests poor learning, limited relevancy, and low quality of schooling. As expected, this factor is particularly impor- tant in the 13­15 age range, which corresponds to primary completion and lower secondary. Additionally, the rate of return analysis shows that learning and quality are particularly prob- Central American Education Strategy 49 Figure 25. Private Spending per Student per Year, 2000­2002 400 350 300 year per 250 200 student 150 per $ 100 US 50 0 El Salvador Guatemala Nicaragua primary secondary Source: Table A.11. lematic at the lower secondary level and for low-income groups. These low perceived rates of return on the labor market can also represent a demand constraint, in particular if they are combined with the high opportunity costs of schooling in the 13­18 age-range. Lack of School Places. Finally, pure supply-related factors (such as presence of a school or grade) account for only a minor share of the explanation for non-attendance, with the exception of Nicaragua in the 13­15 age range where this lack of supply is more acute. The impact of these constraints, however, is bound to be underestimated in household surveys. The very low transition from grade 6 to 7 in Honduras points to still severe supply-side constraints in lower secondary in that country. To a minor extent, the low transition rates of Nicaragua and Guatemala, also suggest supply-side constraints. With little perceived benefit of attending lower secondary schools, high private direct costs, bad schools, and a pressing need to work, it is no surprise that secondary enrollment rates remain so low. These constraints vary somewhat across urban and rural areas and socio- economic groups, with lack of supply being more of an issue in rural areas and for upper sec- ondary education and economic related reasons more of an issue for lower income families. The Case for Public Intervention A necessary first step in the analysis of how to expand secondary education is a discussion of the role of the public versus the private sector in secondary education. Governments Have a Key Role to Play in Promoting Broad-based Expansion in Secondary. East Asian countries offer insight into the important role of governments in ensuring that the evolution of education systems in developing countries supports national growth and com- petitiveness. In East Asia, governments invested in education in a sequential fashion, first expanding primary to generate access and quality for all children, then invested in secondary 50 A World Bank Country Study to assure the same at that level, and now are investing in the tertiary level. This sequential investment has guaranteed speed and balance of transitions between education levels. Independent of the important impact on national growth and competitiveness, other Latin American, Asian, and OECD countries, enjoy wide support for mass schooling for reasons of equality of opportunities and social cohesion. This consensus necessitates a strong governmental role in a sense to comply with its social contract with its citizens. Mass schooling is also frequently seen as integral to the construction of a national identity, in countries as diverse as Cuba and the United States. Broad-based Expansion in Lower Secondary Will Require Public Intervention in Central Amer- ica. In Central America, like elsewhere, public intervention will be necessary to promote the necessary broad-based growth in lower secondary. There are at least two reasons for this. First, as shown above, private rates of return to lower secondary are low in all coun- tries, creating a major disincentive for enrollment, especially when coupled with the increasing opportunity costs of attending school in the 13­18 age range. A public inter- vention should not be decided on the basis of the private rate of return but on the basis of the difference between the social and private rate of return, where the social rate of return includes private benefits and external effects (or externalities).85 Social rates of return are, however, very difficult to measure. As such, we are just left with two pieces of information: (a) low private rates of return for lower secondary suggesting low quality and relevance; and (b) quantitative evidence that broad-based secondary education has a strong impact on the accumulation of human capital and national growth, suggesting substantial positive exter- nalities. In this context, we can probably assume that social rates of return are higher than private ones and, therefore, that, without public intervention, a socially suboptimal equi- librium level of lower secondary education will be reached. A public intervention may then act at different levels to ensure the social optimum: (a) implement policies which improve the quality and relevance of lower secondary encouraging higher demand from families and the labor market for this subcycle; and/or (b) reduce the opportunity costs of schooling by providing demand-side subsidies and/or developing alternative secondary education modal- ities which make it possible to combine work with schooling. Second, rates or return to secondary are shown to be particularly low for low-income groups, while direct private costs are particularly high for these same groups, creating addi- tional disincentive to attend for low-income people. Therefore, there will be scope for public intervention on pure equity grounds. Public intervention may encourage higher participation of low-income groups in secondary education by implementing policies that improve the quality of the schools that they attend and/or reducing the burden of private costs through demand-side subsidies or other policy measures (other public policies such as improving the access of the poor to highly paid jobs will also be useful but go beyond the education sphere). Public Intervention in Secondary Education is Still Too Limited and Selective in Central America. Such an approach has not generally been adopted in Central America, largely because of a lack of widespread social consensus that all children should, and have the right to, attend secondary (or at least basic education86), and too little acknowledgement of the critical role of secondary education for national growth and competitiveness. Interestingly, some of the Central American socialist movements, particularly in Nicaragua, emphasized the impor- tance of mass education and literacy for these reasons, although this consensus has broken 85. On this point, see also Hammer (1996). 86. Primary plus lower secondary. Central American Education Strategy 51 down in the post-socialist societies (or if it exists it applies to the primary level of schooling rather than extending into secondary). This lack of consensus has resulted in a still too lim- ited and "selective" public intervention in secondary education. El Salvador was somewhat more successful in managing its public intervention because a broad consensus emerged around the need for compulsory and universal basic education in the country. With new research confirming the critical role of secondary education for national growth and com- petitiveness we hope that there will be renewed broad commitment to secondary (or basic) education for all Central Americans with a resulting new role for the government. Secondary education should be given increasing priority in the national budget. National education account analysis would be needed to assess the relative proportion of public and private financ- ing of secondary education. Unfortunately, we only have partial information on private financ- ing (household spending in public and private schools) and public financing (public financing in public schools), which limits the completeness of our analysis. Our limited information sug- gests that private financing represents between 50 and 80 percent of total secondary education spending in Central American countries (see Table 6). This is a substantial share which indicates that higher priority should be given to secondary education in the national budget. Public intervention should ensure a favorable environment to equitable secondary education coverage. Much larger proportions of students are enrolled in public schools at the lower sec- ondary level than the upper secondary level underscoring the importance of public intervention at this level (see Table 7).87 Despite this, public enrollment in secondary is composed primarily of Table 6. Total Education Expenditure in Secondary Education by Source of Financing (US$ thousands) El Salvador, Guatemala, Nicaragua, 2002 2001 2002 Total education expenditure 219,000 215,000 52,200 Estimated public expenditure 108,000 40,400 13,900 Estimated private expenditure 111,000 174,300 38,400 Percent of estimated private exp. 51% 81% 74% Source: Table A.10 and household surveys. Table 7. Public and Private Enrollment Share in Secondary Education (percent) El Salvador, 2002 Guatemala, 2000 Honduras, 2003 Nicaragua, 2001 7­11 7­9 10­11 7­11 7­9 10­11 7­11 7­9 10­11 7­11 7­9 10­11 Public sector 80 84 72 46 55 24 78 82 68 73 75 66 (80) (86) (68) (75) (79) (68) Private sector 20 16 28 54 45 76 22 18 32 29 25 34 (20) (14) (32) (25) (21) (32) Notes: Between brackets: Data of the MINED, 2002 for El Salvador and 2003 for Honduras. Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2003; Nicaragua, EMNV 2001. 87. That lower-secondary is increasingly becoming part of "basic education" (Holsinger and Cowell 2000). 52 A World Bank Country Study students from the middle-upper income quintiles. Only in El Salvador is there a more equitable distribution of students in public secondary due to widespread lower secondary (see Figure 26). There have been few interventions in Central America that aim to overcome this inequitable enrollment distribution by reducing the opportunity costs or private direct costs of schooling. An exception is the recent development of alternative secondary modalities in rural areas,88 which contributed to a reduction of opportunity costs of schooling. Private costs in the public sector are as high or higher than the public subsidy for the poorest quin- tile in Nicaragua and Guatemala (see Figure 27). Guatemala and Nicaragua need to estab- lish a much more favorable environment to equitable secondary education coverage. Greater Private Financing of Upper Secondary May Be an Appropriate Option in Guatemala, Honduras, and Nicaragua. In Guatemala, Honduras, and Nicaragua ministries of education are not likely to have enough funds for the expensive tasks of expanding both lower and upper secondary in the short term.89 Because of this these countries may wish to explore more pri- vate financing options. Increased private financing may also be appropriate in that because few students are making it through lower secondary, limited expansion of upper secondary is a reasonable goal for the short term. In addition, private rates of return are high in upper secondary education, and may therefore make up for the higher opportunity costs at that level. High private rates of return may also suggest that the price elasticity of demand for upper secondary education will be lower than for lower education (in other words, higher direct costs, as paid in the private sector, would have less of an impact on the demand for schooling at that level). Increasing private financing does not mean abandoning or cutting public support, which would remain important for creating an environment that is con- ducive to private sector participation and expansion, and the mobilization of private resources from companies, NGOs, and so forth. As a larger and more diverse pool of students graduate from lower secondary schools, the government would then need to take increasing responsibility to ensure wider access to upper secondary. Scope for larger public intervention in upper secondary in El Salvador. In El Salvador, by con- trast, there is scope for a larger public intervention in upper secondary, especially in terms of cost-effective modalities, demand-side subsidies, mixed public-private schemes, and so forth. Efforts should, however, be made to minimize the crowding-out of the existing pri- vate financing (in other words, upper-income people who are willing to pay for upper sec- ondary should continue to do so). Suggested Policy Options Only a Comprehensive Policy Package Will Successfully Overcome the Range of Constraints Fac- ing Secondary Enrollment. The barriers that prevent more youth from enrolling in secondary are multiple and they will need to be addressed using multiple interventions. Furthermore, the constraints vary somewhat from country to country as do the fiscal possibilities for secondary expansion according to the different stages each country is at in terms of access to and quality 88. Such as Telesecundaria in Guatemala and Educatodos in Honduras. 89. As also recently confirmed by Porta and Arcia (2003) in their financial gap analysis for Nicaragua. Figure 26. Public and Private Enrollment by Income Quintiles in Secondary El Salvador Guatemala 50 60 45 ) 50 (%) 40 %( 35 40 30 25 30 Enrollment enrollment of 20 of 15 20 10 Percent Percent 10 5 0 0 I II III IV V I II III IV V Income Quintiles Income Quintiles public enrollment 2002 public enrollment 1995 Public enrollment Private enrollment private enrollment 2002 private enrollment 1995 Honduras Nicaragua 60 60.00 Central 50 50.00 (%) (%) 40 40.00 American 30 enrollment 30.00 Enrollment of of 20 20.00 Education Percent 10 Percent 10.00 0 0.00 I II III IV V I II III IV V Income Quintiles Income Quintiles Strategy Public enrollment Private enrollment Public enrollment 2002 Public enrollment 1995 Private enrollment 2002 Private enrollment 1995 Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2003; Nicaragua, EMNV 2001. 53 54 A World Bank Country Study Figure 27. Public and Private Expenditure Per School-Age Population in Public Secondary Education Average Nicaragua First quintile Average First quintile Guatemala Average Salvador First quintile El 0 20 40 60 80 100 120 140 US $ public exp per capita private exp per capita Source: El Salvador, EPHM 2002; Guatemala, ENCV 2000; Nicaragua, EMNV 2001. of secondary. As mentioned previously, we recommend that in Nicaragua, Guatemala, and Honduras a minimum goal should be to generate broad-based expansion at the lower sec- ondary level. In El Salvador we recommend broad-based expansion of upper secondary. Poli- cies will clearly need to vary somewhat across countries based on these differences. Distance Education and Other Flexible Delivery Mechanisms Can Be Very Promising, Although They Need to Be More Fully Assessed. Flexible delivery mechanisms demonstrate good poten- tial to minimize work, economic, and supply-side constraints. Flexible delivery mechanisms use new pedagogical approaches designed to better serve rural and low-income communities as well as working individuals. They meet the needs of low-income communities by offering a free or low-cost alternative to traditional secondary. Many expand access to rural commu- nities that have not had access to secondary school without the high costs of building and staffing a school. And they allow students to continue working or raising a family while also studying and earning their secondary degree.90 Flexible delivery mechanisms would be partic- ularly advisable in Honduras (due to the existing fiscal constraints), as well as in Nicaragua and Guatemala, where they would constitute a valuable tool to improve coverage with equity. Typically, these programs are offered outside the traditional school day, require fewer hours to complete, do not have traditional teachers, and are geared more toward skill rather than academic training. One well-known example in Central America is the Honduran Edu- catodos. Thisprogramoffersbasiceducationtoyoungpeopleandadultsoutsidethetraditional study system, delivered by trained facilitators, who work with study groups outside normal work hours, using approved texts and taped programs. Another example is the Tutorial Learn- 90. A discussion of the benefits of alternative models for secondary education is also provided in Figueredo and Anzalone (2003). Central American Education Strategy 55 ing Systems project (or SAT--Sistema de Aprendizaje Tutorial), founded in Colombia in the 1970s, an alternative model for high school education, oriented toward the rural sector, based on the participative presence methodology, a comprehensive curriculum and content geared toward developing abilities and competencies relevant to student lives. Another example still is Telesecundaria in Guatemala, a distance education program, based on television, which only requires one facilitator per grade. Unfortunately, these flexible secondary programs are less common in urban areas. While scattered populations and lack of teachers and schools make flexible delivery particularly important in rural areas, urban youth often face intense pressure to work and could therefore also benefit greatly from these services. These programs have had some success in improving secondary enrollment, particu- larly in marginalized areas (Box 2 in Appendix C provides a synthesis of alternative sec- ondary models in El Salvador and Honduras). Although promising, distance education programs such as Educatodos and SAT have not received sufficient assessment. A key ques- tion remains as to the quality of the education offered through distance education programs. If these delivery mechanisms do not ensure quality, they, in effect, set up a parallel and infe- rior education system for the poor. This should be vigorously avoided. One way to measure this would be to analyze the relative rates of return to traditional and alternative secondary programs. Developing a strong testing and accreditation system will help ensure more uni- form quality standards across secondary education options. Along these lines, a key chal- lenge of these programs will also be to ensure continuous quality of outcomes as programs go to scale. This may require additional per pupil expenditure. Additionally, it will be important to assess and expand flexible programs for marginal urban areas where the opportunity costs of schooling are particularly high due to more dynamic labor markets. While assessments of and improvements to distance education programs should be sup- ported in all countries, it is clear that they should not replace traditional secondary education. Other complementary interventions to ensure wider access to traditional secondary education will be needed. Demand-side subsidies (financed by public but also private funds) could ease the access to traditional schooling for low-income students. They would so by addressing eco- nomic constraints and opportunity costs. On the other hand, cost-efficient interventions such asmixedschemesofpublicfinancingandprivatedelivery,ruralschoolnetworks,school-based management and basic education centers could provide wider access to traditional secondary education to both middle and low income students in urban and/or rural areas. They would do so by addressing supply and learning constraints (mixed schemes of public financing/pri- vate delivery and basic education centers) and supply constraints (rural school networks and community-managed schools). We review below these different options. Demand-side Subsidies and Other Interventions to Expand Coverage May Be Needed. Demand-side subsidies, such as targeted scholarship programs and conditional cash trans- fers have been described above. At the secondary level they could resolve some of the eco- nomic barriers to secondary, and also compensate for the opportunity costs of schooling for low-income students, easing the work constraint.91 There are few examples of condi- tional and targeted cash transfers in Central America, and most of them have been con- centrated at the primary level. Similarly, relatively little use of scholarships has been made. Two exceptions are two scholarship programs set up in Costa Rica and El Salvador. A 91. Related to this, a recent study done on Nicaragua (Laguna 2003) states the importance of extend- ing the conditional cash transfer program of the country also to the secondary level. 56 A World Bank Country Study World Bank evaluation concluded that the upper secondary school scholarship program in El Salvador (which also seeks to mobilize private sector participation in the scholarship program) was a success, with clearly established pre-selection, selection, awarding, and monitoring processes. The key problem facing the program, however, is its sustainability given the lack of funds to continue financing the scholarships.92 El Salvador should be encouraged to further develop scholarship or conditional cash transfer programs and the lessons extracted from its first experience with scholarships used for the design and imple- mentation of similar programs in other Central American countries. Targeted vouchers, which provide a cash payment to families to be used at a school of their choice, could also improve the quality of services by providing school choice and by the same token a possibility of "exit" from inefficient schools.93 It is not clear that vouch- ers are an appropriate option for the Central American countries where much of the pop- ulation lives in areas where schools are too few and far between to offer any meaningful choice. These programs have also been found to exacerbate social segmentation which would be extremely undesirable. Public information and awareness campaigns could, again, play a key role in showing the importance of a full basic education cycle for a child development and opportunities on the labor market. Mixed Schemes of Public Financing/Private Delivery Have Good Potential if There is a Cost- Efficient Private Sector. In urban areas, an analysis undertaken on El Salvador shows that there is substantial potential for promoting mixed schemes of public financing-private delivery, taking advantage of idle capacity and higher cost-efficiency of the private sector (higher academic results, lower recurrent costs). This option has the potential of address- ing the supply constraint in urban areas and easing the learning constraint for middle or low-income students. The sizable role of the private sector in secondary education in all four countries makes this option a valid one, although an assessment of its capacity and cost-efficiency would be needed to establish its potential.94 If there is little extra room in private schools, it may also be envisaged to finance some further expansion or even new private schools, assuming private schools are indeed more cost-efficient than public ones. There are also few examples of public-private partnerships in Central America. Some examples of partnerships exist in Costa Rica and El Salvador, but with limited coverage. Box 4 describes the experience of El Salvador. 92. The hopes for participation of the private sector in helping financing the program have not materialized. 93. The secondary education targeted voucher program introduced in Colombia in 1992, the so-called PACES, can illustrate this type of demand-side intervention. The available empirical evidence on the Colombian PACES suggests that, in the 20 percent that are participating municipalities, the program pro- vided an effective way of increasing educational attainment for the poor students benefiting from the vouchers. These positive results seem to have been the result of a variety of factors, among which munic- ipal participation based on sound criteria, such as limited public school capacity and important private sector excess capacity; participation of schools with educational quality comparable to public schools; and effective targeting of vouchers to poor communities. Political economy considerations (lack of govern- ment's commitment, teacher union's opposition) and administrative difficulties (delays in disbursements, burden of program monitoring) led to the discontinuation of the program at the national level in 1997, but this, however, survived at the subnational level. 94. A recent study of Porta, Arcia and Laguna (2004) shows that private schools have better results than public schools in Nicaragua (controlling for socio-economic factors), which may suggest higher cost- effectiveness. Central American Education Strategy 57 The Scope for Rural School Networks Should Be Analyzed. In contrast, in rural Box 4. Public/Private Partnerships in areas, where there is little private supply, El Salvador school networks up to grade 9 or even 11 or 12 may represent an interesting option The current partnership approach developed in El Salvador consists of providing public for addressing secondary education cov- resources to private organizations, mostly not- erage, addressing the supply constraint for-profit education NGOs or religious groups, to also in upper secondary, although not help establish and administer schools. In some necessarily the quality constraint. cases, MINED pays for the teachers and the counterpart institution provides the rest (infra- structure, materials, administration). In other School-based Management is Also a Pos- cases, the scheme involves contracting out the sibility in Secondary Schools. Secondary administration of the school, with MINED pro- level community-managed schools exist viding most of the teachers and some material resources and parents providing the rest. in El Salvador and Nicaragua. Local Examples of these partnerships are: teacher contracting and community monitoring may make this model partic- · the Fé y Algería schools which have presence in most of Latin America; ularly attractive in terms of increasing · the schools administered by the Dominicas coverage in rural areas, as has been seen in de la Anunciata Congregation; and community-managed primary schools. · the Maristas schools in poor neighborhoods. It is less clear, however, how this model Source: World Bank 2004e. would also result beneficial for improv- ing coverage in urban areas, where there is higher teacher availability and less tra- dition of community participation. While community involvement in school management will help in keeping the schools running (and there is positive evidence of this in Nicaragua), it is not clear if it would be successful at monitoring teacher effort or instigating necessary pedagogical improvements. SBM models that give teachers greater responsibility and skills might be a more promising SBM model for secondary schools. Basic Education Centers Have Good Potential to Improve the Transition to Secondary.Finally, basic education centers that integrate primary with lower secondary in the same school, have good potential to help support the transition to lower secondary. This innovation has been piloted in El Salvador and Honduras with mixed results so far (more favorable in El Salvador than in Honduras). Basic education centers are typically introduced along with legislation establishing a compulsory basic education cycle (usually made up of primary and lower secondary) as well as a more integrated curriculum. They could therefore address both supply-side constraints (up to grade 9), in urban and rural areas, and quality constraints of lower secondary education by improving the relevance of that subcycle. By combining to some extent primary and secondary grades, unit costs of secondary school- ing should also be lower (Lewin and Caillods 2001). Basic education centers should be fur- ther encouraged in Honduras and developed in Nicaragua and Guatemala, using the lessons of the El Salvador case. Learning Improvements Are Essential. Improving both access to secondary and the quality of secondary are fundamental to expanding enrollment in this cycle. Similarly, both sup- ply and demand constraints have to be addressed. The earlier section on learning provides suggestions for that area although other learning/quality improvements are specific to the 58 A World Bank Country Study secondary level such as ensuring that teachers are trained in the secondary level subject that they teach. Additionally, urgent interventions are needed to drastically improve the qual- ity and relevance of lower secondary, including more innovative pedagogical approaches and movement towards a basic education cycle. Budget Shares for Secondary Education and Public Financing for Education Should Be Increased in El Salvador, Nicaragua, and Guatemala. Budget simulation analysis for El Salvador. A simulation analysis for El Salvador suggests that this country will need to spend between 0.9 and 1 percent more of its GDP by 2015 on secondary education to achieve universal coverage in lower secondary (that is, also includ- ing full transition of the poor) and wide coverage in upper secondary (gross enrollment rate of 70 percent). Models for this simulation include cost-effective options such as pri- vate subsidized schools in urban areas and an expansion of school based management in rural areas (already applied to secondary education through EDUCO), as well as quality improvements and expansion of the targeted scholarship program (see Table A.17). These are advised strategic policy options for El Salvador (see above). Even in the least expensive model, the share of secondary education in the public education budget would have to more than double by 2015 in El Salvador. To be able to address the multiple imperative challenges of its educations sector, that is also including across the border quality improvements and universal completion.95 El Sal- vador would need to increase public education expenditure in terms of GDP to levels close to the Latin American average (about 4.5 percent; World Bank 2004b) by 2015. This is assuming that reaching universal completion will also require a small increase of the primary share in terms of GDP, in spite of the gains in efficiency through reduced repetition, because of the necessary quality improvements and the enormous challenge of keeping in school the most disadvantaged groups. Guatemala and Nicaragua. Specific simulations for Nicaragua and Guatemala have still not been finalized. Both will, however, need to increase their secondary education budget share and their overall education budgets as a percentage of GDP to achieve their education targets. This is particularly the case for Guatemala which has enormous challenges in primary and secondary.96 In both countries, there may be some scope for decreasing the public subsidy to tertiary education (although this is easily said than done given the political pressures that countries generally face to keep tertiary heavily subsidized). Budget Shares for Secondary Education Should Be Increased and Efficiency of Public Financing in Education Drastically Improved in Honduras. Budget simulation analysis for Honduras. A budget simulation analysis for Honduras suggests that this country will need to spend between 0.3 and 0.5 percent more of its GDP by 2015 on secondary education to achieve broad-based coverage in lower secondary (gross enrollment rate of 70 percent). These models also include efficiency improvements in secondary (lower repetition97), an increase in non-traditional delivery in rural areas and further development of 95. The current completion rate is already binding in El Salvador. 96. As also confirmed by very recent and still preliminary simulations in Guatemala which show that the education budget share in terms of GDP will have to increase from 2.4 percent in 2003 to about 3.6 percent by 2015, with a doubling secondary education share (from less than 10 percent in 2003 to about 20 percent by 2015). See Di Gropello (2005). 97. Repetition is particularly high in secondary in Honduras. Central American Education Strategy 59 basic education centers (key strategic policy options for Honduras; see above) and limited crowding-out of private enrollment (see Table A.18). A substantial expansion in upper sec- ondary (gross enrollment rate of 50 percent) would require another 0.5 percent of the GDP (in a scenario with efficiency gains and limited crowding-out of private enrollment). Even in the least expensive simulation, the share of secondary education in the public education bud- get would almost double by 2015 (by only expanding lower secondary, the share of secondary education would increase by 50 percent). Drastic reductions in the repetition rate in primary and the quality improvements trig- gered by the Education For All Fast Track Initiative would make it possible to reach universal completion by 2015 at a decreasing primary share of the GDP (World Bank 2004c). However, the share would need to increase in the short run and no allowances are made for the increasing marginal cost of covering disadvantaged groups. Overall, Honduras will probably need an addi- tional 1 percent of its GDP to finance its education reform by 2015 (and that would only cover an expansion of lower secondary). Drastic efficiency improvements will therefore be needed at all levels to address these education challenges within the already substantial education budget (which reaches 7 percent of GDP), including key improvements in teacher management, reduc- tions in administration costs and reductions in the public subsidy to tertiary education. We now turn to a matrix synthesizing this agenda for action in education in Central America. CHAPTER 4 The Regional Matrix T he matrix that we present below summarizes the agenda for action for Central America, focusing on the three interrelated areas of learning, primary completion and secondary coverage. Following the structure of the previous analysis, this summary follows an outcome-based approach, through which each of the interven- tions/policies addressed in the main report (through Chapters II to V) is analyzed in the context of reaching a specific outcome. For each area, the matrix reports some key statis- tical findings (Chapter I), the chief constraints blocking improvement in each of the areas (Chapters II, III, and IV), the main interventions that the countries have implemented in these areas thus far (for which we also have some evidence of impact; Chapters II, III, IV, and V), and finally, suggested policy options to help the region address core challenges (Chapters II, III, IV, and V). It is important to note that the matrix is designed to be reader-friendly and to reflect the findings in the body chapters of the main report. It is not designed to be either particularly nuanced, or exhaustive. For greater nuance it is critical that the reader examine the body chapters of this report. The matrix includes interventions and policy options that address each of the three regional priorities identified (learning, primary completion and secondary coverage). We did not attempt to categorize them according to types of findings/challenges (such as coverage, efficiency, equity) or constraints by regional priority because they gener- ally affect several of these findings/challenges and constraints. For instance, bilingual educa- tion programs support the specific needs of indigenous children who are underserved in the nation's education system and as such could be grouped as an equity enhancing policy. How- ever, they also clearly affect coverage and efficiency in that it is hoped that indigenous com- munities would be more likely to enroll and keep their children in school if they feel more comfortable with the school and believe the education their children receive is useful and rel- evant. Essentially, the matrix can only be read vertically. 61 62 AWorld Bank Table 8. Regional Agenda for Central American Countries Country Findings Constraints Existing Interventions Suggested Policy Actions Primary and Secondary: Improving Learning Study Low educational Low culture of evaluation Regional quality standards at primary 1) Improve scope of national assessment sys- achievement in all Weak teacher education. level (not fully applied, however). tems: start evaluation at grade 1; dissemi- countries, as measured Education for All Fast Track Initiative nate results (report cards); align national by several indicators. Low teacher effort and standards to regional ones to provide basis insufficient instructional in Honduras and Nicaragua aimed at increasing primary completion for well-functioning regional accreditation Standardized testing: time. through general quality improve- system; participate in international exams. Learning outcomes are low Traditional pedagogical ments (management of the sectors, 2) Develop selection criteria for teachers and intermediate or low vis-à-vis practices and lack of capac- performance-based delivery, etc.) accreditation exams which may require to country's standards. ity to use other methods. and other interventions (see below). move towards a system of full teacher certi- Insufficient share of non- National education assessment sys- fication in all countries, like done in El Repetition: salary recurrent expendi- tems have been introduced, with Salvador and Costa Rica. High in all countries in ture in education budget standardized exams used for diag- 3) Develop mechanisms which link salary primary. Particularly high (inadequate classroom nostic purposes (these systems are increases to teacher performance and effort, in grade 1 (about 30%). resources). still limited in scope, lack of results analyzing the feasibility of introducing salary High age-for-grade distortion Poor economic means, lack dissemination). scales promoting higher effective hours of in all countries (average age of parental education, Strengthening of in-service teacher work or team-based merit-pay schemes simi- at grade 6 > = 12.8; at cultural barriers. training (more structured, flexible lar to the one applied in El Salvador. grade 7 > = 14). and decentralized programs). 4) Improve monitoring of teachers' absences Curricular reforms, towards a con- and work hours by involving communities Rates of return: structivist approach in all countries organized in school governments (and, if Substantial but lower than (not fully implemented). possible, flexible contacting). expected for the level of 200 official schooling days in 5) Introduce an official 200 days schooling year educational attainment. Honduras and El Salvador. in Nicaragua and Guatemala or consider other means of increasing instructional time Interventions related to equity of (longer days). learning are developed below. Equity of learning: 6) More emphasis on classroom practices in both Learning is also unequally pre-service and in-service training; develop distributed across urban-rural innovative ways of organizing in-service areas, socio-economic groups teacher support (assess "teachers' micro cen- and indigenous/non- tres" applied in Nicaragua); develop effective indigenous populations. system of school supervisors to ensure continu- ous teacher support in the application of new Persistent urban-rural and methodologies (assess asesores pedagogicos of socio-economic gap in terms El Salvador). of repetition in all countries. 7) Implement publicity campaigns to increase Unequal private rates of return awareness on importance of education and across socio-economic groups. help families support children's academic success in school. 8) Ensure higher shares of non-salary recurrent expenditure (at least 20 percent of total edu- cation budget) to provide sufficient school and classroom resources and quality teacher education and training facilities. Policy actions related to equity of learning are developed below. Primary: Universal Completion Central High, and increasing, Gross Over-age (due to repetition Education for All Fast Track Initia- 1) An analysis of low-income high-performing and Net Enrollment rates and/or late entrance). tive in Honduras and Nicaragua schools suggests that policies to attract and American (GER: > = 100%; NER: > = 80%), Low quality of the cycle aimed at increasing primary com- retain competent teachers in poor areas, but internal efficiency is (see above), with direct and pletion to 100% through general promote interactive pedagogical techniques still too low (due to low indirect (through repetition quality improvements, rural school and encourage a substantive involvement of performance of the poor), Education and over-age) impact on networks, and schooling of dis- teachers and parents in school management although improving. completion. advantaged student populations will be particularly useful to improve the (indigenous, over-age, etc.). performance of the poor. Completion/Survival: Substantial enrollment increase 2) Consolidate community-based schooling in Strategy Gross primary completion rate in rural areas in primary promoted rural and poor areas, maintaining high levels still < 80% for all countries; by multigrade and, above all, com- of community empowerment. To enhance < 70% for Nicaragua and munity-based schooling (EDUCO, impact on student flows and learning out- 63 (continued) 64 Table 8. Regional Agenda for Central American Countries (Continued ) Findings Constraints Existing Interventions Suggested Policy Action AWorld Guatemala. Gap for level of In particular, low learning Centros Autonomos, PRONADE, PRO- comes, and improve sustainability, other Bank income per-capita. Projected of the poor (low-income HECO). Some positive impact of com- specific actions include: provide pedagogical probability of survival of the schools have less textbooks munity-based schooling also on strength to the models, including ensuring Country 7 year cohort to grade 6 and less skilled teachers; retention, repetition and learning higher levels of teacher pedagogical auton- between 68% and 83%. lack of parental education, outcomes (less so in Nicaragua). omy and support; ensure a better alignment linguistic and cultural barri- Some positive impact of multigrade of teachers' monetary benefits with the tra- Study Equity of completion: ers are not conducive to teaching (grades 1 to 6) in rural areas ditional system; while maintaining flexible Although improving, inequity learning). in all countries (improved student contracting, implement a merit system, is still substantial. Urban-rural Poverty is also a direct con- retention and learning outcomes, which provides professional development gap generally small and straint to enrollment (high decreased over-age). opportunities. decreasing in primary coverage private costs in % of house- Limited accelerated education, auto- 3) Develop and strengthen multigrade school- but still substantial for primary hold income). matic and flexible promotion pro- ing by providing continuous pedagogical completion (between 40 and In rural areas: difficult grams implemented or to be support, including ad hoc textbooks and 60% less in rural area). access in remote areas, lack implemented. teacher training. Socio-economic differences of teachers, especially skilled Still limited bilingual programs in 4) Develop and strengthen bilingual education have tended to shrink and are teachers, incomplete schools three countries (have improved programs by providing increased access to now generally low in terms of (Nicaragua, Honduras). learning outcomes in Guatemala). indigenous populations, continuous peda- primary coverage but are Insufficient share of non- gogical support, relevant resources and still substantial for primary Monetary incentives for teachers in salary recurrent expendi- specialized teachers. completion (between 55% difficult and rural areas. ture in education budget 5) Combine monetary incentives for teachers to and 75% less for the lowest (constraint for the applica- Varying impact of conditional cash work in rural areas with other policies that quintile compared to the tion of programs for the transfers in Honduras and Nicaragua. encourage high-quality teachers to work and top one). poor and disadvantaged). remain in rural and difficult areas (such as Although improving, the gap improved teacher education or scholarships between indigenous and non- for local teachers). indigenous is substantial. 6) Consider putting official start of primary at 6 and enforce this policy. 7) Implement publicity campaigns to increase awareness on importance of education, pro- mote enrollment and help families support children's academic success in school. 8) Implement combined education-heath inter- ventions focused on nutritional status in pre- primary and primary. 9) Assess the scope for conditional cash trans- fers; abolish school fees. 10) Improve allocation of funds, towards more non-salary recurrent expenditure, to ensure provision of school resources and quality teacher education. 11) Improve reporting and coordination of exter- nal funds (less duplications of activities in pri- mary, coordinated interventions to reach common targets). 12) Increase public financing for education in proportion of GDP (except in Honduras). Secondary: Broad-Based Expansion (lower secondary in Guatemala, Honduras, Nicaragua; upper secondary in El Salvador) Coverage: Over-age in grade 6 and 7. Some initial positive impact of, the 1) Assess and develop flexible delivery systems Still too low, although increasing. Low learning in primary. still too limited, distance education such as distance education programs (see Educatodos) (also for urban areas). The GER ranges between 39% programs (El Salvador, Honduras Age/work related reasons. and 61%; the NER between 28% [Educatodos and SAT for indigenous 2) Consider promoting basic education centers Central and 47%. Secondary gap for level Economic constraints communities], Guatemala [Telese- (see experience of El Salvador). of income per-capita. (especially high private cundaria], etc.). 3) Explore the potential for public-private part- costs to attend secondary in Basic education centers have devel- nerships in financing and delivery (analysis of American Internal Efficiency: Guatemala and El Salvador). oped with different levels of success cost-efficiency of private schools). Low transition rate from grade 6 Low quality/relevance of the (in El Salvador and Honduras). 4) Innovate with rural school networks, up to to 7 in three countries. Gross cycle (in particular lower sec- Some limited experience with grade 9, or even 11 or 12. Education secondary completion rate ondary where private rates scholarships. 5) Assess and analyze the potential for commu- < = 40% in all countries. Projected of return are low). Community-based schooling up to nity-based schooling for grades 1­9 or, even, probability of survival of the 7­17 Lack of supply in lower sec- grade 9 (El Salvador) or 11 (Nicaragua), secondary, in rural (and urban) areas. cohort to grade 11 is between ondary in three countries. Strategy with substantial impact on enroll- 6) Develop targeted scholarship programs and 45% and 69%. In upper secondary in all ment. Lack of evidence on the impact conditional cash transfers, and/or other countries. on student retention and learning policies to deal with private costs (such as outcomes. subsidized texts or fees for the poor). 65 (continued) 66 AWorld Bank Country Study Table 8. Regional Agenda for Central American Countries (Continued ) Findings Constraints Existing Interventions Suggested Policy Action Equity: Lack of teachers, skilled Some examples of public-private 7) Improve quality and relevance of lower sec- Urban-rural gap very substantial teachers and supply in partnerships. ondary, by applying the interventions men- (between 40 and 70% less for the rural areas. Monetary incentives for teachers in tioned above for learning, promoting a basic NER; and between 70 and 90% Low budget priority for difficult and rural areas. education reform and assessing the potential less for the GCR). secondary education. for new pedagogical approaches such as the development of Tele-aprendizaje. Socio-economic differences very substantial for secondary 8) Increase public financing for secondary edu- enrollment and completion rates cation (also share of the overall education (between 80% and 99% less for the budget); decrease public subsidy for tertiary lowest quintile compared to the education; encourage private financing in top one for the GCR). upper secondary (less so in El Salvador). Indigenous gap. APPENDIXES APPENDIX A Comparative Tables 69 70 A World Bank Country Study Educational Outcomes Table A.1. GER and NER in Pre-Primary, Primary and Secondary (MINED) Pre-Primary Primary Secondary Year GER NER GER NER GER NER El Salvador 2002 48 43 93 82 (1­9) 48 24 (10­11) 105 85 (1­6) 61 47 (7­11) Guatemala (a) 2001 55 41 103 85 33 22 (7­12) 2003 na na 106 89 38 23 Honduras (a) 2001 57 na 108 na 39 (7­11) na 2003 54 na 119 na 54 (7­11) na 48 (7­12) na Nicaragua (a) 2001 26 26 104 82 54 36 (7­11) 2002 28 28 109 86 57 38 2003 29 29 106 84 59 39 Costa Rica 2002 87 86 (b) 105 99 66 59 (7­12) Latin America 2000 58 na 130 97 86 64 Countries with similar per capita income Peru 2000 59 128 104 81 61 Colombia 2000 37 112 89 70 57 Jordan 2000 31 101 94 88 76 Romania 2000 73 99 93 82 80 Algeria 2000 3 112 98 71 62 Paraguay 2000 83 111 92 60 47 Ecuador 2000 69 115 99 57 48 Philippines 2000 na 113 93 77 53 Bolivia 2000 46 116 97 80 68 Note: (a) Previous years have been kept for comparison with other indicators; (b) Only includes 6-year old. Source: Official Statistics of the MINED; The World Bank/WDI (2003). Table A.2. GER and NER in Pre-Primary, Primary and Secondary (Household Surveys) Pre-Primary Primary Secondary GER NER GER NER GER NER El Salvador (2002) 44 41 95 85 (1­9) 47 30 (10­12) 105 87 (1­6) 64 52 (7­11) Guatemala (2000) 23 21 99 78 39 28 (7­11) 31 25 (7­12) Honduras (2002) 37 30 105 85 50 35 (7­11) Nicaragua (2001) 34 28 111 81 58 39 (7­11) Source: Household Surveys (El Salvador: EHPM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001). Central American Education Strategy 71 Table A.3. Standardized Test Scores Grade 3 Grade 6 (or 4) Grades 11 (and 9) Year Spanish Math Spanish Math Spanish Math El Salvador 1996 (a) na 4.0 2.2 2.9 1998 (a) 4.2 4.0 2.7 1.6 4.5 (2.4) 5.0 (1.1) 1999 5.2 5.2 2001 (b 498 488 890 865 2002 (b) 1705 (1301) 1672 (1264) 2003 1689 1678 Guatemala 1998 (c) 55.1 46.2 National 2000 61.5 48.7 2001 na na OREALC/ National 2002(d) 51.8 38.9 (58.3) (46.0) Honduras 1998 (e) 41 43 47 40 1999 42 43 na na 2000 41 43 46 39 2002 38 44 43 39 Nicaragua 2002 (f) 247 251 244 248 Costa Rica 1999 (g) 79.1 67.8 2001 78.0 71.1 74.2 (69.2) 70.3 (55.6) Notes: (a) Scores from a scale of 0 to 10; (b) Scores follow the following scale: grade 3 (300­450: basic; 451­600: intermediate; 601­700: superior); grade 6 (700­850: basic; 851­1000: intermediate; 1001­1100: superior); grade 9 (1100­1250: basic; 1251­1400: intermediate; 1401­1500: superior); "bachillerato"-- grade 11 (1500­1650: basic; 1651­1800: intermediate; 1801­1900: superior); (c) Scores from a scale of 0 to 100; tests conduced from 1998 to 2001 were not equated and hence inferences cannot be made as to whether student achievement has improved or declined over the years; (d) Scores from a scale of 0 to 100; (e) Scale to assess the level of educational achievement: 0%­30%: Very low; 31%­59%: Low; 60%­79%: Average; 80%­100%: Good; (f) Results are standardized using a mean of 250 and a standard deviation of 50. For Spanish grade 3, from 0­270: basic; from 271­320: intermediate; from 321­431: proficient; for Math grade 3, from 0­260: basic; from 261­310: basic; and from 311­413: proficient. For Spanish grade 6, from 0­270: basic; from 271­330: intermediate; from 331­435: proficient; for Math grade 6, from 0­310: basic; from 311­380: basic; and from 381­450: proficient; (g) Scores from a scale of 0 to 100. Grade 6 and 9: pass rate: 65%; Grade 11 ("bachillerato"): pass rate needs to be 70% including the "presentation" score at the end of the year. Source: El Salvador: The World Bank (2004b) and MINED official data; Guatemala: The World Bank (2004a); Honduras: The World Bank (2004c), UMCE/UPN/SE (2003a); Nicaragua: MECD (2003); Costa Rica: MEP/DPN (2001). 72 A World Bank Country Study Table A.4. Repetition Rate (different sources) RR RR First First Grade Source Grade Primary RR Primary Secondary RR Secondary El Salvador MINED, 2002 16% 6% (1­9); 7% (1­6) 3% 3% (7­11) HS, 2002 22% 11% (1­6) 4% 3% (7­11) Estimates (a) 2002 35% Guatemala MINED, 2000 na 15% na 4% (7­9); 1% (10­12) 2003 na 14% na 3% (10­12); 1% (10­12) HS, 2000 21% 12% 6% 5% (7­11) Honduras MINED, 2001 14% 9% na 16% (7­12) Estimates (a) 2001 30% Nicaragua MINED, 2002 15% 9% 9% 6% (7­11) HS, 2001 12% 8% 9% 6% (7­11) Estimates (b) 2002 30% Costa Rica MINED, 2002 (c) 11% (16%) 6% (8%) 11% (10%) 16% (13%) Latin America, WDI 2000 12% Countries with similar per capita income (d) Peru UIS 2000/01 11% 5% Colombia UIS 2000/01 5% 4% Jordan UIS 2000/01 Romania UIS 2000/01 3% 2% Algeria UIS 2000/01 13% Paraguay UIS 2000/01 8% 1% Ecuador UIS 2000/01 2% 4% Philippines UIS 2000/01 2% 2% Bolivia UIS 2000/01 4% 4% Notes: (a) Based on estimates produced on the basis of age for grade disagregation. In Honduras, household survey data were used; while in El Salvador, MINED official data were used; (b) Estimates based on Arcia (2003); (c) urban (rural); (d) Come from UIS (2003). Central American Education Strategy 73 Table A.5. Private Rates of Return Cohorts: 18­60 18­30 31­45 46­60 El Salvador, 2001 Average 9.4 6.8 10.8 11.4 Primary 6.2 2.2 6.5 9.4 Lower Secondary 4.8 4.5 6.0 7.5 Upper Secondary 11.9 8.9 14.3 18.0 Tertiary 18.9 17.9 18.9 13.0 Guatemala, 2000 Average 11.5 9.9 12.0 13.7 Primary 8.2 5.7 7.5 12.3 Lower Secondary 6.2 7.5 7.6 1.9 Upper Secondary 24.3 21.4 26.0 33.0 Tertiary 10.4 7.0 10.9 9.4 Honduras, 2002 Average 10.4 8.7 11.2 11.7 Primary 8.7 6.4 9.9 10.1 Lower Secondary 7.9 6.0 10.8 11.7 Upper Secondary 13.8 13.3 12.4 15.1 Tertiary 13.3 10.9 13.9 12.4 Nicaragua, 2001 Average 9.3 7.6 10.2 10.4 Primary 5.8 4.4 6.3 7.3 Lower Secondary 6.2 4.2 7.8 9.8 Upper Secondary 10.7 11.6 11.6 2.2 Tertiary 18.2 17.5 16.3 24.0 Costa Rica Average 10 Latin America Average 12 Asia Average 9.9 OECD Average 7.5 Source: El Salvador, EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001; SANIGEST (2003); and Psacharopoulos and Patrinos (2002). 74 A World Bank Country Study Table A.6. Completion Rates at the Primary and Secondary Level (a) Primary CR (%) Secondary CR (%) El Salvador (2002) 55 (grade 9) 40 (grade 11) 75 (grade 6) 22 (grade 12) Guatemala (2001) 57 20 (grade 9) (b) (2003) 64 25 (grade 9) Honduras (2001) 69 19 (grade 11) (2001) 14 (grade 12) (b) (2003) 80 31 (grade 11) Nicaragua (2002) 69 33 (grade 11) Costa Rica (2001) 89 na Latin America (2001) Brazil 90 61 (grade 8) Chile 101 61 (grade 8) Jamaica 95 61 (grade 8) Mexico 100 61 (grade 8) Paraguay 78 61 (grade 8) Peru 97 61 (grade 8) Ecuador 96 Bolivia 72 Colombia 85 Countries with similar income per capita (1995­2001) Peru 97 Colombia 85 Jordan 104 Romania 98 Algeria 91 Paraguay 78 Ecuador 96 Philippines 92 Bolivia 72 Notes: (a) PCR: Ratio between proxied graduate students in grade 6 and 12 year old school-age popula- tion; (b) 2003 likely to be over-estimated in primary. 2001 added for comparison with other countries. Source: MINED Official Data; Di Gropello, Dubey, Winkler (2004); WDI. Central American Education Strategy 75 Table A.7. Primary Completion Rate for Different Cohorts (a) 18 years cohort 25 years cohort 35 years cohort El Salvador, 2002 76 68 54 Guatemala, 2000 56 45 36 Honduras, 2002 71 65 57 Nicaragua, 2001 55 56 42 Costa Rica, 2000 85 80 82 Notes: (a) PCR = Proportion of 18, 25 and 35 year cohort having completed primary education. Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001; Costa Rica, EHPM 2000. Education Expenditure Table A.8. Public Education Expenditure in Proportion of GDP (percentages) (a) 1995 1996 1997 1998 1999 2000 2001 2002 2003 El Salvador 2.3 2.5 2.6 2.7 2.9 3.1 3.3 3.2 Guatemala 1.7 1.6 1.8 2.1 2.4 2.4 2.7 2.5 2.6 Guatemala (b) 2.0 2.3 2.2 2.5 2.3 2.4 Honduras 4.3 4 4.2 4 5.4 6.1 7.2 7.2 Nicaragua (c) 4.9 4.9 5.6 5.4 6.8 6.7 6.4 6.5 6.9 Nicaragua (d) 2.8 2.7 3 2.9 3.8 3.7 3.5 3.8 4 Costa Rica (e) 3.5 3.9 3.8 3.8 3.7 4.1 4.5 4.8 Notes: (a) Public expenditure on education includes all public funds channeled to the educational sys- tem through public institutions, therefore including both national and external public funds. (b) New classification (only education expenditure, and excludes social security and public debt). (c) with offi- cial GDP data: (d) with "unofficial" GDP data. There is some agreement that GDP figures may be under- estimated by as much as 70%, this second data series reflects this underestimation (e) Public expenditure includes public funds channeled only through the Ministry of Education (including both national and external public funds). Sources: El Salvador: WB WDI and LAC Database; Ministerio de Hacienda; CEPAL population estimates Guatemala: Banguat, Ministry of Finance, and Planning Unit, Ministry of Education; 2001­2003: CIEN. New classification: MFP and Bank of Guatemala. Honduras: 1990­1996: "Estudio Sectorial Plan Dece- nal"; 1997­1998: UNAT 2001; 1999­2002: SIAFI; IMF Financial Statistic different volumes Nicaragua: 1995­2000: MCHP and Central Bank, as reported in WB 2001 ("Nicaragua Public Expenditure Review"); 2001­2002: Directorate of Finance, MECD. Costa Rica: Central Bank, Ministry of Finance and ICEC (National Statistical Office). 76 A World Bank Country Study Table A.9. Economic Disaggregation of Expenditure (in 2002 US$ millions, unless otherwise specified) recurrent recurrent (salaries) (non-salaries) capital total El Salvador (a) 320 63 88 471 Guatemala (b) 467 (c) 33 500 Guatemala (d) 327 46 0 373 Honduras(e) 403 (c) 62 465 Honduras (f) 370 15 11 396 Nicaragua (g) 60 30 14 104 Nicaragua (h) 101 (c) 14 115 Costa Rica (in percent) (i) 91.6 8.4 (l) 100 (a) Total public expenditure. The relative amounts of salaries is somewhat underestimated by not accounting for the salary transfers to EDUCO schools, which are allocated under the "goods and ser- vices" category and, as such, cannot be disentangled from non-salary recurrent expenditures; (b)Total public expenditure; (c) Total recurrent expenditure; (d) Only MINED expenditures. Assumes 80% of cur- rent transfers go to salaries; (e) Total public expenditure; (f) Only MINED expenditures. 92% of transfers go to salaries and 2% to non-salary recurrent costs. Global allocations have been pro-rated. (g) Expendi- ture of the MINED on pre-primary, primary and secondary; (h) All public expenditures of the MINED, except tertiary education; (i) Only MINED expenditures. Higher education costs allocated assuming 92% go to salaries; (l) Includes investment costs. Sources: see Table A.8. Central American Education Strategy 77 Table A.10. Functional Disaggregation of Expenditure (in 2002 US $ millions, unless otherwise specified) Pre- primary Primary Basic Secondary Tertiary Admin. Other Total El Salvador (a) 31.6 247.6 27.9 (b) 27.1 27.8 22.0 (c) 384 El Salvador (d) 37.5 284.5 49.2 37.7 (e) 62.0 (f) 471 Guatemala (g) 36.8 250 40.4 61.5 36.5 74.5 (i) 500 Guatemala, 2001(g) 37.0 259.9 40.4 74.5 36.9 71.1 (i) 520 Honduras (h) 16.0 150.1 70.1 67.1 74.5 18.2 (i) 396 Honduras, 2001 (h) 15.1 156.2 69.2 72.0 55.1 12.3 (i) 380 Nicaragua (l) 4.1 72.0 13.6 43.0 (m) 7.1 3.9 (n) 144 Nicaragua (o) 6.0 83.0 13.9 43.0 (m) 7.1 5.0 (n) 158 Costa Rica (%) (p) 41.8 (q) 24.2 20.2 7.6 6.2 100 (a) All recurrent public expenditure channeled through public institutions is included; (b) Includes grades 10­12; (c) Includes the portion of recurrent costs financed through external funds, which could not be allocated to any of the levels; (d) Includes all public expenditure channeled through public institutions; (e) Includes amount directed to Education Development; (f) Includes investment expendi- ture that cannot be allocated across levels, as well as the Teacher Welfare and Strengthening of Access to Education programs; (g) Includes all recurrent public expenditure channeled through public institu- tions + unallocated capital expenditure; (h) Includes all recurrent public expenditure chan- neled through the Ministry of Education + unallocated capital expenditure. 2002 is budget; (i) Includes all investment and recurrent expenditures that cannot be allocated across levels; (l) Includes all recurrent public expenditure channeled through the Ministry of Education; (m) Includes all university expenditure (transfers and non transfers); (n) Includes adult education, special education and teacher training; (o) Includes all public expenditure channeled through the Ministry of Education; (p) Only MINED public expenditures; (q) Includes pre-primary. Sources: see Table A.8. 78 A World Bank Country Study Table A.11. Private Spending per Student by Income Quintile (as percentage of annual household income) Nicaragua Guatemala El Salvador Total Public Total Public Total Public Primary Q.I 0.8 0.8 7.6 6.9 6.5 6.1 Q.II 1.0 0.9 2.7 2.3 3.4 3.1 Q.III 1.2 1.1 2.0 1.8 3.1 2.7 Q.IV 1.7 1.2 1.9 1.5 2.8 2.2 Q.V 2.2 1.0 2.0 0.6 2.5 1.3 Total 1.4 0.9 1.9 1.0 2.5 1.8 Secondary Q.I 3.5 3.3 69.1 47.5 15.7 14.3 Q.II 2.6 2.5 17.8 14.9 7.4 6.6 Q.III 3.2 2.6 10.9 7.7 5.7 4.9 Q.IV 3.3 2.6 7.1 4.7 5.1 3.9 Q.V 3.2 1.8 4.0 1.8 3.8 2.3 Total 4.1 2.8 8.6 4.7 5.4 3.9 Source: El Salvador, EPHM 2002; Guatemala, ENCV 2000; Nicaragua, EMNV 2001 Central American Education Strategy 79 Teachers Table A.12. Hourly Salaries Across Teachers and Professions with Similar Educational Attainment Monthly Income from Main Occupation (US$) Net Honduras, El Salvador, Nicaragaua, Guatemala, 2002 2002 2001 2000 Category Average Average Average Average TEACHERS Primary teachers 2.2 0.6 Secondary teachers 2.3 1.0 Public sector primary teachers 0.5 Public sector secondary teachers 0.5 Teachers (a) 3.4 2.1 Teachers (b) 6.2 Public sector teachers (a) 3.8 2.2 Public sector teachers (b) 3.7 Public Employees (c) With primary 0.6 1.0 With upper secondary 1.7 3.4 0.8 1.9 With tertiary university 3.0 4.1 1.8 2.9 With tertiary non-university 3.7 0.6 Private Employees With primary 0.7 0.5 With upper secondary 1.7 0.9 1.3 With tertiary university 9.6 2.9 2.4 With tertiary non-university 2.3 0.9 Notes: (a) With tertiary non-university education or lower. For Guatemala: all teachers; (b) With tertiary university. For Guatemala: all teachers; (c) All employees for Honduras. Source: El Salvador: EPHM 2002; Guatemala, ENCV 2000; Honduras, EHPM 2002; Nicaragua, EMNV 2001. 80 A World Bank Country Study Community-based School Management Table A.13. Teacher Attendance and Work Hours Nicaragua El Salvador Honduras Guatemala Variable Non-Auto Autonomous Public EDUCO Public PROHECO Public PRONADE Work Hours 27.3 29.0* 33.7 36.8* 39.7 36.4 -- -- Teacher Absences (a) -- -- 1.4 1.2* 13.9 17.4 0.19 0.15* Teacher Absences (b) 0.34 0.35 -- -- 1.94 1.54* -- -- Days Worked in School -- -- -- -- -- -- 111.0 113.0* School Closings -- -- -- -- 29.5 20.5* -- -- Notes: (a) According to teachers or parents; (b) according to students. Data are not comparable across countries due to measurement differences. In all countries the data are taken from third/fourth grade questionnaires. Asterisk means that comparison is significantly different at p<0.01, p<0.05 or p<0.10 level. Source: Nicaragua data, 2002; Sawada and Ragatz, 2004; UMCE data, 2003; PRONADE data, 2001; Marshall, 2004a, 2003b. Table A.14. Third/Fourth Grade Test Score Averages By School Type Guatemala Honduras El Salvador Nicaragua Model: World Bank 2002 HCRG 2002 Control PRONADE Control PRONADE Control PROHECO Control EDUCO Control AUTO Spanish 0.04 -0.37* 0.02 -0.16* -0.04 0.03 1.75 1.73 -0.03 0.02 Mathematics 0.05 -0.39* 0.01 -0.07 -0.04 0.03 3.74 3.59 -0.01 0.01 Science -- -- -- -- -0.12 0.08* -- -- -- -- Average SES 3.3 2.3* 0.31 -0.46* 0.49 -0.42* 0.66 0.28 4.1 4.5* Central Average Parental Education 5.4 3.0* 2.5 1.7* 3.2 2.5* 0.53 0.50 5.5 6.2* Notes: For El Salvador average SES refers to percentage of homes with electricity, while Average Parental Education is the percentage of mothers with basic American education. Asterisk means that comparison is significantly different at p<0.01, 0.05 or 0.10 level. Source: World Bank data 2002; Marshall, 2004b; Nicaragua data 2002; UMCE data 2003; Sawada and Ragatz, 2004. Education Strategy 81 82 A World Bank Country Study Table A.15. Teacher Education and Experience Nicaragua El Salvador Honduras Guatemala Variable Non-Auto Autonomous Public EDUCO Public PROHECO Public PRONADE Percent Teachers with University 44.5 38.2* 30.0 74.0* 6.4 2.9* 2.5 0.0* Percent Teachers without Certificate -- -- -- -- 3.9 54.8* -- -- Teacher Experience 5.8 7.2* 9.5 3.9* 7.6 2.2* 9.7 4.6* Notes: Asterisk means that comparison is significantly different at p<0.01; p<0.05, or p<0.10 level. Source: Nicaragua data 2002; Sawada and Ragatz (2004); UMCE data 2003; Marshall, 2004b. Factors Associated to Non-attendance in Secondary Table A.16. Reasons for Non-attendance in the 13­15 and 16­18 Age Ranges El Salvador Guatemala Nicaragua 13­15 16­18 13­15 16­18 13­15 16­18 School related reasons % % % % % % No place available 1.5 0.7 2.4 1.4 No grade 0.1 0.1 6.7 2.4 No school 1.6 1.1 School too far 2.4 (a) 1.4 (a) 1.5 0.9 1.9 0.4 No teachers 1.1 0.3 Subtotal 2.4 1.4 4.7 2.8 12.1 4.5 School/family related reasons Lack of interest 34.7 27.4 22 19 23.5 21.5 Lack of money, too expensive 25.4 19.7 29.7 23.5 2.1 2.4 Total 62.5 48.5 56.4 45.3 37.7 28.4 Family related reasons Household work 5 9.6 11.4 17 20.5 31.7 Need to work 12.5 25.7 22.8 29.2 1.1 0.1 Subtotal 17.5 35.3 34.2 46.2 21.6 31.8 Family problems 5 8.5 24.7 18.2 Parents do not want 3.9 0.9 Total 26.4 44.7 34.2 46.2 46.3 50 Student related reasons Age 1 1.2 0.6 0.6 Completion of studies 0.9 0.6 6.2 7.8 Pregnant 0.5 1 2.4 7.4 Illness 2.2 1.5 Total 4.6 4.3 9.2 15.8 Other 11.1 6.8 4.8 4.2 6.8 5.8 (a) Includes no place available. Source: El Salvador, EPHM 2002; Guatemala, ENCV 2000; Nicaragua, EMNV 2001. APPENDIX B Secondary Targets Simulation Results 83 84 A World Bank Country Study Table A.17. El Salvador-Secondary Targets Simulation Results 2002 2005 2008 2011 2015 Scenario 1: With same delivery structure in public sector, declining share of private sector in urban area and increased scholarship coverage Share of urban private/public sector in 2015: (lower secondary) 88% public; 12% private; (upper secondary) 88% public; 12% private Share of public sector enrollment covered by scholarship in 2015: 20% lower secondary; 20% secondary (base year: 0% lower secondary, 4% upper secondary) Lower Secondary Costs/Education Budget 12.6 14.3 16.5 19.0 26.0 Upper Secondary Costs/Education Budget 6.0 8.0 10.1 14.6 22.6 Secondary Costs/GDP 0.60 0.73 0.89 1.20 1.61 Scenario 2: Same as scenario as above with public subsidy to private urban schools to keep current enrollment share (80% public, 20% private in urban lower secondary; 66% public, 34% private in urban upper secondary) Lower Secondary Costs/Education Budget 12.6 14.2 16.2 18.6 25.0 Upper Secondary Costs/Education Budget 6.0 7.5 9.9 13.1 19.3 Secondary Costs/GDP 0.60 0.71 0.86 1.17 1.47 Scenario 3: With different delivery structure in public sector, declining share of private sector in urban areas and increased scholarship coverage Lower Secondary: target public delivery structure: EDUCO: 85% (vs. 37% in 2002); distance educa- tion: 15% (vs. 15% in 2002). Upper Secondary: target public delivery structure: Traditional schools: 70% (vs. 54% in 2002); distance education: 30% (vs. 46% in 2002). Share of public sector enrollment covered by scholarship in 2015: 20% upper basic; 25% secondary Share of urban private/public sector in 2015: 88% public; 12% private Lower Secondary Costs/Education Budget 12.6 14.0 16.0 18.2 24.9 Upper Secondary Costs/Education Budget 6.0 7.9 10.8 14.9 23.9 Secondary Costs/GDP 0.60 0.72 0.90 1.22 1.61 Scenario 4: Same as scenario above but with public subsidy to private urban schools to keep enrollment share (80% public, 20% private in urban lower secondary; 66% public, 34% private in upper secondary) Lower Secondary Costs/Education Budget 12.6 13.8 15.6 17.6 23.6 Upper Secondary Costs/Education Budget 6.0 7.6 10.0 13.4 20.4 Secondary Costs/GDP 0.60 0.70 0.84 1.14 1.45 Source: The World Bank (2004b). Central American Education Strategy 85 Table A.18. Honduras-Secondary Targets Simulation Results (US$ millions) Aver. Aver. Aver. Aver. 2002 2003­2005 2006­2008 2009­2011 2012­2015 2015 Scenario 1: without efficiency gains and same delivery structure in lower secondary · Net Coverage Rate target in 2015: 70% (base: 41%) in grades 7­9 and 50% (base: 22%) in grades 10­12 · School-Age Population growth rate: 2.64% over the time-period · GDP real growth rate: 2003: 3.5%; 2004­2015: 4% · Inflation growth rate: 2003: 7%; 2004: 6%; 2005­2015: 5% · Share of SE education budget/GDP: 6.1% constant over the time-period GDP 6473 7929 10355 13484 18390 20885 SE Education Budget 396 485 633 825 1125 1278 Total Costs Lower 42 57 87 128 187 240 Total Costs Upper 26 37 60 95 150 201 Total Costs Secondary 68 94 147 223 337 441 Costs/Education Budget 17.2% 19.3% 23.2% 27.0% 29.9% 34.51% Costs/GDP 1.05% 1.18% 1.42% 1.65% 1.83% 2.11% Scenario 2: Same as scenario 1 but with efficiency gains in both levels · Repetition rate decreases from 16% in 2002 to 8% in 2015 (lower) · Repetition rate decreases from 18% in 2002 to 9% in 2015 (upper) Total Costs Lower 42 57 84 121 176 223 Total Costs Upper 26 36 58 90 139 185 Total Costs Secondary 68 93 142 211 315 408 Costs/Education Budget 17.2% 19.2% 22.4% 25.5% 28.0% 31.59% Costs/GDP 1.05% 1.17% 1.37% 1.56% 1.17% 1.95% Scenario 3: Same as scenario 2 but with changing delivery structure in lower secondary · Proportion of different public providers in 2015: plan basico: 26%; CEB: 39%; Educatodos: 30%; Telebasica: 5% (from a base of: plan basico: 56%; CEB: 32%; Educatodos/SAT: 10%; Telebasica: 1%) Total Costs Lower 42 51 71 100 144 187 Secondary Total Costs Secondary 68 87 129 190 283 372 Costs/Education Budget 17.2% 17.9% 20.4% 23.0% 25.1% 29.1% Costs/GDP 1.05% 1.10% 1.24% 1.40% 1.53% 1.78% Source: The World Bank (2004c). APPENDIX C Boxes 87 88 A World Bank Country Study Box A.1. Curricular and Pedagogical Reforms in Central American Countries El Salvador: The education reform, which started in 1994/1995, promoted a curriculum change including both primary and secondary. A constructivist curriculum was introduced and pedagogical practices were made consistent with the new approach. Four main subjects are taught at the basic level: language and literature, math, natural sciences, and social studies. New textbooks were produced for all grades. The school year for preprimary, basic and secondary education consists of 200 days distributed over 11 months. The education reform program promoted 5 hours a day of classes in basic education, 40 hours in the academic track of secondary education, 44 hours in the first two years and 30 hours in the third year of the technical/vocational track. A recent assessment undertaken by Schiefelbein (2004) and Kraft (2004) shows that too little emphasis is put on reading and writing skills in the curriculum of the first three grades of primary education. Additionally, teaching practices remain traditional and contribute to explain the still unsatisfactory learning outcomes in students of disadvantaged socio-economic background. Among the causes mentioned for this stagnation in pedagogical practices are the absence of teach- ing guides to operationalize the new curriculum and the insufficient pedagogical preparation of teachers. Finally, the textbooks that accompanied the new curriculum were never evaluated from the learning perspective. The quality of the grades 7 to 9 is particularly low also due to frontal ped- agogical techniques and the lack of an integrated and pertinent curriculum. There is also some evidence that effective class time is substantially lower than the official one.98 Guatemala: The education reform started in 1996 also promoted the introduction of a new curriculum, with key emphasis on multicultural education and bilinguism Spanish-Maya. The curriculum intro- duces active and participatory pedagogical practices in pre-primary and primary, with an inter- disciplinary focus. New textbooks were produced. To deepen the new approach, a new curricular reform, also extended to lower secondary, is under way since 2003. Primary school teachers are supposed to teach five hours per day, or 25 hours per week for 180 days. Full-time secondary school teachers should teach 22.5 hours per week. Political instability and teacher opposition have complicated the effective application of this new curriculum. However, they have been successes in the new bilingual approach (see below). There is evidence that most teachers do not teach more than four hours per day, maybe even less, and an average of 133 days per year.99 Honduras: At the beginning of the 1990s, a curriculum based on "Rendimiento Basicos" for pre-primary, pri- mary and lower secondary was approved. New textbooks were also produced for the same levels. The official school calendar establishes a minimum of 200 schooling days per year, covering 10 months, and 30 class hours per week (primary). However, little pedagogical change was introduced in the curriculum and applied in the classroom. Recent evidence,100 although only gathered on a sample of about 120 schools, shows that traditional dictation techniques are still widely used. Therefore, in 2002, a new curriculum, with a constructivist focus, was designed for pre-primary and basic education (Curriculm Nacional Basico). New textbooks for pre-primary and basic education are currently under validation and production; and teacher training to help with the application of the new curriculum under way. There is evidence of much lower effective class time than the official one.101 98. See Shiefelbein (2004), which reports figures of about 500 class hours a year (however, based on old data). 99. See Rubio, F. (2000), "Presentation on the Status of Bilingual Education," MEDIR Project, USAID. 100. See Di Gropello and Marshall (2004). 101. See Honduras-Secretary of Education (2002), "Todos con educacion: Honduras 2003­2015," which reports up to 333 yearly hours (versus the official 1200 ones) in a worse case scenario. Central American Education Strategy 89 Box A.1. Curricular and Pedagogical Reforms in Central American Countries (Continued) Nicaragua: Curricular reform was key to the education reform and gradually implemented from 1993 to 1998 for primary education, with a humanistic-constructivist focus. New textbooks were produced and the technical staff of the Ministry fully replaced. The official school calendar establishes 180 schooling days per year and 5 class hours per day (primary). The curricular reform was largely aimed at eliminating from the education system the focus, con- tents and structure that had been established under the Sandinista, with the result that it paid too little attention to technical aspects and to teacher involvement in the process of change. The reform was overall an improvement over the past situation, but lacked consistency in the articu- lation of its different internal components and in its focus, with consequences on the teaching- learning process. Textbooks were largely financed by external donors, with the result that they still do not fully reflect the national, rural and multicultural environment. Finally, teachers still do not seem to have fully internalized the constructivist approach. Recent evidence suggests that more than 90 percent of teachers keep on applying traditional techniques where students are seen as passive actors.102 Teacher training has not been fully planned to address this challenge, although this is gradually changing. Source: Own elaboration, based on Arrien (1998). 102. See MECD (2004), "Politicas de educacion basica y media," Version preliminar. 90 A World Bank Country Study Box A.2. Flexible Delivery of Secondary Education in Central America Distance Education in El Salvador. Distance Education is part of MINED's adult education pro- grams. Adult education is offered to students over 15 years old that cannot attend daily school. It includes four programs that cover 150,071 persons in 2002. Here we only focus on the Distance Education program which covers 23,291 upper basic and secondary students. The objective of the Distance Education program, which is also supported by the World Bank's Secondary Education Project, is to expand the coverage of upper basic and secondary education targeting the working population mostly in rural areas. The program includes curricular improvements and the provi- sion of material and training to the teachers. It is executed in basic and secondary centers and offered for the third cycle and the academic track of secondary education. The third cycle consists of 5 teaching modules while the secondary education module consists of six modules. Students pay about US$4 to buy each module book. Classes are held in the morning of Saturday of Sunday. Teachers usually are those that teach in the daily school and have been trained in distance edu- cation. They receive an additional salary of US$70 per month. The directors of schools that offer distance education also receive an additional salary of US$80 per month to compensate for their administrative duties related to this modality. The distance education program is growing rapidly. From the 19,500 students attended in 2001, it grew to 23,300 students in 2002. Two-thirds of those enrolled are below 25 years old. MINED plans to continue to expand the program, as resources permit. The program still needs to be fully evaluated. Distance Education in Honduras. The Educatodos program started in 1995 with the support of USAID, as an alternative scheme for supplying basic education to young people and adults outside the traditional study system. By 2000, Educatodos had 88,000 students enrolled, mainly in the first six grades of primary education. However, recently, Educatodos has been expanding into the sec- ondary range. In 2001 it had 7,510 participants from 7 to 8 grades, and was in the process of estab- lishing new groups for 7 to 9. Educatodos graduates from the third cycle receive the SEP certificate for the Ciclo Comun. The Educatodos scheme is delivered by trained facilitators (some paid, some volunteers) who work with study groups outside normal work hours, using approved texts and taped programs. The program has developed an integrated curriculum for 7 to 9 grades which conforms to the SEP guidelines and has been officially approved. The cost per student of this scheme is esti- mated by USAID at US$68 a year, plus the capital and operating expenses of a motorcycle for the pro- moter. The scheme is relatively cheap because it does not hire professional teachers, and many of the facilitators are not paid. This reduced teacher assignment may, however, impact negatively on the quality of the educational outcomes. The Tutorial Learning System--TLS (Sistema de Aprendizaje Tutorial--SAT) is an alternative model for high school education, oriented toward the rural sector. It has been implemented in commu- nities in the Honduran Mosquitia region by the Bayan Indigenous Socio-Economic Development Association (Asociación de Desarrollo Socio Económico Indígena--Bayan) since 1996. Once again, the SEP recognizes the degrees granted to the graduates of these programs a valid high-school qualification. TLS is still a small scale operation. In 2003, it was operating in indigenous commu- nities in the Atlantic Coast departments of Atlántida, Gracias a Dios and Colón, with 70 student groups with an average of 22 students per group. However, during 2004­2006 the TLS program plans to expand in other depressed indigenous regions such as the Lenca area (Lempira, Sta Bar- bara and Copan) reaching a student population of 5,000 in 220 groups. Although the initial evidence on the quality of outcomes of these two programs in Honduras is heartening, and provides a shameful comparator for the much more expensive main programs of the Ministries of Education, a full evaluation of these two programs is still needed. Source: World Bank (2004b and 2004c). References AED (Academy for Educational Development). 2004. "Calidad y Equidad: Renovando la Visión de Largo Plazo en la Eduación Preescolar y Básica de El Salvador." USAID, unpublished document. Arcia, G. 2003. "The Financing of Public Education in Nicaragua." Consulting report for the World Bank. Arias, O. 2004. "Poverty in El Salvador during the 1990s: Evolution and Characteristics." Background paper for The World Bank Poverty Assessment of El Salvador. Arríen, J., ed. 1998. La Educación y la Reforma de la Educación en Cinco Países Centroameri- canos. Managua: PREAL. Bommier, A., and S. Lambert. 2000. "Education Demand and Age at School Enrollment in Tanzania." Journal of Human Resources 35(1):177­203. Bourguignon, F., F. Ferreira, and P. Leite. 2002. "Ex-Ante Evaluation of Conditional Cash Transfer Programs: The Case of Bolsa Escola." Working Paper No. 516, William David- son Institute, Michigan. Costa Rica Ministerio de Educación Pública/Departamento de Pruebas Nacionales. 2001. "Informes Nacionales sobre Resultados de las Pruebas Nacionales de la Educación For- mal, ciclo II y III." San José, Costa Rica. Cummings, S., and S. Tamayo. 1994. "Language and Education in Latin America: An Overview." Human Capital Working Paper No. 13068, The World Bank, Washing- ton, D.C. Di Gropello, E. 2005 (forthcoming). "Review of Education Sector Expenditures in Guatemala." Background paper for the Guatemala Public Expenditure Review, The World Bank. Di Gropello, E., and J. Marshall. 2004. "Teachers' Incentives in Rural Honduras." The World Bank. Processed. 91 92 A World Bank Country Study Di Gropello, E., A. Dubey, and D. Winkler. 2003. "Reaching Basic Education for All in Latin America." The World Bank. Processed. Dutcher, N. 1997. "The Use of First and Second Languages in Education: A Review of Inter- national Experience." Working Paper No. 18803, The World Bank, Washington, D.C. ECLAC. 1994. "El Gasto Social en America Latina: un Examen Cuantitativo y Cualitativo." Cuadernos de la CEPAL. Santiago: CEPAL. ------. 1996. "El Gasto Social en America Latina: un Examen Cuantitativo y Cualitativo." Cuadernos de la CEPAL. Santiago: CEPAL. Edwards, J. 2004. "Education and Poverty in Guatemala." In "Guatemala--Equity and Stu- dent Achievement in Primary Education." Volume II, Report No. 22691, The World Bank, Washington, D.C. Edwards, J., B. Fuller, and S. Parandekar. 1997. "Primary Education Efficiency in Hon- duras. What remains to be done?" Latin America Human Development Working Paper Series No.7, The World Bank, Washington, D.C. Figueredo, V., and S. Anzalone. 2003. "Alternative models for secondary education in devel- oping countries: rationale and realities." American Institute for Research. Processed. Fuller, B., and D. Holsinger. 1993. "Secondary Education in Developing Countries." ESP Discussions Papers Series No.7, The World Bank, Washington, D.C. Fuller, B., and M. Riverola. 1998. "Nicaragua's Experiment to Decentralize Schools: Views of Parents, Teacher and Directors." Working Paper Series on Impact Evaluation of Education Reforms, Paper No. 5, The World Bank, Washington, D.C. Gershberg, A. I. 2003. "Empowering Parents while Making them Pay: Autonomous Schools and Education Reform Processes in Nicaragua." Paper written for the project funded by the Tinker Foundation. Hammer, J. 1996. "The Public Economics of Education." Policy Research Department, The World Bank, Washington, D.C. Processed. Holsinger, D., and R. Cowell. 2000. "Positioning Secondary School Education in Devel- oping Countries." In Series Policies and Strategies for Secondary education, Inter- national Institute for Educational Planning, UNESCO, Paris. Honduras--UMCE/Universidad Pedagógica Nacional Francisco Morazan/SE. 2003a. Informe Nacional de Rendimiento Academico 2002. Tegucigalpa. ------. 2003b. Informe sobre factores asociados al rendimiento académico Honduras 2002. Tegucigalpa. IFPRI. 2003. "Proyecto PRAF/BID Fase II: Impacto Intermedio." Sexto Informe. Jimenez, E., and Y. Sawada. 1999. "Do Community-Managed Schools Work? An Evaluation of El Salvador's EDUCO Program." The World Bank Economic Review 13(3):415­441. Nicaragua Reform Evaluation Team. 1996. "Nicaragua's School Autonomy Reform: A First Look." Working Paper Series on Impact Evaluation of Education Reforms No. 1, The World Bank, Washington, D.C. King, E., B. Özler, and L. B. Rawlings. 2001. "Nicaragua's School Autonomy Reform: Fact or Fiction?" Working Paper Series on Impact Evaluation of Education Reforms No. 19, The World Bank, Washington, D.C. Kraft, R. 1998. Rural Education Reform in the Nueva Escuela Unitaria of Guatemala. Wash- ington, D.C.: AED. Laguna, J. R. 2003. "Años de Estudio y Superación de la Pobreza en Nicaragua. El Caso de la Red de Protección Social." Universidad de Chile. Processed. Central American Education Strategy 93 Lavy, V. 1996. "School Supply Constraints and Children's Education Outcomes in Rural Ghana." Journal of Development Economics 51:291­314. Lewin, K., and F. Caillods. 2001. "Financing the Development of Secondary Education in Developing Countries." Prospects 31(1). Loayza, Fajnzylber, and Calderón. 2002. "Economic Growth in Latin America and the Caribbean Stylized Facts, Explanation and Forecasts." International Monetary Fund's Seminar Series No.2003, Washington, D.C. Lloyd, C., B. Mensch, and W. Clark. 2000. "The Effects of Primary School Quality on School Dropout among Kenyan Girls and Boys." Comparative Education Review 44(2):113­147. Lloyd, C., S. El Tawila, W. Clark, and B. Mensch. 2003. "The Impact of Educational Qual- ity on School Exit in Egypt." Comparative Education Review 47(4):444­467. Marshall, J. 2003a. "If You Build It Will They Come? Primary School Quality and Grade Attainment in Rural Guatemala." Stanford University. Processed. ------. 2003b. "School Attendance and Academic Achievement: A Simultaneous Equa- tions Approach." Stanford University. Processed. Nicaragua--Ministerio de Educación, Cultura y Deportes. 2003. "Informe de Resultados 2002." Managua. OECD. 2002. Education at a Glance--OECD Indicators. Paris. OEI. 2003. "Organización y Estructura de la Formación Docente en Iberoamérica." Doc- umento de trabajo. ------. 2004. "Diagnostico de los Sistemas de Evaluación Educativa de Centroamérica." Consulting report for OEI, San Salvador. Patrinos, H., and E. Velez. 1996. "Costs and Benefits of Bilingual Education in Guatemala: A Partial Analysis." Human Capital Development Working Paper No. 74, The World Bank, Washington, D.C. Paxson, C., and N. Schady. 2002. "The Allocation and Impact of Social Funds: Spending on School Infrastructure in Peru." World Bank Economic Review 16(2):297­319. Parker, C. 2004. "Teacher Incentives and Student Achievement in Nicaraguan Autonomous Schools." The World Bank, Washington, D.C. Processed. Porta, E., and G. Arcia. 2003. "Necesidades Presupuestarias y Brecha Financiera para la Implementación de la Política Educativa de Mediano y Largo Plazo (2004­2015)." MECD, Managua. Porta, E., G. Arcia, and J. E. Laguna. 2004a. "Otro vistazo a la Autonomía Escolar de Nicaragua: Aceptación y percepción en 2004." MECD, Managua. ------. 2004b. "Análisis de los factores asociados con el rendimiento académico en 3° y 6° grado de primaria." MECD, Managua. Porta, E. and J. E. Laguna. 2004. "Análisis de la Rentabilidad de la Educación en Nicaragua." MECD, Managua. Processed. PREAL. 2002. "El Tiempo de Aprendizaje en America Latina." Serie Políticas, Santiago, Chile. ------. 2003. "Time to Act: A Report Card on Education in Central American and the Dominican Republic." Task Force on Education Reform in Central America. Psacharopoulos, G., and H. Patrinos. 2002. "Returns to Investment in Education: A Fur- ther Update." World Bank Policy Research Working Paper No.2881, The World Bank, Washington, D.C. SANIGEST. 2003. "Fortalecimiento del Sector Educativo en Costa Rica: Situación Actual y Opciones de Politica." Consulting report for the World Bank. 94 A World Bank Country Study Sawada, Y., and A. Ragatz. 2004. "Decentralization of Education, Teacher Behaviour and Outcomes: The Case of El Salvador's EDUCO Program." The World Bank, Washing- ton, D.C. Processed. Schiefelboin, E. 2004. "Se Reduce la Repetición, Deserción y Extraedad, pero Continua el Bajo Aprendizaje, Especialmente en los Grupos de Menores Ingresos." In Academy for Educational Development, "Calidad y Equidad: Renovando la Visión de Largo Plazo en la Eduación Preescolar y Básica de El Salvador." Skoufias, E., and S. Parker. 2001. "Conditional Cash Transfers and Their Impact on Child Work and Schooling: Evidence from the PROGRESA Program in Mexico." Discussion Paper No. 123, International Food Policy Research Institute, Washington, D.C. UIS (UNESCO Institute for Statistics). 2003. Global Education Digest: Comparing Educa- tion Statistics across the World. Montreal. UNESCO. 2003. "Primer Foro Centroamericano de Educación para todos 2003." Informe Final Relatoría. Vawda, A., and H. Patrinos. 1999. "Producing Educational Materials in Local Languages: Costs from Guatemala and Senegal." International Journal of Educational Development 19:287­299. World Bank. 2003. "Closing the gap in education and technology." World Bank Latin American and Caribbean Studies, Washington, D.C. ------. 2004a. "Guatemala--Equity and Student Achievement in Primary Education." Volumes I and II, Report No. 22691, Washington, D.C. ------. 2004b. "El Salvador--Education Strategy Note." Unpublished document. ------. 2004c. "Honduras--Education Strategy Note." Unpublished document. ------. 2004d. "Honduras. Development Policy Report. Accelerating Broad Based Growth." Report No. 28222, Washington DC. ------. 2004e. "Central America Education Strategy Paper." Report No. 29946, Wash- ington, D.C. Available at http://www.worldbank.org/laceducation ------. 2004f. "Strengthening the Foundation of Education and Training in Kenya." Sec- tor Report No. 28064-KE, Washington, D.C. Younger, S. 2000. "The Incidence of Public Services and Subsidies in Peru: Evidence from Household Surveys." Working Paper No. 103, Cornell Food and Nutrition Policy Pro- gram, Ithaca, NY. ------. 2003. "Benefits on the Margin: Observations on the Marginal Benefit Incidence." World Bank Economic Review 17(1):89­106. Central America Education Strategy is part of the World Bank Country Study series. These reports are published with the approval of the subject government to communicate the results of the Bank's work on the economic and related conditions of member countries to governments and to the development community. The main objective of this regional report is to provide an in-depth diagnosis of where Central American countries stand along several education dimensions, underscoring the most urgent and serious challenges, and suggesting policy options to address them. This is the first attempt at providing a comprehensive quantification of educational outcomes in four comparable Central American coun- tries: El Salvador, Honduras, Guatemala, and Nicaragua. The report identifies three urgent priorities for improving Central American education systems: improving learning, reaching universal primary completion, and expanding secondary coverage. In response to identified barriers to learning, the report details some key areas of policy intervention, which include the improvement of national assessment systems, the enhancement of teacher perform- ance, and the diversification of teaching methodologies. The main policy recommendations for improving completion are to strength- en and improve the existing supply-side interventions (multigrade schooling, school-based management, bilingual education, and so forth), while also applying effective demand-side policies and broader interventions. Finally, when analyzing specific policy options to increase secondary coverage, the report concludes that only a comprehensive policy package (including higher budget shares for secondary education and interventions such as flexible delivery mechanisms, demand-side subsidies, and basic education centers) will successfully overcome the range of constraints facing secondary enrollment. World Bank Country Studies are available individually or by sub- scription, both in print and online. ISBN 0-8213-6258-5 TMxHSKIMBy362587zv,:':#:;:) THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org