CY) ~~~~~ -I ~ ~ ~ ~ ~ ~ ~ ~ ~ I E ~ ~ K ~ A ~ ~ ~ m ~ ~ w ). . 1 Achieving Universal Primary Education by 2015 A Chance for Every Child Barbara Bruns, Alain Mingat, and Ramahatra Rakotomalala THE WORLD BANK Washington, D.C. © 2003 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved 1 2 3 4 06 05 04 03 The findings, interpretations, and conclusions expressed herein are those of the authors and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data induded in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The World Bank encourages dissemination of its work and will normally grant permission prompdy. 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, telephone 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, World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. ISBN 0-8213-5345-4 Library of Congress Cataloging-in-Publication Data has been applied for Cover photo: Students writing in a war-torn classroom, Baukau, East Timor. Photo by Alex Baluyut/World Bank Photo Library. CONTENTS: Acknowledgments vii Acronyms ix Executive Summary 1 Chapter 1. The Global Challenge of Education for All 23 Why Is Universal Primary Education So Important? 26 Why Universal Primary Education Must Mean Universal Primary Completion 29 Chapter 2. The Global Scorecard: Progress since Jomtien 37 What Is the Primary Completion Rate? 37 Data Sources and Methodological Issues 38 Advances in Primary Completion during the 1990s 42 The Global Prospects for Universal Primary Completion by 2015 56 Chapter 3. What Will It Take to Achieve Universal Primary Completion by 2015? 61 Determinants of EFA Progress 61 Implications for Accelerating EFA Progress 66 Chapter 4. Costing the MDG of Universal Primary Completion 71 Service Delivery 74 System Expansion 78 System Financing 79 Country-Level Simulation Results 82 Aggregate Results 101 Requirements by Region 107 Estimating the Global Costs of Reaching the Education MDG 109 Chapter 5. Implications for Countries and Donors 113 Importance of a Flexible Approach 113 Considerations for Developing Countries 117 Considerations for the Donor Community 118 The EFA Fast-Track Initiative 122 Conclusion 125 Bibliography 127 Technical Annexes 131 BOXES 1 Global "Education for All" Goals 2 2 EFA Indicative Framework 18 3 First EFA Fast-Track Group, 2002 19 1.1 Millennium Development Goals 24 1.2 Global "Education for All" Goals 25 2.1 Primary Completion Progress in Uganda 45 2.2 Primary Completion Progress in Brazil 52 3.1 Accounting Framework for Spending on Primary Education 67 4.1 The Incremental Costs of HIV/AIDS for Universal Primary Completion 77 5.1 Key Education Policy Options 116 5.2 First EFA Fast-Track Group, 2002 123 :TABLES 1 Prospects for Universal Primary Completion by 2015 5 2 Benchmarks for Primary Education Efficiency and Quality 7 3 Estimated Annual Financing Gap by Region 11 1.1 Proportion of Children Completing Primary School: Regional Averages and Selected Countries 34 1.2 Comparison of Gross Enrollment Ratio, Net Enrollment Ratio, and Primary Completion Rate for Selected Countries, 2000 34 2.1 Length of the Primary Cycle in 155 Developing Countries, circa 2000 39 2.2 Primary Completion Progress by Region, 1990-2000, Country-Weighted 43 2.3 Primary Completion Progress by Region, 1990-2000, Population-Weighted 44 2.4 Africa: Changes in Primary Completion Rates during the 1990s 46 2.5 East Asia and the Pacific: Changes in Primary Completion Rates during the 1990s 49 2.6 Europe and Central Asia Changes in Primary Completion Rates during the 1990s 50 2.7 Latin America and the Caribbean: Changes in Primary Completion Rates during the 1990s 51 iv CONTENTS 2.8 Middle East and North Africa: Changes in Primary Completion Rates during the 1990s 55 2.9 South Asia: Changes in Primary Completion Rates during the 1990s 56 2.10 Prospects for Universal Primary Completion by 2015 59 3.1 Key Education System Parameters for Adjusted Sarmple of 49 Countries, Grouped by Relative EFA Success 63 3.2 Regression Analysis of Key Parameters 65 4.1 Benchmarks for Primary Education Efficiency and Quality 73 4.2 Alternative Scenarios for Domestic Resource Mobilization 81 4.3 India: MDG-2015 Financing Gap under Alternative Policy Measures 86 4.4 India: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 88 4.5 Pakistan: MDG-2015 Financing Gap under Alternative Policy Measures 90 4.6 Pakistan: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 92 4.7 Armenia MDG-2015 Financing Gap under Alternative Policy Measures 94 4.8 Armenia MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 96 4.9 Niger: MDG-2015 Financing Gap under Alternative Policy Measures 98 4.10 Niger MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 99 4.11 Possible Costs of Achieving UPC in Afghanistan, 2000-2015 101 4.12 All 47 Countries: MDG-2015 Financing Gap under Alternative Policy Measures 102 4.13 Al 47 Countries: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 104 4.14 Estimated Annual Financing Gap to Achieve the Education MDG, by Region (Scenario C2) 108 4.15 A Global Estimate of the Annual Incremental Costs to Achieve the Education MDG and Likely Financing Gap 111 5.1 Official Development Assistance to Basic Education in Sub-Saharan Africa, by Donor, 1998-2000 119 CONTENTS v F I G U R E S 1 Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted 3 2 Primary Completion Progress in Europe and Central Asia, East Asia and the Pacific, and Latin America and the Caribbean Regions, 1990-2015, Country-Weighted 4 3 Primary School Completion Rates and Gross Enrollment Ratios in a Sample of Low-Income Countries, circa 1999/2000 7 4 Domestic and External Financing Required to Achieve the Education MDG in 47 Countries, 2001-2030 10 1.1 Average Educational Attainment of Adult Population by Region, 2000 27 1.2 Proportion of Adults Who Can Read and Write Easily by Highest Grade Attained, Togo and Niger 30 1.3 Sample Schooling Profiles 31 1.4 Primary Gross Enrollment Ratios and Completion Rates, Selected Countries, 1999 31 1.5 Schooling Profiles Disaggregated by Income 33 1.6 Schooling Profiles Disaggregated by Gender and Income 33 2.1 Primary Completion Progress by Region, 1990-2000, and Projected Trends (Country-Weighted) 42 2.2 Global Progress in Primary Completion, 1990-2000 and Projected Trends (Country-Weighted) 57 2.3 Primary Completion Progress in Europe and Central Asia, East Asia and the Pacific, and Latin America and the Caribbean Regions, 1990-2015, Country-Weighted 58 2.4 Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted 58 3.1 Primary School Completion Rates and Gross Enrollment Ratios in a Sample of Low-Income Countries, circa 1999/2000 62 3.2 Class Size in Relation to Teacher Salary 66 3.3 Evolution of Average Teacher Salary in Primary Education, by Region and Subregion, 1975-2000 69 4.1 Domestic and External Financing Required to Achieve the Education MDG in 47 Countries, 2001-2030 105 4.2 Domestic and External Financing Required to Achieve the Education MDG in Sub-Saharan Africa, 2001-2030 107 vi CONTENTS ACKNOWLEDGMENTS. This study is the work of a team led by Alain Mingat and Barbara Bruns. Team members included Ramahatra Rakotomalala, who developed the simulation model used and carried out the simulations for the African countries; Ashutosh Dubey, who supervised the data collection and simulations for the countries outside Africa; Nicholas Wilson, Saida Mamedova, and Jose Carlos Orihuela, who devel- oped the new database of primary completion rates; Jee-Peng Tan, who con- tributed to data collection and analysis for the African countries; and Andrew Clark and Hongyu Yang, who assisted in data collection and analysis. Dina Abu- Ghaida carried out a review of relevant literature. Julie Wagshal tirelessly supported the team and handled the report processing through all stages with superb effi- ciency. Many other colleagues, especially Emanuela di Groppello, Lianqin Wang, Michael Drabble, Douglas Lehman, and Nancy Vandycke, helped in collection of the country-level data. We are grateful to Catherine Sunshine for copyediting; to John Mosier and Al Barkat of Intermax, Inc. and the team at Dohatec New Media for outstanding work in developing the CD-ROM; and above all to Mark Ingebretsen for superb production assistance. We are grateful to Eduard Bos of the World Bank for the population data used in this analysis, to Sanjeev Gupta and Erwin Tiongson of the International Mone- tary Fund for fiscal data, and to our colleague Don Bundy and his collaborators at the Imperial College, U.K., for data and assistance in analyzing the impact of HIV/AIDS on the costs of achieving universal primary education. Many other World Bank colleagues helped with the analysis of country-level results, including Philip Goldman, Alan Wright, Brigitte Duces, Michelle Riboud, Keith Hinchliffe, Gary Theisen, Hena Mukherjee, Amit Dar, Christine Allison, Paud Murphy, Bruce Jones, Alexandria Valerio, Safaa El Kogali, Benoit Millot, Regina Bendokat, Kei- ichi Ogawa, Ayesha Vawda, Ousmane Diagana, John Newman, Peter Moock, Chris Thomas, Chris Shaw, and Tanaporn Poshyananda. Colleagues at numerous bilateral agencies also generously shared existing data and aided with data collec- tion in the field. Without all of their help, this work would not have been possible. Finally, we are indebted to the governments of the Netherlands and Norway, which generously supported this research. We are especially grateful to Evelyn Herfkens, then minister of development cooperation in the Netherlands, for hosting the April 2002 international conference on Education for All in Amsterdam, at which this work was first presented. The research was carried out under the direct supervision of Ruth Kagia, Birger Fredriksen, and Josef Ritzen, and the team owes much to their vision in launching this exercise, their intellectual guidance throughout, and their unflag- ging support. The technical work was overseen by an advisory panel led by Shanta Devarajan and including Jamil Salmi, Maureen Lewis, Elizabeth King, Deon Filmer, Emmanuel Jimenez, and Margaret Miller, which contributed much to the conceptual framework and methodology. Discussions with a large number of other colleagues helped to sharpen our thinking-notably Hanke Koopman and vii Jeannette Vogelaar of the Netherlands Cooperation Agency; Len Good, Tom Wal- lace, and Susan Moir at the Canadian International Development Agency (CIDA); Harry Hagan and Steve Packer at the U.K. Department for International Develop- ment (DFID); Clay Lowery, Brian Crowe, Stephen Krasner, Buff Mackenzie, and Greg Loos of the U.S. government; Michael Hofmann of the German Ministry for Cooperation and Development; Serge Tomasi, Paul Coustere, and Jean-Claude Balmes from the French Ministry of Foreign Affairs; Olaf Seim from the Norwe- gian Ministry of Foreign Affairs; Chris Coiclough of Sussex University, U.K.; Sir John Daniel, Abimanyu Singh, and Khawla Shaheen at UNESCO; Cream Wright at UNICEF; Lant Pritchett at Harvard University; Nancy Birdsall and Steve Radelet at the Center for Global Development; Brian Ames at the International Monetary Fund; Gene Sperling of the Basic Education Coalition; Phil Twyford, Anne Jellema, Oliver Buston, and Patrick Watt of the Global Campaign for Edu- cation; and Luis Crouch, Carolyn Winter, Bob Prouty, Peter Buckland, Hans- Martin Boehmer, Sukai Prom-Jackson, Kin-Bing Wu, and Eric Swanson at the World Bank. Any data inconsistencies, conceptual flaws, or other errors, however, are the sole responsibility of the authors. We dedicate this book to our spouses, Miguel, Helene and Fara, and to our own school-aged children, Elena, Pedro, Bakoly and Mamisoa, who share our vision of a world in which every child has the chance to complete primary school. viii ACKNOWLEDGMENTS ACRONYMS: AFR Africa Region AIDS Acquired immune deficiency syndrome DAC Development Assistance Committee (of the OECD) EAP East Asia and the Pacific Region ECA Europe and Central Asia Region EFA Education for All IBRD International Bank for Reconstruction and Development (of the World Bank Group) IDA International Development Association (of the World Bank Group) 6-7 Group of Seven 6-8 Group of Eight GOP Gross domestic product GER Gross enrollment ratio GNI Gross national income HIV Human immunodeficiency virus LCR Latin America and the Caribbean Region MDB Multilateral development bank MDG Millennium Development Goal MNA Middle East and North Africa Region MRY Most recent year MTEF Medium-term Expenditure Framework NER Net enrollment ratio N60 Nongovernmental organization OECD Organisation for Economic Co-operation and Development PCGDP Per capita GDP ix PCR Primary completion rate PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper PTR Pupil-teacher ratio SAR South Asia Region SWAP Sector-wide approach UNAIDS Joint United Nations Programme on HIV/AIDS UNESCO United Nations Educational, Scientific, and Cultural Organization UNICEF United Nations Children's Fund UPC Universal primary completion x ACRONYMS Executive Summary Few global goals have been as consistently and deeply supported as the notion that every child in every country should have the chance to complete at least a primary education. The 1990 World Conference on Education for All in Jomtien, Thailand set this goal to be achieved by 2000. The World Education Forum in Dakar in 2000 reaffirmed and extended the Jomtien commitment, bringing a welcome emphasis on schooling quality while acknowledging that universal primary com- pletion had not yet been reached (box 1). Universal primary completion and gen- der equity in primary and secondary education were affirmed again in that same year as Millennium Development Goals (MDGs). Education, and particularly primary education, is a goal in and of itself, but it is also a powerful driver of progress toward the other MDGs. More equitable dis- tribution of education is correlated with lower poverty and inequality and faster economic growth (Birdsall and Londofio 1998). Greater education for girls has strong positive impacts on the health of infants and children, immunization rates, family nutrition, and the next generation's schooling attainment (World Bank 2001). New data from Africa show that education for girls and boys may be the single most effective preventive weapon against HIV/AIDS (World Bank 2002b). Primary education also contributes to better natural resource management, includ- ing conservation of the tropical rain forest (Godoy and Contreras 2001). Increas- ingly, however, research suggests that many of these positive externalities associated with primary education require that a minimum threshold of five or six years of schooling be attained-hence the importance of ensuring primary school comple- tion, and not just primary school access. Combined with sound macroeconomic policies, education is fundamental for the construction of globally competitive economies and democratic societies. Edu- cation is key to creating, applying, and spreading new ideas and technologies which in turn are critical for sustained growth; it augments cognitive and other skills, which in turn increase labor productivity. The expansion of educational opportunity is a `win-win" strategy that in most societies is far easier to implement than the redistribution of other assets such as land or capital. Ultimately, education builds what Amartya Sen (1999) calls "human capabilities"-the essential and individual power to reflect, make choices, seek a voice in society, and enjoy a better life. In short, education is one of the most powerful instruments known for reduc- ing poverty and inequality and for laying the basis for sustained economic growth, sound governance, and effective institutions. Yet the world remains far from the core Education for All (EFA) goal-universal primary school completion. This study assesses whether universal primary comple- DAKAR WORLD ED)UCA11ON FORUM.' C(OA,.S PITI *!Us M E V DE IP VE GOF 0S Expand and improve comprehensive early childhood care and education, especially for the most vulnerable and disadvantaged children. Ensure that by 2015 all children, particularly E nsurie that, 7cy 2015, children everywvbcre, girls, children ir difficult o(rcLunstances, and boys and girls alike, will be able to comlplete those belonginlg to ethnic minorities, have a fil7 course of primary schooling. access to and complete frec and compulsory primary education of good quality. Ensure that the learning needs of young peo- ple and adults are met through equitable access to appropriate learning and life skills programs. Achieve a 50 percent improvement in levels of adult literacy by 2015, especially for women, and equitable access to basic and continuing education for all adults. alminate gtcndct dispatitrir';s , primary an=J R:Hrirate gender dlisuptity i n pnrmry and secondary education hY 2005, and achieve sLcoi dafy cducation, pn refiably by 2005, gender equality in education by 2015, with a and at al lixels of education no later than focus on ensu-ing girls' full and equal access 2015. to and aclievement in basic education of good quality. Improve all aspects of the quality of educa- tion and ensure excellence of all so that rec- ognized and measurable learning outcomes are achieved by all, especially in literacy, numeracy, and essential life skills. tion can be achieved by 2015, the target date set by the Millennium Development Goals. Specifically, it asks: o How close is the world to achieving the millennium goal of universal primary completion? o Is it achievable by 2015? o If so, what would be required to achieve it, in terms of both education policy reform and incremental domestic and international financing? ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN A new World Bank database developed for this study shows that over the 1990s the average rate of primary school completion in the developing world (on a country- weighted basis) improved only from 72 to 77 percent, far short of the progress needed to ensure achievement of the education MDG of universal primary com- pletion. On a population-weighted basis, buoyed by China's high reported com- pletion rate, the global picture looks slightly better, rising from 73 to 81 percent over the decade. On either basis, however, the global average masks large regional differences in both the distance from the MDG and the progress made over the last decade, as can be seen from figures 1 and 2. Sub-Saharan Africa has the lowest completion rate by far, with barely half of all school-age children completing primary school; it is followed by South Asia, with an average completion rate of about 70 percent. The Middle East and North Africa showed a disturbing pattern of stagnation over the 1990s, with the average completion rate remaining around 74 percent. The Europe and Central Asia region (92 percent) is closest to the goal of universal pri- mary completion, followed by Latin America and the Caribbean (85 percent) and East Asia and the Pacific (84 percent). Moreover, within every region, trends at the country level diverge sharply, with rapid progress registered in some countries, stagnation in others, and declines else- where. For example, while the global average completion rate for girls improved FIGURE 1 Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted Primary completion rate (percent) 1 0 0 - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 80 |MNA - " _ SAR 60 AF 40- 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG Current trend Source: Annex figure B.5. EXECUTIVE SUMMARY 3 FIGURE 2 Primary Completion Progress in Europe and Central Asia, East Asia and the Pacific, and Latin America and the Caribbean Regions, 1990-2015, Country-Weighted Primary completion rate (percent) 100 5 90 ECA 80 - EAP LCR 70 - 60- - l l 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG Current trend Source: Annex figure B.5. more than that for boys over the 1990s, it still lags that of boys, at 76 percent com- pared to 85 percent. Serious gender disparities are evident in at least 13 countries, where girls' completion rates trail those of boys by more than 10 percentage points. While countries such as Tunisia, Bangladesh, and Sri Lanka have made impressive progress in narrowing the gender gap, in other countries it has widened, or nar- rowed only because of declines in boys' completion rates rather than improvement in girls'. Overall, though, the trends over the 1990s provide some encouraging evidence that where political will is strong, effective reforms are adopted, and international support is adequate, dramatic progress in increasing primary completion rates is possible. A significant number of countries, from Brazil and Nicaragua in Latin America to Cambodia in East Asia to South Africa and The Gambia in Africa, reg- istered improvements in the primary completion rate of 20 percentage points or more in less than a decade. This holds out hope that any developing country whose completion rate is currently 70 percent or higher could meet the MDG by 2015, provided it can achieve and sustain the rate of improvement registered by these high-performing countries. On the other hand, progress is clearly fragile. Thirteen middle-income and 15 low-income countries saw their completion rates stagnate or decline over the 1990s. The case of Afghanistan (which dropped from an already low 22 percent in 1990 to an estimated 8 percent in 1999) is obvious and dramatic. But other coun- tries losing significant ground include Zambia, the Republic of Congo, Albania, Cameroon, Kenya, Madagascar, Qatar, Iraq, the United Arab Emirates, Bahrain, and Venezuela. 4 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD THE GLOBAL PROSPECTS FOR UPC BY 2015 At the trend rate of progress achieved over the 1990s, by 2015 the global primary completion rate will not exceed 83 percent. On a population-weighted basis, the world would come closer to achieving the MDG, with about 9 out of every 10 chil- dren globally completing primary school. But, as figures 1 and 2 indicate, under- lying this global average would be a wide gulf in performance across regions. Ultimately, the MDG will not be attained unless every child in every country has the chance to complete primary school, and change will have to happen at the level of national education systems in order to reach the goal. Therefore, the focus of this analysis is the country-by-country prospects for reaching universal primary com- pletion (UPC) by 2015. According to the best available estimates, 37 of 155 developing countries have achieved or have virtually achieved universal primary completion and another 32 are "on track" to reach the goal on trend rates of progress achieved over the 1990s (table 1). Some 86 countries, however, are at risk of not reaching the goal unless progress is accelerated. They include countries that are making good progress but will fall short of the goal because their completion rates started from a very low base, as well as countries with higher completion rates that have registered declin- ing trends or stagnation during the 1990s; these 43 countries are labeled "off track." Another 27 countries must be considered "seriously off track": on current trends, their completion rates will not exceed even 50 percent by 2015. Of the 70 countries that are off track or seriously off track, 51 are low-income countries. Table 1 ............................................................. Prospects for Universal Primary Completion by 2015 Low-income * Middle-income All Developing Progress Rating Countriesa Countriesb Countries On track . 22 47 69 Achieved UPC . 11 26 37 On track to achieve UPC by 2015 11 21 32 Offtrack 51 19 70 Off track to achieve UPC by 2015 28 15 43 Seriously off track 23 4 27 No data available 9 7 16 At risk, subtotal 60 26 86 Total 82 73 155 a. Countries eligible for lending from the International Development Association (IDA) and "blend" countries eligible for IDA and IBRD lending, plus non-member low-income countries such as the Democratic People's Republic of Korea. b. Countries eligible for lending from the International Bank for Reconstruction and Development (IBRD), plus non-member middle-income developing countries. EXECUTIVE SUMMARY 5 Finally, there are 16 countries for which no data are available, and at least some of these, such as Somalia, Liberia, and Myanmar, are very likely at risk as well. This picture is not encouraging. But a significant share of the at-risk countries could reach the goal, if they could match the average rate of progress of 3 percent- age points per year observed in the best-performing countries over the 1990s. At this rate of progress, all of the middle-income and more than two-thirds of the low-income at-risk countries would reach the MDG. This goal is achievable and should be the focus of country policy and international assistance. However, the countries lagging furthest behind-many in Sub-Saharan Africa, and many countries scarred by conflict-would need to improve at even faster rates, for which there is little historical precedent. Some of these countries are making impressive progress in extremely difficult contexts. But it is dear that the worldwide attainment of universal primary completion by 2015 will require an even stronger combination of political will, sustained and deep reform, faster diffusion of best practices, and intensified financial effort than has been mar- shaled to date. WHAT WILL IT TAKE TO ACHIEVE UNIVERSAL PRIMARY COMPLETION BY 2015? To answer this question, we focused on the 55 largest low-income' countries in the world, which are home to 75 percent of all children out of school globally. These are countries whose fragile domestic resource base and institutional weaknesses make them the priority arena for a global effort to support the achievement of uni- versal primary completion. Building on pioneering work by Colclough and Lewin (1993) and other researchers,2 we analyzed primary completion rates and gross enrollments as a function of characteristics of the education system that have long been identified as key: the resources allocated to primary education; average teacher salaries and unit costs; spending on complementary non-teacher-salary items; average class size (pupil-teacher ratio); and average rate of grade repetition. Even in this relatively small sample, there was enormous variance across countries in the fiscal commit- ment to primary education and in these indicators of the structure and costs of their education service delivery, as can be seen from table 2. The sample exhibited great variance in system outcomes as well, with primary completion rates ranging from 20 to 100 percent, and gross enrollment ratios ranging from 30 to 120 percent. Very notable in figure 3 is the variance in the rela- tionship between schooling enrollments and completion rates, which provides a strong argument for the importance of tracking primary completion directly. The diagonal line in the graph represents perfect one-to-one mapping between the gross enrollment ratio (GER) and the primary completion rate (PCR), but.very 1. Countries with gross national income (GNI) per capita of US$885 or less in 2000. 2. See, for example, Mehrotra (1998), Colclough and Al-Samarrai (2000), and Delamonica, Mehrotra, and Vandemoortele (2001). 6 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2 ,................................................................... Benchmarks for Primary Education Efficiency and Quality SAMPLE MEAN IN . t ~~~1 999t2000, . . . Highest- Sample Range Adjusted Completion 2015 Variable in 1999/2000 Sample Countries Benchmarks Service delivery Average annual teacher salary (as multiple of per capita GDP) 0.6-9.6 4.0 3.3 3.5 Pupil-teacher ratio 13: 1-79: 1 44:1 39:1 40: 1 Spending on inputs other than teachers (as percentage of primary education recurrent spending) .1-45.0 24.4 26.0 33 Average repetition rate (percent) * 0-36.1 15.8 9.5 10 or lower System financing Government revenues (as percentage of GDP)a 8.0-55.7 19.7 20.7 14/16/18b Education recurrent spending (as percentage of government revenues) 3.2-32.6 17.3 18.2 20 Primary education recurrent spending (as percentage of total education recurrent spending) 26.0-66.3 48.6 47.6 50c Private enrollments (as percentage of total) 0-77.0 9.4 7.3 10 a. Government current revenues, excluding grants. b. Staggered targets proportional to per capita GDR c. For six-year primary cycle; otherwise prorated for length of cycle. FIGURE 3 Primary School Completion Rates and Gross Enrollment Ratios in a Sample of Low-Income Countries, circa 1999/2000 Primary completion rate (percent) 110- Group 1 100 - 90 - 80 - 70 - 60 - 50- 40- Gru 3 30 - A 201 Gru- 10- 0 0 20 40 60 80 100 120 140 Gross enrollment ratio (percent) * Group 1 * Group 2 A Group 3 * Group 4 2 few of these low-income countries have achieved this. Instead, three stylized group- ings may be observed, which we used to deepen the analysis: Group 1 Relatively successful countries, with high GER (85 percent or above) and high PCR (70 percent or above). Group 2 High inefficiency countries, with high GER (80 percent or above) but low PCR (60 percent or lower). Group 3 Low coverage countries, with low GER and PCR (both 60 percent or lower). Group 4 Countries falling in between the defined ranges, presenting milder versions of these patterns. When education spending and service delivery characteristics were analyzed for the three stylized groups, several clear patterns emerged. The relatively success- ful countries in Group 1: * Devote a higher share of their gross domestic product (GDP) to public primary education * Have unit costs that fall in the middle of the range-not too high and not too low * Pay teachers an average annual wage of about 3.3 times per capita GDP * Have slightly higher spending on complementary, non-teacher-salary inputs * Have an average pupil-teacher ratio of 39:1, and * Have average repetition rates below 10 percent. The Group 2 and Group 3 countries deviated widely from these average val- ues, and in very distinct ways. Group 2 countries have significantly lower average spending and strikingly higher repetition-28 percent on average, compared to below 10 percent for Group 1. Group 3 countries have dramatically higher unit costs-about 70 percent higher than the other groups'-driven by very high aver- age teacher salaries. It appears from the experience of these Group 2 and 3 countries that deviating very far from the patterns observed in the more successful countries (for example, pupil-teacher ratios of 75:1 or 13: 1, rather than 39:1 or 40:1) has forced their edu- cation systems into unhealthy adjustments and poor outcomes. The analysis sug- gests that the relatively balanced parameters observed in the Group 1 countries may offer a set of indicative benchmarks to guide service delivery and financing reforms. Bringing key service delivery and domestic financing parameters into line with benchmarks drawn from higher-performing countries offers a clear strategy for creating a higher-quality learning environment for children, associated with lower repetition, higher retention in school, and, consequently, a higher rate of pri- mary completion. Transparent parameters such as these also reveal each country's degree of domestic fiscal commitment to the goal of universal primary completion. Any global strategy for accelerating EFA progress must take this into account, encour- aging more domestic effort where it is low, and taking care not to penalize coun- tries currently showing stronger commitment. These findings also imply that the road to universal primary completion for different countries will vary, depending on how their costs and structure of service delivery compare with the indicative benchmarks. For example, the high cost 8 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD structure of Group 3 countries makes achieving universal primary completion pro- hibitively expensive; the high repetition and dropout rates of Group 2 countries make it virtually impossible. The inescapable conclusion-reaffirming what Coiclough and Lewin (1993) posited a decade ago-is that the attainment of uni- versal primary completion depends even more crucially on education system reform than on incremental financing. COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION It follows that the soundest basis for estimating the global financing requirements for achieving the education MDG is to aggregate these from country-level analysis that takes into account the reforms needed for a viable strategy in each country context. We used a simulation model to do this, estimating the costs of achieving universal primary completion in the 47 countries in our sample that have not yet achieved the goal, under different scenarios of gradual policy reform toward the benchmarks. Depending on each country's initial situation, a gradual process of either increase or decline in average teacher salaries, the pupil-teacher ratio, average repetition, and each of the other variables is programmed to occur between 2002 and 2015, at the same time as the evolution of student flows is projected in light of the latest data on population trends. This framework focuses on the quality and quantity of primary education sup- ply, but also recognizes that demand-side issues (household budget constraints, direct and opportunity costs of schooling, the social value attached to educating girls or children with disabilities, and so forth) are important determinants of school attendance and completion. Accordingly, our cost estimates assume that primary education is completely free to users (no tuition, book charges, teacher supplements, or contributions to construction from the community, for example), and we make explicit budgetary provisions for additional subsidies and incentives to overcome demand-side constraints for the most disadvantaged children, includ- ing a special provision for stipends to HIV/AIDS orphans. We assume these pro- grams would be tailored to the specific country context. We assume a public sector responsibility forfinancing the bulk of primary schooling, but not necessarily pub- lic provision. Indeed, increased service delivery through community schools, alter- native schools, nonprofit private schools, and schools run by nongovernmental organizations (NGOs) is in many developing countries a key strategy for achieving more efficient use of public resources and more equitable geographic coverage. The gradual reforms in all parameters to 2015 influence the efficiency of stu- dent flows, the domestic resources available for primary education, and the progress toward universal primary completion, in effect producing 47 country- specific strategies for achieving the MDG. Under these scenarios, the countries analyzed would expand their education system coverage 30 percent by 2015 (with a doubling of enrollments in Africa). Average spending per student would more than double in real terms, reflecting the impact of economic growth on average teacher salaries, the significant increase in schooling quality implied by the bench- mark allotment for non-salary inputs, and our provision for additional targeted support to AIDS orphans. Increased efficiency of student flows resulting from EXECUTIVE SUMMARY 9 FIGURE 4 . Domestic and External Financing Required to Achieve the Education MDG in 47 Countries, 2001-2030 Millions of 2000 constant U.S. dollars 45,000 - 40,000 - 35,000 - 30,000 - 25,000 - 20,000 - 15,000- 10,000 - 5,000 - 0- 2000 2005 2010 2015 2020 2025 2030 - Total spending on primary education Total domestic financing for primary education Financing gap these quality improvements would substantially accelerate the progress toward uni- versal primary completion by 2015. But even with increased fiscal effort in many of the countries in line with the targets for domestic resource mobilization, the simulations show that these countries, as a group, would not be able to achieve the goal without sustained and significant external financial support. Over the period to 2015, we simulated an increase in these 47 countries' own financing for primary education from a base of about $8.5 billion in 2000 to about $21 billion per year in 2015.3 Even this significant a domestic effort would not cover the total incremental costs of reaching the education MDG. Our simulations showed a financing gap over the period, rising from about $1.0 billion in the ini- tial year to a peak of $3.6 billion in 2015 when full primary coverage and quality are achieved. At the peak, this financing gap represents 15 percent of total expen- ditures. Thereafter, the financing gap would decline steadily, to an estimated 3-5 percent of total expenditures in 2030. As Table 3 shows, the bulk of the external support-more than 75 percent of the total, or close to $1.9 billion per year-would be needed in Africa. The simu- lations show that all 33 Sub-Saharan African countries in this low-income sample would face a financing gap in achieving universal primary completion. The exter- nal funding required would also represent a much larger share of their total financ- ing needs-as high as 36% in the peak year of 2015, before declining to about 6% of total requirements by 2030. 3. Unless otherwise noted, dollar amounts in this book are 2000 constant U.S. dollars. 10 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 3 ................................................. Estimated Annual Financing Gap by Region (millions of 2000 constant U.S. dollars) Latin Percentage America East Asia Middle East Europe and of Total Type of South and the and the and North Central Financing Financing Afnca Asia Caribbean Pacific Africa Asia Total Gap Recurrent 1,127 97 14 30 21 34 1,323 55 Operation 841 97 14 30 21 34 1,037 43 AIDS 286 0 0 0 0 0 286 12 Capital 725 300 34 6 49 0 1,114 45 Total 1,852 397 48 36 70 34 .2,437 100 Note: Numbers may not sum to totals because of rounding. The four South Asian countries we studied would require about $397 million per year in external funding; the three low-income countries analyzed in Latin America and the Caribbean would face a gap of $48 million per year; two countries in East Asia would require external support of about $36 million per year; the one Middle Eastern country in the sample would need $70 million per year; and the three countries analyzed in the Europe and Central Asia region would have a com- bined financing gap of about $34 million per year. An important finding is that about 55 percent of the external financing needed would be for recurrent budget support, and only 45 percent for capital support (new school construction). Since construction investments are generally easiest for donors to mobilize, we assurme that all of the new construction needed in these countries would be financed externally. But the simulations make clear that an even larger volume of external support would be needed for recurrent budget requirements. Under our target parameters, virtually all countries in the sample would increase their domestic financing for EFA, and would finance 90 percent of the incremental recurrent costs of achieving the goal themselves. But the bigger constraint to achieving the goal will be the availability of external financing for recurrent expenses, not capital. The financing gap estimated in this study is a lower-bound estimate of the global costs of attaining the education MDG, for several reasons. First and most crucially, our simulations in essence captured the incremental costs of expanding pri- mary education systems in these countries to reach the goal by 2015. They did not capture the important needs-particularly in these very low income countries-for rehabilitation and upgrading of the current system. Our data set did not permit a detailed appraisal of the adequacy of existing dassroom and administrative infra- structure or the adequacy of system functioning in each country, an appraisal that would be required to estimate the costs of needed upgrading, rehabilitation, and capacity building to complement the expansion costs we estimated. Given the EXECUTIVE SUMMARY 11 precarious functioning of the education system in very many of the countries in our sample, it can be assumed that these needs are substantial. Because these invest- ments are needed immediately, moreover, our simulation results for the first few years of the projection period particularly underestimate the true needs for external financing in these countries. Second, although our sample included all of the most populous low-income countries-accounting for 94 percent of all children out of school in low-income developing countries-there are about 20 small low-income countries and several conflict-affected countries that were not analyzed. Moreover, we only estimated financing requirements through six grades of primary schooling; countries whose official primary cycle is longer than six years will face financing requirements that we did not capture. A full costing of the external needs would have to include all countries and reflect the full length of the primary cycle in each. Third, this costing exercise simulated a reform path to the MDG for each country that assumed system reforms would be initiated immediately, and pursued steadily to 2015. In reality, there will be many cases where it is politically impossi- ble to launch all needed reforms at the same time, where the pace of implementa- tion will not always be linear, and where there is a need for the education system to deliver better service immediately, while key reforms-particularly on the resource mobilization side-may take longer to legislate and implement. To the extent that external assistance can facilitate such processes, transitional external financing requirements may be higher than the simulation estimates. However, the record on aid effectiveness also clearly points to the pitfalls of external assistance as a substi- tute for country commitment to needed reforms. Finally, this costing exercise focused on the Millennium Development Goal of universal primary completion by 2015, and not on the full set of Education for All goals established at the Dakar conference. Developing countries are committed to pursuing all six Dakar goals, and the incremental costs to attain some of them- especially the elimination of gender disparities in secondary education, the achievement of a 50 percent improvement in adult literacy by 2015, and the expansion of early childhood care and education targeted to the most vulnerable children-will be significant. The financing framework introduced in the present study provides for balanced spending on all levels of education, and not only pri- mary education, and would therefore provide some fiscal space for education sys- tems to pursue the broader Dakar goals. But parallel efforts to the current study are needed for a full costing of the Education for All agenda, and especially to provide guidance on the "good practice" policies, service delivery parameters, and addi- tional external financing that would be needed for developing countries to attain the Dakar goals in full. . ESTIMATING THE GLOBAL COSTS OF THE EDUCATION MDG Despite these limitations, the current study does represent one of the most careful efforts to date to analyze and cost a strategy for attaining the education MDG of universal primary completion. In a world where both developing and developed countries face competing priorities and budget constraints, we insist on the impor- 12 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD tance of a global strategy-such as the one oudined here-that seeks to achieve the goal at minimum adequate cost, rather than "at any cost." In this vein, we tried to generate a plausible estimate of the likely costs of achieving the education MDG (through five or six years of schooling) in all developing and transition countries, building on our detailed analysis of 47 low-income countries. "Scaling up" our analysis to include estimated needs for rehabilitation and expansion of system infrastructure (based on more comprehensive data for a smaller sample of countries) increased the total incremental costs of achieving universal primary completion by about $1.1 billion per year. Since our analysis showed financing needs for primary education already in excess of these coun- tries' capacity to finance them domestically, all of these additional costs were added to the estimated external financing gap. This increased the overall financ- ing gap for these 47 countries by roughly 45 percent, to about $3.5 billion per year. As the rehabilitation needs are all concentrated in the early years of the period, they would increase the external financing needs in those years especially dramatically. Extending the estimate from the 47 countries we analyzed to the full group of 79 low-income countries increased the estimated financing gap by an additional 8 percent-a relatively modest amount, since our sample countries account for such a large share of the total school-age population in low-income countries. Thus, the total incremental costs of achieving the education MDG (through five or six years of schooling) in all low-income countries, including all needs, would total an estimated $9.7 billion per year over the period to 2015, of which about $3.7 billion per year would need to come from official development assistance. This is about 50 percent higher than the $2.4 billion annual gap we projected. Estimating the likely costs and financing gaps for the 47 middle-income coun- tries that have not yet reached the MDG is more difficult, however. Although these countries are already much closer to the goal of universal completion, have more scope for domestic financing of primary education, and have more favorable demographic trends, their unit costs are much higher, due.to lower pupil-teacher ratios and the higher dollar costs of teacher salaries and other inputs. Based on current unit costs and enrollment data, but applying population and economic growth projections, we estimate that the incremental costs of reaching the education MDG in the middle-income countries would be in the range of $23- 28 billion per year, compared to baseline spending on primary education estimated at about $80 billion in 2000. However, this estimate is not stricdy parallel to our estimate for the lower income countries, because it assumes no changes in service delivery efficiency or domestic financial commitment to the goal. Without country-by-country analysis, it is impossible to say how these population, cost, and financing factors would bal- ance out, what the most appropriate reform trajectories for these countries would be, or what residual external financing needs would remain. The one study so far that has applied our methodology (with regionally appro- priate benchmark parameters) to 10 middle-income countries in Latin America and the Caribbean found that these countries should be able to finance the limited amount of school-level expansion needed to reach the primary education MDG, EXECUTIVE SUMMARY 13 without an external gap, if they also adopt policies to improve the efficiency of stu- dent flows and devote reasonable domestic budget allocations to primary educa- tion (di Gropello, Dubey, and Winkler 2002). However, other studies-without assumptions on efficiency or financing reforms-have generated estimates of the financing gap for middle-income countries in the range of $4 billion per year. We believe that, just as in the countries we analyzed, there is clear scope in middle-income countries to increase resource mobilization and improve efficiency in service delivery. Without careful country-by-country analysis of the type we have done, however, the most that can be said is that the incremental costs of reaching the education MDG in middle-income and transition countries could be as high as $23-28 billion per year, and, of this, the need for external financing might range between $1 billion (with appropriate policy reforms) to $4 billion, per year. Summing these with our scaled-up estimates for the low-income countries results in a global estimate that roughly $33-38 billion per year in additional spending on primary education will be needed in developing countries between now and 2015 if the education MDG is to be met. This is the annual average of a spending increase that would take place gradually over the period, but it dearly connotes a significant challenge. The increase relative to current spending levels will be much higher for the low-income countries than for the middle-income and transition countries. We estimate that even with optimal policy reforms and strong domestic fiscal commitment to achieving the goal, countries themselves will not be able to generate the resources needed. We estimate that $5-7 billion of this total spending increase would need to come through external aid. This estimate is anchored in careful country-by-country analysis. It is also shaped by an explicit focus on achieving the goal at minimum and sustainable global cost. But even this conservative estimate is many times higher than aid flows currently available for primary education, especially for the lowest income coun- tries. It will take strong effort and commitment from development partners to mobilize this incremental funding, and equal effort from developing countries to use it well. IMPLICATIONS FOR COUNTRIES AND DONORS At the Monterrey conference on development finance in 2002, the donor commu- nity pledged increased development support channeled in a new and more selective framework to those countries with both sound policies and a willingness to be held accountable for clear results. At the Dakar conference in 2000, the donor commu- nity made a commitment that no developing country with a "credible plan" for achieving EFA would fall short of the 2015 goal for lack of external support. Our analysis suggests that a relatively small set of key parameters are important deter- minants of primary completion rate progress and therefore core elements of a "credible" or sound policy framework in education. Using these "indicative param- eters" to guide education planning could bring increased technical rigor, trans- parency, and financial discipline to the process. Such a framework could help ensure that policy actions, new investments in school expansion, domestic resource 14 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD mobilization, and external assistance all lead to progressive improvements in sys- tem functioning, measured against clear benchmarks. However, this indicative framework is clearly not sufficient for a credible EFA plan, and must not be applied rigidly. First, the system-wide average values on which these parameters rest do not guarantee that the underlying distribution is efficient or equitable-particularly in large federalized education systems such as those of India or Nigeria. In India, for example, while the national average is 52 pupils per teacher, the pupil-teacher ratio is as low as 30:1 in some states and as high as 60:1 in others, reflecting serious disparities in education access and quality across the country. Addressing these regional disparities-which could not be cap- tured in our simple simulation model-will clearly be costly and will require con- certed action at the federal, state, and district levels. A credible EFA plan for any country must go beyond the national average benchmarks and also focus on sub- national variance in education financing and service delivery. Second, while the indicative benchmarks can provide a useful point of refer- ence for all countries, there will be many cases where they are culturally, institu- tionally, or financially inappropriate. The ultimate value of this framework is as a guide to the direction of reform, not as a dictate regarding where it should end. Third and most importantly, the indicative framework can help ensure that education systems have adequate overall resources and a healthy mix of core inputs. But it cannot guarantee the effective management of those resources. In a great many developing countries, achieving better management of education resources- at the central level, at the school level, and in the classroom-is as large a challenge as mobilizing more resources. Indeed, as primary education systems in many of these countries will more than double in size over the coming decade, the manage- ment challenges will become even more acute. At the central level, ministries of education must achieve greater equity and efficiency in allocating financing and deploying personnel across different regions and across schools, as well as between administrative support services and school- level delivery. The share of resources absorbed into central administration in many systems is very high, with litde value added for system quality or student learning. Across different regions, schools with similar enrollments often differ widely in the number of teachers and other resources deployed to them, with no formal ration- ale but with clear implications for quality and equity. Similarly, expenditure track- ing analyses frequently find that only a fraction of the overall education resources allocated to schools actually reaches them, and often too late in the school year to be used productively. Finally, national systems to assess student learning and mon- itor progress at the classroom and school level are crucial for holding education actors accountable and stimulating system-wide improvement. Yet they exist in very few of the countries in our sample. Management capacity at the school level is also crucial. The quality of school leadership makes the difference between an orderly environment where teachers perform and children can learn, and a chaotic environment marked by rampant absenteeism, poor school maintenance, disappearance of books and materials, and poor relations with parents and the community, as seen in all too many education EXECUTIVE SUMMARY 15 systems. Simple and often costless actions such as assigning the best teachers to the early grades, adapting the school calendar to the needs of the community, and making sure that teachers show up on time and work a full week can greatly boost student attendance and learning. Effective management at the school level makes these happen. And ultimately, it is management in the classroom that transforms education resources into student learning. Research shows that after controlling for student characteristics, learning outcomes can differ greatly even across equally resourced classrooms in the same school. What teachers do matters more for student learning than any other single factor. Teachers must use class time effectively; they must make creative use of learning materials; they must have the capacity to adapt their teaching practice to individual students' learning needs; and, above all, they must be motivated to devote time and hard work to proving that "every child can learn." In many developing countries, teachers' incentives, capacity, and practice are all greatly in need of strengthening. Specific policies to address these management issues at all levels of the educa- tion system must equally be core elements of a credible EFA plan. But the first step toward a quality school system is to ensure adequate resources, allocated in an effi- cient balance against core system parameters. Without this, few other policy objec- tives or programs can be implemented or sustained. Adopting this policy and financing framework would have several key implica- tions for developing countries: o The criteria for a "credible plan" would be less ambiguous and more technically rigorous. * Countries' own commitments to EFA could be evaluated more transparently, as the allocation of a "fair share" of domestic fiscal resources to primary education. * Steady improvement in service delivery parameters could be a quid pro quo for continued external support. o The EFA process would be focused more sharply on key outcomes, especially the primary completion rate and student learning progress, and more accurate and timely measurement of these would be required. a Countries and their partners would both be more clearly accountable for ensuring that external funding catalyzes tangible progress toward EFA and is not wasted in ineffective delivery systems. o Countries' overall domestic resource mobilization and spending, not only education ministry spending, would become subject to EFA monitoring. The implications for international development partners are equally strong. The simulation results show clearly that even with a maximum domestic effort, most low-income countries will not be able to achieve universal primary comple- tion by 2015 without changes in both the level and nature of external support. Making good on the international community's commitment at Dakar would require development partners to take six basic steps. First, they must significantly increase donorfunding for primary education. The average external financing needed for just the 48 low-income countries we analyzed is about $2.5 billion per year between now and 2015-almost a tripling of current 16 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD aid for primary education to these countries and about a fourfold increase in the level of donor support to the 33 Sub-Saharan African countries in the sample. Second, donors should ensure better targeting of "EFA priority" countries. Cur- rent patterns of aid to education are not prioritizing countries in greatest need. The countries analyzed have an average primary completion rate of only 57 percent, yet receive only about 10-15 percent of current official development assistance going to education. Third, the mix ofdonor assistance should be changed. Donors need to shift a larger share of external assistance to recurrent budget support. In turn, recipient countries need to show greater budgetary transparency and monitoring of outcomes. Fourth, donors can improve the efficiency of aid transfers. A significant share of donor assistance typically supports technical assistance contracts, consultancies, seminars, study tours, and other expenditures that-no matter how valuable-do not count directly against the "core" resource requirements for EFA estimated in our simulations, about 55 percent of which would be for recurrent costs and notably for teacher salaries and appropriate demand-side interventions. Similarly, the unit construction costs we assumed (averaging about $8,000 per classroom for the sample) are far lower than those many donors report. Shifting to community- based construction of new schools and classrooms to lower unit costs is essential for reaching the MDG but will require flexibility on the part of donors. Fifth, donors should transfer funds via new mechanisms. The stability and pre- dictability of external assistance is crucial if countries are to take on recurrent expenditures (such as hiring of additional teachers) that are not easily compressed if external support fluctuates. On the other hand, it is not easy for bilateral donors, subject to their own political processes and budget constraints, to make long-term funding commitments. Greater use of pooled donor assistance and direct budget transfers in the context of sector-wide approaches (SWAPs) and other program- matic support could help match donor assistance more effectively to countries' core financing needs and ensure a more stable and predictable flow of funding. Finally, there is an urgent need for more effective monitoring of progress, increased research, andfaster diffusion of knowledge about what works. The costs of EFA monitoring, data collection, international research, and global and local activ- ities to diffuse new knowledge are not included in the estimated financing gap, but these investments in the global public good should be considered core responsibil- ities of the international community. The road to EFA will for many countries be an enormous challenge. Accumulated country experience and international research can play an important role in smoothing it. THE EFA FAST-TRACK INITIATIVE Building on the above analysis, a new compact for primary education designed to accelerate global progress toward the education MDG was endorsed by the Devel- opment Committee of the World Bank and International Monetary Fund in April 2002 and by the G-8 in its action plan for education at the June 2002 summit in Kananaskis, Alberta, Canada. The new compact, called the EFA Fast-Track Initia- tive, is the first proposal to emerge since the Monterrey conference that aims at EXECUTIVE SUMMARY 17 accelerating MDG progress using the Monterrey framework of increased develop- ment support in exchange for increased accountability for results. The new initia- tive is supported by all major bilateral donors for education and by UNESCO, UNICEF, the World Bank, and the regional development banks, all of which have joindy formed the EFA Fast-Track Partnership. At the heart of the Fast-Track Ini- tiative are: * A commitment by developing countries to accelerate efforts to EFA Indicative Framework achieve universal primary education EFA o nd.ca...e F.raewo.......... rk cost-effectively, within an "EFA * Average annual teacher salary indicative framework" (box 2); and (as multiple of per capita GDP) * A commitment by donors to provide sustained incremental * Pupil-teacher ratio financing (as much as possible on a * Share of recurrent spending on grant basis), where credible plans to inputs other than teachers accelerate progress in primary, education exist. * Average repetition rate In June 2002, a first set of 18 low- * Education share of government income countries was invited to join the recurrent budget initiative and to submit their EFA plans, *Primary education share of including baseline indicative framework indicators and annual targets, for donor education recurrent budget financing. The 18 countries (box 3) are diverse regionally and in terms of their proximity to universal primary comple- tion; together, they account for an estimated 18 million children without access to education. This first set of countries was invited to consider committing to the Fast-Track Initiative on the basis of two simple and transparent criteria: (a) they have formally adopted national Poverty Reduction Strategy Papers (PRSPs) that integrate their education plans into overall national development priorities; and (b) they have education sector plans in place, agreed with the donors. The ration- ale for these two criteria is that having these elements in place should allow fast- track support to catalyze measurable progress more quickly. It should be noted that the Fast-Track Initiative is aimed at accelerating MDG progress in, and learn- ing lessons from, countries that are currently on track to reach the goal as well as supporting countries that are off track. A second set of five high-priority countries was also invited to join the initia- tive, but with a different status initially, as they did not yet meet the two criteria. These "Big Five" countries are deemed high priority because they account for the largest numbers of children without access to primary education globally-about 50 million of the 113 million children in total estimated to be out of school. The spirit of the Fast-Track Initiative is that country commitment to sound sector pro- grams integrated into broader poverty reduction strategy as well as commitment to appropriate policy actions in line with the EFA indicative framework are important for effective use of development resources. "Analytical Fast-Track" support aims to help these countries reach that status. India is the first of the "Big Five" countries to meet the two criteria, and the government is considering participation in the FTI. ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD In countries with PRSPs and sector plans in place, the Fast-Track process involves a complemen- First EFA Fast- tary in-country analysis to benchmark education sys- Track Group, 2002 tem performance relative to the EFA indicative ....................... framework; to set appropriate annual targets for their Albania country context; and to refine estimates of the exter- Bolivia nal financing needs for accelerated progress in pri- Burkina Faso Ethiopia mary education, consistent with the implementation The Gambia of appropriate reforms and the medium-term expen- Ghana diture framework established in their PRSP. Guinea Guyana Although for the first set of countries these adjust- Honduras ments have been set out in "Fast-track proposals," it Mauritania is expected that the process of identifying priority Mozambique policy actions to align system functioning with the Niaeragua indicative framework benchmarks will increasingly Tanzania be mainstreamed into the development of those plans Uganda in the first place and separate FTI proposals will not Vietnam Republic of Yemen be needed. The first FTI proposals have represented a Zambia more comprehensive assessment of financing needs Analytical Fast-Track Countries than we costed, as they include rehabilitation Bangladesh requirements. The estimated expansion needs, how- Democratic Republic of Congo ever, may be compared with the financing gap esti- Nidgeria mates presented here. Pakistan An important part of the process is also careful assessment of the physical and institutional capacity to execute increased primary education investment and expenditure. The Fast-Track Initiative implies a major expansion of the management challenge for systems that are generally perceived to be weakly managed today. But this cannot be an argument against such expansion; it simply means that attention to capacity building and institutional support must be an equal part of the partnership effort. Finally, the estimated needs are compared with the pipeline of existing donor commitments for primary education in each country, including general budget support under Poverty Reduction Support Credits (PRSC) or other multisector programs. It should be recalled that the financing gaps estimated in the present study are gross financing gaps, with no adjustment for the current level of external assistance to the primary sector. As of March 2003, ten of the first 18 countries invited to join the Fast-Track Initiative submitted proposals for consideration. The Fast-Track partners commit- ted, upon verification of the estimated financing gaps against implementation plans, to ensure that these gaps are filled for the next three years, contingent on countries' continued progress in executing the accelerated program and improving sector functioning in line with their indicative framework targets. The partners also agreed to meet regularly to review implementation, harmonize their education assistance to Fast-Track countries, and decide on additional proposals. Intensified collaboration among donor representatives at the dient country level is a key part of this process. EXECUTIVE SUMMARY 19 In addition to transparent annual monitoring of their progress against indica- tive framework targets, recipient countries also committed to monitoring key out- comes such as the net intake rate into first grade for girls and boys, the primary completion rate for girls and boys, and student learning achievement, although it is understood that these outcome indicators can be slow to reflect progress. CONCLUSION Universal primary completion is crucial for national economic and social advance- ment. It is a goal that all developing countries are committed to achieving by 2015, but one that will not be reached without a significant acceleration of current progress. Faster progress requires the bridging of substantial policy, capacity, and data gaps in many developing countries, in addition to financing gaps. The lack of external financing in some cases is not as binding as the constraints imposed by lack of capacity or the policy framework. This study focuses on two of these gaps-the education policies that in many countries are needed for faster progress, and the incremental financing required to support this progress. The data we used did not permit us to analyze issues of insti- tutional capacity in any depth, despite the obvious importance of capacity for the implementation of policies and investments and the attainment of desired out- comes. Nor does this study focus on the data gap per se, although the research was hampered by the limited, inconsistent, and outdated education statistics available in the countries analyzed, and the new primary completion database we developed is an effort to provide a better basis for monitoring MDG progress. Our projections may be considered a minimum estimate of the incremental financing needed to achieve the MDG in the 48 low-income countries (including Afghanistan) currently furthest from the goal, within a framework of country com- mitment and gradual but effective policy reform. Although the $2.5 billion per year core external funding requirement we estimate is conservative, it is nonethe- less many times higher than the current level of aid for primary education to these countries. Our conclusion is that both the policy and implementation challenge for low-income countries and the financing challenge for their development part- ners will be significant if the education MDG is to be met. Finally, however important a goal it may be, primary completion is not the only challenge facing education systems in the developing world. Rather, it is just the first step toward a system of lifelong learning for all citizens, which is as rele- vant for the poorest countries as it is for the wealthiest. All countries, no matter how far they are today from universal primary completion, must simultaneously invest in and promote the balanced development of all levels of their education sys- tems. In a globally integrated and highly competitive world economy, no country can any longer consider primary schooling a terminal level of education for its labor force. Indeed, it is important that expanded donor support for primary edu- cation under the EFA Fast-Track and other initiatives be matched by efforts to help countries expand lower secondary education, in anticipation of a growing wave of primary graduates. 20 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD But increasing the share of children who complete primary schooling is the essential first step. In a borderless world, where the gulf between the rich, educated, and empowered and the poor, stagnating, and powerless increasingly poses threats to all, the achievement of universal primary completion-like the other MDGs- is of global interest. The new EFA Fast-Track Initiative, if launched successfully and expanded steadily to reach all of the at-risk developing countries, offers the possibility of boosting rates of primary completion progress to the levels necessary to reach the goal. Few global goals have been as consistently and deeply supported as the notion that every child in every country should have the chance to complete primary school. With global effort, it could become a reality. EXECUTIVE SUMMARY 21 CHAPTER The Global Challenge One hundred eighty-nine countries have committed themselves to eight Millen- nium Development Goals aimed at eradicating extreme poverty and improving the welfare of their people by the year 2015 (box 1.1). The second of the goals is "Achieve universal primary education," with the specific target of ensuring that, "by 2015, children everywhere, boys and girls alike, will be able to complete a full course of primary schooling." It echoes a commitment made by many of the same countries in Jomtien, Thailand in 1990 to achieve universal primary education by the year 2000. The Jomtien commitment was reaffirmed and extended at the World Education Forum in Dakar in 2000 (box 1.2). As the Dakar forum acknowledged, the Jomtien goal was not met. Many coun- tries dearly remain far from the target. However, until now it has not been possible to assess where individual countries stand in relation to the target with any accuracy, for lack of internationally comparable data on primary completion rates. In the absence of such data, it has been difficult to evaluate the global prospects for reaching the tar- get by 2015 or to estimate the incremental actions and financing that would be required. Given strong international interest in these issues, several previous studies have attempted to analyze the likelihood of the education MDG being met by pro- jecting trends to the goal based on enrollment data rather than completion rates, and by employing a number of different methodologies for estimating the incremental costs. Resulting estimates of the incremental global financing requirements have var- ied widely, from approximately $7 billion to $15 billion annually. This study seeks answers to three questions: * How close is the world to achieving the millennium goal of universal primary completion? * Is it achievable by 2015? and * Ifso, what would be required to achieve it, in terms of both education policy reform and incremental domestic and internationalfinancing? The approach here differs from all previous studies in that it is based on direct measurement of primary completion rates, rather than relying on conventionally available gross and net enrollment ratios, which are a poor proxy for schooling completion rates. We draw on the first effort to create an internationally standard- ized database of primary completion rates in 155 countries and trace the evolution of these rates from 1990 to the most recent year possible. (See chapter 2 for a defi- nition of the completion rate, the methodology used to calculate it, and a discus- sion of data and technical issues.) Using these new data, we examine the countries and regions in which the greatest progress has been registered since Jomtien and those in which completion rates have stagnated or declined. We analyze the prospects for reaching the MDG with no change in current trends. 23 Millennium Development Goals GOALS AND TARGETS INDICATORS -GOAL 1: ERADtCATE IEXTREME POVERTY AND HUNGER TARGET 1. Halve, between 1990 and 2015, the proportion of people 1. Proportion of population below $1 per day whose income is less than $1 a day. 2. Poverty gap ratio (incidence x depth of poverty) 3. Share of poorest quintile in national consumption TARGET 2. Halve, between 1990 and 2015, the proportion of people 4. Prevalence of underweight children (under 5 years of age) who suffer from hunger. 5. Proportion of population below minimum level of dietary energy consumption GOAL 2: ACHIEVE UNIVERSAL PRIMARY EDUCATION TARGET 3. Ensure that, by 2015, children everywhere, boys and girls : 6. Net enrollment ratio in primary education alike, will be able to complete a full course of primary schooling. 7. Proportion of pupils starting grade 1 who reach grade 5 8. Literacy rate of 15- to 24-year-olds GOAL 3: PROMOTE GINOER EQUALITY AND EMPOWER WOMEN TARGET 4. Eliminate gender disparity in primary and secondary 9. Ratio of girls to boys in primary, secondary, and tertiary education preferably by 2005 and at all levels of education no education later than 2015. 10. Ratio of literate females to males among 15- to 24-year-olds 11. Share of women in wage employment in the nonagricultural sector 12. Proportion of seats held by women in national parliaments GOAL 4: REDUCE CHILD MORTALITY. TARGET 5. Reduce by two-thirds, between 1990 and 2015, the * 13. Under-5 mortality rate under-5 mortality rate. 14. Infant mortality rate 15. Proportion of 1-year-old children immunized against measles GOAL 5: IMPROVE MATERNAL HEALTH TARGET 6. Reduce by three-quarters, between 1990 and 2015, the 16. Maternal mortality ratio maternal mortality ratio. 17. Proportion of births attended by skilled health personnel GOAL 6: COMBAT HIVWAIS, MALARIA, AND OTHER DISEASES TARGET 7. Have halted by 2015, and begun to reverse, the spread 18. HIV prevalence among 15- to 24-year-old pregnant women of HIV/AIDS. 19. Contraceptive prevalence rate 20. Number of children orphaned by HIV/AIDS TARGET 8. Have halted by 2015, and begun to reverse, the 21. Prevalence and death rates associated with malaria incidence of malaria and other major diseases. 22. Proportion of population in malaria risk areas using effective malaria prevention and treatment measures 23. Prevalence and death rates associated with tuberculosis 24. Proportion of TB cases detected and cured under DOTS (Directly Observed Treatment Short Course) GOAL 7: ENSURE ENVIRONMENTAL SUSTAINABILITY TARGET 9. Integrate the principles of sustainable development into 25. Proportion of land area covered by forest country policies and programs and reverse the loss of 26. Land area protected to maintain biological diversity environmental resources. , 27. GDP per unit of energy use (as proxy for energy efficiency) 28. Carbon dioxide emissions (per capita) (Plus two figures of global atmospheric pollution: ozone depletion and the accumulation of global warming gases) TARGET 10. Halve, by 2015, the proportion of people without 29. Proportion of population with sustainable access to an sustainable access to safe drinking water. improved water source TARGET 11. By 2020, have achieved a significant improvement in * 30. Proportion of people with access to improved sanitation the lives of at least 100 million slum dwellers. 31. Proportion of people with access to secure tenure (Urban/rural disaggregation of several of the above indicators may be relevant for monitoring improvement in the lives of slum dwellers) GOALO-01MILPA A AL- PARTNERS1P F-ORDEVOPMENT- . 24 Global "Education for All" Goals DAKAR WORLD EDUCATION FORUM GOALS MILLENNIUM DEVELOPMENT GOALS Expand and improve comprehensive early childhood care and education, especially for the most vulnerable and disadvantaged children. Ensure that by 2015 all children, particularly Ensure that, by 2015, children everywhere, girls, children in difficult circumstances, and boys and girls alike, will be able to complete those belonging to ethnic minorities, have a full course of primary schooling. access to and complete free and compulsory primary education of good quality. Ensure that the learning needs of young peo- ple and adults are met through equitable access to appropriate learning and life skills programs. Achieve a 50 percent improvement in levels of adult literacy by 2015, especially for women, and equitable access to basic and continuing education for all adults. Eliminate gender disparities in primary and Eliminate gender disparity in primary and secondary education by 2005, and achieve secondary education, preferably by 2005, gender equality in education by 2015, with a and at all levels of education no later than focus on ensuring girls' full and equal access 2015. to and achievement in basic education of good quality. Improve all aspects of the quality of educa- tion and ensure excellence of all so that rec- ognized and measurable learning outcomes are achieved by all, especially in literacy, numeracy, and essential life skills. Next, building on and extending the research of others, such as Colclough and Lewin (1993), Oxfam (Watkins 1999), and Delamnonica, Mehrotra, and Vande- moortele (2001), we identify a set of key education policy and domestic financing parameters that can explain countries' differential MDG progress. We find that edu- cation systems in the low-income countries that have either achieved 100 percent primary completion or are relatively close have some basic features in common. Finally, we ask: If those countries currendy lagging behind were to reform key featur es of their education systems to more closely approximate the systems of more suiccessful countries, could universal primary completion be achieved by CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 25 2015? And, if so, what would be the incremental domestic and international financing requirements under this scenario? This study is the first attempt we know of to analyze the challenge of Educa- tion for All through both the development of a new monitoring indicator and direct collection of the most recent available educational enrollment and public finance data. Because of time limitations, the analysis concentrates on the 47 low- income countries with populations over 1 million4 that have not yet achieved uni- versal primary completion. A full estimate of the global cost of achieving the education MDG would have to include the countries not analyzed here. However, the countries we studied are home to 75 percent of all children out of school glob- ally. These countries are far from the goal, with an average primary completion rate of only 57 percent, and their poverty, fragile domestic resource base, and institu- tional weakness make them priority claimants on international support. The bulk of the incremental external resources and effort for global achievement of universal primary completion will very clearly be needed here. A major effort was made for this study to update the global picture of progress to date through the direct collection of data from a large number of developing coun- tries. In cases where recent country data were not available or only partially available, we used published UNESCO data. In all cases, new data were checked for consistency against any available household survey data and UNESCO sources. The exercise pointed to serious issues of accuracy and consistency in education enrollment data for many countries, and in some cases required us to make estimates that diverged from official enrollment statistics when these were inconsistent with population, household survey, or past data, but always using consistency with other data as a guide. Given the ambitious scope of this work and the relatively short time in which it was carried out, this study is only a small first step in what we hope will be a new analytical direction for EFA. The technical annexes and CD-ROM accompanying this volume include the simulation model and all of the raw data, sources, and assumptions used for the countries analyzed, as well as the country-by-country sim- ulation results. We will have succeeded if this work inspires governments coimnit- ted to Education for All as well as national and international education researchers to focus on primary completion rates, and to revisit the policy framework, country data, and simulations presented in this report. The road to EFA will for many countries be an enormous challenge. Accumulated country experience and expanded international research can play an important role in easing the way. WHY IS UNIVERSAL PRIMARY EDUCATION SO IMPORTANT? Education is one of the most powerful instruments known for reducing poverty and inequality and for laying the basis for sustained economic growth. It is funda- mental for the construction of democratic societies and dynamic, globally compet- itive economies. For individuals and for nations, education is the key to creating, applying, and spreading knowledge. 4. Low-income countries were defined as IDA-eligible in 2001 (that is, countries with GNI per capita of $885 or less in 2000). 26 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Primary education develops the capacity to learn, to read and use math, to acquire information, and to think critically about that information. Primary educa- tion is also the gateway to all higher levels of education that train the scientists, teachers, doctors, and other highly skilled professionals that every country, no mat- ter how small or poor, requires. Microeconomic research has established unequivo- cally that education improves individual incomes; Psacharopoulos and Patrinos (2002) estimate an average global private return on primary education of 27 per- cent. Research also indicates the contributions of primary education to better natu- ral resource management, including conservation of the tropical rain forest (Godoy and Contreras 2001) and more rapid technological adaptation and innovation. Broad-based education is associated with the faster diffusion of information in the economy, which is crucial for increased productivity among workers and citizens in traditional as well as modern sectors (Porter 1998; Hanushek and Kimko 2000). When a large share of children do not complete primary education, the pro- ductivity of the labor force, the potential for knowledge-driven development, and the reservoir of human potential from which society and the economy can draw are all fundamentally constrained. As figure 1.1 shows, in several developing regions the average level of schooling of the labor force is still less than a complete primary education. At the start of the new millennium, adults average just 0.8 years of for- mal education in Mali and Niger, 1.1 years in Mozambique and Ethiopia, 2.0 years in Nepal, and 2.5 years in Bangladesh (Barro and Lee 2000). Research strongly suggests that such low levels of human capital are funda- mentally inadequate for sustained economic development, stable democratic insti- tutions, or poverty reduction. Azariadis and Drazen (1990) were the first to postulate that countries may be trapped in a low-returns equilibrium until their FIGURE 1.1 Average Educational Attainment of Adult Population by Region, 2000 Average years of schooling completed 12 - 10- 9.68 9.76_ 8- 6- 5.44 ~6.06 4.57 4 - -3.52 2 U .. Sub-Saharan South Asia Middle East Latin Transition Advanced Africa and North America economies countries Africa and the Caribbean Region Source: Barro and Lee 2000. CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 22 level of human capital accumulation rises above six years of schooling. Once this threshold is passed, countries seem to achieve a higher steady-state macroeconomic growth path. The latest empirical studies on the question of the impact of educa- tion on economic growth all report a positive association (Barro 1999b; de la Fuente and Domenech 2000; Hanushek and Kimko 2000). On democratization, Barro (1999a) finds in a study of more than 100 countries between 1960 and 1995 that the propensity for democracy rises both with primary schooling and with a smaller gap between male and female primary attainment. The role of primary education in reducing poverty and income inequality is even more strongly established than is its contribution to overall economic growth. Illiteracy is one of the strongest predictors of poverty, and unequal access to educa- tional opportunity is one of the strongest correlates of income inequality. A large body of research points to the catalytic role of primary education, "the people's asset" (O'Connell and Birdsall 2001), for those individuals in society who are most likely to be poor: girls, ethnic minorities, orphans, people with disabilities, and people living in rural areas. Extending adequate-quality primary education to these vulnerable groups is crucial in order to equip them to contribute to and benefit from economic growth. Data from the International Adult Literacy Survey (OECD and Statistics Canada 2000) indicate a high correlation between country levels of income inequality and inequality in the distribution of literacy, suggesting that more evenly spread levels of human capital are associated with greater income equality. Recent research by Birdsall and Londofio (1998) confirms that these factors are closely linked: more equitable distribution of education promotes faster economic growth as well as reducing inequality. Birdsall and Londoflo have shown that the degree of inequality in the distribution of education has a strong and robust nega- tive effect on growth, independent of the average level of education and also inde- pendent of factors such as trade openness and varying natural resource endowments. The implication is clear: the expansion of educational opportunity is one of the most powerful tools governments have to simultaneously promote income equality and growth-a "win-win" strategy that in most societies is far easier to implement than the redistribution of other assets such as land or capital. Ultimately, the case for universal primary education goes beyond economic arguments. Education provides people with what Nobel laureate Amartya Sen (1999) calls "human capabilities"-the essential and individual power to reflect, make better choices, seek a voice in society, and enjoy a better life. Education, and particularly primary education, also promotes achievement of all of the other Millennium Development Goals: poverty reduction, gender equity, child health, maternal health, lower HIV/AIDS and other communicable diseases, and environ- mental sustainability. Indeed, a substantial body of research documents that education-and espe- cially education for girls-is one of the strongest drivers of improvement in fertil- ity, health, and nutrition outcomes. Girls' education has documented impacts on infant and child mortality and enhanced family welfare. A recent study of 63 coun- tries concluded that gains in women's education made the single largest contribu- tion to declines in malnutrition during 1970-1995, accounting for 43 percent of 28 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD the total progress (Smith and Haddad 2000). Another study, using data on 100 countries, found that an additional year of female education reduced total fertility rates on average by 0.23 births, while a three-year increase in the average educa- tional level of women was associated with as much as one child less per woman. It also found that mothers who have completed primary education are 50 percent more likely to ensure that their infant children are immunized than mothers with no education (World Bank 2001). And very recent research indicates that for girls and boys, education may be the single most effective preventive weapon against HIV/AIDS. New data from high seroprevalence countries show that the better educated have lower rates of infection, especially among younger people (Gregson, Waddell, and Chandiwana 2001; Kelly 2000; Vandemoortele and Delamonica 2000). In sum, progress toward the goal of universal primary education will unquestionably have strong complementary effects on achievement of the other millennium goals. Universal primary completion is by no means the only challenge facing educa- tion systems across the world. It is only the first step toward the ultimate goal of lifelong learning for all citizens, which is as relevant for the poorest countries of the developing world as it is for the countries of the Organisation for Economic Co- operation and Development (OECD). All countries, no matter how far they are today from universal primary completion, must simultaneously invest in and pro- mote the balanced development of all levels of their education systems. In a glob- ally integrated and highly competitive world economy, no country can any longer consider primary schooling a terminal level of education for its labor force. But increasing the share of children who complete primary schooling is the essential first step. In a borderless world, where the gulf between the rich, educated, and empowered and the poor, stagnating, and powerless increasingly poses threats to all, the achievement of universal primary completion is of global interest. This book lays out a strategy for accelerating progress toward that goal by 2015. WHY UNIVERSAL PRIMARY EDUCATION MUST MEAN UNIVERSAL PRIMARY COMPLETION To date, efforts to achieve Education for All have focused heavily on getting chil- dren enrolled in school, rather than on improving either completion rates or stu- dent learning outcomes. This is problematic for several reasons. First, a growing body of research suggests that completion of at least five to six years of schoolin4 is a critical threshold for sustainable mastery of basic competen- cies, such as literacy and basic numeracy. Literacy surveys conducted in African countries and elsewhere indicate that a high share of the adults who have com- pleted less than five or six years of primary schooling remain functionally illiterate and innumerate for the rest of their lives (figure 1.2). The strong implication is that from a human capital perspective there is a substantial difference between getting all children enrolled in primary school and ensuring that all children complete the five- or six-year primary cycle. Especially striking in the data is the very limited impact on lifelong literacy from as many as three years of schooling. It is plausible that many of the other direct benefits and externalities of education are similarly CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 29 FIGURE 1.2. Proportion of Adults Who Can Read and Write Easily by Highest Grade Attained, Togo and Niger Percent 100 - 90 - 80 - 70 - 60 - 50 - 40- 30 - 20 - 10D - 10- - 0- I I I I I IIIiIIII 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Highest grade attained - Togo Niger Source: UNICEF Multiple Index Cluster Survey (MICS) data, 2000, and authors' estimates. linked to the completion of a relatively high threshold number of five to six years of schooling, which in fact represents the length of the primary cycle in most coun- tries. So, formulating the MDG target in terms of universal primary completion, rather than universal primary enrollment, makes strong sense from the standpoint of human capital formation. Second, schooling enrollment ratios, whether on a gross or net basis, are poorly correlated with the rate of primary school completion. In virtually every developing country, the horizontal line of an average enrollment ratio masks the underlying reality of a curve-shaped schooling profile in which many more children begin school than complete it. Schooling profiles (such as those shown in figure 1.3) con- structed from household survey or education enrollment data give a good picture of how access to schooling can differ from retention along the primary cycle for schooling cohorts in the recent past. These data show, for example, that in both Brazil and Indonesia in the late 1980s access to primary schooling was already fairly universal, with close to 100 percent of children starting grade 1. However, the pat- tern of retention in school was very different, resulting in only about 60 percent of children completing five grades in Brazil, compared to 90 percent in Indonesia. A crucial issue from the standpoint of EFA monitoring is the fact that a single gross enrollment ratio, or even a single net enrollment ratio, can be consistent with a number of different schooling profiles. As a result, there is no consistent correla- tion between either gross or net enrollment ratios and the primary completion rate. 30 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD FIGURE 1 .3, Sample Schooling Profiles Proportion of 15- to 19-year-olds who have completed each grade Brazil 1996 India 1992-3 Indonesia 1997 Attainment profile, ages 15-19 Attainment profile, ages 15-19 Attainment profile, ages 15-19 Proportion Proportion Proportion 1.0- 1.0 - 1.0 - 0.8- 0.8- 0.8- 0.6- 0.6- 0.6 0.4 0.4 - 0.4 - 0.2 0.2 - 0.2 - 0- 0 - 0 - 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 Grade Grade Grade Source: Filmer and Pritchett 1999. Figure 1.4 shows the substantial disparities between gross enrollment ratios and primary completion rates for a sample of developing countries. What is signif- icant from an analytical standpoint is not so much that a disparity exists, but that there is no constant relationship underlying the gap. FIGURE 1.4. Primary Gross Enrollment Ratios and Completion Rates, Selected Countries, 1999 Prlmary gross enrollment rate Primary completion rate I Niger| | Niger l Burkina Faso | | I Burkina Faso Ethiopia| I Ethiopia| Burundi I Burundi Tanzani w--~ Tanzania Guinea-Blssau _ Guinea-Bissau | Cha Chad I Ghana _Ghana Benin ci The Gambia The Gambia Ker ya Gny a Rwanda Ni=ogua Ncaragua India jIndia Uganda |j Uganl Honduras *o nduras| Indonesia * _ Indonesia Bolivia _ Bolivia Bangladesh _ Bangladesh Malawi _ l Malawi 100 50 0 50 100 Percent Source: Annex table A.2. CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 31 Disparities between the primary gross enrollment ratio and the completion rate arise for many reasons: children enter school early (below the official schooling age), or, more commonly in developing countries, they start school late. They may repeat grades. Another common pattern is that children drop out of school before the end of the year, because of their own or other family members' illness or their families' need for their labor, and return to reenroll in the same grade the following year. Finally, schools may be incomplete and not offer all grades locally. All these factors contribute to the fact that gross enrollment ratios are typically 10-60 per- cent higher than primary completion rates. Important gender differences become evident in the comparison of GERs against completion rates, as well. In some countries, for example in the Caribbean, girls' GERs may be somewhat lower than those of boys but their completion rates higher. Elsewhere, as in several West African countries, girls' enrollments may show only a slight disparity with boys', but girls' completion rates are very significantly lower. Completion rate data greatly enhance understanding of the gender issues that exist in educational opportunity. As important as analyzing the overall completion rate is decomposing it for different groups. In every country, completion rates are lowest for children from poor families and children in rural areas. Household survey data, as in figure 1.5, show how the schooling profile for children in the lowest income quintile can lag that of children from higher income groups. Moreover, in some countries, as noted, gender equity is also an issue, and completion rates are sharply lower for girls than for boys. In such countries the combined impact of family income and gender can produce a dramatic disparity between schooling completion rates for girls from the poorest families and boys from the wealthiest families-as in Nepal (figure 1.6). As table 1.1 shows, even in countries where the current GER is close to 100 percent, the proportion of rural children, and especially rural girls, actually com- pleting the primary cycle can be extremely low. It would be a tragedy for these vul- nerable groups if countries such as Togo took 115 percent GER to mean that Education for All had been achieved. Similar issues exist with respect to net enrollment ratios (NERs). Although one might expect net enrollment ratios, which exclude overage students, to be more closely correlated with primary completion rates, this is not the case. Table 1.2 pro- vides an example of the variance in completion rates among countries with an identical 81 percent NER. Therefore, although the net enrollment ratio is useful for monitoring the proportion of the official school-age population that is not cur- rently enrolled-the "out of school population"-it is not a good substitute for direct measure of the primary completion rate as the basic indicator of progress toward the education MDG. In the search for something more reliable than gross enrollment ratios and in the absence of alternatives, the net enrollment ratio has in fact been proposed as the key indicator for monitoring progress toward the education MDG. But in addition to the fact that it does not capture actual primary completion, the net enrollment ratio presents another disadvantage: the target of 100 percent net enrollment in primary school is an unrealistic goal. It would require that every single child enter primary 32 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD FIGURE 1.5. Schooling Profiles Disaggregated by Income Proportion of 15- to 19-year-olds who have completed each grade Mali 1995-96 Pakistan 1991-92 Colombia 1995 Attainment profile, ages 15-19 Attainment profile, ages 15-19 Attainment profile, ages 15-19 Proportion Proportion Proportion 1.0- 1.0 - 1.0 - 0.8 0.8- 0.8 0.6- 0.6- 0.6- 0.4- 0.4- 0.4- 0.2- 0.2- 0.2- 0- 0- 0 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 Grade Grade Grade - Richest - Middle - Poorest Source: Filmer and Pritchett 1999. FIGURE 1.6. Schooling Profiles Disaggregated by Gender and Income Proportion of 15- to 19-year-olds who have completed each grade Nepal 2001 Attainment profile, ages 15-19 Proportion Proportion 1.0 - 1.0 - 0.8- 0.8- 7 \ 0.6- 0.6- 0.4 - 0.4- 0.2- 0.2- 0 0- 1 3 5 7 9 1 3 5 7 9 Grade Grade - Male - Female - Rich and Male - Poor and Female Source: Filmer 2000 and updates. CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 33 Table 1.1 Proportion of Children Completing Primary School: Regional Averages and Selected Countries I PERCENTAGE OF AGE GROUP COMPLETING PRIMARY CYCLE Primary Gross Region and Country Enrollment Ratio (%) Total Rural Rural Girls Africa 77 45 Niger 31 20 12 7 Burkina Faso 45 25 16 10 Guinea 62 34 25 11 Benin 86 39 27 14 Mauritania 88 46 42 38 Mozambique 78 36 21 14 Madagascar 90 26 12 11 Togo 115 68 57 46 South Asia 100 70 Latin America and the Caribbean 113 85 - - Middle East and North Africa 95 74 . - Not available. Source: Annex table A.2 and authors' estimates from World Bank education country status reports. school at exactly the official schooling age, proceed through the cycle with zero repetition, and stay in school with no disruptions, resulting in a 100 percent on- time completion rate. If such a perfect cohort flow could be achieved, the net enrollment ratio would in fact be equal to the primary completion rate: both would be 100 percent. However, although virtually all children in OECD coun- tries complete primary school, primary NERs rarely reach 100 percent. Indeed, the Table 1.2 Comparison of Gross Enrollment Ratio, Net Enroilment Ratio, and Primary Completion Rate for Selected Countries, 2000 Gross Enrollment Net Enrollment Primary Completion Country . Ratio Ratio Rate (%) El Salvador . 111 81 . 80 Mongolia 92 81 . 66 Togo 115 81 68 Source: Annex table A.2 and UNESCO data. 34 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD average net enrollment ratio across the OECD is only 94 percent, and this ratio has been remarkably stable over the past 30 years of educational development in indus- trial countries (Brossard and Gacougnolle 2001). What this net enrollment ratio tells us is that even in the most advanced coun- tries, some children start school a litde early or late, some may struggle to get through the primary curriculum, and some may be held back a grade at one point or another, but with the right support and, above all, a school system-wide ethos that "every child can learn," very close to 100 percent of children eventually com- plete primary schooling. This more flexible concept is a more realistic-and sub- stantively meaningful-goal for developing countries as well. It puts the onus on school systems to prepare teachers with diverse pedagogical strategies to meet chil- dren's different learning needs. It requires school systems to allocate resources so that special support is provided to slower learners, children with physical or emo- tional disabilities, or children for whom consistent school attendance is jeopard- ized by poverty or family health crises. It requires school systems to put in place systems of learning assessment to ensure that children's grade progression actually reflects adequate mastery of the primary curriculum. These are in fact the substantive goals of EFA. They should be measured by an indicator that captures a country's progress over time in delivering this kind of quality educational service to its children. That indicator should allow for the uneven path to primary school completion that is a reality for many children in the developing world, given constraints to on-time enrollment (if an overcrowded local school lacks places, for example) and constraints to stable attendance. The methodology proposed in this report for calculating a primary completion rate that provides a direct measure of the share of children who complete primary education (regardless of their age at completion) provides such an indicator. The next chapter analyzes international progress in improving primary completion rates over the decade of the 1 990s, while subsequent chapters focus on the policies that underlie this progress. Important aspects of the story in many countries are policy reforms and pedagogical innovations to create primary schools that more flexibly meet the needs of children, both inside the classroom and outside of school. CHAPTER ONE * THE GLOBAL CHALLENGE OF EDUCATION FOR ALL 35 CHAPTER The Global Scorecard: X Progress since Jomtien Shifting from primary enrollments to completion rates as the basic measure of MDG progress may seem like moving back the goal post in a game already lost. Universal primary completion is undeniably a more challenging goal than uni- versal primary enrollment. Of 155 developing countries, about half have already built sufficient schools and places to educate 100 percent of their pri- mary school-aged children. But only 37 of those countries today retain 100 per- cent of children in school through primary graduation. Unlike universal access, universal completion cannot be achieved without ensuring schooling quality, students' learning progress, and household demand for education-all of which are interlinked. Nonetheless, countries from Brazil and Nicaragua in Latin America to Cam- bodia in East Asia to South Africa, Guinea, and The Gambia in Africa have proven over the 1990s that dramatic progress in increasing primary completion rates is possible-provided that political will is strong, effective reforms are adopted, and international support is adequate. Improvements of 20 percentage points in the primary completion rate in less than a decade-or more than 2 percentage points increase per year-have been registered in these countries and a number of others (annex tables B.3 and B.4). On the other hand, progress is not automatic, nor is sustained status assured. In a troubling number of countries, primary completion rates have slid backwards since Jomtien. Thirteen middle-income and 15 low-income countries saw their completion rates stagnate or decline over the 1990s (annex tables B.5 and B.6). The case of Afghanistan (which dropped from an already low 22 percent in 1990 to an estimated 8 percent in 1999) is obvious and dramatic. But several middle- income Gulf states, Latin American countries such as Venezuela and Belize, and African countries such as Zambia, Republic of Congo, Cameroon, Kenya, and Madagascar have also lost ground. Later in this chapter we explore global, regional, and country-by-country trends in primary completion rates more fully. But first, we turn to the definition of the primary completion rate and associated methodological issues: How it is calculated? How reliable is it as a statistic? And why hasn't it been compiled before? WHAT IS THE PRIMARY COMPLETION RATE? The primary completion rate (PCR) is a flow measure of the annual output of the primary education system. It is calculated as the total number of students success- fully completing (or graduating from) the last year of primary school in a given 37 year, divided by the total number of children of official graduation age in the pop- ulation. It is an application of the OECD methodology for measuring secondary school completion rates to the primary level. It should be emphasized that the primary completion rate is not the same as the "cohort survival rate" estimated by UNESCO. That indicator measures the survival to grade 5 among those children who enroll in school. But this has the important limitation of not reflecting the sometimes large share of children, espe- cially in low-income countries, who do not have access to primary school. The PCR measures the proportion of all children of official graduation age who complete primary school in a given year. As the numerator in the primary comple- tion rate counts all children completing the final grade of primary school, it will typically include overage children who either started school late or have repeated one or more grades of primary school, but are now graduating successfully. In coun- tries where there is some repetition yet the dropout rate is low, the primary comple- tion rate can, in a particular year, exceed 100 percent. However, completion rates consistently above 100 percent can be assumed to reflect data weaknesses, in either reported enrollment statistics or age-specific population estimates. The primary completion rate focuses on capturing the share of children who ever complete the cycle; it is not a measure of "on-time" primary completion. An on-time completion rate could also be calculated, by netting overage children out of the numerator. But data for this are not readily available. More fundamentally, though, the philosophy of this study is that the key number of policy interest to countries from a human capital standpoint is the share of children who eventually obtain a primary-level education. It should be understood that even though overage children may appear in the numerator of the primary completion rate, they appear only once. Since children are counted only when they actually graduate from, or complete, the cycle, steady- state monitoring of the completion rate will give an accurate picture of trends over time in the average share of students in a population cohort who complete primary school. As school system flow distortions eventually decline, the share of overage or underage children will be reduced as well. However, as noted in chapter 1, they may never disappear altogether. In education systems that are not in a steady state-that is, where either the size of the school-age population is changing or the coverage of the education sys- tem is expanding rapidly-the current primary completion rate may not be a good reflection of the likely future completion rate for the cohort now entering primary school. But if the primary completion rate is monitored over time it will reflect these trends and give a good sense of progress toward the MDG goal of universal primary completion. As such, it may be used to set meaningful targets. DATA SOURCES AND METHODOLOGICAL ISSUES Primary completion rates are calculated from the same two basic data sources used to compute gross and net enrollment ratios: (a) enrollment data from national ministries of education, and (b) United Nations population data. 38 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD The grade-specific enrollment data required for the primary completion rate are collected in all countries and published by the UNESCO Institute for Statis- tics. We used this database to calculate a baseline (1990) value of the PCR for all countries. However, since developing the most up-to-date picture possible of where countries currently stand in relation to the MDG target was a priority of this exer- cise, we collected enrollment data for the most recent year possible directly from national education ministries, through World Bank task teams. In most cases, that meant the year 2000. When it was impossible to obtain more recent data, we relied on published UNESCO data, most often for 1997. In some cases, the only avail- able data were for even earlier years. The primary completion rate is both conceptually and practically a fairly straightforward education statistic. But there are some methodological and data issues. The first is the differing length of the primary cycle across countries (which also affects gross and net enrollment ratios). For this study, primary education is defined as UNESCO's ISCED (International Standard Classification of Educa- tion) Level 1: "the beginning of systematic apprenticeship of reading, writing and mathematics; the start of compulsory education; primary education; first stage of basic education." As UNESCO notes, this stage in most countries is covered in a five- or six- year cycle. About 45 percent of countries have six-year, and another 13 percent of Table 2.1 ............................................................................... Length of the Primary Cycle in 155 Developing Countries, circa 2000 Percentage of Years in Primary Number of Developing Typical Cycle Countries Countries Regions/Countries 3 2 . a . Russia, Armenia 4 26 18 Europe and Central Asia Africa 5 20 13 South Asia East Asia and the Pacific 6 70 45 Africa East Asia and the Pacific Latin America and the Caribbean Middle East and North Africa 7 21 14 Africa Latin America and the Caribbean 8 12 8 Europe and Central Asia Africa 9 2 1 Libya, Rep. of Yemen 10 2 1 Jordan, West Bank/Gaza a. Less than 1 percent. Source: UNESCO Institute for Statistics. CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 39 countries have five-year, primary cycles. But in 24 percent of countries, the official primary cycle is longer. As countries develop economically and educationally, the length of compulsory education is typically extended to the next cycle, which UNESCO defines as Level 2: "lower secondary education" or the "second stage of basic education." However, in many countries this integrated cycle of Levels 1 and 2 of compulsory education is called primary education. In some countries this cycle is seven years, and in four countries it is nine or 10 years. Two of these four countries, the Republic of Yemen and Lebanon, have decided to consider the primary cycle as the first six years of schooling for EFA monitoring purposes. We have accordingly done so in this study. About 20 percent of countries have a shorter primary cycle, usually four years. Many lusophone and former Soviet Union countries follow this pattern. In 2000, Armenia adopted a three-year primary cycle. The Russian Federation recently extended its three-year cycle by an additional year, effective 2003. Our database calculates the primary completion rate based on the official cycle length in each country. Obviously, it is considerably easier to get 100 percent of children through three grades of primary school than through eight. But since gross enrollment and net enrollment ratios are also estimated on the basis of coun- tries' official cycle length, this ensures that the primary completion rates are consis- tent and therefore comparable with those series. A second issue arises from the fact that not all countries report the number of children completing primary school. Typically, this requires the collection of end- of-year enrollment data, and many low-income countries only report enrollments at the beginning of the year. For countries where actual primary graduates are reported, the primary completion rate estimate uses that number. In these cases, accounting for dropouts is not an issue, as students who drop out during the course of the year naturally do not appear in the end-of-year enrollment numbers. For countries that do not report end-of-year enrollments, we calculated a "proxy primary completion rate" defined as follows: Proxy primary completion rate = (the total number of students in the final year of primary school, minus repeaters) divided by (the total number of children of official graduation age in the population). The reasoning is that the repetition rate in the final year of primary schooling typically does not change dramatically from one year to the next, although steady improvement over time can occur. Thus, the share of children repeating the grade this year, having failed it the previous year, is a reasonable approximation of the share of students who are likely to fail the grade this year. Subtracting this number from the total number of children enrolled in the grade at the beginning of the year gives a reasonable approximation of the number of children who will successfully complete the grade and graduate from primary school in the current year. When estimating a proxy primary completion rate, ideally one would also make an adjustment for students who drop out during the year. Where estimates of dropout existed, they were used. However, data on dropout rates in the final year of schooling typically were not available, and thus most of the proxy primary comple- tion rates present an overstatement of the true primary completion rate, and 40 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD should be taken as an upper-bound estimate. For a few countries in the sample, even recent data on repetition rates could not be obtained; for those countries, the proxy PCR further overstates the true primary completion rate. The third and final issue concerns the population data used in the denomi- nator of the primary completion rate. All population data used in this report were taken from the United Nations/World Bank population database used for all World Bank work, including the calculation of the World Development Indicators. This series is compiled for all countries from national census data, with regular review and adjustments of national data by a panel of international demographers using emerging demographic and household survey data, med- ical registries, and other agreed sources. This data series includes total popula- tion estimates, population estimates by age and gender, and projections through 2015. The age-specific population estimates for boys and girls needed for calcu- lating the primary completion rate are readily available in this database. This data series is the best internationally comparable set of population esti- mates available. However, there is a somewhat higher risk of error in the age- specific data used for the primary completion rate than in the overall popula- tion estimates, which may have discouraged the systematic estimation of PCRs in the past. And for countries that have not carried out national censuses for some time or have experienced war, mass migration, or other major disloca- tions, these estimates may not be very accurate even though they are the best available. In sum, strenuous efforts have been made to develop an internationally consis- tent set of estimates of the primary completion rate in 155 developing countries. But this first set of primary completion rate and proxy estimates needs to be regarded as just that-an initial data set that can be improved greatly in terms of both robustness and timeliness if national governments and international partners work together to refine them. Even a cursory review will point up many gaps in the data, particularly for small countries, earlier years, and gender breakdowns, and obvious anomalies and estimates that are suspect. There is work to be done to encourage countries to collect end-of-year data on the number of graduates, to allow true PCRs to be estimated for all countries. There is scope for more systematic quality assurance and maintenance of this database through collaboration with the UNESCO Insti- tute for Statistics. And there is the prospect of improved population data becoming available soon, as more countries complete the detailed age-specific analysis of 2000 censuses or carry out new household surveys. Thus, the PCR and proxy estimates in this report are only a point of depar- ture. But they represent the most direct effort to date to measure progress toward the MDG target of universal primary completion and to provide a basis for future monitoring. Within the limitations of available data, the completion rates pre- sented here reasonably capture the reality of primary education system coverage and student attainment in many of the 155 developing countries measured. With a collaborative international effort to improve the quality of this database, it could be an increasingly valuable tool. CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 41 ADVANCES IN PRIMARY COMPLETION DURING THE 1 990S Since Jomtien in 1990, the global average completion rate for the developing world has improved only from 72 to 77 percent (see Annex figure B.1). Underly- ing this average, as figure 2.1 shows, is substantial variance across regions in both the distance from UPC and the progress made over the 1990s. Sub-Saharan Africa has the lowest completion rate by far, with barely half of all school-age children completing primary school; it is followed by South Asia, with a regional average completion rate of about 70 percent. The Middle East and North Africa showed a disturbing pattern of stagnation over the 1990s, with the average com- pletion rate remaining around 74 percent. The Europe and Central Asia region (92 percent) is closest to the goal, followed by Latin America and the Caribbean (85 percent) and East Asia and the Pacific (84 percent). Table 2.2 presents changes in median completion rates, as well as regional means, because in many cases the means are skewed by a few extremely high- or low-performing countries, some of whose data are questionable. In Africa and Latin America, for example, the increase of only 3 percentage points in the regions' median completion rates over the period indicates that increases in their mean completion rates were driven by high numbers for a few countries. On the other hand, in South Asia the median completion rate improved from 50 to 67 percent FIGURE 2.1 Primary Completion Progress by Region, 1990-2000, and Projected Trends (Country-Weighted) Primary completion rate (percent) 100 - ECA 90 -, EAP 80 - 1990 2000 2010 2020 2030 2040 2050 - Primary completion progress Projected trend Source: Annex figure B.3. 42 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2.2 ............................................................................. Primary Completion Progress by Region, 1990-2000, Country-Weighted ... 1990 MOST RECENT YEARa Region Mean Median Range Mean Median Range Africa 49 42 11-135 55 45 19-117 East Asia and 78 89 39-99 84 90 54-108 the Pacific Europe and 89 90 67-100 92 93 77-109 Central Asia LatinAAmericaand 76 86 28-112 85 89 40-110 the Caribbean Middle East and 75 75 32-102 74 76 30-104 North Africa South Asia 64 50 22-111 70 67 8-112 All developing 72 81 11-135 77 83 8-117 countries IDA-eligible 50 45 11-112 62 64 8-117 countries IBRD-eligible 84 89 43-135 87 92 44-111 countries a. Usually 1999/2000. over the decade-an impressive trend that is largely masked in the comparison of means by the dramatic decline in Afghanistan. Both the means and medians for low-income countries improved more than those for middle-income countries. On a population-weighted basis (table 2.3) the global progress is more encour- aging, with the global completion rate increasing from 73 to 81 percent over the decade. The population-weighted average is dramatically more positive for East Asia (97 percent compared with 84 percent on a country-weighted basis), reflect- ing the high reported completion rate in China. For the Middle East and North Africa, the population-weighted average of 83 percent is also significandy higher than the country-weighted average, influenced by Egypt's weight. For the remain- ing regions there is relatively litde difference. But in Sub-Saharan Africa, the aver- age completion rate of 51 percent on a population-weighted basis is even more discouraging than the country-weighted mean. Globally, even though completion rates for girls improved more than those for boys over the decade, girls' average completion (76 percent) continues to lag that of boys (85 percent). Every region showed a significant increase in girls' completion rates, with the 14 percentage point improvement in Latin American countries the most impressive change. On the other hand, it is sobering that the population- weighted completion rate for boys actually declined in the Africa region over the decade, was stagnant in Europe and Central Asia, and was virtually stagnant in East Asia and in the Middle East and North Africa. Only the South Asia and Latin CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 43 Table 2.3 I..................................................................................... Primary Completion Progress by Region, 1990-2000, Population-Weighted 1 990 MOST RECENT YEARa Region Girls Boys Both Girls Boys Both Africa 43 57 50 46 56 51 East Asia and 92 97 96 98 98 97 the Pacific Europe and 85 95 90 93 95 93 Central Asia Latin America and 71 64 69 85 81 83 the Caribbean Middle East and 71 84 78 78 86 83 North Africa South Asia 59 77 68 63 84 74 All developing 65 79 73 76 85 81 countries a. Usually 1999/2000. America regions showed a significant increase in the share of boys completing pri- mary school. Latin America's increase of 17 percentage points over the decade was even stronger than the increase for girls in that region. AFRICA Data for African countries are presented in table 2.4, sorted from highest to lowest completion rate in the most recent year available. The first five countries listed have achieved universal primary completion according to available data and our working definition of universal completion as 95 percent or higher. South Africa has reached the goal since Jomtien. By contrast, Zambia, which had essentially achieved universal primary completion in 1990, has since then suffered a substan- tial decline to 83 percent (in 1995). Both in 1990 and today, Sub-Saharan Africa is the region with the lowest average completion rate, at 55 percent. More than half the countries in the region for which data are available are less than halfway to the MDG target. The encouraging news, however, is that a substantial number of African countries have been able to increase completion rates over the 1990s, and some-such as Guinea, Eritrea, Mali, and Mauritania-have made truly impressive progress from very low starting levels. The Gambia, starting from a slightly higher base, has made even faster progress, with the primary completion rate increasing by more than 3 percentage points per year from 1991 to 2000. Malawi's rate of increase in the first half of the decade was extremely high, 4 percentage points per year, but it is doubtful that it has been sustained. On the other hand, Uganda's improvement (see box 2.1) may accelerate further over the next two years, as the wave of children who entered school following the elimination of fees in 1996 make their way to graduation from the seven-year system. 44 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD > m l Primary Completion Progress in Uganda In 1990, only an estimated 39 percent of Ugandan children completed the seven-grade pri- mary cycle. By 2001 that share was 65 percent and growing rapidly, as the result of a bold reform in 1996 that eliminated primary school tuition fees for up to four children per fam- ily. President Museveni's dramatic action removed a key obstacle for poor families and emphasized the importance of education. The impact on demand for primary schooling was immediate and tremendous. In 1997 total primary school enrollments jumped from 3.4 million to 5.7 million children, with the greatest increases coming among girls and the poorest children. By 1999 the wealth bias that had characterized access to primary education was all but eliminated, and by 2000 there was vir- tually no gap between male and female net enrollment ratios (89.3 percent vs. 88.8 percent). The primary system has had to scramble to deal with the swell in enrollments, however. Pupil-teacher ratios shot up from 40:1 to 60:1 by 1999, and unqualified teachers had to be deployed to many areas, until the Ministry of Education could ramp up the production of addi- tional trained teachers. Input ratios for textbooks and materials also deteriorated. Although the government acted quickly to reallocate spending to primary education and to mobilize additional donor support, the loss of tuition income at the school level and the huge influx of new students led to palpable declines in schooling quality. With any large enrollment expansion, a decline in average student learning outcomes can be expected. But in Uganda, the drop has been precipi- tous: between 1996 and 1999, the share of students receiving a satisfactory score fell from 48 per- cent to 31 percent in mathematics, and from 92 percent to 56 percent on the English oral test. Uganda's experience-and the earlier experience of Malawi with a similar elimination of tuition fees-provides strong evidence that schooling demand in low-income countries is more elastic than previously estimated. But Malawi, where schooling quality has continued to erode Uganda: Primary GER and PCR, 1990-2001 (percent) 18 71 74 61 65 39 40 1990 1992 1994 1996 1998 2000 - Gross enrollment ratio (GER) L - Primary completion rate (PCR) - - -- Projected GER Projected (PCR) CHAPTER TWO o THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 45 since the elimination of fees in 1995, has already provided sobering evidence that cnroUlmcnt gains, and especially completion rate progress, will not persist if schooling quality does not mcet minimum standards. For poor parents in particular, the opportunity costs of children's school attendance are high, and parents will not keep children in primary school dirough to completiol unless they perceive that school conditions are inimially adequate, the curriculum content is rJl- evant, and students are learning. The fiscal priority Uganda has givcn to primaty education and the ministry's systematic actions since 1997 to irnprove quality have attracted. substantial donor support, but there is still considerable progrcss to be made. Ugana'as cxperience shows, however, that rapid progress in primnary completion is possible, with bold actions to climinate demand-side constraints where they exist and strong complementary attention to schooling quality. Table 2.4 Africa: Changes in Primary Completion Rates during the 1990s . PRIMARY COMPLE'rION RATE * 1990 MOST RECENT YEAR * Eligible Years . for in IDA/IBRD primary Country lending cycle Girls Boys Both Year Girls Boys Both Year Cape Verde IDA 6 119 115 117 1997 Zimbabwe IDA 7 94 100 97 1990 111 116 113 1997 Mauritius IBRD 6 135 136 135 1990 108 115 111 1997 Botswana IBRD 7 126 102 114 1990 107 96 102 1996 SouthAfrica IBRD 7 81 72 76 1990 100 95 98 1995 Namibia IBRD 7 80 59 70 1990 94 86 90 1997 Zambia IDA 7 84 110 97 1988 75 90 83 1995 Swaziland IBRD 7 71 1990 85 78 81 1997 SaoTome and IDA 4 * 84 2001 Principe Gabon IBRD 6 77 66 71 1991 80 79 80 1995 Gambia, The IDA 6 35 45 40 1991 70 2000 Lesotho IDA 7 82 45 64 1990 83 55 69 1996 Nigeria IDA 6 62 82 72 1990 61 73 67 2000 Uganda IDA 7 30 49 39 '1990*-.- 65 .2001 Ghana IDA 6 54 71 63 1990 . - 64 1999 Togo *IDA 6 26 55 41 1990 52 73 63 1999 Tanzania IDA . 7 45 46 46 .1989. 60 58 59 .1997 Kenya .IDA 8 57 69 63 1990 57 58 58 1995 4,O ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2.4 (continued) Africa: Changes in Primary Completion Rates during the 1 990s PRIMARY COMPLETION RATE __ _ __ _ __ __ _ 1990 MOST RECENT YEAR Eligible Years . * for in IDABRD primary Country lending cycle. Girls Boys Both Year Girls Boys Both Year Malawi IDA * 8 22 38 30 1990 40 61 50 .1995 Mauritania . IDA 6 26 41 34 1990 43 48 46 1998 Equatorial Guinea IBRD . 5 - . 43 48 46 1993 Congo, Rep. of IDA 6 55 68 61 1990 60 28 44 *2000 Cameroon IDA 6 52 61 57 1990 39 . 46 43 1999 Burundi IDA 6 43 49 46 1990 - - 43 1998 Senegal IDA 6 35 56 * 45 1989 34 48 41 2000 C6te d'Ivoire IDA 6 32 55 44 1990 33 48 40 1999 Congo, Dem. IDA 6 35 60 48 1990 34 45 40 .2 0 0 0 Rep. of Benin IDA 6 15 31 23 119901 30 47 39 1998 Mozambique IDA 5 23 36 30 1990 22 50 36 1998 Eritrea IDA 5 20 23 22 1991 31 40 35 1999 Sudan IDA 8 - 33 38 35 1996 Guinea IDA 6 9 24 16 1990 24 44 34 2000 Comoros *IDA 6 *32 38 35 1991 34 32 33 1993 SierraLeone IDA 7 - - - - 30 36 32 *2000 Guinea-Bissau . IDA * 6 12 21 16 1988 24 *40 *31 *2000 Madagascar .IDA 5 35 33 34 1990 26 26 26 1998 Burkina Faso * IDA . 6 . 14 24 . 19 *1990 20 * 30 * 25 *1998 Rwanda .IDA 6a 35 33 34 1990 27 30 28 *2000 Ethiopia *IDA 6 18 25 22 1990 12 36 24 1999 Mali *IDA 6 9 14 1 1 1990 18 29 23 1998 Niger . IDA 6 13 23 18 1990 15 23 20 1998 Central African * IDA 6 19 37 * 28 1990 * - * 19 12 0 0 0 Repub:ic Chad IDA 6 7 31 19 1990 9 29 19 2000 Angola IDA 4 35 42 39 1990 Liberia IDA 6 Seychelles IBRD 6 Somalia IDA 8 -Not available. a. Rwanda 1990 data are for a seven-grade primary cycle; 2000 data are for the new six-grade cycle. CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 47 These examples provide clear evidence that a path of rapid and sustained improvement in primary completion rates can be achieved by countries no matter what their starting level of education system coverage, and no matter what their level of income per capita. EAST ASIA AND THE PACIFIC Five countries in the East Asia and Pacific region have achieved universal primary completion. But in general, the stellar pace of education progress seen in East Asian countries over the 1970s and 1980s has not been sustained since Jomtien. With the exception of Cambodia, which has registered tremendous improvement from a low base in the last several years, and to a lesser extent the Lao People's Democratic Republic, primary completion rates in the East Asia and Pacific region have largely stagnated or declined over the past decade (table 2.5). According to these estimates, none of the East Asian and Pacific countries that had yet to achieve universal pri- mary completion as of Jomtien has done so since, except possibly Vietnam. EUROPE AND CENTRAL ASIA European and Central Asian developing countries inherited an extensive education infrastructure, and at the beginning of the transition in 1990 were characterized by virtually universal primary enrollments and the highest average primary comple- tion rate of any region-89 percent. The most recent data (table 2.6) still show Europe and Central Asia (ECA) as the region closest to universal primary comple- tion. Eleven of the 23 countries in the region for which data are available have achieved the goal and it can be considered within reach of all the ECA countries, most of which are also helped by a relatively short official primary cycle. However, the region faces very significant quality issues, increasing evidence of demand constraints, and the challenge of adapting educational content and goals to the needs of more open societies and market economies. With the severe eco- nomic dedines that accompanied transition, along with the social, demographic, and political upheaval the region has experienced, many countries have had great difficulty maintaining the inherited education infrastructure, let alone improving the quality of education delivered. Many of the former Soviet republics relied heav- ily on subsidies from Moscow to develop and maintain their education systems; the withdrawal of these, plus the sharp economic declines of the 1990s, have placed education systems in the region under stress. The dislocations of the 1990s have had a particularly strong impact on the poorest countries, in Central Asia, the Caucuses, and the Balkans, all of which are now struggling to maintain a basic level of education services in the face of public finance constraints and institutional weaknesses. Very high officially reported enrollments and completion rates in several Central Asian countries-especially Kazakhstan, Turkmenistan, Kyrgyz Republic, and Uzbekistan-are not included in our database, as they are not corroborated by household survey data which reveal troubling dedines in schooling attendance. Population data for many countries in the region are also problematic, in some cases based on censuses from 1980. Available data show Armenia, Georgia, Moldova, and Tajikistan as the ECA countries farthest from the MDG, and of these, only Moldova appears to have 48 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2.5 .......................................................................................I....... East Asia and the Pacific: Changes in Primary Completion Rates during the 1 990s * PRIMARY COMPLETION RATE * * 1990 MOST RECENT YEAR Eligible for Years in IDA/IBRO Primary . Country Lending Cycle Girls Boys Both Year G Girls Boys Both Year China IBRD 5 95 107 99 1990 106 111 108 1996 Vietnam IDA 5 - 98 104 101 2001 Samoa IDA 8 92 105 99 1997 Korea, Rep. of IBRD 6 96 96 96 1990 98 95 96 2000 Fiji IBRD 6 - - - 93 97 95 1992 Philippines IBRD 6 91 88 89 1989 92 1996 Indonesia IDA 6 93 93 92 1990 92 90 91 2000 Thailand IBRD 6 90 95 93 1990 90 2000 Malaysia IBRD 6 91 91 1990 90 89 90 1994 Vanuaru IDA 6 *90 89 89 1990 81 92 86 1992 Mongolia IDA 4 82 1998 Cambodia IDA 6 32 46 39 1997 70 2001 Lao PDR IDA 5 56 1995 64 73 69 2000 Solomon Islands IDA 6 59 70 65 1990 54 77 66 1994 PapuaNewGuinea IBRD 6 51 55 53 1990 53 64 59 1995 Timor-Leste, Dem. Rep. IDA 6 - - - - 53 55 54 .2001 Kiribati IDA . 7 Korea, DPR n. a. 4 Marshall Islands IBRD 6 . Micronesia, Fed. States of. IBRD 6 - - - - - - - - Myanmar * IDA . 5 Palau IBRD 8 - - - - - - Tonga IDA . 6 - - - -.- - n. a. Not applicable. - Not available. registered progress over the 1990s. Albania's completion rate declined over the 1990s. Overall, the region is characterized by a growing gap in performance between states that are developing rapidly, such as the Baltic and Central European countries-with the Czech Republic, Latvia, Lithuania, and Hungary all showing completion rate progress-and those such as Albania, Armenia, Georgia, and the Central Asian countries, whose economic problems are increasingly reflected in the education sector. CHAPTER TWO * THE GLOBAL SCORECARD; PROGRESS SINCE JOMTIEN 49 Table 2.6 Europe and Central Asia: Changes in Primary Completion Rates during the 1990s PRIMARY COMPLETION RATE 1990 MOST RECENT YEAR Eligible Years for in IDA/IBRD Primary Country Lending Cycle Girls Boys Both Year Girls Boys Both Year Czech Republic IBRD 4 86 92 89 1992 107 110 109 1995 Hungary IBRD 4 90 97 93 1989 102 102 102 1995 Azerbaijan IDA 4 101 99 100 1998 Romania IBRD 4 91 100 96 1989 98 99 98 1996 Slovak Republic IBRD 4 92 100 96 1992 97 96 97 1996 Poland .IBRD . 8 98 102 100 1990. 97 96 96 1995 Serbia and IDA 8 68 * 77 72 1990* - . 96 *2000 Montenegro Russia IBRD 3 - - - - - - 96 .2001 Croatia *IBRD . 8 86 85 86 1992 96 95 96 2001 Lithuania .IBRD . 4 84 92 88 1992 94 97 95 1996 Ukraine * IBRD . 3 .94a . 2002 Belarus . IBRD 4 94 100 97 1992 * 92 * 95 93 1996 Turkey .IBRD . 5 82 99 90 1990 89 95 92 1994 Bulgaria *IBRD 4 87 93 90 *1990. 92 92 92 .1996 Macedonia, FYR IBRD 8 84 94 89 .1992. 87. 94 91 .1996 Albania : IDA 8 .92 102 97 .1990. 95 84 89 *1995 Bosnia and IDA 4 - *- - - 88 .1999 Herzegovina Estonia IBRD 6 91 95 93 1992 86 89 88 1995 Latvia IBRD 4 75 77 76 1992 84 87 86 1996 Armenia IDA 4 95 70 82 1996 Georgia IDA 4 82 1998 Moldova IDA 4 65 68 67 1991 79 1999 Tajikistan . IDA 4 . - - - 75 80 77 *1996 Kazakhstan .IBRD . 4 - - - - - - - Kyrgyz Republic IDA 4 Slovenia .IBRD 4 93 .106 99 1992. - . - . - . - Turkmenistan . IBRD 4 Uzbekistan IDA 4 - - - - -Not available. a. Staff estimate. 50 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD LATIN AMERICA AND THE CARIBBEAN Twelve of 30 countries in Latin America and the Caribbean have achieved UPC and three more countries are on the cusp of doing so. The remaining 15 countries have varied widely in their progress since Jomtien (table 2.7). Two countries, Nicaragua and Brazil, have raised primary completion rates by more than 20 percentage points over the decade, and for Brazil especially, with an eight-year primary system, this is truly impressive progress (box 2.2). Bolivia has similarly increased completion through an eight-grade primary system from 55 to 72 percent over the decade. Only slightly less dramatic is the progress in El Salvador, Costa Rica, and Peru, all of which started with a higher base and increased their completion rates by more than 15 percentage points over the decade. Colombia has also shown signifi- cant improvement. However, a few countries in Latin America have shown the reverse trend. Countries that appeared to have universal primary completion within close reach, such as Venezuela, Guyana, and Belize, have seen completion rates decline by as much as 8 percentage points over the decade. Ecuador and Honduras show basi- cally stagnating trends. Table 2.7 ....................................................................................................... Latin America and the Caribbean: Changes in Primary Completion Rates during the 1990s PRIMARY COMPLETION RATE 1990 * MOST RECENT YEAR Eligible Years : for in IDA/IBRD Primary Country Lending Cycle Girls Boys Both Year Girls . Boys Both Year St. Kitts and Nevis IBRD 6 j - . 104 115 110 2001 Grenada IDA . 7 - .- .- .104 107 106 .2001 St. Lucia IDA 7 106 117 112 1990 104 109 106 2001 Dominica . IDA 6 - .- - i107 99 103 .2000 Mexico IBRD 6 89 88 89 1990 93 85 100 2000 Antigua and IBRD 7 95-100a 2000 Barbuda Cuba n. a. 6 95 jQQa. 2001 Chile IBRD 6 97 92 94 1990 99 2000 Uruguay IBRD 6 98 93 95 1990 101 95 98 2000 Peru . IBRD 6 . . 85 *1988* . 98 *2000 Ecuador IBRD 6 98 99 99 1992 96 96 96 1999 Argentina IBRD 7 95 91 96 2000 Jamaica IBRD 6 94 87 90 1990 98 91 94 2000 Trinidad and IBRD 5 93 95 94 1990 94 94 94 2000 Tobago CHAPTER TWO * THE GLOBAL SCORECARD PROGRESS SINCE JOMTIEN Si Table 2.7 (continued) Latin America and the Caribbean: Changes in Primary Completion Rates during the 1 990s PRIMARY COMPLETION RATE 1990 * MOST RECENT YEAR . Eligible Years for in IDA/IBRO Primary Country Lending Cycle Girls Boys Both Year Girls Boys Both : Year Panama IBRD 6 *87 88 87 71990 94 *2000 CostaRica IBRD 6 74 72 73 1990 - - 89 2000 Guyana *IDA 6 93 90 92 1990 95 84 89 *2000 Colombia IBRD 5 83 61 72 1990 85 2000 : : : : : :1990::.: St. Vincent IDA 6 84 85 84 2001 Belize IBRD 6 90 89 90 1990 83 81 82 1999 El Salvador IBRD 6 62 59 61 1989 80 2000 Paraguay IBRD 6 65 65 65 1990 78 2000 Venezuela IBRD 5 96 87 91 1990 79 77 78 1999 Bolivia IDA 8 55 1990 72 2000 Brazil IBRD 8 54 42 48 1990 72 1999 Honduras IDA 6 63 68 66 1991 67 2000 Nicaragua IDA 6 50 40 45 1990 70 61 65 2000 Dominican IBRD 8 - - - 67 56 62 *2000 Republic Guatemala .IBRD 6 39 46 43 1991 52 2000 Haiti IDA 6 27 29 28 1990 40 1997 n.a. Not applicable. - Not available. a. Staff estimate. 111I0 Primary Completion Progress in Brazil In the space of one decade, Brazil has increased the share of children who complete primary school from 48 to 72 percent, one of the 10 fastest rates of improvement observed in our global sample. How Brazil did it provides a good picture of how sensitive primary comple- tion rates are to schooling quality, particularly for the poor. Before 1990, entry to primary school was practically universal in Brazil (recall the schooling profile pictured in figure 1.3), but less than half of all children completed the eight-year cycle. Dropout was worst in the poor northeast region, and especially in rural schools. But after 1995, Brazilian education policy under Minister Paulo Renato Souza focused strongly on improving the quality of primary education overall, and especially for the poor. 52 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD ( w°S1 72continued First, Brazil passed a major constitutional change in the distribution of fiscal resources for education. The FUNDEF reform established a yearly per-pupil spending floor to ensure a minimum standard of education for all children. National tax revenues began to cross- subsidize the states and municipalities least able to mobilize their own fiscal resources for education, thereby reducing the gulf in teacher salaries and school quality between richer and poorer regions. Innovative programs also channeled increased resources directly to schools, so they could implement their own development plans. Second, the federal ministry strengthened its role in norm setting and quality assur- ance. It established the first national student assessment system and developed new national curriculum guidelines that stressed problem solving, independent learning, and critical thinking. A national commission set new quality standards for textbooks and learning materials and the ministry made on-time delivery of adequate learning materials to schools a visible national priority. New legislation was passed to sort out the confused and overlapping roles of the federal, state, and municipal governments in primary educa- tion, with the federal ministry for the first time assuming clear responsibility for guaran- teeing equity and quality. Third, heavy emphasis was placed on upgrading teacher quality-and teacher motiva- tion. Higher qualifications were set for teacher certification, and the hiring of teachers with- out competitive examination, part of old-style patronage politics, was disallowed. A federally funded in-service teacher training program, using cost-effective distance delivery and high- quality materials, helped raise the share of primary teachers with a complete secondary edu- cation. Teacher salaries and pensions were increased, and civil service reforms began to allow for dismissal based on performance. Brazil: Primary Education Access by Income Quintile, 1992-2001 Pecentage of children enrolled 100- 99 95- - 979 93L-- 94 90 - 87 85 83 80 - 75- l75 7 0 - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1992 2001 -4- Richest quintile -c Fourth quintile -4- Third quintile S Second quintile c Poorest quintile Source: Ministry of Education and Culture, Brazil. CHAPTER TWO o THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 53 Fourth, innovative demand-side programs such as bo/sa escola:, which channcls subsis- tence grants to low-income mothers whose children stay in school, and home visits from health agents who also check on school attendance helped send the message to very poor families that "every child belongs in school" and "Lvery child can succeed." In short, adequate and more equitable financing and a concerted program of quality improvements, efficiency enhancements, and demard-sidc intervcntions have combined in Brazil to raise the primary completion rate from 48 to 72 percent in the space of one decade. As the figure shows, the bulk of the progress has come from improvement in the quantity and quality of services dclivered to children in the poorest income quintilcs. Brazil's experi- ence shows that with political commitment and comprehensive strategies, simultaneous rapid progress in educational quality and equity is possible. Nonetheless, the major story in the region is the significant number of countries that have progressed rapidly, some from a very low base, suggesting that this region may have many lessons to share. Estimates for this region are also more reliable than for the other regions, as most Latin American and Caribbean coun- tries report actual primary completion, and educational statistics are reasonably good. In terms of gender equity, in Guatemala, El Salvador, St. Lucia, and Bolivia completion rates for girls lag those for boys, but an equally common pattern in Latin America is the opposite: boys' completion rates are lower than girls', some- times significantly so, as in Uruguay, Jamaica, Mexico, and Nicaragua. MIDDLE EAST AND WORTH AFRICA Only two of the 19 countries in the Middle East and North Africa have achieved universal primary completion-Jordan and the Arab Republic of Egypt, whose sta- tistics indicate a dramatic 22 percentage point improvement over the decade. In general, data for this region are of questionable quality and most of the completion rates are proxy rates. The available data show substantial progress in Tunisia, which increased from 75 to 91 percent, as well as improvements in Kuwait, Algeria, Oman, Saudi Arabia, and Morocco, the last from a low base. Although the average completion rate for the region as a whole changed very little over the decade, this region more than any other is marked by great underlying variation at the country level (table 2.8). Fully half of the countries for which two data points exist show declining completion rates over the period. Bahrain, which in 1990 reported 100 percent completion, has apparently suf- fered a decline to 91 percent since then. Although the data showing a tremendous drop in Qatar must be considered questionable, completion rates in Iraq, Syria, the Islamic Republic of Iran, and Djibouti all appear to have fallen. Gender disparities in completion rates are evident in Morocco (47 percent completion for girls, 63 percent for boys), in Egypt (92 percent for girls, 104 per- cent for boys), in Djibouti (24 percent for girls, 36 percent for boys), and dramat- ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2.8 Middle East and North Africa: Changes in Primary Completion Rates during the 199Os PRIMARY COMPLETION RATE - 1990 MOST RECENT YEAR Eligible Years for in IDA/lBRD Primary Country Lending Cycle Girls Boys Both Year Girls Boys Both Year Jordan IBRD 6 102 101 102 1990 106 102 104 2000 Egypt, Arab Rep. IBRD 5 70 84 77 1990 92 104 99 1996 Iran, Islamic Rep. IBRD 5 88 101 94 1990 89 95 92 1996 Bahrain n.a. 6 101 100 101 1990 99 85 91 1996 Tunisia IBRD 6 g 70 80 75 1990 90 93 91 1996 Algeria IBRD 6 76 89 82 1990 88 93 91 1996 Syrian Arab Rep. g IBRD 6 92 103 98 1990 86 95 90 1996 United Arab n.a. 6 98 90 94 1990 86 76 80 1996 Emirates Oman . n.a. 6 63 70 67 1989 76 76 76 1996 Kuwait n.a. 4 55 57 56 1991 71 69 70 1996 Lebanon IBRD 5 70 1996 Saudi Arabia n.a. 6 56 64 60 1990 69 68 69 1996 Yemen, Rep. IDA 6 38 77 58 2000 Iraq IBRD 6 57. 69 63 g1990i 52 63 57 .1995 Morocco gIBRDg 6 g35 58 47 1991 47 63 55 1996 Qatar :n.a. 6 g74 74 74 1990 43 45 44 1995 Djibouti IDA 6 24 40 32 .1990g 24: 36 30 g1999 Libya n.a. g 9 West Bank/Gaza IDA 10 . - - - - . - , - . - n.a. Not applicable. - Not available. ically so in the Republic of Yemen (38 percent for girls, 77 percent for boys). Tunisia, on the other hand, has made clear progress in narrowing the gender gap over the decade, and Egypt and Morocco also appear to have made some progress. In a troubling number of cases in this region, however, gender parity has improved only because boys' completion rates have fallen. SOUTH ASIA The average primary completion rate for the South Asia region (70 percent) is the second lowest in the world, above only that of Africa. Even the higher population- weighted average of 74 percent still means that only three out of every four children CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 55 Table 2.9 South Asia: Changes in Primary Completion Rates during the 1990s * PRIMARY COMPLETION RATE e 1990 MOST RECENT YEAR Eligible Years for in IDA/IBRD Primary Country Lending Cycle Girls Boys Both Year Girls Boys * Both Year Maldives IDA 5 111 111 111 1992 110 113 112 1993 SriLanka IDA 5 94 106 100 1990 114 108 111 2001 India IDA 5 61 78 70 1992 63 88 76 1999 Bangladesh IDA 5 47 54 50 1990 72 68 70 2000 Nepal IDA 5 29 67 49 1988 58 70 65 2000 Bhutan IDA 7 59 2001 Pakistan IDA 5 30 57 44 1989 59 2000 Afganistan *IDA * 6 14 29 22 1989 . 15 8 1999 -Not available. in the region complete a primary education, which for most of the countries is only a five-year system. Two countries in the region have achieved universal primary completion: Sri Lanka and the Maldives (table 2.9). Three others have made very impressive progress: Nepal, Bhutan, and Bangladesh. Bangladesh has also made exceptional progress in gender equity, with girls' completion rates now apparendy higher than boys'. The largest country in the region, India, shows encouraging signs that strong efforts at educational improvement since the mid-1990s, especially at the state and district levels, are beginning to produce results. However, this rate of progress will need to be accelerated if the country that is home to the largest number of children out of school globally is to meet the MDG. Trend analysis is not possible for Pakistan, but it is evident that the completion rate started the decade from a low base and with a substantial gender disparity and that progress has been minimal. The data also show the terrible erosion of primary education in Afghanistan during the 1990s, especially for girls. However, early reports for 2002 indicate a massive return of children to Afghan primary schools, with a very high share of them girls. Data for countries in this region are quite lim- ited and, for most countries, of poor quality. Inconsistencies between official enrollment data and household surveys are not uncommon. THE GLOBAL PROSPECTS FOR UNIVERSAL PRIMARY COMPLETION BY 2015 Figure 2.2 provides a graphic picture of global prospects for achieving the educa- tion MDG by 2015 on current trends. At the current rate of progress, by 2015 the 56 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD FIGURE 2.2 Global Progress in Primary Completion, 1990-2000 and Projected Trends (Country-Weighted) Primary completion rate (percent) 100- 90 - 60- 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG - Current trend Source: Annex figure B. 1. global primary completion rate will not exceed 83 percent. On a population- weighted basis, the world will come closer to the MDG, with about nine out of every 10 children globally completing primary school (see annex figure B.2). But as figures 2.3 and 2.4 show, these global averages substantially conceal the gulf that would exist across regions in 2015. Three regions-Europe and Central Asia, East Asia and the Pacific, and Latin America and the Caribbean-would come close to the goal, if not achieve it. Three other regions-South Asia, Middle East and North Africa, and Africa-would be left behind, and in the case of Africa, significantly behind. On current trends, in 2015 only 60 percent of all African children will complete a primary education. Ultimately, the MDG will not be attained unless every child in every country has the chance to complete primary school, and change will have to happen at the level of national education systems in order to reach the goal. Therefore, the focus of this analysis is the country-by-country prospects for reaching universal primary completion by 2015. Summary results of an exercise to project the country-by-country prospects for achieving the education MDG are shown in table 2.10. Among the world's 155 developing countries, the best available data indicate that 37 countries (including 11 low-income countries) have achieved universal primary completion. At the trend rates of progress registered over the 1990s, another 32 countries can also be expected to reach the goal. Even though, as the previous sections showed, primary completion rates clearly can go down as well as up, for the purposes of a baseline estimate we labeled as "on track" the countries whose past trends, if continued, would be sufficient to reach the goal. Thus, 69 countries (including 22 low-income ones) are likely to reach the goal by 2015. The other 86 countries, however, are at risk of not reaching the goal. Forty- three of these countries, labeled "off track," are countries that are short of the goal CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 52 FIGURE 2.3 Primary Completion Progress in Europe and Central Asia, East Asia and the Paciic, and Latin America and the Caribbean Regions, 1990-2015, Country-Weighted Primary completion rate (percent) 100 - _ 3 90 ECA 80 EAP LCR 70 - 60 - 1990 1995 2000 2005 2010 2015 - - Required trend to achieve MDG Current trend Source: Annex figure B.5. FIGURE 2.4, Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted Primary completion rate (percent) 100 - ."''" , 60- M N SAR 40 - 40 -lll 1990 1995 2000 2005 2010 2015 --°- Required trend to achieve MDG Current trend Source: Annex figure B.5. but could be expected to reach it with a reasonable acceleration of progress or reversal of a mild declining trend. But another 27 countries, considered "seriously off track," will not even reach 50 percent primary completion by 2015 on current trends. Of these 70 off-track and seriously off track countries, 51 are low-income countries. A final set of 16 countries has no available data, including countries such 58 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 2.10 ............................................................. Prospects for Universal Primary Completion by 2015 Low-income Middle-income All Developing Progress Rating Countries" Countriesb Countries On track 22 47 69 Achieved UPC 11 26 37 On track to achieve UPC by 2015 11 21 32 Off track 51 19 70 Off track to achieve UPC by 2015 28 15 43 Seriously off track 23 4 27 No data available 9 7 16 At risk, subtotal 60 26 86 Total 82 73 155 a. IDA-eligible and "blend" countries (eligible for IDA and IBRD lending), plus non-member low-income countries such as Cuba. b. IBRD-eligible plus non-member middle-income developing countries. as Somalia, Liberia, and Libya. We consider this subgroup also at risk, for a total of 86 countries at risk of not achieving the MDG. What would happen if all of the countries currently off track could accelerate their progress to achieve the trend improvement rate of 3 percentage points per year registered by the best-performing countries over the 1990s?1 Although this would mean achieving and sustaining historically high rates of progress, there are three reasons why this scenario might be possible. First, in the context of a national commitment to the MDGs, countries might reassess their past performance and focus on accelerating it. Second, there is a growing number of success cases over the 1990s from which to learn. Third, linear projections of historical rate improve- ment in some cases lead to implausible projections, particularly in the case of declining trends (Qatar and Afghanistan would have primary completion rates below zero in 2015). If all 70 of the off-track countries were able to increase their primary comple- tion rates from 2003 to 2015 at an average rate of 3 percentage points per year, the world would come much closer to meeting the MDG target, but it would still fall short. Under this accelerated improvement scenario, however, all 19 of the middle- income countries-whether off track or seriously off track-would reach the goal. Thirty-five of the at-risk low-income countries would also meet the MDG. This scenario should be considered an achievable goal. However, the 16 low-income countries that are furthest from the goal would need to achieve and sustain an even faster rate of progress in order to reach universal i. This was calculated by finding the median of the trend improvement rates for the 10 best-performing IBRD countries and the 12 best-performing IDA countries, which represented the entire subset of countries averaging more than 2 percentage points per year improvement over the 1990s. CHAPTER TWO * THE GLOBAL SCORECARD: PROGRESS SINCE JOMTIEN 59 primary completion by 2015. Thirteen of the countries would need to raise pri- mary completion rates by more than 4 percentage points per year; and three coun- tries (including Afghanistan) would require a sustained increase of more than 5 percentage points per year. The prospects for these countries-heavily concen- trated in Sub-Saharan Africa, many scarred by conflict-are sobering at best. As can be seen from annex tables B.3 and B.4, only six countries have regis- tered trend rates of improvement in the primary completion rate of more than 3 percentage points per year. In two of these cases, the trend was observed over less than a full decade, and in several cases, the data are somewhat questionable. In short, for the 16 countries furthest from the goal to reach it by 2015 will require completion rate progress at historically unprecedented rates. But there is some basis for hope that the trend rates of primary completion progress will increase in coming years. A good number of developing countries have achieved dramatic gains in primary completion over the past decade. Faster diffusion of their experiences and knowledge about reform strategies that work may help countries at all levels of educational development to accelerate progress. In the next chapter, we analyze key features of the primary education systems in these higher performing countries, and the lessons they hold for countries at risk. 60 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD What Will It Take to AchieveCHAPTER Universal Primary Completion by 2015? - Chapter 2 showed that while a number of developing countries have already achieved universal primary completion and others are on track to do so by 2015, as many as 86 developing countries are at risk of not reaching the MDG on cur- rent trends. Sixty of these are very low income, or IDA-eligible, countries; 26 are middle-income countries. Any global strategy for achieving the MDG must find ways to help these at-risk countries accelerate progress. A first step toward such a strategy is understanding what drives EFA success. Why have some countries achieved universal primary completion so much faster than others? Among those that haven't achieved it, why are some making more rapid progress? A growing body of international experience and research offers the potential for a deeper understanding of the determinants of EFA success. This chapter reviews and extends that work, based on analysis of primary completion. DETERMINANTS OF EFA PROGRESS For the 55 low-income countries with populations over 1 million for which suffi- cient data could be compiled, we compared primary gross enrollment ratios and primary completion rates in a scatter plot (figure 3.1).1 The diagonal line in this figure, where the PCR equals the GER, represents a perfectly efficient student flow through the primary system, in a steady state. Since the completion rate obviously cannot exceed the gross enrollment ratio, all country observations fall below the line. The greater the distance from the line, the less efficient the cohort flow in that country. The closer to the origin a country falls, the lower its primary education coverage. Figure 3.1 shows the wide range in primary completion rates across countries and equally wide disparities in the relationship between gross enrollments and pri- mary completion. To deepen our understanding of the underlying determinants of high primary completion, we compared characteristics of the education systems in "successful" and less successful countries. To ensure that the sample of successful countries was regionally diverse and of a reasonable size (to avoid biasing the results i. The starting sample included all 79 IDA-eligible countries with per capita GNI below $885 in 2000. Time limitations prevented us from analyzing the 16 low-income countries with populations below 1 million. Of the remaining 63 countries, we had to exclude Bosnia and Herzegovina, Liberia, Myanmar, Somalia, Sri Lanka, Serbia and Montenegro, Tajikistan, and Afghanistan for lack of data, although we used an alternative methodology to estimate the financing requirements for Afghanistan, as described in chapter 4. This left us with a sample of 55 countries, of which 8 have achieved or are close to UPC (defined here as completion rate through grade 5 or 6 over 90 percent) and 47 have not. See annex table A.1 for a full list of countries in the sample. 61 FIGURE 3.1 Primary School Completion Rates and Gross Enrollment Ratios in a Sample of Low-income Countries, circa 1999/2000 Primary completion rate (percent) 1 10- Gop 100- 90 - 80- 70 - o 60 - N 50 - Gross enrollment ratio (percent) * Group 1 * Group 2 * Group 3 * Group 4 to the particular institutional features of any one region), we set a relative defini- tion of EFA success as follows: EFA success = GER 85 percent or above and PCR 70 percent or above.2 We called this set of high-performing countries Group 1. Given the wide range of GERs and PCRs in the sample, in order to sharpen the analysis we also set boundary parameters that separated the most extreme of the "unsuccessful" coun- tries into two distinct, stylized groups: * High inefficiency countries: GER 80 percent or above, but PCR 60 percent or lower. These countries were designated Group 2. * Low coverage countries: GER and PCR both 60 percent or lower. This was Group 3. When the sample was sorted on these boundaries, 10 countries fell into the category of relative EFA success (Group 1), 8 in high inefficiency (Group 2), and 7 in low coverage (Group 3). (For a full listing of the country groups see annex table A.3). Twenty-four countries fell in between the defined ranges (Group 4). Follow- ing Colclough and Lewin (1993), for each of the three groups we analyzed the domestic financing available for primary education; spending per student and key underlying cost factors; and the average repetition rate, which is a key driver of the 2. With "EFA success" defined as a completion rate over 90 percent, four of the eight countries in the sample that met the criterion were countries in Eastern Europe and Central Asia. Given the unique institutional legacy of these countries, it would bias the analysis of success factors if these countries retained this weight in the successful group. We omitted these countries from Group 1, but in order to get an adequate sample size had to set the criterion of (relative) EFA success as 70 percent completion rate or higher. European and Central Asian countries with estimated completion rates below 90 percent were retained in the simulation exercise, however. See annex table A.1. 62 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 3.1 Key Education System Parameters for Adjusted Sample of 49 Countries, Grouped by Relative EFA Success Sample Variable Group 1 Group 2 Group 3 Group 4 Average Number of countries . 10 8 7 24 n.a. Gross enrollment ratio (percent) 103 91 48 85 84 Primary completion rate (percent) 83 39 27 53 53 Government revenues (as percentage of GDP)a . 20.7 21.3 17.1 19.5 19.7 Education recurrent spending as percentage of GDP 3.8 2.5 2.6 3.3 3.1 as percentage of government revenues 18.2 15.5 16.9 17.7 17.3 Primary education recurrent spending as percentage of GDP 1.7 1.3 1.3 1.4 1.5 as percentage of total education recurrent spending 47.6 52.4 50.8 47.2 48.6 Unit cost (as percentage of per capita GDP) 9118 .5 18.7 11.8 12.4 Average annual teacher salary 3. 3 3.4 6.9 3.5 4.0 (as multiple of per capita GDP) Spending on inputs other than teachers (as percentage of primary education recurrent spending) 26.0 23.4 27.1 23.3 24.4 Pupil-teacher ratio 39:1 49:1 56:1 40:1 44:1 Private enrollments (as percentage of total) 7.3 10.5 7.7 10.4 9.4 Average repetition rate (percent) 9.5 27.8 19.5 13.3 15.8 n.a. Not applicable. a. Government current revenues, excluding grants. completion rate. The mean values are summarized in table 3.1. The full set of country-level data is found in annex tables A.2 through A.5. These data show that the Group 1 (relative EFA success) countries: * Devote a higher share of national resources to public primary education (1.7 percent of GDP compared with 1.3 percent of GDP in Groups 2 and 3) * Exhibit about average unit costs (spending the equivalent of 11.8 percent of per capita GDP per public primary student, compared with the sample average of 12.4 percent) * Pay annual teacher salaries averaging 3.3 times per capita GDP * Spend slighdy more than average of their recurrent budget on items other than teacher salaries * Have a pupil-teacher ratio of about 39:1, considerably below the averages for Groups 2 and 3, and * Have much lower repetition than the other groups (9.5 percent compared with the sample average of 15.2 percent). Group l's pattern may be summed up as: healthy spending; reasonable unit costs, teacher salaries, and class size; and low repetition. It is interesting to note that Group l's higher spending on primary education derives mainly from higher spending on education as a share of the government budget, and not from higher tax revenues CHAPTER THREE * WHAT WILL IT TAKE TO ACHIEVE UNIVERSAL PRIMARY COMPLETION BY 2015 63 relative to GDP or a higher share of education spending for primary education. In fact, Group 1 devotes a lower share of education spending to the primary level than the other groups, most likely because its relatively high primary completion rates mean more demand for subsequent levels of the education system. The countries in Group 2 have gross enrollment ratios close to those of Group 1, and also above the average for the sample. But their completion rates are only half as high as Group l's and below the sample average. This dramatic gap between enrollments and completion is the principal stylized characteristic of Group 2. Group 2 has the highest revenue-GDP ratio in the sample, but its lower share of total public spending for education results in lower spending on primary education as a share of GDP than in Group 1, even though Group 2 countries devote a higher share of their education spending to the primary level. Group 2's unit costs are the lowest of the sample, reflecting its higher pupil- teacher ratio (49:1, compared with Group l's 39:1) and lower spending on non- salary items. But the defining characteristic of Group 2 is the average repetition rate of 28 percent-at any given moment, more than one of every four primary school children in these countries are repeating a grade. Thus, although unit costs are relatively low, the costs per graduate in these countries are very high. The dropout rate in primary school, although not measured direcdy here, is clearly high in Group 2 countries, reflected in the fact that only 39 percent of chil- dren complete primary school, despite high access. The stylized pattern of Group 2 is therefore inadequate spending on quality and excessive repetition. Group 3 countries are painfuily far from EFA goals by any definition. The first defining characteristic of these countries is extremely low primary coverage. Less than half of all children in these countries have access to schooling, and only one child in four completes a primary education. Group 3 countries mobilize the lowest share of national resources in taxes of any of the groups, which translates into a low share of GDP for primary education, even though Group 3 countries' budget shares for education in general and for primary education in particular are close to those of the other groups. A second defining characteristic of Group 3 countries is their dramatically higher unit costs-60-70 percent above unit costs in Groups 1 and 2. The under- lying driver is also clear. Teacher salaries in Group 3 average almost seven times per capita GDP, about double the ratio of the other countries. The extremely high cost of teachers forces the education system to adjust with very high pupil-teacher ratios (56 students per teacher, compared with 39 in Group 1). Perhaps related to the very large class size, Group 3's repetition rates are also high-more than double those of Group 1, although still lower than in Group 2. The stylized characterization of Group 3 is low primary coverage deriving from a disastrous combination of low spending, high unit costs driven by extremely high teacher salaries, and relatively poor efficiency. We conducted two different batteries of statistical tests to assess the validity of these results. First, we tested whether the differences in variables across our three reference groups were statistically significant (at the 5 percent level).3 Since the samples were small, we used non-parametric tests to compare the distribution of 3. We are grateful to Luis Crouch for assistance in this analysis. 64 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 3.2 ......... ... ............................. Regression Analysis of Key Parameters Coefficients t Stat Intercept 47.86 5.08 1990 PCR 0.40 4.14 Recurrent spending on primary education as % of GDP 12.28 3.90 Average Teacher Salary (as multiple of per capita GDP) (4.49) (4.02) Pupil-teacher ratio (0.01) (0.04) Average repetition rate (%) (0.72) (3.46) R Square 0.81 Observations 44 Note: 1990 PCR values were taken from tables 2.4-2.9. All other regression variables were taken from Annex table A.2. values for each parameter across the three groups. This analysis confirmed that the difference in primary completion rates between Groups 1, 2, and 3 was statistically significant. The difference in primary GER between Group 3 and the other groups was also statistically significant. As expected, the difference in GER between Groups 1 and 2 was not statistically significant. This analysis also confirmed our hypothesis that key explanatory variables were differentially distributed across the three groups to a statistically significant degree. Between Groups 1 and 2, the difference in the repetition rate was significant. Between Groups 1 and 3, there were statistically significant differences in the teacher salary, the pupil-teacher ratio, and the repetition rate. And between Groups 2 and 3, there were statistically significant differences in the teacher salary and unit costs. In sum, this analysis confirmed that the three stylized groups did in fact reflect statisti- cally significant differences in primary completion rates and in four key underlying variables: the teacher salary, the pupil-teacher ratio, the repetition rate, and unit costs. We then used regression analysis to evaluate the explanatory power of these variables in relation to the sample as a whole. We used the baseline PCR in 1990 as a way of controlling for the independent effects that national cultural and histori- cal factors have on the evolution of the education system.4 As a result, the model focuses on the variables that have driven PCR progress over the decade. The regres- sion results are summarized in table 3.2. They show three variables (in addition to the baseline PCR) as statistically significant correlates of differences in primary completion rates. One variable-primary education recurrent spending as a share of GDP-had a strong positive effect. The other two variables-the average teacher salary and the average repetition rate-had a strong negative effect. These results suggest that indeed the notably high average teacher salaries in some countries in the sample has been a constraint to school system expansion and comple- tion rate progress. They also confirm the more intuitively obvious fact that high repe- tition constrains primary completion progress. When the pupil-teacher ratio appears 4. Five countries for which we had no 1990 PCR value were omitted from the analysis. CHAPTER THREE * WHAT WILL IT TAKE TO ACHIEVE UNIVERSAL PRIMARY COMPLETION BY 2015 65 FIGURE 3.2: Class Size in Relabon to Teacher Salary Average class size 100 - 90 - 80 - 70- * 60- U ' 50-m * 4oo~~~~~U- * *-E~; 40 - = 30 - -11amc 20- 0 -I I I I I I I I I 0 1 2 3 4 5 6 7 8 9 10 Average teacher salary (as a multiple of GDP/capita) Source: Mingat 2001. in the regression along with the average teacher salary, its additional explanatory impact is insignificant. But these two variables are correlated, and in alternative regres- sions without the teacher salary variable, the pupil-teacher ratio became statistically significant, similarly with a negative effect. The regression variables explained about 80 percent of the variance in primary completion rates across the sample. When compared with the earlier analysis of the three clusters, the regression results point to spending on primary education as an important factor in comple- tion rate outcomes in general, but not an important differentiator of the three styl- ized groups we analyzed (high GER: high PCR; high GER: low PCR; low GER: low PCR). Those outcome patterns appear to be more clearly linked to how resources are used than to the level of resources available. In sum, these statistical tests confirmed, first, that the three groups we identified on the basis of differential GER and PCR outcomes indeed had statistically significant differences in the distribution of key underlying variables. Second, regression analysis established that three of these key variables are statistically significant factors-both positive and negative-in explaining variance in primary completion rates. * IMPLICATIONS FOR ACCELERATING EFA PROGRESS The above analysis suggests several things. First, the pattern exhibited by the countries with the highest primary completion rates-relatively healthy spending on primary education as a share of GDP, moderate unit costs, and low repetition- may represent a broadly balanced and sustainable pattern of resource allocation that is a necessary condition for EFA progress. While it is far too strong to label the Group 1 average values for some of these parameters-such as 39 students per teacher or teacher salaries of about 3.3 times per capita GDP-as "norms" for a healthy education system, it appears that deviating very far from these values forces education systems into unhealthy adjustments, if financing is constrained. 66 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD (gis3 9i aGccounting Framework for Spending on Primary Education Total public spending on primary education Average teacher salary as a multiple of GDP/capita GDP pupil-teacher ratio x (1 - share of pupils in privately financed schools) X (1 + spending on inputs other than teachers as a multiple of spending on teacher salaries) x total enrollments x school-age population school-age population total population A relatively simple accounting framework, outlined above, helps to show why increases in some of the parameters we examined must necessarily be balanced by decreases in others. This accounting identity has at least four important implications for our analysis. First, and most fundamentally, the accounting identity establishes that the total amount of resources spent on primary education in a given country must equal the per-student spending on teachers and inputs other than teachers times the size of the school-age popula- tion enrolled. If a country wishes to increase the share of the school-age population enrolled, it must either increase its spending on primary education or find economies in the average teacher salary, the efficiency with which teachers are deployed (the pupil-teacher ratio), and/or the spending on inputs other than teacher salaries. The empirical experience in low- income countries shows that it is often impossible to increase overall spending on primary education as more children are enrolled, and that compression of spending on other inputs and increases in the pupil-teacher ratio are very commonly the balancing items. Second, the accounting identity shows that the fiscal cost of enrollment expansion is linked to the size of the privately financed schooling sector. Note that the issue is not the extent of pri- vately delivered schooling, which in many countries is a positive force for schooling expansion since government resources channeled through NGO, community, or religious schools can often help improve the efficiency of education spending. The privatelyfinancedschooling sector in this framnework refers to the school sector that receives no government funding, typically a for-profit sector that serves only an elite segment of the population. Although the relative size of this sector is not usually an explicit policy variable for countries, it is not uncommon for con- straints on public financing of primary education to lead to eroding quality in the public system and to produce a spontaneous shift in enrollments to privately financed schools. Leaving aside the question of whether or not this is good for education policy, the accounting identity shows that this can ease the fiscal costs of achieving universal primary enrollment. Third, the accounting identity indicates that the fiscal pressure a country faces in achieving universal primary enrollment is also a function of the relative size of the school-age (7- to 12-year-old) population in the overall total-or the "schooling dependency ratio." This ratio varies widely across countries. For example, in the year 2000 in our sample, while the dependency ratio for the sample as a whole was 16 percent, it ranged from 9 percent in Georgia to 18 percent in a number of African countries (see annex table A.5). The average for the African countries was 17 percent. The projected evolution of this variable also differs across regions. Whereas for the African countries in our sample, population projections show little decline in the size of the school-age population before 2020, for many of the countries outside Africa it is expected to decline substantially-creating a "demographic bonus" that will make it relatively easier to achieve universal primary enrollment without CHAPTER THREE ° WHAT WILL IT TAKE TO ACHIEVE UNIVERSAL PRIMARY COMPLETION BY 2015 62 (Ni(gtT 3^ continued increasing the share of national resources devoted to education. The accounting identity shows that, other things being equal, Benin and Kenya will have to spend twice as much on education as a share of GDP in 2015 as will Georgia in order to achieve the same degree of primary education coverage and quality. Finally, and importantdy for the present study, while the above accounting identity explains primary education unit costs, it does not explain costs per graduate. If the target of analysis is the primary completion rate (rather than primary enrollments), the above framework must bc complemented with one additional variable: the average repetition rate in primary education, which is the key driver of student flow efficiency, completion rates, and costs per graduate. The expenditure accounting framework shown in box 3.1 explains how these variables are linked. The implications of the accounting framework are clearly borne out by our results. In a resource-constrained setting, if average teacher salaries are very much higher than 3.3 times per capita GDP, there is upward pressure on the pupil-teacher ratio (PTR)-as the data from our sample, graphed in figure 3.2, show. Although from an accounting standpoint, the same ratio of average teacher salary to PTR may be satisfied with many different values for the two variables, from a service delivery standpoint they are not all equally efficient. Research espe- cially points to the adverse impact on student learning when average class size exceeds the range of 40-45 (Lockheed and Verspoor 1991). Thus, very high aver- age teacher salaries balanced by large class size is not an efficient resource mix. On the other hand, empirical evidence does not suggest that lowering class size below the range of 40-45 is an efficient investment, either. Lowering class size is cosdy, and research indicates that within the range of 40-45 to as low as 20 pupils per teacher, declines in class size are not correlated with appreciable gains in student learn- ing (Lockheed and Verspoor 1991). If the pupil-teacher ratio is very much below 40, the system needs to employ many more teachers, and as the accounting framework indicates, if financing is constrained this will cause average salaries to be compressed. This is also evident from figure 3.2. This pattern is also suboptimal, as low teacher salaries are linked to other systemic problems: inability to attract the best and brightest into teaching; inability to attract teachers to work in remote or hardship areas; inability to reward high performance; chronic absenteeism caused by low teacher motivation or the need to work multiple jobs; and, in many countries, over- whelming pressures for teachers to demand fees for private tutoring or other direct payments from parents, to supplement their unsustainably low salaries. The "right" level of teacher salary in a given country is one that is sufficient to attract qualified individuals into teaching and motivate continued good perform- ance, given that the schools are in competition with other sectors for educated workers. The appropriate level will depend on the supply of educated individuals, the demand from all sectors of the economy (and foreign countries), and the com- bined attractiveness of salary and non-salary compensation (such as the shorter work hours, long vacations, and job stability that are common in the teaching pro- fession). Only labor market data for the individual country can determine the appropriate wage for teachers in a given country context. 68 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD FIGURE 3.3* Evolution of Average Teacher Salary in Primary Education, by Region and Subregion, .1975-2000 Average teacher salary as a multiple of GDP/capita 20- 15 - 10- 15 1975 1980 1985 1990 1995 2000 _ Francophone Sahelian African countries -4- Anglophone Africa _ Francophone Africa -_- Asia _ Middle East 4 Latin America Source: Mingat 2001. However, there are some broad global patterns. For example, the average wage paid to teachers, as a multiple of country per capita income, tends to decline as countries develop economically. The average annual wage in the sample of very low income countries we studied was in the range of 3 to 4 times per capita GDP (although with substantial variance, as we have noted). In middle-income coun- tries in Latin America it is in the range of 2 to 2.5 times per capita GDP, and in the OECD it is currently about 1.8 times per capita GDP. Another global pattern is the historical trend recently examined by Mingat (2001), which shows a steady decline in the average teacher wage in developing countries over the last 25 years, from 6.6 to about 3.7 times the per capita GDP (figure 3.3). While disparities across regions still exist, there has been a strong downward movement of wages in the highest wage subregions (francophone Africa and within francophone Africa, the Sahelian countries), moving toward conver- gence with the other regional averages over time. The data presented in table 3.1 make clear that the incremental EFA progress from an additional unit of education spending in both the Group 2 and Group 3 countries is much lower than in Group 1. At Group 2's low internal efficiency, in fact, the challenge of reaching 100 percent primary completion (more than a dou- bling of current completion rates) would be staggering and the costs exorbitant. It would imply construction of nearly 80 percent more schools, with commensurate teacher hiring and other inputs, than will be needed in the Group 1 school systems to achieve the same goal. The ratio is far higher than the average repetition rate (28 percent) in these countries alone, because of the high correlation between repetition and dropout. Research shows that children who have repeated at least one grade are much more likely CHAPTER THREE * WHAT WILL IT TAKE TO ACHIEVE UNIVERSAL PRIMARY COMPLETION BY 2015 69 than non-repeaters to drop out before completing primary school, a probability that is even higher for girls. Countries such as those in Group 2 simply cannot reach the goal of universal primary completion without substantially reducing repetition. In Group 3 countries, very high unit costs limit the impact of additional spending. The equivalent amount of additional financing in a Group 1 country and a Group 3 country (with equal GDP) could bring 160 children into school in the former for every 100 in the latter. The high cost structure of educational provi- sion in Group 3 countries, driven by very high average teacher salaries, has clearly limited the expansion of coverage in the past and unless addressed will continue to constrain the pace of EFA progress in the future. Extremely high pupil-teacher ratios in these countries and relatively high repetition rates only make things worse, as they contribute to relatively low primary completion rates. Although the inter- nal efficiency of Group 3 countries is not quite as low as for Group 2, it still means that of every 100 children who enter school, only a litde more than half complete. For countries in Group 3 to make faster EFA progress, not just one but many parameters of their education systems need to change sharply. A final implication of table 3.1 is that countries in our sample currently exhibit very different levels of domestic commitment to reaching the goal of universal primary completion. Underlying the average values reported in the table is considerable variance in the share of GDP being spent on primary education, from more than 3 percent in Lesotho and Zimbabwe to less than 1 percent in countries such as Pakistan, Lao PDR, and Georgia. Needless to say, an additional 1-2 percent of GDP for primary education could make a huge difference in any country. Any global strategy for accelerating EFA progress must encourage more domestic effort where it is low, and take care not to penalize the countries currendy showing the strongest domestic commitment. Ultimately, this analysis shows that the road to universal primary completion will be quite different for countries in groups 1, 2, and 3, and for the remaining countries, which essentially present milder versions of the same issues. Although none of the education systems in this low-income sample is without problems, on average the balanced education system parameters of Group 1 countries suggest they are better positioned to reach the MDG without major system change.5 But for the other countries in the sample, that is clearly not the case. As the Group 2 countries demonstrate, if the EFA goal were framed as universal primary enrollment, these systems would be close to achieving it. Group 3 countries are much further behind, but even they could eventually get to universal enrollment with adequate financing. But universal primary completion is another story. It is simply unachievable in education systems functioning with internal efficiency this low, no matter how much money is poured in. The inescapable conclusion recon- firms what Colclough and Lewin (1993) asserted a decade ago: the attainment of universal primary education, for most low-income countries, depends even more crucially on education system reform than on incremental financing. 5. It should always be recalled, however, that underlying these average values there may be considerable variance in these parameters both across countries in the group and within individual countries that would, in fact, constrain EFA progress. 70 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD CHAPTER Costing the MDG of Universal Primary Completion A A central implication of chapter 3 is, in essence, the following: If the MDG of uni- versal primary completion is in fact achieved by 2015, it will be because at-risk countries have succeeded in transforming their education systems to function more like the systems of Group 1 countries. A concerted set of reforms would have improved those countries' resource mix and improved the learning environment for children, resulting in lower repetition and higher retention in school and cul- minating in a higher rate of primary completion. In countries where domestic financing for education was low (compared to other countries) it would have been increased, either by allocating more budgetary resources to education or by increas- ing the share of education spending devoted to primary education. And, in coun- tries where domestic resources alone were insufficient, notwithstanding a maximum fiscal effort and strong commitment to financing primary education, external aid would have been available to fill the gap. In this framework, it is clear that the mix of policy actions required for accel- erated MDG progress differs considerably from one country to another. It follows that the domestic and external financing requirements for achieving the MDG will be highly sensitive to the extent to which, and pace at which, appropriate reforms are implemented. Therefore, the soundest basis for estimating global financing requirements is to aggregate these from country-level analysis. We used a relatively simple simulation model to do this for the 47 low-income countries that have not yet achieved 90 percent completion or higher through grade six and for which adequate data could be compiled.' Our sample includes virtually all of the countries that are "seriously off track," many of the countries that are "off track," and nine countries that are currendy "on track" but whose progress might be accelerated with appropriate reforms. The model was developed to test the proposition that accelerated MDG progress could be achieved by bring- ing core financing and service delivery parameters of education systems in at-risk countries into line with the parameters observed in countries that have higher pri- mary completion. A more detailed discussion of the model is presented in annex box A. 1. Key dis- tinguishing features of the model and our approach may be summarized as follows: 1. The simulations are country-specific and the cost of achieving UPC by 2015 is the aggregate of the estimates for the 47 countries analyzed. I. Since the data demands are significant, we focused the analysis on countries relatively far from the goal. Afghanistan also has primary completion below this level, but since we could not obtain adequate data to run the model, we used an alternative methodology to estimate Afghanistan's financing needs (see table 4.11). 7 1 2. The target variable is primary completion-that is, the share of each school-age cohort that completes five or six years of schooling.2 3. The model tests a dynamic path of policy reform to improve education service delivery. 4. The model specifically responds to concerns about student learning through variables to improve quality. 5. The model responds to concerns about demand-side constraints through explicit provision for targeted subsidies to the most vulnerable populations. 6. The model specifically acknowledges the broader resource needs of the education sector by limiting spending on primary education to a sustainable share of the overall education budget. 7. The model distinguishes between recurrent and capital costs, and generates a separate estimate for each component. 8. It is assumed that universal primary completion will be achieved in all countries using public resources, with no user fees or other costs imposed on students. 9. The 2002 United Nations/World Bank projections of the school-age population through 2015 are used. 10. The baseline year for the projections is 1999 or 2000 for all countries. 11. Because of the prevalence of AIDS in many Sub-Saharan African countries, a separate effort was made to estimate the additional costs that AIDS will impose on the attainment of universal primary completion in Africa (box 4.1). Because, as argued earlier, it is inherendy impossible for many countries to achieve universal primary completion without improving the efficiency of educa- tion system functioning, our simulations expressly test how reforming key param- eters of the education system in low-completion countries would affect the costs of achieving the MDG. Targets drawn from the observation of higher-performing countries are set as the goal for all countries in the sample. The model is constructed around four sets of component variables, estimated in sequence: * Enrollment: the number of pupils in publicly funded schools over the period * Service delivery: the recurrent costs of services in publicly funded schools * System expansion: the capital costs of needed classroom construction * Systemfinancing: the volume of domestic resources mobilized for primary education. 2. In order to avoid biasing the financing gap estimates toward countries with longer primary cycles and away from countries with shorter ones, we modeled student flows through the equivalent of six grades in all countries where the primary cycle is longer than six or shorter than five years. In countries where the primary cycle is five years, we retained that, but adjusted budget shares accordingly. 72 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Annex box A. 1 provides more details on the stylized dynamics of the model and the full set of variables used, especially in the calculation of student enroll- ments over the projection period. The key observation is that, in order to achieve universal primary completion by 2015, each country must achieve 100 percent entry of school-age children into grade 1 by 2010 (for a five-grade system) or 2009 (for a six-grade system). For many countries in this sample, this implies a greatly accelerated pace of primary enrollment expansion over the coming years. The fol- lowing sections focus on the model's key policy variables, which relate to service delivery, system expansion, and system financing (table 4.1). Table 4.1 ................................................................... Benchmarks for Primary Education Efficiency and Quality . e SAMPLE MEAN . . IN 1999/2000 Highest- Sample Range Adjusted Completion 2015 Variable in 1999/2000a I Sampleb Countries Benchmarks Service delivery Average annual teacher salary 0.6-9.6 4.0 3.3 3.5 (as multiple of per capita GDP) Pupil-teacher ratio 13:1-79:1 44:1 39:1 40:1 Spending on inputs other than teachers . 0.1-45.0 24.4 26.0 33 (as percentage of primary education recurrent spending) Average repetition rate (percent) 0-36.1 15.8 9.5 10 or lower System expansion . * . - Unit construction cost $6,500-$24,000 - $6,500-$12,600' System financing Government revenues (as percentage 8.0-55.7 19.7 20.7 14/16/18' of GDP)d Education recurrent spending 3.2-32.6 17.3 18.2 20 (as percentage of government revenues) Primary education recurrent spending 26.0-66.3 48.6 47.6 50/42f (as percentage of total education recurrent spending) . Private enrollments (as percentage 0-77.0 9.4 7.3 10 of total) . a.The range includes data from the full sample of 55 countries. b.The adjusted sample excludes European and Central Asian countries. c.Construction costs in constant dollars based on "good practice" average values observed in each region. d.Government current revenues, excluding grants. e. Staggered targets proportional to per capita GDP. f. Benchmark is 50 percent for a six-year primary cycle; 42 percent for a five-year cycle. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 73 SERVICE DELIVERY The model highlights the four most important determinants of primary educa- tion quality and unit costs per primary graduate: the average teacher salary; the pupil-teacher ratio; the share of spending on inputs other than teacher salaries; and the repetition rate. The dynamics of these four variables in the simulations depend on whether their initial values in a country are above or below the indicative frame- work targets. Depending on the direction of movement, adjustments to the same parameters in different country contexts are considered either "quality enhancing" or "efficiency enhancing," as discussed below. TEACHER SALARIES The single largest cost item in any education system is the salary bill for teachers- accounting for more than 70 percent of recurrent spending in virtually all coun- tries in our sample and as high as 95 percent in some of them. Across these countries there is wide variation in average annual salaries, ranging from 0.6 to 9.6 times per capita GDP. Historical and institutional factors clearly influence this variable, evident in our sample from the distinct regional patterns. In the Sahelian African countries, for example, the average is more than six times per capita GDP; in the Eastern European and Central Asian countries, the average teacher salary is lower than GDP per capita. For the simulations, the 2015 target for this parameter is set at 3.5 times per capita GDP, a round figure that is close to the observed aver- age in the highest-completion countries in our sample (3.3). However, because the average level of teacher salaries is the most politically sensitive of all the parameters, we made special assumptions about the pace at which it could be adjusted. For countries below the target, where average salaries need to be raised in the simulation, the political dynamics are easier. Although we programmed all other parameters to reach the target values only by 2015 through gradual movement, in the case of upward adjustment of teacher salaries, we assumed that the reform could be implemented more quickly. Given the positive impact on system quality such a change could have, it would be desirable to implement it as quickly as possible. Unlike other parameters (such as lowering the pupil-teacher ratio, which requires additional classroom construction), it is also technically possible to implement an upward salary adjustment almost immediately. And, given the political popularity of such a move, implementing it sooner rather than later could help consolidate support for the reform program as a whole. More than half of the countries in the sample (28 of the 47) were cases of upward adjustment in teacher salaries. The major constraint to this particular reform is fiscal sustainability, not polit- ical opposition. But our financing framework explicitly assumes that countries' adoption of needed reforms (that is, those consistent with this indicative frame- work) would constitute a "credible plan" for EFA attainment and that any resulting financing gaps would be supported by international donors. So, in order to gain the maximum quality benefits from this reform early in the projection period, and to demonstrate how much this adjustment could maximally add to the costs of EFA, we assumed that upward adjustments of average teacher salaries-where jus- tified, in relation to the reference parameters-are implemented immediately. 24 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Very importandy, we also implicitly assume that such a reform is implemented in an intelligent manner that would maximize the positive impact on schooling quality-for example, by establishing new and higher standards, weeding out the weakest performers, introducing a structure of incentives to reward performance, and putting in place stringent processes for new teacher selection. The size of the upward adjustment, which is very significant in some cases (particularly in the Europe and Central Asia region), raises obvious questions about the realism of assuming that such a change could be implemented for one segment of the civil service in isolation. In these cases the simulation should be taken as simply laying out the potential cost (and international financing) implica- tions of moving toward the benchmark value in this area. Because raising average salaries can be expected to improve the quality of the teaching force as well as reduce absenteeism, stimulate greater accountability for teaching effectiveness, and create incentives for high performance or deployment to remote areas, it is considered a quality improvement in countries with salaries currendy below the target. For countries with teacher salaries above the target level, the adjustment down- ward is considered an efficiency improvement. Since it is legally and politically impossible in most contexts to reduce the salaries of civil servants, the simulations assume that this reform must be implemented in an especially gradual way. It is assumed that a new cadre of teachers is recruited at the pace of new classroom con- struction and paid at the target level of 3.5 times the per capita GDP, and that all recruitment of higher-paid civil service teachers is suspended. A number of coun- tries in francophone Africa and elsewhere have in fact implemented such a reform in teacher contracting and have generally found no shortage of well-qualified can- didates willing to work at the lower salary level, suggesting that the higher salary is not (or is no longer) an efficiency wage in these economies. However, the longer- term impact of this reform on teacher motivation and performance and student learning, as well as its political sustainability, are still open questions and merit fur- ther research. In the simulations we assume that incumbent teachers continue to be paid on their current salary scale, but that over time their weight in the overall salary bill diminishes through retirement. Thus, the average salary approaches the target level. In many countries in the sample, however, it still remains above the 3.5 tar- get by 2015. PUPIL-TEACHER RATIO The range in pupil-teacher ratios across the sample is similarly wide, from 13:1 to 79:1. Although the pupil-teacher ratio is not perfecdy correlated with average class size in most countries, we take it as a reasonable proxy. The target value for this parameter is 40: 1, based on the observed average in the high-completion countries, and also supported by a body of research on class size, as discussed in chapter 3. For countries currently above this level, the downward movement is considered a qual- ity improvement. For countries currently below, the adjustment upward is consid- ered an efficiency improvement. In all countries, the simulations gradually adjust the average pupil-teacher ratio to reach 40:1 by 2015. Although careful teacher CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 75 deployment to achieve a 40:1 pupil-teacher ratio across all schools is one of the most powerful strategies an education system can pursue to promote efficiency and equity, geographic conditions and extremely low population density in some coun- tries, such as Mongolia and the Republic of Yemen in this sample, can make it difficult to achieve in reality. RECURRENT SPENDING ON INPUTS OTHER THAN TEACHERS The amount of resources available for non-teacher-salary items is a crucial factor in education quality. Relatively abundant research indicates that books and other learning materials are highly cost-effective complementary inputs in the learning process. Although less extensively researched, teacher development and supervi- sion, system management, student learning assessment, school maintenance, and other items clearly are also important elements in quality education systems. Yet most countries find that the pressure of teacher salaries means the budget for these other items is constantly squeezed. The recurrent budget share for spending on items other than teacher salaries is the only variable in the study for which we set the target level (33 percent) signifi- cantly higher than the observed average for the high-completion countries (26 per- cent). We did so for three reasons: (a) to signal the crucial importance of increasing the quality of the learning environment in many countries, especially through the provision of more abundant and better-quality books and materials, if universal primary completion is to be reached; (b) to signal that school supervision, student assessment, teacher development, and many other system management functions are in urgent need of upgrading, and this will require considerable professionaliza- tion of these functions and imply additional cost; and (c) to signal that universal primary completion cannot be achieved in most settings without provision of spe- cial assistance to "the last 10 percent of children"-those at greatest risk of not enrolling in school or dropping out, be they girls, very poor children, children from ethnic minorities or remote rural communities, children with disabilities, or simply children falling behind in their learning because of illness or their families' needs for intermittent labor. All of these imply additional costs that school systems must be prepared to absorb. Given the projected growth of one particular set of children at risk-HIV/AIDS orphans-an additional specific provision is made for targeted subsidies to these children in the African countries (box 4.1). The simulation model we used does not go to the level of identifying which specific inputs among the above should be prioritized in a given country; it simply creates budgetary space for a healthy level and appropriate mix of expenditures on schooling quality, efficient system management, and appropriate demand-side interventions. For virtually all of the countries in the sample, the increase in this variable to the target level is considered a quality improvement. Eight countries in the sample are currently above this level. In some cases, this is because the average teacher salary is very low. In other cases, the downward adjustment is considered an efficiency improvement, on the grounds that these budgetary resources-since they are not currently producing the desired outcomes-need to be better spent. It should be noted that our target variable (the share of recurrent spending on inputs and items other than teacher salaries) is not equivalent to "non-salary 76 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD ( K oQ~ 1 The Incremental Costs of HIV/AIDS for Universal Primary Completion The severity differs across countries, but today virtually all of Sub-Saharan Africa is battling HIV/AIDS. The epidemic is already affecting education systems in the region and its impact can be expected only to worsen in most countries between now and 2015. Given HIV/AIDS prevalence in Sub-Saharan Africa and what we know about its implications for teacher supply and student attendance, it is an important element to consider in efforts to cost the achievement ofUPC. There are three main ways in which HIV/AIDS affects primary education systems. First, future growth of the school-age population will be smaller in countries with AIDS prevalence than it would have been otherwise. Second, the stock of teachers will be greatly affected. Teachers who are sick are likely to be absent and substitutes will be needed to avoid disruptions to school functioning. More teachers are likely to die while they are still in service, implying higher personnel turnover, increased need for new recruitment, and greater need for teacher training than would have been the case in the absence of AIDS. Third, the proportion of orphans in the school-age population will be larger in an AIDS-affected country than it would have otherwise been. It has been shown that maternal and double orphans are more likely than non-orphans to drop out of school (Subbarao, Mattimore, and Plangemann 2001). Since the MDG goal is to ensure that all children com- plete at least five or six years of primary education, orphans are likely to need special support if that goal is to be achieved. Demand-side financing such as stipends or other support, tailored locally to best fit the needs of these children, will have to be developed. Our costing exercise made special provisions in order to estimate these incremental impacts of AIDS. The United Nations/World Bank population projections used have been updated by demographers to reflect the impact of HIV seroprevalence on future population growth. To model the impact on teachers, we used new estimates produced by a team at the Imperial College, U.K., of AIDS incidence among teachers. This research shows that (a) teachers are affected in the same proportion as adults in general (the figures used are UNAIDS estimates between 2000 and 2015), and (b) the sickness evolves over 10 years and during that period, on average, teachers are absent 260 days. To estimate the costs of keeping orphans in school, we used UNAIDS-provided estimates of the population of maternal and double orphans in 1999 and projected these to 2015 for 10 countries, using a simulation model developed by the Imperial College. We then extrapo- lated the pattern of evolution in these 10 countries to other countries in Africa by subregion (WestAfrica, East Africa, and southern Africa). Finally, we estimated that $50 per year would be required to maintain each maternal and double orphan in school, a cost estimate that is consistent with some of the recent programs in the region channeling financial support to such children. Overall, these projections showed that the impacts of HIV/AIDS will add at least $287 million per year to the estimated costs of achieving the MDG in the 33 Sub-Saharan African countries in our sample. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 77 spending." Our target variable includes all non-salary spending plus all salaries of personnel other than teachers-administrators, watchmen, cooks, and others employed by the education ministry who are not assigned to classroom teaching. In some countries where total spending on inputs other than teacher salaries is rel- atively high yet system outcomes are still poor, an excess of administrative staff on the payroll is a factor. AVERAGE REPETITION RATE Reported grade repetition in the countries studied ranges from 0 to 36 percent. High repetition is incompatible with the goal of universal primary completion, and the observed average in the higher-completion countries (9.5 percent, which is well below the sample mean) corroborates this. Accordingly, we assume that coun- tries with average repetition rates above 10 percent would adopt policies to bring repetition gradually down to 10 percent by 2015. The repetition rate is often viewed as an education outcome rather than a pol- icy variable, but there is accumulating country experience with tailored strategies that are effective in reducing repetition. These strategies include introducing local language instruction in the critical early grades of primary school; designing the first two grades as a single curriculum block with strong emphasis on basic literacy and numeracy, to provide children more time to master key concepts; assigning the most experienced teachers to the first few grades; and providing cross-peer tutoring for children falling behind. It should be stressed that effective strategies such as these positively affect the repetition rate by increasing students' learning achieve- ment. In contrast, a mandatory policy of universal promotion, which does not pro- duce real effects on student learning, is not considered an effective strategy for reducing repetition. The repetition rate is a key driver of the costs of primary completion, so in high-repetition countries effective policies to reduce it are crucial. For countries with repetition rates already below 10 percent, no change was made. Consistent with this, when universal primary completion is achieved in 2015, the gross enroll- ment ratio for the sample is 107 percent. SYSTEM EXPANSION Unit construction costs, expressed in 2000 constant U.S. dollars per classroom, were based on values recently reported in World Bank project appraisals and other sources for the countries in the sample. In all cases the cost estimate was for a fully furnished and equipped classroom built to adequate standards. Wherever possible, regional average "good practice" cost estimates were used. For the African countries and the European and Central Asian countries, an average figure of $8,000 was used. For South Asian and East Asian countries, an average cost of $6,500 was used. For Haiti, Honduras, and Nicaragua, values of $11,000-$12,000 were used. The highest estimate in the sample-$ 12,600 for the Republic of Yemen-is con- sidered by regional experts a good practice target for the country and is consider- ably below past unit construction costs. 78 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD For the countries outside Africa in the sample, system expansion was projected from actual data on the baseline number of classrooms. However, it was not possi- ble to obtain these data for all of the African countries, and an assumption had to be made that the number of teachers was a rough proxy for the number of func- tioning classrooms. This assumption is somewhat problematic, as virtually all school systems in this sample have some degree of double-shift or even triple-shift utilization of classrooms, which means that the simulations underestimate the true incremental capital costs of reaching the UPC goal. For all of the countries, we projected the number of additional classrooms needed to ensure universal primary completion by 2015 on the assumption of 40 students per classroom and one classroom per teacher. The projections, there- fore, implicitly assume that the existing stock of classrooms is adequate, even though the baseline number of classrooms in some countries in our sample includes those as rudimentary as "classrooms" under a tree. Thus, it is important to note that the very real needs for upgrading current school facilities to the standards we assume for the future are not captured in our simulations. However, we attempt to make an adjustment for this in the final section of this chapter. Similarly implicit in the simulations is the assumption that other system infra- structure (district and central administrative offices, teacher resource centers, teacher training institutes, and so forth) exists at the beginning of the projection period in sufficient quantity and adequate quality to support system functioning. Although our simulation target for recurrent spending on inputs other than teacher salaries is designed to cover operating and maintenance costs for all school system infrastructure-and not only dassrooms-our capital cost estimates are limited to the need for incremental dassroom construction, and do not capture the need for incre- mental expansion of other system facilities. This also dearly underestimates what may be significant needs in the countries analyzed. It was impossible within our time frame to obtain the country-specific baseline data on the quantity and quality of exist- ing school system infrastructure, other than dassrooms, that would be required to project the incremental needs in these other areas. However, in the final section of this chapter, we also make a rough estimate of these needs for the sample. Thus, the classroom construction requirements projected in these simulations must be understood as a minimum estimate of the total capital costs these countries will likely need to incur in order to achieve fully functional school systems capable of realizing universal primary completion. More detailed, country-specific work is needed to estimate infrastructure and rehabilitation needs more precisely. It can be assumed that in this set of countries these additional needs are significant. SYSTEM FINANCING DOMESTIC RESOURCE MOBILIZATION The financing blockofthe model estimates the domestic resource flows forprimaryedu- cation over the projection period. In 1999/2000, public spending on primary education ranged from 0.2 to 3.3 percent of GDP in the countries studied, a huge range. In this costing exercise, it is assumed that external financing will only be available to those CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 29 countries that show evidence of a strong domestic commitment to achieving universal primary completion by allocating a fair share of national resources to the goal. But what constitutes a "fair share"? In order to ensure that our target parame- ters for domestic resource mobilization did not penalize the poorest countries with the most fragile tax bases, we decomposed the share of GDP spent on primary edu- cation into three underlying variables and set separate targets for each: * The revenue-GDP ratio, reflecting differences in the overall size of the public sector and the national resource base * The share of domestic revenues allocated to education, an indication of public priority given to education * The share of the education budget allocated to primary education, an indication of specific commitment to universal primary completion. Since the lowest-income countries typically have more difficulty mobilizing tax revenues than do wealthier countries, we staggered the target values for the revenue- GDP ratio in 2015-either 14 percent, 16 percent, or 18 percent of GDP, depending on the level of per capita GDP. For the second variable-the share of government revenues devoted to education-we set a target of 20 percent, a round number reasonably close to the 18.2 percent average for the high-completion countries in our sample. For the third variable-the primary education share of total education spending- we set a target of 50 percent for countries with a six-year primary cycle, again a round number slightly above but consistent with the reference countries. For coun- tries with a five-year primary cycle, a pro-rated share of 42 percent was used. Where countries' current values were lower than these targets, they were adjusted upward-in essence, asking the country to increase the domestic resources it is mobilizing for EFA. But there were several country cases where spending on one or more of these subcomponents currendy exceeds the targets. While from the standpoint of an EFA costing exercise it is tempting to maintain these levels-which would unquestionably aid in reaching the goal-we were con- cerned that some of the spending patterns may not be sustainable over the medium term. And, if the higher levels of resource mobilization were retained, the external financing requirements estimated for these countries in the simulation would be correspondingly lower. This seemed a perverse outcome, effectively penalizing the countries with the highest domestic commitment to EFA attain- ment and rewarding those doing less. Thus, we opted for a scenario (C2) that put all countries on an equal financial footing by instituting the target values, even when this forced an artificial decline in spending on primary education from current levels. However, because the over- all financing estimates are so sensitive to these variables, and because one of these-the tax-GDP relationship-is exogenous to the education sector, we also ran two sensitivity analyses, in which countries' higher spending levels were assumed to persist. Under scenario C 1, we allowed higher-than-target spending on education and primary education to be maintained. Under scenario C3, we retained higher-than-target tax-GDP ratios. The assumptions in each scenario are summarized in table 4.2. As expected, scenarios Cl and C3 did change the results for particular countries and lowered the size of the overall financing gap. 80 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 4.2 ................................................................... Alternative Scenarios for Domestic Resource Mobilization TARGETS FOR 201 5 UNDER THREE . ALTERNATIVE SCENARIOS Simulation Variable C1 C2 C3 Government revenues as 14/16/18a 14/16/18 14/16/18, but if percentage of GDP current share exceeds 18% it stays unchanged Public spending on 20-26b 20 20 education as percentage of government recurrent revenues, excluding grants Primary education If current share is 50/42 50/42 spending as percentage . higher than of total public recurrent 50/42 it stays education spending' unchanged a. Staggered targets proportional to per capita GDR b. Values below 20 percent are increased to a target of 20 percent by 2015. Values above 26 percent are reduced to a target of 26 percent by 2015. Values in the range of 20-26 percent remain unchanged. c. The target is 50 percent for six-year primary systems and 42 percent for five-year systems. PRIVATE ENROLLMENTS AS A SHARE OF TOTAL The share of enrollments in privately financed schools has an important impact on public sector financing requirements. A target for the share of private enrollments was set at 10 percent in these simulations. This share is relatively close to the observed average for the high-completion group (7.3 percent) and for the sample (9.4 percent), but the rationale for this target was more conceptual than empirical. The conceptual framework of this report is that attainment of universal pri- mary completion is a responsibility of national governments and that the children in any country that are currently out of school are those least able to contribute to the costs of education. As countries progress toward universal primary completion, the target populations are increasingly poor, remote, and marginalized. Cost recov- ery and cost sharing are less appropriate financing strategies for these populations than for any other segment. Therefore, we assume that no user fees or other costs are imposed on public school students, and on top of this we make explicit provi- sion for targeted subsidies to the most vulnerable groups. Government responsibility tofinance universal primary completion, however, does not imply that all schooling must or should be publiclyprovided. To the con- trary, the target parameters we use are very consistent with service delivery arrange- ments that channel government financing to private providers, especially to NGO or community-run schools. For simplicity, enrollments in these alternative schools are classified as "public" in our simulations, since they are publicly funded. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 81 However, we assume that in every country the uppermost income decile does have the capacity to contribute to the financing of primary education. In virtually all countries an elite private school sector exists, serving from 5 to 15 percent of primary students. To avoid having scarce public resources subsidize elite groups in a setting where EFA has not been achieved, we assume that 10 percent of primary enrollments in all countries modeled will be privately financed. In countries where the current share of enrollments in private schools is below this level, the increase to 10 percent in the simulation is a resource gain. In many countries in the sample, the private share of primary enrollments is currently above 10 percent, usually reflecting the limited supply or poor quality of public schooling. Since our simulation exercise explicitly models a scenario of qual- ity improvement and expansion in public primary education, it may be expected that some shift in enrollments back to public schooling would occur. However, an alternative scenario in which a significant number of students are served by private providers that are publicly subsidized is also consistent with the simulations. In the latter case, enrollments in for-profit private schools are assumed to be no more than 10 percent of all enrollments, but an unspecified number of children could be enrolled in nonprofit private schools, financed with public education resources. . COUNTRY-LEVEL SIMULATION RESULTS For each country in the exercise, the adjustment from initial parameters to the full set of target parameters effectively generates a threefold strategy of: * Quality improvement * Efficiency improvement * Increased domestic resource mobilization. The specific elements of the strategy in each of these three broad areas depend upon the country's initial conditions, the number of parameters that would require adjustment toward the benchmarks, and the direction of the adjustment required. It should be noted that the combined effect of the above strategies is the achieve- ment of an equitable primary education system, implicit in the goal of universal primary completion. In order to demonstrate in each country case the relative need for either qual- ity improvement, efficiency improvement, or increased domestic resource mobi- lization, the model generates separate results in each area. For each country, these are summarized in an analytical table that shows the hypothetical financing gap under each of the three sets of policy measures. These disaggregated financing gaps are hypothetical because in reality it would be ill-advised as well as unrealistic to try to implement one or another of the scenarios in isolation. First, there are clear interaction effects among these different reforms which demand that, from a tech- nical standpoint, actions be taken concurrently. For example, it would be very dif- ficult to achieve reductions in the repetition rate (an efficiency reform) in the absence of actions to improve quality (lowering class size, increasing spending on textbooks and teacher training, and so forth). Second, from a financial standpoint it would be completely unsustainable to implement quality reforms that generate high incremental costs (such as an increase in average teacher salaries) without the 82 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD key counterbalancing efficiency reform of an increase in the pupil-teacher ratio, where that is below the target. Nonetheless, presenting the results in this format helps to clarify the impact of the various reforms needed in a specific country con- text. For each country, the "status quo" parameters are contrasted with: Scenario A (quality reform): the change in parameters and resulting annual financing gap when only quality measures are implemented Scenario A +B (quality plus efficiency reforms): the change in parameters and resulting annual financing gap when both quality and efficiency measures are implemented Scenario A+B+C (quality efficiency, andfinancing reforms): the change in parameters and resulting annual financing gap when quality and efficiency and systemfinancing parameters are adjusted. The model also generates a second analytical table for each country that com- pares the total cost estimates for reaching the MDG target for that country with the potential financing sources. Recurrent and capital costs are presented sepa- rately. Potential financing sources are also disaggregated, as follows: * Domestic resources: recurrent, capital, and total * Gap for externalfinancing: recurrent, capital, and total. The simulations assume that all of the recurrent costs in each country case will be covered as much as possible by domestic resources. Only if they exceed domes- tic financing are the remaining recurrent costs presented as part of the gap for external financing. This is in recognition of the fact that donor assistance is more often channeled to capital costs than to recurrent budget support. Finally, for the African countries, the special incremental costs that can be attributed to the impact of AIDS in these countries are presented as a separate line item. We were not able to obtain sufficient data to extend the AIDS analysis to all countries in the sample, but this could certainly be done in the future. The full set of analytical tables for the 47 low-income countries modeled is pre- sented in annex C. The results for four sample countries are discussed below in order to demonstrate how the financing estimates are generated, to show how the indica- tive framework can serve as a diagnostic tool for countries seeking to accelerate MDG progress, and to highlight some of the important limitations of this exercise. INDIA The largest country in the sample is India, with approximately 100 million chil- dren of primary school age. Although the officially reported gross enrollment ratio in 1999 was 100 percent, it is commonly estimated that 20-30 million primary- age children in India are not in school. By these estimates, India alone accounts for as much as one-quarter of the estimated 113 million children worldwide not attending primary school-or one-quarter of the global challenge of achieving Education for All. The proxy primary completion rate we estimated for India (76 percent in 1999) confirms that about 25 percent of children do not complete the five-year primary cycle. Current data on the schooling profile are not available, but a reasonable estimate is that roughly 5-10 percent of children never enter pri- mary school (mainly rural children, scheduled tribes, scheduled castes, and girls), CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 83 and of those who enter school, about 20 percent drop out before completing, pro- ducing a primary completion rate on the order of 76 percent. This is broadly con- sistent with national survey data showing that about 20 percent of children aged 6-10 are not attending school (World Bank 2002e). Survey data also suggest that either the official 100 percent GER estimate is overstated, or that primary school repetition is higher than officially reported and there are a significant number of overage children in the primary schools. For the purposes of our India simulation, we assumed the latter. There are clearly areas of classroom overcrowding in India and small population areas where schools are not available. However, overall it appears that the Indian gov- ernment's policy commitment to site a school within one kilometer of every community with more than 300 people has achieved a very high degree of access to primary schooling. The greater issue is the high dropout rate before completion due to low schooling quality and high household demand for child labor. Although a 76 percent completion rate was sufficient to place India among the Group 1 countries we used to estimate the target parameters (see chapter 3), it is clear that India, at the low end of that group, still has a substantial way to go to meet the MDG target of universal primary completion. The primary com- pletion rate increased over the 1990s from an estimated 70 percent3 to 76 per- cent, which is undeniable progress, but this trend rate (about 0.9 percentage point per year) would put India's PCR at only 90 percent in 2015. More encouraging are household survey data that indicate strong progress on gender equity, with the share of rural girls aged 6-10 enrolled in primary school increasing from 55 to 75 percent between 1993 and 1999, a truly remarkable achievement. The continuing challenge for a very large, ethnically diverse, and federal country such as India is both to accelerate overall primary education progress and to ensure that gains are evenly distributed across a highly decentralized and, as of today, unequal education system. The service delivery and financing parameters we focus on vary considerably across different states and districts in India. Teacher salaries, for example, are negotiated at the state level in India, but are pegged to national benchmarks and vary widely in relation to state-level per capita GDP, a phenomenon that has led states such as Rajasthan and Bihar to introduce para-teachers at lower wages. Our simulation, which relies on a single target ratio of salary to per capita GDP, cannot capture the differing degrees of salary adjustment (either upward or downward) that may in fact be needed in many parts of India. Similarly, while the pupil-teacher ratio averaged nationally is 52:1, it ranges from below 30:1 in some states to 59:1 in others. This in effect results in an underestimate by the model of the true number of teachers that 3. Because of discrepancies in official enrollment data, we estimated the 1990 completion rate for India on the basis of data from the National Family Health Survey, rather than official enrollment statistics, as the latter produced a value that India experts considered artificially high. The 1999 completion rate is calculated according to our standard methodology. 84 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD would be required to achieve the target ratio of 40:1 in all parts of the country by 2015 if, in reality, teachers cannot be redeployed or students reassigned across states. Tables 4.3 and 4.4 summarize the simulation results. Under the quality enhancement simulation generated for India (scenario A), the key actions would be the hiring of additional teachers to reduce the number of pupils per teacher from 52 to 40 by 2015; a slight increase in average teacher salaries, from 3.4 to 3.5 times per capita GDP; and a substantial increase in spending on inputs other than teacher salaries, which would increase from 23 to 33 percent of the recurrent pri- mary budget (table 4.3). The combined effect of the three quality adjustments and the impact of growth on factor costs is that per-pupil spending almost triples over the period to 2015 in constant dollars (from $35 in 2000 to $101 in 2015), with a particularly large increase in spending on complementary inputs to improve school quality, such as books, materials and system management, and possibly demand- side subsidies to target populations. Under the quality plus efficiency enhancement scenario (A+B), the major change would be policy actions to reduce repetition gradually over the period from the starting level, estimated at 20 percent based on data from states and districts, to the target of 10 percent. This improvement in system internal efficiency would "finance" some of the costs of quality improvement, reducing the estimated financ- ing gap. However, financing gap estimates generated by this model for India should be interpreted carefully and taken as an indicative exercise only. While it may give an overall sense of the broad policy priorities and the direction of change required, a modeling exercise based on national average indicators may significantly underesti- mate the true costs of attaining the indicative framework parameters, since in a decentralized education system, quality "surpluses" in some parts of the country that push up the national averages are not transferable to lower-quality states and districts. An excess of classrooms in Kerala in reality will not reduce the need to build more schools in Bihar, but in a simple simulation model such as the one we used, this is effectively what happens. Virtually every country in the sample faces similar equity issues in service delivery and financing across different subregions, but in federal education systems these issues are harder to resolve both in principle and in practice. Although it may not be easy politically, in a unitary system disparities across regions in the pupil- teacher ratio or spending allocations can be managed with teacher redeployment and adjusted allocation rules. In federal systems, the scope for such administrative redeployment and/or fiscal redistribution typically does not exist, at least in the short term. Reforms of the "rules of the game" in these areas at the federal and state levels may take years to negotiate and enact. Thus, the India simulation underestimates the true magnitude of reforms and new investments needed in some subnational entities in order to bring service deliv- ery quality and efficiency in all parts of the country to our benchmarks. More precise estimates for India should be developed through modeling exercises at the subna- tional level and aggregation of these resource gaps. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 85 Table 4.3 ............................................................................. India: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES * * * Average Annual G*vernment Revenuesc Primary Private * . . Spendingon TeacherSalary(as * Education Enrollments Annual Pupils per Inputs Other * Multiple of Per Average As% of . for . Recurrent (As % of Financing Policy Scenarioa Teacher than Teachersb Capita GDP) Repetition Rate GOP Education Spendinge Total) Gap" Status quo 52 23.2% 3.4 20.0% 21.2% 12.4% 32.1% , 12.5% 146 A only 40 33.3% 3.5 2,470 A + B 40 33.3% 3.5 10.0% . 1,782 "Best C1 40 33.3% 3.5 10.0% 16.0% 20.0% % 420% 100 . 67 practice: C2 16.0% 20.0% 67 A + B + . C3 . * * 21.2% 20.0% ' * _._ 28 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice" b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. A simulation exercise at the national level is, however, important on the resource mobilization side, given the crucial role of federal authorities in assuring fiscal equity across decentralized entities. In this case, it shows that India's fiscal parameters are quite far from the targets. While government revenues as a share of GDP (21.2 per- cent) exceed the 16 percent target we set for countries with its level of per capita GDP, spending on education is very low compared with the indicative benchmarks. The share of the consolidated state and federal recurrent budget spent on education in 1999, at 12.4 percent, is substantially below our target value of 20 percent. The share of education spending allocated to the primary level, 32.1 percent, is also low, compared to the target of 42 percent for a five-year primary system.4 Thus, even though the tax-GDP ratio in India is quite high, the combined effect of low alloca- tions for education in general and for primary education in particular is that public spending on primary education in India amounts to only 1.0 percent of GDP, compared to 1.7 percent of GDP for high-completion countries. Thus, the indicative framework points to insufficient allocation of public resources as a root cause of India's incomplete primary education coverage and the quality problems that lead a relatively high share of children to drop out before completing primary school. The United Nations/World Bank population projections for India show fertil- ity declines after 2010 resulting in a stable primary school-age population between 2000 and 2015. This helps to ease the financing requirements for meeting the MDG substantially. But to lower the pupil:classroom ratio and improve school quality, the simulations indicate that India will need to spend at least $435 million per year on classroom construction. There may also be significant short-term needs for upgrading and rehabilitation of existing schools and core system infrastructure that are not captured in our simulations, for India or for any other country. The simulation exercise also points to a need for increased recurrent expendi- tures to achieve universal primary completion-with considerably higher per- student spending on additional teachers, books, better system management, demand-side stipends, and other inputs. Under the quality and efficiency scenar- ios we model, these improvements in school quality and policy actions to reduce repetition effectively produce a steady improvement in the efficiency of student flows. The private share of enrollments also declines slightly, from 12.7 percent currendy to 10 percent by 2015. Although the total financing needed over the period is close to $8 billion per year, the simulation indicates that if India's resource allocation for education and for primary education were to increase to the target values by 2015, domestic resources could cover the great bulk of these needs (table 4.4). However, a financing gap of between $200 and $500 million per year would remain, concentrated in the early years of the period (see annex table A.8). Annualized over the entire projection period, India's external gap would be $67 million per year. 4. In some states, the official primary cycle is still only four years, so this share could legitimately be slightly lower. However, since the government has established eight years as the official duration of compulsory schooling, it is reasonable to analyze system completion rates and costs through grade 5. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 82 co Table 4.4 ................................................................................................................................................................ India: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 constant U.S. dollars) FINANCING SOURCES * . . . . GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost : Resources *.:. Item Period Scenario Mobilized . Recurrent Capital Total Recurrent * Capital Total Recurrent Capital Total Cumulative, Cl 115,278 109,754 6,525 116,279 109,754 5,524 115,278 0 1,001 1,001 2001-2015 C2 115,278 . 109,754 6,525 116,279 109,754 5,524 115,278 0 1,001 1,001 c: ____*_. C3 115,858 109,754 6,525 116,279 109,754 6,104 115,858 0 422 422 Annual Cl1 7,685 7,317 435 7,752 7,317 368 7,685 0 67 67 C2 7,685 7,317 435 7,752 7,317 368 7,685 0 67 67 L> * . C3 7,724 7,317 435 7,752 7,317 407 7,724 0 28 28 ,~ t *.Annual C1 0 0 0 0 0 0 _ o * C2 0 0 0 0 0 C3 0 0 0 0 0 0 Annual C1 7,317 435 7,752 7,317 368 7,685 0 67 67 o E . * C2 7,317 435 7,752 7,317 368 7,685 0 67 67 C3 7,317 435 7,752 7,317 407 7,724 0 28 28 Note "Best practice" policies refer to the combination of scenarios A + B + C. Shaded cells denote no change from values directly above. The model generates a single target for domestic revenue mobilization for pri- mary education, but the financing of primary education in India is a concurrent cen- tral government and state responsibility, with only about 12 percent of total spending at the central government level. In a decentralized fiscal context such as this, the indicative framework financing targets can in effect only be achieved if: (a) states currently spending below the indicative targets are able to increase their spending; (b) the central government increases its spending on education and transfers the increment to the neediest states; or (c) states with higher fiscal capacity mobilize increased resources and the central government develops the fiscal intermediation capacity to equalize education resources across states. Analyzing the fiscal and political feasibility of these or other options in the Indian context is beyond the scope of this study. But the case of Brazil may be instruc- tive. After decades of severe disparities in education spending and quality in a context of decentralized education financing, Brazil made major national strides after a 1997 constitutional reform set an equal per-student funding floor for primary education across the country (box 2.2) and redistributed resources across states and municipal- ities accordingly. Our simulation for India points to reform of primary education finance as a key issue for the country in order to achieve the MDG by 2015. Finally, it should be noted that our simulation results for India are significandy different from the estimates for India in an earlier World Bank financing gap cal- culation for the education MDG (Devarajan, Miller, and Swanson 2002). In that estimate, India accounts for more than $2 billion of an estimated global financing gap of $ 10 million to $15 million per year. But that estimate assumed no change in unit costs or system efficiency, and made no assumptions about the capacity for domestic resource mobilization The three hallmarks of our approach-adjusting for population trends, examining the potential for efficiency improvements, and, most importantly, establishing an expectation that scarce donor assistance will not substitute for an appropriate commitment of domestic resources to EFA-gready affect the size of our estimated financing gap for India and a number of other countries in our sample. PAKISTAN As with India, official education statistics for Pakistan are limited and present internal inconsistencies. In Pakistan's case, though, it is clear that a significant share of children do not today have access to primary school: of the estimated 19.2 mil- lion children in the school-age population in 2000, only 12.5 million are enrolled, for a GER of 65 percent. Although data are sketchy, we estimate a completion rate in 2000 of around 59 percent. Pakistan also has a school-age population that is projected to continue to grow, and would be about 15 percent higher in 2015 than in 2000. As table 4.5 shows, service delivery in the public system in Pakistan is currently far from our benchmarks. Although average teacher salaries, at 3.6 times per capita GDP, are relatively close to the target, spending on inputs other than teacher salaries, at 19 percent of recurrent spending, is well short of the 33.3 percent target. Teacher salaries would decline slighdy and spending on other inputs would rise substantially under the quality improvement scenario for Pakistan. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 89 0 Table 4.5 ................................................................................. Pakistan: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES * * * Average Annual * Government Revenuesc Primary Private * * Spending on Teacher Salary (as . Educatfon Enrollments Annual Pupils per Inputs Other Multiple of Per . Average As % of % for Recurrent (As % of Financing Policy Scenarioa Teacher than Teachersb Capita GOP) Repetition Rate GDP Education Spendingd. Total) Gape Status quo 32 19.3% 3.6 6.2% 16.7% 10.2% 51.8% 29.4% 285 A only 32 33.3% 3.5 450 A + B 40 33.3% 3.5 6.2% 261 'Best C1 40 33.3% 3-5 6.2% 16.0% 20.0% 42.0% 10.0% 204 practice: C2 16.0% 20.0% 204 A + B + C3 . * e , 16.7% 20.0% . * _*._ 173 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A + B + C is considered "best practice". b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. The reported pupil-teacher ratio, at 32:1, is quite low compared to the bench- mark of 40:1. It is unusual to see such a low pupil-teacher ratio in a country that has not reached universal primary coverage, but there has been a significant shift of enrollments to private schools in Pakistan in recent years, driven by the erosion of quality in the public system. While about 29 percent of total primary enrollments are now in fully private-that is, privately financed-schools, it appears that employment in the public schooling sector has not experienced any corresponding retrenchment. The share of privately financed primary enrollments in Pakistan today is quite extraordinary for a low-income country and means that the subsec- tor is serving far more than the elite in Pakistan. Under the quality plus efficiency (A + B) scenario, Pakistan would steadily increase the pupil-teacher ratio in the public system to 40:1 by 2015. This could in effect be accomplished in two very different ways. One route would be a substan- tial improvement in the management and quality of the public schools that would provoke a spontaneous shift of students back to the public system. But a second possible route would be a greater reliance on privately managed schools for service delivery, given evidence of these schools' higher efficiency. Under the latter route, the government would transfer capitation grants to nonprofit private schools that serve low-income populations. Even though the services would remain privately managed, by our definition the enrollments in these schools would be "public." The rationale for public subsidization is that as Pakistan, like other countries in our sample, moves to increase its primary comple- tion rate, it must reach increasingly poor, rural, and disenfranchised populations. Expecting these groups to finance the full cost of their primary education is not only inequitable, it would also, we know from research, impede the attainment of the goal. It is important to note that this assumes no subsidies for private schools serving the highest-income students, so there would still remain a fully private (that is, privately financed) sector accounting for 10 percent of total primary enrollments in 2015. As table 4.5 shows, even under the status quo case, Pakistan would have a resource gap of $285 million annually, because of the projected increase in the school-age population and the fact that not all children are enrolled today. The table also shows that if Pakistan were to implement the quality measures only, this financing gap would swell to $450 million per year. However, if the complemen- tary efficiency measure of an increase in the pupil-teacher ratio were also achieved, the gap would be lowered to $261 million per year. On the resource mobilization side, Pakistan's ratio of government revenues to GDP is currendy at the target value of 16 percent. However, the share of the budget going to education, about 10 percent, is far below the 20 percent target. And, even though spending on primary education-52 percent of the education budget in Pakistan-is above the target of 42 percent for a five-year primary sys- tem, the net effect of these patterns is a low ratio of education spending to GDP, at 0.9 percent currently. Thus, when the indicative targets for domestic resource mobilization are intro- duced in the simulations, Pakistan's spending on primary education increases as a share of GDP. But an estimated financing gap of $204 million per year would CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 91 N Table 4.6 ...........................................................................................I......................................................................... Pakistan: MDG-2015 Cost Estimates and Sources of Financing under 'Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 constant U.S. dollars) FINANCING SOURCES . . .Domestic COST OF . GAP FOR EXTERNAL * * . Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost . Resources .. Item Period Scenario * Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cumulative, Cl 15,919 16,748 2,224 18,972 15,919 0 15,919 829 2,224 3,053 2001-2015 * C2 15,919 16,748 2,224 18,972 15,919 0 15,919 829 2,224 3,053 _______*._. C3 16,373 16,748 2,224 18,972 16,373 0 16,373 375 2,224 2,599 Z; u .Annual C1 1,061 1,117 148 1,265 1,061 0 1,061 55 148 204 C2 1,061 1,117 148 1,265 1,061 0 1,061 55 148 204 C3 1,092 1,117 148 1,265 1,092 0 1,092 25 148 . 173 ,~ = .Annual Cl 0 0 0 0 0 0 2 Q O-. *. C2 0 0 0 0 0 0 C3 0 0 0 0 0 0 Annual Cl 1,117 148 1,265 1,061 0 1,061 55 148 204 E C2 1,117 148 1,265 1,061 0 1,061 55 148 204 C3 1,117 148 1,265 1,092 0 1,092 25 148 173 Note "Best practice" policies refer to the combination of scenarios A + B + C. Shaded cells denote no change from values directly above. remain, with about three-quarters of this total needed for new classroom construc- tion (table 4.6). ARMENIA Of the low-income countries in Europe and Central Asia in our starting sample, only three appear not to have achieved universal primary completion (through sixth grade)-Armenia, Georgia, and Moldova-although data for all countries in this region are considered problematic. According to official statistics, Armenia's gross enrollment ratio in the year 2000 was over 100 percent, but the completion rate through sixth grade was below 70 percent.5 The core service delivery parameters in Armenia, Georgia, and Moldova all deviate sharply from the benchmarks in a pattern that is common to ECA coun- tries: the number of teachers employed (relative to the student population) is far higher than in other countries and the average teacher salary is far lower. In Arme- nia, the average teacher salary would increase dramatically, from 0.6 to 3.5 times per capita GDP, as a quality measure in the simulation. But as a corresponding effi- ciency measure, the current 13:1 pupil-teacher ratio would rise to 40:1, also a tremendous adjustment-even if phased in gradually over a 15-year period. Given no projected growth of the school-age population, the clear implication is that the number of teachers employed would decline gradually but significantly over the period. Although this could probably be handled through attrition and selective retrenchment linked to the introduction of new certification or performance stan- dards, the management challenge would be significant. The realism of such a dra- matic increase in teacher salaries must be considered questionable as well, in a broader civil service context, and in light of current fiscal pressures in Armenia (recall, however, that our approach assumes that incremental external resources will finance any resulting gaps). But the simulation serves to illuminate the root causes of Armenia's key educational issues: excess staffing; inadequate salaries leading to low teacher motivation, absenteeism, and informal shifting of costs to families; and high operating and maintenance costs for an inefficient number of schools and classrooms, which divert resources from other needed areas such as modernization of curriculum and learning materials, teacher retraining, and system management. On the resource mobilization side, the simulations show that in addition to an inefficient pattern of spending, Armenia's level of spending on primary education as a share of GDP is inadequate. Armenia's revenue-GDP ratio is close to the tar- get, but the budget share for education, at 15 percent, is below the 20 percent benchmark. This is the principal financing variable that would adjust in the simu- lation, as the share of education spending going to the primary level is already at the target. The upward adjustment in the budget share for education by 2015 generates significant additional resources. Nonetheless, given the magnitude of the spending increases required for increased teacher salaries, the model projects a financing gap 5. Armenia's primary cycle is officially three years, but for the purposes of this simulation, we adjusted values for all countries to the equivalent of a six-year primary system. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 93 Table 4.7 ................................................................................ Armenia: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES . C: FINANCING MEASURES * * * Average Annual G Government Revenuesc Primary Private * . Spending on Teacher Salary (as Education Enrollments * Annual Pupils per Inputs Other Multiple of Per Average . As% of %for Recurrent (As% of Financing Policy Scenarioa Teacher than Teachersb Capita GDP) Repetition Rate GDP Education Spendingd Total) Gape Status quo 13 52.9% 0.6 0.1% . 15.8% 15.1% 51.3% 0.0% 0 A only 13 33.3% 3.5 61 A + B 40 33.3% 3.5 0.1% 12 'Best C1 . 40 33.3% 3.5 0.1% 16.0% e 20.0% 50.0% 10.0% 15 practice": C2 16.0% 20.0% 15 A + B + C3 . * * . 16.0% 20.0% . . 15 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (C1, C2, and C3). The combination of scenarios A + B + C is considered 'best practice". b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. of about $15 million per year over the period, notwithstanding efficiency gains from the huge increase in the pupil-teacher ratio that is assumed. All of the gap would be for recurrent financing, given the projected slow growth of the primary school-age population and-in relation to the simulation benchmarks-the tremendous scope for absorbing increased enrollments into the existing physical plant. Obviously, even where the numbers of students and classrooms appear theo- retically in balance, there will still be infrastructure needs-for school consolida- tion, rehabilitation, and even additional classroom construction in response to migration. It is impossible to factor these into the relatively simple model we used. Nonetheless, a result such as this signals that the bulk of the incremental costs of reaching the MDG in Armenia will be of a recurrent nature, rather than costs for expansion. As in other countries, however, capital costs for system rehabilitation, which we could not capture, may be significant. The year-by-year financing projections show that Armenia's needs for external financing would be heaviest in the first years of the simulation period and would decline to zero by about 2008, as the reforms to improve efficiency and to increase domestic financing take hold. Thus, although Armenia's financing gap does not appear large when averaged over the entire 2000-2015 period, for the initial five years the need for external financing would be significant, averaging about $40 million per year. This is because our simulation assumes that actions to improve the level and structure of teacher remuneration are introduced as early as possible, in order to maximize the potential impact on school quality and also to reduce teachers' informal demands on households for support, which may con- strain student attendance. While the teacher compensation reform is assumed to be implemented quickly, however, achieving school consolidation and increases in average class size and mobilizing increased domestic financing for primary educa- tion would take longer. As annex table A.8 shows, the simulations produce a similar pattern of rela- tively high needs for external financing in the initial years for the other two ECA countries in our sample, Georgia and Moldova. Georgia's gap averages about $35 million per year and Moldova's about $15 million per year up to 2005. In all cases, the gap begins to decline thereafter and is eliminated by 2009. The simulations for the ECA countries in our sample provide a framework for identifying the key reform directions and tradeoffs facing these countries. They also indicate that the needs for external financing in this region are likely to be larger at the outset of the period than later. But the simulation benchmarks are so far from the institutional and political reality of these countries that the financing gap estimates for these countries should be interpreted as illustrative only. It should also be recalled that we modeled the costs of achieving universal primary comple- tion through six grades of schooling, rather than the shorter cycle these countries actually have. More detailed modeling is needed to cost reform trajectories toward targets that are more realistic for these countries-such as average class size in the range of 25-30 and average teacher salaries in the range of 2-2.5 times per capita GDP, and three- or four-grade primary completion. And while, recalling box 3.1, these alternative parameters would have a neutral effect on the overall financing gap, they would lower the size of the gap in the initial years of the simulation. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 95 Table 4.8 .................................................................................................................................................................... Armenia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 constant U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL *Domestic COST OF MDG-201 5 DOMESTIC RESOURCES . FINANCING Cost . Resources . Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cumulative, C1 491 712 0 712 491 0 491 221 0 221 2001-2015 C2 491 712 0 712 491 0 491 221 0 221 _____* * C3 . 491 . 712 * 0 . 712 491 0 491 221 0 221 Annual C1 33 47 0 47 33 0 33 15 0 15 C2 33 47 0 47 33 0 33 15 0 15 C3 33 47 0 47 33 0 33 15 0 15 -r .Annual C1 0 0 0 0 0 0 C2 0 0 0 C3 0 0 0 0 0 0 Annual C1 47 0 47 33 0 33 15 0 15 0 E *. C2 47 0 47 33 0 33 15 0 15 C3 47 0 47 33 0 33 15 0 15 Note "Best practice" policies refer to the combination of scenarios A + B + C. Shaded cells denote no change from values directly above. NIGER The primary completion rate in Niger, at 20 percent in 2000, is one of the lowest in the world. Only one child in three in Niger is enrolled in primary school (GER 33 percent), and only one child in five completes it-and among girls the comple- tion rate is one in 10. Niger, Mali, Chad, Ethiopia, and a handful of other Sub- Saharan African countries are unquestionably the setting for EFA's greatest challenges. For these countries to meet the MDG, education system expansion and simultaneous quality improvement will need to occur at a pace no country has ever seen. In the case of these very low completion countries, in fact, there is a clear question of the realism of our simulations, which model the resource requirements of a trajectory of system expansion leading by definition to 100 percent comple- tion in 2015. Even if financing were unconstrained, this trajectory may simply be physically and institutionally impossible. But the simulation exercise serves to illu- minate the issues that Niger, and several other African countries in our sample, must confront. Niger's incipient education system is relatively close to the normative targets in a few areas: the pupil-teacher ratio is 37:1 instead of 40:1, and repetition averages 13 percent instead of 10 percent. Under the quality scenario, the share of recurrent spending on items other than teacher salaries would increase from 26 percent cur- rently to the target of 33.3 percent (table 4.9). The underlying driver of Niger's inability, in the 40 years since independence, to offer all children a primary education is found in the very high unit cost of public primary education, due to extraordinarily high average teacher salaries-9.6 times the per capita GDP, compared with the benchmark of 3.5 times per capita GDP. Recognizing this, the Ministry of Education launched a bold reform in 2000 to suspend the recruitment of civil service teachers and to establish a new cadre of contract teachers at a more sustainable salary rate. Since 2000, the pace of teacher recruitment and enrollment expansion has accelerated tremendously. Like Senegal and other countries that have adopted this approach, Niger has been able to recruit new teachers with the same level of formal qualification as the existing force, and in fact has had an excess supply of candidates. In the simulation, we model the expan- sion of the new teacher cadre at the target salary of 3.5 times per capita GDP, while providing for gradual attrition of the higher-paid teaching force. Reflecting this mix, by 2015 the average teacher salary in Niger declines to a value of 4.3. The other issue for Niger can be seen from the resource mobilization targets. Unlike many countries in the sample, Niger is currently making a huge fiscal effort in support of primary education: the government allocates 31.5 percent of domes- tic resources to education and 62 percent of the education budget to primary edu- cation, both values well above the benchmarks. But Niger's slight resource base and undeveloped economy are clear in the very limited ability of the country to mobi- lize tax revenues-only 9.1 percent of GDP-compared with the target of 14 per- cent for Niger's level of GDP. Our simulations require that Niger gradually increase government revenues to reach 14 percent of GDP by 2015, which may be difficult to achieve. At the same time, though the country's strong fiscal commitment to EFA is laudable, the 31.5 percent budget share for education cannot be considered sustainable, and the 62 percent allocation to primary education will also need to CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 92 S Table 4.9 ............................................................................. Niger: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenuesc Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils per Inputs Other Multiple of Per Average As % of % for Recurrent (As % of Financing Policy Scenarioa Teacher than Teachersb Capita GOP) Repetition Rate GOP Education Spendingd Total) Gape Status quo 37 25.9% 9.6 13.0% 9.1% 31.5% 62.0% 4.0% 135 A only 37 33.3% 9.6 146 A + B 40 33.3% 4.3 10.0% 48 "Best Cl 40 33.3% 4.3 10.0% . 14.0% 26.0% 50.0% 10.0% 46 practice": C2 * 14.0% 20.0% . 53 C3 . 14.0% 20.0% . . 53 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A + B + C is considered "best practice". b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table 4.10 ................................................................................................................................................................ Niger: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 constant U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL *Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost . Resources . . : Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cumulative,, C1 796 1,078 403 1,481 796 0 796 282 403 685 2001-2015 C2 684 1,078 403 1,481 684 0 684 394 403 797 c ______ ._ C3 684 1,078 403 1,481 684 0 684 394 403 797 Annual C1 53 72 27 99 53 0 53 19 27 46 C2 46 72 27 99 46 0 46 26 27 53 . 'u l C3 46 72 27 99 46 r0 46 26 27 . 53 ,.Annual . C1 * , I ..O.1* cl 0, * C2 1 1 O 0 1 1 C3 11 0 0 Y Annual C1 73 27 100 53 0 53 20 27 47 0o E * C2 73 27 100 46 0 46 28 27 54 C3 73 27 100 46 0 46 28 27 54 Note "Best practice" policies refer to the combination of scenarios A + B + C. Shaded cells denote no change from values directly above. decline, as primary completion increases and a larger share of children reach sec- ondary and tertiary levels. Thus, about 50 percent of the $99 million annually that Niger would need to spend on the trajectory of accelerated progress we model would need to come from external financing sources. About $27 million of the total would be for school con- struction, which we assume could be completely financed by donors. But Niger would need an equal amount from the outside world in recurrent budget support (table 4.10). Given Niger's relatively low HIV seroprevalence, the impacts on the system from HIV/AIDS (to replace sick teachers and provide support to maternal or dou- ble orphans) would add only an estimated $1 million per year to recurrent costs. But this amount would also need to be externally supported. In sum, Niger is a clear case of a country where, even with maximum domestic resource commitment and reform progress on the key issue of teacher salaries, a significant financing gap remains. The absence of external support would place a binding constraint on Niger's progress toward universal primary completion. The fact that the country has already started on the path of reform toward the indicative framework bench- marks suggests that Niger is a case of a country with a "credible plan" for EFA. AFGHANISTAN The reconstruction of primary education in Afghanistan will be a massive chal- lenge for the country and international partners over the coming decade. It was impossible to obtain sufficient public finance data to carry out a full simulation for Afghanistan. Population, education enrollment, and service delivery data are also scarce, outdated, and inconsistent. Nonetheless, because Afghanistan's needs will clearly contribute to the global costs and external financing requirements of achieving EFA, we used the target parameters for service delivery to try to estimate what the order of magnitude of resource requirements for Afghanistan would be in the same framework. The last population census for Afghanistan was conducted in 1978 and all published population statistics since then are extrapolations of these census data. However, these are based on a wide variety of assumptions and lead to population estimates between 20.6 and 26.9 million. We used an estimate at the midpoint of this range. We estimated the total resource cost of achieving 100 percent completion in Afghanistan by 2015 based on the following assumptions: * Repetition rate over the planning horizon is 10 percent (no baseline data available) * In the absence of salary data, figures for average teacher salaries in Pakistan were used * Construction costs for a fully equipped classroom are $6,500 (constant 2000 dollars), the average we used for other countries in South Asia * Non-teacher-salary expenditures are 33 percent of recurrent expenditure on primary education * The pupil-teacher ratio is 40:1. 100 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 4.11 ..................................................................... Possible Costs of Achieving UPC in Afghanistan, 2000-2015 (millions of 2000 constant U.S. dollars) Period Recurrent Costs Capital Costs Total Cumulative 613 827 1,440 Annual 41 55 96 Under these assumptions, estimated resource requirements to rebuild and improve education service delivery in Afghanistan during 2000-2015 would be as shown in table 4.11. In the absence of fiscal data, we made no assumptions about the scope for domestic resource mobilization and instead present these as the total requirements, which for the next several years at least may well need to be financed almost entirely from donor sources. Even with the need to reconstruct a largely defunct school system, however, it should be noted that about 40 percent of the require- ments will be for recurrent budget support. AGGREGATE RESULTS The simulation results were aggregated into an estimate of the average annual financing gap for this set of 47 low-income countries (table 4.12). Scenario C2 (the base case) shows the highest financing requirements, reflecting the reduction of higher spending levels we imposed for some countries on grounds of sustain- ability. Under this scenario, a total of $2.4 billion per year in external financing would be required to meet the MDG target by 2015 in only the 47 countries stud- ied. If the resource requirements for Afghanistan, estimated using a different methodology, are included, the total for these 48 countries would be about $100 million per year higher A discussion of how the aggregate results were constructed from the different policy scenarios follows. Table 4.12 shows that under Scenario A, which introduces only quality improvement measures, a financing gap of about $7.5 billion per year is generated for the sample as a whole, close to $2 billion of which is for India. The reduction of this gap to $4.3 billion per year under Scenario A + B shows the impact of introducing efficiency measures, such as increasing class size where it is currendy below 40 and reducing repetition where it is above 10 percent. When the target parameters for domestic financing are also introduced, the gap is cut roughly in half, to about $2.1 billion per year under scenario C2. Under scenario Ci, which maintains higher spending shares for primary education in those countries cur- rently spending more, the financing gap is slightly lower, about $2 billion per year. The financing gap shrinks significandy under scenario C3 to less than $1.6 billion per year. Scenario C3 maintains higher overall government spending in those countries whose tax-GDP ratio currently exceeds our targets. Maintaining this level of domestic resource mobilization would clearly aid in reaching the CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 101 Table 4.12 ......................................................................................... All 47 Countries: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues . Primary Private Spending on Teacher Salary (as . Education Enrollments Annual Pupils per Inputs Other Multiple of Per Average As % of % for Recurrent (As % of Financing Policy Scenarioa Teacher than Teachersb Capita GDP) Repetition Rate GDP Education Spendingd Total) Gaap Status quo 13-79 0.1-45 0.6-9.6 0-36% ! 8-56 1.4-32.6 26-66 0-77 A only 36 33.7% 4.5 7,489 A + B 40 33.3% 3.8 8.2% 4,348 "Best C1 . 40 33.3% 3.7 8.2% 15.2% 21.1% 48.6% 10.0% 2,033 practice": C2 15.2% 20.0% 2,151 A + B__.C3 . *_*_ *_ 20.3% 20.0% * _*1,563 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A + B + C is considered "best practice". b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. MDG, and in countries such as Angola, Nigeria, and the Republic of Yemen, where domestic revenue mobilization is currently in the range of 35-50 percent of GDP and greatly exceeds our 18 percent maximum target, scenario C3-or some- thing closer to it-may be a more appropriate fiscal scenario than the C2 scenario. However, since even in these countries it is not clear that current levels of public revenue mobilization can be sustained over the long period we analyzed, we focus our discussion on the results under scenario C2. Table 4.13 compares these cost estimates to potential sources of financing. As can be seen, under scenario C2 the total cost of achieving universal primary com- pletion in these 47 countries would average about $16 billion per year over the period. About 90 percent of this cost ($14.8 billion per year) would be recurrent. Incremental school construction requirements for the sample would be about $1.5 billion per year. Under the indicative parameters, these countries would increase their domestic resource mobilization for EFA from about $8 billion-$9 billion in 2000 to $23 bil- lion per year by 2015. Averaged over the period, domestically mobilized primary education funding would total about $13.8 billion per year (under scenario C2). We assume that all of the domestic resources are applied first to the recurrent budget requirements. But approximately $1.1 billion per year in recurrent needs would remain uncovered. Since the countries' domestic resources are not adequate to cover their recurrent budget needs, virtually all of the incremental capital costs would need external financing, about another $1.1 billion. (About $0.4 billion of capital costs could be financed in part by a few countries in the sample.) The special exercise we undertook to estimate the impact of HIV/AIDS on MDG attainment in Africa indicated that the additional costs for these education systems could be on the order of $286 million per year. These costs are all of a recurrent nature-for providing subsistence support to maternal and double orphans and for recruiting and training additional teachers. The Sub-Saharan African countries in our sample clearly will be ill prepared to bear these additional costs, so they are added to the gap for external financing. Thus, the total external financing gap for these 47 countries is estimated to average about $2.4 billion per year over the period. Including Afghanistan, it would average $2.5 billion per year. An important finding is that about 55 percent of the external financing needed would be for recurrent budget support, and only about 45 percent for capital sup- port (new classroom construction). Since construction investments are generally easiest for donors to mobilize, we assume that all of the new construction needed in these countries would be financed externally. But the simulations make clear that an even larger volume of external support would be needed for recurrent budget requirements. Under our target parameters, virtually all countries in the sample would increase their domestic financing for primary education, and would finance 85 percent of the total cost of achieving the MDG themselves. But the biggest constraint to achieving the goal will be the availability of external financing for recurrent expenses, not capital. Figure 4.1 provides a graphic picture of these financing requirements to 2015 and their likely evolution thereafter. It should be remembered that underlying this evolution are substantial shifts in education system coverage and quality over the CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 103 Table 4.13 ............................................................................................................................................................................ All 47 Countries: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios FINANCING SOURCES GAP FOR EXTERNAL Cost. Domestic COST OF MDG-201 5 DOMESTIC RESOURCES . FINANCING Cost . .Resources * Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent * Capital Total Cumulative, C1 214,897 222,665 22,728 245,393 208,811 6,085 * 214,896 13,853 16,643 30,496 2001-2015 C2 213,124 222,665 22,728 245,393 207,109 6,014 213,123 15,556 16,714 .32,270 c ______ *.__ C3 234,632 222,665 22,728 245,393 I 212,430 9,521 * 221,951 10,234 13,208 23,442 3 Annual C1 14,326 14,844 1,515 16,360 13,921 406 14,326 924 . 1,110 2,033 C2 14,208 14,844 1,515 16,360 13,807 401 14,208 1,037 1,114 2,151 . C3 15,642 14,844 1,515 16,360 14,162 * 635 14,797 682 881 1,563 Annual C1 286 286 0 0 286 286 c1 ea 0. *. C2 286 286 0 0 285 286 ____._. C3 . 286 . 286 63 . . 63 223 *. *. 286 Annual C1 15,130 1,515 16,646 13,921 406 14,326 1,210 1,110 2,319 . .15,130 1,515 16,646 13,807 401 13,208 1,323 1,114 2,437 C3 15,130 1,515 16,646 13,226 . 635 14,860 . 905 881 1,785 Note "Best practice" policies refer to the combination of scenarios A + B + C. Shaded cells denote no change from values directly above. FIGURE 4.1 Domestic and External Financing Required to Achieve the Education MDG in 47 Countries, 2001-2030 Millions of 2000 constant U.S. dollars 45,000 - 40,000 - 35,000 - 30,000 - 25,000 - 20,000 - 15,000 - 10,000 - 5,000 - _ * * ' ' _ _ ' _ _ _ _. , 2000 2005 2010 2015 2020 2025 2030 _ Total spending on primary education _ Total domestic financing for primary education - Financing gap period. Under our scenario of MDG attainment, these 47 countries increase pri- mary enrollments from 229 million children in 2000 to 301 million children in 2015. The increase is almost entirely driven by progress in Africa, where primary enrollments would almost double, from 71 million to 136 million children. This will be a huge management challenge for education systems that are perceived to be weakly managed today. In the other regions, where starting coverage is higher and the population of school-age children is projected to stabilize or decline, total primary enrollments will barely change, increasing from 159 million to 164 mil- lion by 2015. Total expenditure on primary education in all 47 countries increases in our simulations to about $25 billion per year in 2015, or about $76 per child (in con- stant 2000 dollars). This compares to a starting level of about $32 per student. This increase in unit costs in real terms reflects both the impact of economic growth on factor costs, notably teacher salaries, and the increase in schooling qual- ity required to achieve universal primary completion. Underpinning the increase in quality is an important shift in the composition of spending toward non-salary inputs. Per-student spending on inputs other than teacher salaries triples in real terms over the period. Figure 4.1 shows very clearly that as primary education costs increase, due to expanding enrollments and improvements in quality, the indicative framework tar- gets also require countries to increase their domestic spending on primary educa- CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 0OS tion. These 47 countries' own financing for primary education increases from a base of about $8.5 billion in 2000 to about $21 billion in 2015. Notwithstanding this significant domestic effort, there is a financing gap over the period to 2015, and it reaches a peak in 2014, when total financing requirements approach $25 billion and the financing gap rises to $4 billion. Figure 4.1 also shows that after 2014 the external financing gap will decline, for four main reasons: (a) construction needs will decline sharply from $1.5 billion per year to the lower pace of expansion required by natural growth of the school-age population; (b) the demographic transition expected in most of the countries will cause the share of children aged 7-12 in the population to drop, which, other things equal, will reduce the share of national resources needed to finance primary educa- tion; (c) continued GDP growth will boost the tax-GDP ratio for these countries to levels higher than the targets we assumed; and (d) the secular decline in the level of teacher salaries relative to per capita GDP observed with economic development and expansion of the formal labor market (recall figure 3.3) will set in. Indeed, these dynamic effects would in reality probably affect our target parameters before 2015. On the other hand, the figure shows that the financing gap for these very low income countries will not disappear entirely. By 2015, these countries-most of which are currently very far from the goal of universal primary education-will have achieved a transformation of their primary education systems in terms of quantity, quality, and efficiency. They will also have increased their own financing for primary education and will be on the road to sustainable and self-financing sys- tems. After 2015, the dependence on foreign financing will gradually dedine, from 15 percent of total expenditures to about 3-5 percent. However, it is sobering to realize that the primary education systems in these very low income countries will require some degree of external assistance for a long time to come. The size of that assistance relative to domestic financing varies substantially across the countries in our sample-with the greatest needs, not surprisingly, in the poorest of the Sub-Saharan African countries. External financing needs as a share of total primary education expenditure for the African countries rise from 28 per- cent at the outset to a peak of 36 percent in 2014, before declining to 6 percent by 2030. This higher dependency reflects both the African countries' weaker economies and their greater distance from universal primary education coverage. While the very high dependence of these 33 countries on external support for education over the period to 2015 is troubling, the projections clearly indicate a path toward self-sustaining education systems thereafter (figure 4.2). It should also be recalled that some of these countries may have the capacity to contribute more domestic resources to the financing of primary education than we assumed under the C2 resource mobilization scenario profiled here. Under our alternative fiscal scenario C3, in which higher tax-GDP ratios are retained for countries that are in fact cur- rently above our targets, the financing gap for the African countries is about 25 per- cent lower over the period, averaging $1.3 billion, rather than $1.8 billion per year. Under this scenario, in the peak year of 2014 the financing gap represents 26 percent of total primary education expenditure, rather than 36 percent under scenario C2. The time profile of external financing requirements to attain the MDG merits further analysis. We modeled only one basic scenario, setting 2015 as the date for 106 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD FIGURE 4.2. Domestic and External Financing Required to Achieve the Education MDG in Sub-Saharan Africa, 2001-2030 Millions of 2000 constant U.S. dollars 14,000 - l 12,000 - 10,000 - 8,000 - 6,000 -=~ 2,000' 2000 2005 2010 2015 2020 2025 2030 - Total spending on primary education -6- Total domestic financing for primary education -* Financing gap achievement of universal primary completion in all countries in the sample, irre- spective of their actual trends. For many countries, particularly in Africa, this was an ambitious target. But for some others, it actually forced a slowing of their trend rate of progress. An alternative scenario might set all countries on at least the trajec- tory of "maximum feasible" progress to the goal that has been observed empirically (about 3 percentage points per year increase in the primary completion rate). This would effectively shift forward some of the external financing gap to the years before 2015 and flatten the curve. REQUIREMENTS BY REGION The regional breakdown of the financing gap in table 4.14 shows that, including costs for the impact of AIDS, about 75 percent ($ 1.9 billion) of the external sup- port would be needed for 33 countries in Sub-Saharan Africa. In South Asia, four countries (India, Pakistan, Bangladesh, and Nepal) would need $397 million per year in external funding. In Latin America and the Caribbean, three countries (Nicaragua, Honduras, and Haiti) would need $48 million per year. In East Asia and the Pacific, two countries (Lao PDR and Cambodia) would need $36 million per year. In the Middle East, one country (the Republic of Yemen) would need $70 million per year, and in Europe and Central Asia, three countries (Armenia, Geor- gia, and Moldova) would require about $34 million per year. Only one country in our sample (Mongolia) showed no financing gap in these simulations, largely CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 107 Table 4.14 Estimated Annual Financing Gap to Achieve the Education MD6, by Region (Scenario C2) (millions of 2000 constant U.S. dollars) Latin Percentage America East Asia Middle East Europe and of Total Type of South and the and the and North Central . Financing Financing Africa Asia Caribbean Pacific Africa Asia Total * Gap Recurrent 1,127 97 14 30 21 34 1,323. 55 Operation 841 97 14 30 21 34 .1,037* 43 AIDS 286 . O 0 0 0 286. 12 Capital 725 300 34 6 49 0 1,114 45 Total 1,852 397 48 36 70 34 j2,437 100 Note: Numbers may not sum to totals because of rounding. because the school-age population is not projected to increase and the baseline pupil-teacher ratio is far below the benchmark we used. All results for individual countries may be found in annex C, and the simulation model and baseline data for all countries may be found in the CD-ROM accompanying this book. The financing gap estimated in this study must be understood as a lower-bound estimate of the global costs of attaining the education MDG, for several reasons. First and most crucially, these simulations estimated the incremental costs of expanding primary education systems from the baseline numbers of classrooms and teachers in each of the 47 countries in 2000 to the numbers that would be needed in order to reach the goal by 2015. But they did not capture the important needs- particularly in these very low income countries-for rehabilitation and upgrading of existing classrooms; expansion or upgrading of other system infrastructure such as teacher support centers, district offices, and teacher training institutes; key "soft- ware" investments such as curriculum design, management information systems, student assessment systems, textbooks, and so forth; and training and capacity building for teachers, school directors, and system adrministrators to bring all of them up to an adequate level of functioning. Our data set did not permit the detailed appraisal of existing school-level and system-level infrastructure or the ade- quacy of current system functioning that would be required to estimate the costs of needed upgrading, rehabilitation, and capacity building in each of these 47 coun- tries, in order to complement the expansion costs we estimated. Given the precari- ous functioning of the education system in very many of the countries in our sample, it can be assumed that these needs are substantial. Because these invest- ments are needed immediately, moreover, our simulation results for the first few years of the projection period particularly underestimate the true needs for external financing in these countries. Second, although our 47-country sample includes all of the most populous low-income countries and accounts for 94 percent of all children without access to primary education in the low-income world, there are about 20 small low-income 108 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD countries and several conflict-affected countries that we could not analyze, for lack of data. Moreover, we also only estimated financing requirements through six grades of primary schooling for some countries whose official cycle is longer. A full costing of the external needs in low-income countries would have to include all of the countries and reflect the full length of the primary cycle in each. Third, this costing exercise simulated a path to the MDG for each country that assumed that system reforms would be initiated immediately, and pursued steadily to 2015. In reality, there will be many cases where it is politically impossi- ble to launch all needed reforms at the same time, where the pace of implementa- tion will not always be linear, and where there is a need for the education system to deliver better service immediately, while key reforms-particularly on the resource mobilization side-may necessarily take longer to legislate and implement. To the extent that external assistance may facilitate such processes, transitional external financing requirements may be higher than the simulation estimates. However, the record on aid effectiveness also clearly points to the pitfalls of external assistance as a substitute for country commitment to needed reforms. Finally, this costing exercise focused on the millennium education goal of uni- versal primary completion by 2015, and not on the full set of Education for All goals established at the Dakar conference. Developing countries are committed to pursu- ing all six of the Dakar education goals, and the incremental costs to attain some of them-especially the elimination of gender disparities in secondary education, the achievement of a 50 percent improvement in adult literacy by 2015, and the expan- sion of early childhood care and education targeted to the most vulnerable children- will be significant. The financing framework introduced in this study provides for increased spending on all levels of education, not only primary education, and would provide some fiscal space for education systems to pursue the full set of Dakar goals. But parallel efforts to the current study, to research the "best practice" poli- cies, service delivery parameters, and external financing needed for attainment of the other Dakar goals, are needed for a full costing of the Education for All agenda. ESTIMATING THE GLOBAL COSTS OF REACHING THE EDUCATION MDG Despite these limitations, the current study does represent one of the most careful efforts to date to analyze and cost a strategy for attaining the education MDG. In a world where both developing and developed countries face competing priorities and budget constraints, we insist on the importance of a global strategy-such as the one outlined here-that seeks to achieve the goal at minimum adequate cost, rather than "at any cost." Current patterns of education spending, where they clearly are not producing results today, should not be the basis for MDG costing. "External financing gaps" should reflect a true residual need after sound national policies and resource commitments are in place, and should not substitute for these. Even the conservative estimate put forward here is many times higher than aid flows currently available for primary education; it will take strong effort and commitment from development partners to rnobilize this incremental funding, and equal effort from developing countries to use it well. CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 109 Is it possible, then, to generate a plausible estimate of the likely costs of achiev- ing the education MDG-through five or six years of schooling-in all developing and transition countries, building on the detailed analysis done here for 47 low- income countries?6 We try in this section to do so. Since our cost estimates are for only 47 countries, and cover only expansion costs at the classroom level, "scaling up" our analysis to a truly global estimate requires four steps. (See table 4.15.) First, we need to estimate the system rehabilitation needs for which we were unable to get country-by-country data, at both the classroom level and the system infrastructure level. We obtained more detailed data for a small subset of countries in our sample and found that, on average, about 30 percent of existing infrastructure was estimated in need of replacement or upgrading. The estimated annual cost of this rehabilitation-assumed to be carried out over a three year period-equaled roughly 50 percent of the total primary education recurrent budget over that period. On this basis, the total additional requirement for our 47 countries over the first three years would be slightly over $3.9 billion per year, or $11.6 billion in total. These are one- time expenditures, however. Averaged over our 15 year projection period, they add $0.8 billion per year, or 10 percent, to the estimated annual incremental cost to reach the MDG of $8 billion, and add 33 percent to our estimated annual financing gap. The second step is to estimate infrastructure expansion needs at the system level, since our analysis focused on classroom expansion. We roughly estimate that system infrastructure (teacher resource centers, district offices, central ministry facilities, and so forth) should expand at an equal pace with classroom construction and in general should not exceed 20 percent of the costs of classroom expansion. This would add another $4.6 billion to the $23 billion in capital requirements we estimated for the period to 2015, or $0.3 billion per year. This represents an addi- tional 12 percent increase in the estimated financing gap. The third step is to extend this comprehensive estimate to countries we did not analyze. Our 47 countries account for about 94 percent of the out-of-school popula- tion in low-income countries. Scaling up our estimated incremental costs to cover the total needs for all low-income countries is relatively straightforward, as unit costs and appropriate system parameters are similar. Scaling up from the 47 countries we analyzed to the full group of 79 IDA countries would increase the incremental costs by an estimated $0.6 billion per year, or 7 percent. If we assume the same capacity for domestic resource mobilization in the low income countries as in our sample coun- tries, our estimated financing gap would increase by a further $0.2 billion per year. In sum, we estimate that the incremental cost of achieving the education MDG in all low-income countries, including all needs, would total about $9.7 bil- lion per year over the period to 2015, of which about $3.7 billion per year would need to be financed externally. This is about 50 percent higher than the gap we estimated for the 47 countries in our sample. 6. We obtained detailed data on costs and parameters for five or six years of primary schooling in all countries in the sample, regardless of the official length of the primary cycle. The scaled-up costs estimated in this section similarly correspond to the equivalent of getting all children through five or six years of primary education in all countries. To the extent that countries actually have longer primary education cycles, these costs underestimate the true costs of reaching the education MDG. 110 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD The fourth and final step-projecting the likely costs and financing gaps for the 47 middle-income countries that have not yet reached the MDG-is more dif- ficult. On the one hand, the middle-income countries are much closer to the goal, with an average primary completion rate of 87 percent (on a country-weighted basis), compared with 62 percent for the low-income countries. These more diver- sified economies also have more scope for domestic resource mobilization. With appropriate domestic commitment, the upper-middle-income countries in particular-with a tax-GDP ratio averaging 23 percent compared to 19 percent for our sample-should be able to finance a substantial part of the costs of universal primary education. Demographic factors are also more favorable: the school-age population is a lower, and typically declining, share of the overall population, which makes it easier for a fixed share of national income to cover education system needs. On the other hand, there are several offsetting factors. The share of the overall education budget typically available for primary education is lower in middle- income countries, given higher enrollment ratios in secondary and tertiary educa- tion and funding pressures from these levels. Most importantly, the unit costs of primary education in middle-income countries are much higher in dollar terms, because of lower pupil-teacher ratios and the higher (dollar) costs of teacher salaries, construction, and other inputs. Pupil-teacher ratios in middle-income countries tend to be lower than 30:1 and are often under pressure from teachers' unions to decline further. It would be diffi- cult in most middle-income countries to establish 40:1 as an appropriate target, although high-performing education systems in Singapore and South Korea provide a clear example of how cost-effective such a policy can be. An even bigger factor is the average teacher salary: while teacher salaries in middle-income countries are lower in per capita GDP terms than in our sample countries, they are much higher in dollar terms. The dollar value of non-salary inputs and construction costs is also higher, reflecting the average level of incomes and prices in these more developed economies. Overall, primary education unit costs in middle-income countries are in the range of$ 180-220 per student, or 5-6 times the unit cost in our sample. Table 4.15 .......................................................................................................................... A Global Estimate of the Annual Incremental Costs to Achieve the Education MDG and Likely Financing Gap ESTIMATED ESTIMATED INCREMENTAL COST EXTERNAL GAP (ANNUAL AVERAGE) (ANNUAL AVERAGE) Estimate for 47 countries . $8 billion . $2.4 billion Rehabilitation . $0.8 billion $0.8 billion System expansion $0.3 billion $0.3 billion Other low-income countries . $0.2 billion $0.2 billion All low-income countries $9.7 billion $3.7 billion Middle-income countries . $23-28 billion $1-$3 billion All developing countries $33-38 billion $5-$7 billion CHAPTER FOUR * COSTING THE MDG OF UNIVERSAL PRIMARY COMPLETION 111 Based on current unit costs and enrollment data, but applying population and economic growth projections, we estimate that the incremental costs of reaching the education MDG in the middle income countries would be in the range of $23-28 billion, compared to spending on primary education estimated at about $80 billion in 2000. However, this estimate is not strictly parallel to our estimate for the lower income countries, because it assumes no changes in service delivery efficiency or domestic financing. Without country-by-country analysis, it is impossible to say what the possible impact of appropriate policy reforms on these costs might be, nor to estimate the potential for increased domestic resource mobilization to con- tribute to their financing. The one study so far that has applied our methodology (with regionally appropriate benchmark parameters) to 10 middle-income coun- tries in Latin America and the Caribbean found that the countries should be able to finance the limited amount of school-level expansion needed to reach the MDG, without an external gap, if they also adopt policies to improve the efficiency of student flows and devote reasonable domestic budget allocations to primary education (di Gropello, Dubey, and Winkler 2002). (The study did not evaluate rehabilitation needs or infrastructure expansion needs at the system level.) Against this, one can set an earlier World Bank analysis of MDG attainment that assumed no change in unit costs, student flows, or domestic financing. In this analy- sis, the estimated financing gap for middle-income countries was in the range of $4 billion per year (Devarajan, Miller, and Swanson 2002). However, we believe that this overstates the likely financing gap because, just as in the countries we analyzed, there is clear scope in middle-income countries to increase resource mobilization and improve efficiency in service delivery. The most that can be said without country-by-country analysis of the type we have done is that the incremental costs of reaching the education MDG in middle-income and transition countries are likely to be in the range of $23-28 billion per year. Of this, the need for external financing may be in the range of $1 billion per year, with appropriate policy reforms, to $3 billion per year. Summing these estimates with our scaled up estimate of the incremental costs and financing gap for low-income countries results in a global estimate that roughly $33-38 billion per year in additional spending on primary education will be needed to reach the goal. It should be kept in mind that this is the annual aver- age of a spending increase that would take place gradually over the period to 2015, from the roughly $90 billion that developing countries are spending today on pri- mary education to a projected $160 billion in 2015. We estimate that between $5 and $7 billion per year of this total amount needed will not be able to be generated domestically by these countries and would need to come through external aid. This range is fairly wide, but it is anchored in the most careful country-by- country analysis available. Our belief is that an extension of our methodology to the middle-income developing countries, with an explicit focus on achieving the MDG at minimum and sustainable global cost, would result in a global financing gap at the lower end of this range. It would also prompt a refocusing of external assistance for education on the lowest-income countries currently furthest from the goal of universal primary completion. 112 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD CHAPTER Implications for Countries and Donors At the World Education Forum in Dakar in 2000, the international community pledged that no developing country with a "credible plan" for achieving EFA would fall short of the 2015 goal for lack of external support. At the Monterrey conference on development finance in 2002, the donor community pledged increased development support channeled in a new and more selective framework to countries with both sound policies and a willingness to be held accountable for clear results. This study was written in the spirit of those commitments, to try to understand which policies are essential drivers of universal primary completion and what would constitute a credible plan for achieving the education MDG. Our analysis of primary completion rates across a sample of 55 low-income countries showed that a relatively small set of key parameters are important deter- minants of primary education outcomes: overall spending on primary education; average class size; average teacher salaries; spending on inputs other than teacher salaries; and the rate of repetition. When the overall level of resources is adequate and the distribution is balanced, education systems have the basic ingredients they need to perform well. If resource allocation skews any of the core parameters too much in one direction or the other (for example, average class size of 10 or average class size of 70), other areas are forced into compensating adjustments that are almost always unhealthy. Many chronic problems of low quality, inefficiency, and inequity-for low-performing systems are always inequitable-can be traced to imbalances in these key elements, and the unhealthy adjustments that primary education systems make. IMPORTANCE OF A FLEXIBLE APPROACH As important as these core elements are, however, there are at least three reasons why this framework is not sufficient for a credible EFA plan, and must not be applied rigidly: * System-wide averages do not guarantee an efficient underlying distribution. * Target parameters may not be optimal in different country contexts. * The overall level and mix of resources do not guarantee the transformation of those resources into quality schools and higher student learning. AVERAGE VALUES The analysis in this report hinges on system-wide average values for the core parameters. Yet a reasonable average value does not guarantee an efficient or equi- table underlying distribution-particularly in large federalized education systems such as those of India or Nigeria. A system-wide average of 40 students per teacher 113 could reflect a very efficient underlying pattern in which teachers in urban areas working in double-shift schools teach more than 40 students each while teachers in sparsely populated rural areas working in multigrade schools teach fewer. Or it could reflect the exact opposite-an excess of teachers stationed in the desirable urban areas and a serious shortage of teachers and schools in rural zones, which is very commonly seen in developing countries. The two patterns will produce very different student learning outcomes and completion rate progress. Our analysis implicitly assumes the efficient and intelligent underlying allocations that would cause these core parameters to be associated with primary completion progress. But a credible EFA plan would have to make this explicit. TARGET PARAMETERS The target parameters used in this exercise can provide useful points of reference or benchmarks for all countries. But there will be many cases where they are culturally, institutionally, or financially inappropriate, and rigid adherence to any particular target values must be avoided. The ultimate value of this framework is as a guide to the direction of reform, not as a dictate regarding where it should end. In the case of Armenia, for example, the current pupil-teacher ratio of 13:1 and average teacher salary of 0.6 times per capita GDP are not conducive to a quality primary education system. Comparison with the reference parameters of 40:1 and 3.5 times per capita GDP points clearly to the directions in which the system needs to move. But it may be just as adequate and more feasible politically for Armenia to aim for a pupil-teacher ratio of 30:1 and an average salary of 2.5 times per capita GDP. From the standpoint of resource allocation, this balance is equivalent to our target parameters. The key is to recognize when the starting imbalance is untenable and then to move in the right direction. TRANSFORMATION OF RESOURCES INTO LEARNING OUTCOMES The "indicative framework" presented in this study can ensure that education sys- tems have adequate overall resources and a healthy mix of core inputs. But it can- not guarantee the next-and crucial-step: the management of those resources to transform them into student learning. Paying teachers on average an adequate wage does not automatically produce the high standards, careful recruitment, qual- ity in-service training, and performance management needed to turn those teach- ers into an effective force in the classroom. The 33 percent of total recurrent spending we target for spending on inputs other than teacher salaries is ample enough for an education system to cover many important needs-from design of a relevant curriculum to national student learning assessments to proper building maintenance-but there are also many ways such resources can be spent badly, with little impact on MDG progress. Even more important than mobilizing more resources for primary education is improving the management of resources: at the national level, at the school level, and in the classroom. At the national level, ministries of education must achieve greater equity and efficiency in resource allocation and personnel deployment between administrative support services and schooling delivery, across different regions and across schools. The share of resources absorbed into central adminis- 114 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD tration in many systems is very high, with little value added for system functioning or student learning. Across different regions, schools with similar enrollments often differ widely in the number of teachers deployed to them, with no formal rationale but with clear implications for system quality and equity. Similarly, expenditure tracking studies frequently find that only a fraction of the overall edu- cation resources allocated to schools actually reaches them, and often too late in the school year to be used productively. Finally, crucial central support functions are often handled poorly or are nonexistent. National systems to assess student learn- ing and monitor progress at the classroom and school level are essential for holding education actors accountable and stimulating system-wide improvement, yet they exist in very few of the countries in our sample. Education statistics are rarely audited and often years out of date, and thus cannot serve as a tool for manage- ment decisions on an ongoing basis. Management capacity at the school level is also crucial. The quality of school leadership makes the difference between an orderly environment where teachers perform and children can learn, and a chaotic environment marked by rampant absenteeism, poor school maintenance, disappearance of books and materials, and poor relations with parents and the community, as seen in all too many education systems. Simple and often costless actions such as assigning the best teachers to the early grades, adapting the school calendar to the needs of the community, or mak- ing sure that teachers show up on time and work a full week, can greatly boost stu- dent attendance and learning. Effective school-level management makes these happen. Ultimately, it is management in the classroom that transforms education resources into student learning outcomes. Research shows that after controlling for student characteristics, learning outcomes can differ greatly even across equally endowed classrooms in the same school. What teachers do matters more for learn- ing outcomes than any other single factor. Teachers must use class time effectively; they must make good use of learning materials; they must have the capacity to adapt their teaching practice to individual students' learning needs; and, above all, they must be motivated to devote time and hard work to proving that "every child can learn." In most of the countries in our sample, teachers' incentives, capacity, and practice are greatly in need of strengthening. In short, good policies, innovative programs, and effective management in a great many areas must accompany a good core distribution of resources in a high- performing education system. Box 5.1 gives an idea of the wide range of manage- ment challenges a primary education system must address and outlines good-practice policies from across the developing world. These are aimed at ensur- ing that adequate resources translate into cost-effective expansion of schooling cov- erage, effective teaching and learning, and the flexibility in service delivery and other support needed to keep girls, the poorest, disabled, and other vulnerable chil- dren in school. A new World Bank paper (2002c), drawing lessons from a number of country case studies, provides more insights into what has worked and offers a set of principles to guide effective program and policy interventions. These are a crucial complement to the resource allocation framework we analyzed as elements of a credible EFA plan. CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 115 ~~ Key Education Policy Options GOAL POLICY CHOt$ES MEANS Low-cost and carefully * Lower-cost designs and construction material targeted expansion . Community-based construction * Streamlined preservice training (that is, shorter formal training, more hands-on training in classrooms, distance delivery) * Locally recruited teachers * Incentives for teacher deployment to remote and rural areas More cost-effective use *Double-shift schools of existing school . Multigrade schools infrastructure * Teacher redeployment and efficient class size Greater private * Simple regulatory framework for private providers (that is, accreditation system and collection of provision and basic statistics) financing of education a Grants to cost-effective nonpublic providers lighter system * Planning for HIV/AIDS impact management * School mapping (and later, more sophisticated education management information systems) * Review of role, selection, and training of school heads * Control of teacher absenteeism * Equitable funding across schools (per-student allocations) ........................................................................................................................................ Quality teaching * Emphasis on literacy and numeracy skills and clear learning goals for students * Student-centered interactive teaching methods * Ongoing professional development in content areas and pedagogical skills . Teacher networks and resource centers * Quality teacher manuals . Mother-tongue instruction in initial years * Increased hours and/or days of instruction * Salary structure that rewards teaching performance and rural or difficult postings c Quality instructional * Local teaching materials materials . Timely and equitable distribution of low-cost learning materials (textbooks) to schools and students Ir * Curriculum revision to improve relevance * Distance education (for example, radio education) lighter accountability * Simple school monitoring and reporting system (including private schools) mechanisms * Periodic assessment of student learning outcomes * Stakeholders empowered in school affairs Institutional * Reinforced management functions (planning, budgeting, staffing) strengthening - Greater school autonomy ........................................................................................................................................ Promote education of * Targeted stipends for girls girls * Labor-saving technologies, water points, and school-based childcare facilities to ease girls' household work * Schools located closer to communities * Separate latrines for girls .Recruitment of more female teachers and administrators * More mothers involved in school committees m ci Ensure school * Elimination of school fees affordability * Textbooks and school supplies provided free * Additional stipends for poor households and AIDS orphans :7 E Make schooling * Parents involved in school councils with decisionmaking power afttractive to parents * School calendar compatible with local economic activity and communities * School environment improved with latrines, water, electricity * School health and nutrition programs * Early childhood development programs * Nonformal education programs for youths and adults * Community libraries (eventually Internet centers) 116 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD . CONSIDERATIONS FOR DEVELOPING COUNTRIES The above are very important caveats. But they do not negate the fact that the first step toward a quality school system is to ensure adequate resources, allocated in a healthy balance across core system parameters. Without this, few other policy objectives and programs can be implemented or sustained. Placing the EFA planning process within this type of policy and financing framework-with internationally agreed values or ranges for target parameters- would mean that countries would compare their performance on a set of key domestic resource mobilization and service delivery parameters to the benchmarks observed in better-performing education systems. When a country's initial parameters deviate significantly from the bench- marks, a clear criterion for a "credible EFA plan" would be commitment to a grad- ual yet well-defined process of reform, to bring those areas of system performance into line. Progress would be evaluated annually, and in a very transparent manner, as the initial parameters, benchmark values, and appropriate annual targets for progress would be clear. This kind of technical rigor, transparency, and financial discipline has been missing from EFA planning to date-which has in many cases consisted of "wish lists" of actions that are neither prioritized, realistically costed, nor, in many cases, physically feasible. The analytical framework proposed here would help ensure that policy actions, new investments in school expansion, domestic financing, and external assistance are sustainable and lead to progressive improvements in system functioning, measured against clear benchmarks. Key outcomes, such as the pri- mary completion rate and learning outcomes (when standardized assessment sys- tems in these countries become more widespread), would also be tracked. Sustained and predictable external financing would be the quid pro quo for steady progress in improving these core indicators of system functioning and progressive improvement in outcomes. Some key implications of this approach for developing countries are as follows: * The criteria for a "credible plan" would be less ambiguous and more technically rigorous. * Countries' own commitments to achieving the MDG could be evaluated more transparendy, as the allocation of a "fair share" of domestic fiscal resources to primary education. * Steady improvement in service delivery would be a quid pro quo for continued external support. * The EFA process would be focused more sharply on key outcomes, especially the primary completion rate, student learning progress, and gender parity, and more accurate and timely measurement of these would be required. * Countries and their partners would both be accountable for ensuring that external funding catalyzes tangible progress toward universal primary completion and is not wasted in ineffective delivery systems. CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 117 * Countries' overall domestic resource mobilization and spending, not only education ministry spending, would become subject to transparent monitoring. CONSIDERATIONS FOR THE DONOR COMMUNITY The implications of this study for international development partners are equally strong. Our results clearly show that even with a maximum domestic effort, these 47 countries plus Afghanistan will not be able to achieve the education MDG by 2015 without: * A significant increase in donor funding for primary education * Better targeting of donor assistance to "EFA priority" countries * A change in the mix of assistance * Gready increased efficiency of aid transfers * Transfer of funds via new mechanisms; and * More effective monitoring of progress, increased research, and faster diffusion of knowledge about what works. INCREASED AID FOR PRIMARY EDUCATION The average external financing needed for the 48 low-income countries we studied, including Afghanistan, at $2.5 billion per year over the next 12 years (constant 2000 dollars) represents a significant increase over current aid for primary educa- tion to these countries. Although it is difficult to compile solid country-level data, we estimate that this is almost triple the level of external support for primary edu- cation these countries currently receive. If the numbers are disaggregated regionally, the even greater disparity between current levels of assistance and estimated needs for Sub-Saharan Africa becomes clear. For the 33 African countries that account for $1.9 billion of the overall $2.4 billion per year gap (excluding Afghanistan), estimated disbursements of official development aid for primary education over the last three years have averaged only about $500 million per year. New commitments have averaged little more than $600 million per year (table 5.1). In other words, the annual external disburse- ments to these countries would have to almost quadruple. For the 13 countries outside Africa with projected financing gaps, the estimated needs would average $586 million per year over the period, almost 50 percent higher than the estimated current level of disbursements of about $400 million per year. BETTER TARGETING TO "EFA PRIORITY" COUNTRIES The preceding analysis shows that, although our financing gap estimate is lower than some previous global estimates, it would nonetheless require a substantial increase in donor funding for primary education in these countries to fill it-and a fourfold increase for Africa. This will not be easy to achieve. On the country side, although the scenario for accelerated MDG progress we model explicidy assumes that institutions and policies in the education sector (and fiscal management more generally) gradually become stronger through implementation of reforms and shifts in resource allocation, there is no question that institutional "capacity gaps" currently constrain the level of assistance to many of these countries. 118 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Table 5.1 ............................................................................................................... Official Development Assistance to Basic Education in Sub-Saharan Africa, by Donor, 1998-2000 (commitment basis, millions of current U.S. dollars) Donor 1998 1999 2000 International Development Association Total education 1,201.5 534.8 468.7 Africa education 372.3 194.1 159.7 Africa basic education 218.3 131.0 58.8 Other multilateral development banks Total education 1,274.7 773.7 1,335.5 Africa education 868.9 309.5 1,041.3 Africa basic education 28.2 . 110.6 * 116.2 Development Assistance Committee countries Total education 4,459.2 5,014.3 3,541.6 Africa education 2,328.4 . 1,259.2 1,405.6 Africa basic education 422.2 418.4 378.8 All donors Total education 6,935.4 6,322.8 5,345.8 Africa education 3,569.6 1,762.8 2,606.6 Africa basic education 668.7 660.0 553.8 Note: The World Bank (IBRD and IDA) reports flows to primary education," whereas the Development Assistance Committee (DAC) of the OECD reports flows to "basic education." DAC countries and African Development Bank regional classifications cover continental Africa and not Sub-Saharan Africa, as reported by the World Bank. Therefore the regional totals are approximate. Sources: World Bank Business Warehouse; OECD DAC Database. But the current low level of external support for primary education may also reflect relatively unfocused commitment to the MDG on the donor side. When our estimated gap of $2.4 billion per year is compared against total assistance for educa- tion reported by international donors and multilateral banks, which averages about $7 billion per year, it looks relatively small (annex table D. 1). That assistance-for all levels of educational development and all developing (low- and middle-income) countries-serves many different, and potentially important, purposes. But, very clearly, it is not today giving priority to the countries where achievement of the MDG is at greatest risk without external support. Only about 20 percent of donor assistance for education is currently channeled to basic education (annex table D.2). At Dakar, the international community pledged that no developing country's EFA progress would be constrained for lack of external support. If that commit- ment is to be upheld, the development partners will need to either: Mobilize a significant real increase in funding for primary education, channeled largely to Sub-Saharan Africa, or CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 1 19 * Achieve an equally dramatic reallocation of existing assistance, from other levels of education to primary education, and from middle-income countries to low-income, especially African, countries. A CHANGING MIX OF ASSISTANCE According to our analysis, close to 60 percent of the external financing require- ments for these countries over the next 13 years will be of a recurrent nature. Although recent data show that World Bank/IDA disbursements to these countries for primary education have begun to approach this mix, it is not clear that this is true for other donors. Assistance from all external partners has traditionally given heavy priority to capital investments, such as school construction or equipment. A dear message of the simulation results is that if donors want to see the MDG of universal primary completion attained, a relatively high share of external assistance will have to be channeled to recurrent budget support. And for greater gender par- ity and supporting HIV/AIDS orphans, in particular, recurrent budget support for stipend programs will be critical. Of course, donors will be reluctant to go this route without clear acceptance by countries that their entire budget-and domestic resource mobilization effort- will be on the table for discussion and monitoring. This is already happening in the PRSP and PRSC processes that most of the countries in our sample are engaged in. Uganda is a good example of a country whose receptivity to greater budgetary transparency-and outcome monitoring-has been directly associated with a large increase in the share of development assistance channeled to the country as budget- ary support. GREATER EFFICIENCY OF EXTERNAL SUPPORT There is a related and overarching need for more efficient transfers of develop- ment assistance. The average annual gap we estimate can be considered as the minimum cost of achieving EFA under the most optimistic (although gradual and achievable) scenarios of policy reform and efficient aid flows. Another way of conceiving of the $2.4 billion is as the net or "core resource transfer" that would be needed. But it is clear that $2.4 billion in reported donor assistance does not today equal $2.4 billion of core resource transfer to recipient countries. For example, we estimate that roughly $286 million per year (2000 constant dollars) would be needed on average over the period by the 33 African countries in our sample to address the impact of AIDS on their education systems. A significant part of this derives from the projected need to provide subsistence support (food, cash, clothing) to AIDS orphans to prevent them from dropping out of school, a cost we estimated at $50 per orphan per year based on data from some existing programs. This represents the actual transfer to the child, without any specific pro- vision for the overhead and administrative costs required to channel support. But it may be noted that various donor/NGO programs of this nature operating today generally cost $100-$150 per child in order to achieve the equivalent of a $50 net transfer to the beneficiary. Although these programs usually include counseling 120 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD and other services that cannot be classed simply as overhead, it does give a sense of the margins that can exist between the "core resource transfers" we costed and the total costs of assistance. Similarly, the unit construction costs we assumed are lower than those many donors report. If the aid-financed construction of a fully equipped classroom in Africa costs close to $24,000, as reported by some donors, it should be realized that only $8,000 of that amount (the unit cost we used for African countries in the sample) would count against our estimated capital financing gap. On the other hand, involving communities in school construction has been shown in many countries to lower unit costs, and this could result in lower capital costs than we estimated and effectively narrow the gap. However, as we assumed that virtually all school construction over the period would be financed externally, shifting more systematically to community-based approaches would require flexibility on the part of donors. Finally, a significant share of donor assistance typically supports technical assis- tance contracts, consultancies, study tours, seminars, and other activities. No mat- ter how laudable or even crucial these may be for inspiring education reforms or building capacity, they cannot be counted directly against the "net" resource needs we estimate. The costs that enter into our financing gap, outside of construction needs, are very largely the core, local cost requirements for running a primary edu- cation system. It is impossible to say with any precision what equivalent "gross" level of devel- opment assistance would be needed to fill our estimated $2.4 billion net gap. It could conceivably be 50-100 percent higher. But the efficiency of transfers varies considerably across donors. And the readiness of international partners to move in coming years toward greater use of pooled assistance and direct budgetary transfers in the context of SWAPs, PRSCs, or other program lending could gready improve the efficiency with which each dollar of future external assistance offsets the esti- mated financing gap in these countries. NEW TRANSFER MECHANISMS A renewed push for EFA by 2015 will require major changes from business as usual for both at-risk countries and their development partners. The policy and financing framework laid out in this report could afford a basis for gauging coun- tries' commitments to EFA and guiding service delivery reforms. The quid pro quo for monitorable improvements in service delivery quality and efficiency and for key results such as increasing primary completion rates and student learning would be adequate, sustained, and predictable levels of financing from interna- tional partners. The stability and predictability of external assistance is crucial if countries are to take on recurrent expenditures (such as hiring of additional teachers) that are not easily compressed when external support fluctuates. On the other hand, it is not easy for bilateral donors, subject to their own political processes and budget constraints, to make long-term funding commitments. Greater used of pooled donor assistance and direct budget transfers in the context of SWAPs and other CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 121 programmatic support could help match donor assistance more effectively to countries' core financing needs and ensure a more stable and predictable flow of funding. Both parts of this compact would require dient countries and international partners to collaborate in new ways. Countries would need to accept benchmarks for system performance and a non-bureaucratic yet participatory way of involving development partners in monitoring budgets and progress. Donors would need to forego many of the trappings of the way aid is channeled today in favor of resource transfers that would permit them to monitor only overall outcomes, and not their national share of procured inputs. Whether this flow of assistance could be coordinated across agencies as they currendy operate or would merit a specific new mechanism for pooling EFA support in new ways cannot be said. It is clear, however, that whether existing or new arrangements are used, they must be made to function better, with lighter and more effective coordination and with an increase in the efficiency with which resource transfers effectively meet countries' financing needs. GLOBAL RESEARCH AND KNOWLEDGE DIFFUSION A potentially important byproduct of a more technically rigorous (and data- intensive) EFA planning process could be a deeper and wider base for global research on how primary education systems improve. An internationally accessible database of key expenditure and service delivery variables for a large number of countries would clearly result from any mainstream adoption of the policy and financing framework proposed in this report. Verified and internationally compa- rable data on these parameters and education outcomes could be a major boon to education researchers. Under the most optimistic scenario, this could produce a deeper understanding of how to accelerate primary completion rate progress that would benefit those countries currently furthest from the MDG. Although the costs of EFA monitoring, data collection, international research, and global and local activities to diffuse new knowledge are not included in the estimated financ- ing gap, these investments in the global public good should be considered core responsibilities of the international community. . THE EFA FAST-TRACK INITIATIVE Building on the above framework, a new compact for primary education designed to accelerate global progress toward the education MDG was endorsed by the Development Committee of the World Bank and the International Monetary Fund in April 2002 and by the G-8 in its action plan for education at the June 2002 summit in Kananaskis, Alberta, Canada. The new compact, called the EFA Fast-Track Initiative, is the first global proposal to emerge since the Monterrey conference that aims at accelerating MDG progress using the Monterrey frame- work of increased development support in exchange for increased accountability for results. The new initiative is supported by all major bilateral donors for educa- tion and by UNESCO, UNICEF, the World Bank, and the regional development 122 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD banks, all of which have jointly formed the EFA Fast-Track Partnership. At the heart of the Fast-Track Initiative are: * A commitment by developing countries to accelerate efforts to achieve universal primary education cost-effectively, within a transparent global accountability framework (the EFA indicative framework outlined in this study); and * A commitment by donors to provide sustained incremental financing (as much as possible on a grant basis), where credible plans to accelerate progress in primary education exist. In June 2002, a first set of 18 low-income countries was invited to join the ini- tiative and to submit their EFA plans, including baseline "indicative framework" indicators and annual targets for 2003, for donor financing. The 18 countries (box 5.2) are diverse regionally and in terms of their proximity to universal primary com- pletion; together, they account for an estimated 18 million children without access to education. This first set of countries was invited to consider committing to the Fast-Track Initiative on the basis of two simple and transparent criteria: (a) they have formally adopted national poverty reduction strategies (PRSPs) that integrate their education plans into overall national development priorities; and (b) they have education sector plans in place. The rationale for these two criteria is that having these elements in place should allow the Fast-Track Initiative to catalyze measurable progress more quickly. It should be noted that the initia- tive is aimed at accelerating MDG progress in, and learning lessons from, countries that are currendy on First EFA Fast- track to reach the goal as well as countries that are Track Group, 2002 off track. The first group of countries includes one- ....................... Vietnam-that has virtually achieved the goal and oth- Albania ers such as The Gambia and Uganda that are considered Bolivia on track. In countries such as these, the aim of the Fast- Burkina Faso Track Initiative is to help countries reach the goal more Ethiopia The Gambia quickly and in the process generate lessons and demon- Ghana stration effects for other countries. Guinea A second set of five high-priority countries was also Guyana Honduras invited to join the initiative, but with a different status Mauritania initially, as they did not yet have sector plans and/or Mozambique national poverty reduction strategies in place. These Nicaragua Niger "Big Five" countries are deemed high priority because Tanzania they account for the largest numbers of children with- Uganda out access to primary education globally-about 50 Vietnam million of the 113 million children in total estimated Republic of Yemen Zambia to be out of school. The spirit of the Fast-Track Initia- Analytcal Fast-Track Countries tive is that country commitment to sound sector pro- Bangladesh grams integrated into a broader poverty reduction Democratic Republic of Congo strategy as well as commitment to appropriate policy India actions in line with the EFA indicative framework are Nigeria Pakistan important for effective use of development resources. CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 123 "Analytical Fast-Track" support aims to help these countries reach that status. India is the first of the "Big Five" countries to meet the two criteria, and the government is considering participation in the Fast Track. In countries that have plans in place, the Fast-Track process involves a com- plementary in-country analysis to benchmark the education system's performance relative to the EFA indicative framework; to appropriate annual targets for their country context; and to refine estimates of the external financing needs for accel- erated progress in primary education, consistent with the implementation of appropriate reforms and the country's medium-term expenditure framework. Although for the first set of countries these adjustments and targets have been set out in "Fast-Track proposals," it is expected that the process of identifying prior- ity policy actions to align system functioning with the indicative framework benchmarks will increasingly become a natural part of the development of a PRSP and a credible education sector plan and separate FTI proposals will not be needed. The first FTI proposals have represented a more comprehensive assess- ment of financing needs than we costed, as they include rehabilitation require- ments. The estimated expansion needs, however, may be compared with the financing needs estimated in this study. An important part of the process is also careful assessment of the physical and institutional capacity to execute increased primary education investment and expenditure. The Fast-Track Initiative in all cases implies a major expansion of the management challenge for systems that are generally perceived to be weakly man- aged today. But this cannot be an argument against such expansion; it simply means that attention to capacity building and institutional support must be an equal part of the partnership effort. Finally, the estimated needs are compared with the pipeline of existing donor commitments for primary education in each country, including general budget support under PRSCs or other multisector programs. It should be recalled that the financing gaps estimated in the present study are gross financing gaps, with no adjustment for the current level of external assistance to the primary sector. As of March 2003, ten of the first 18 countries invited to join the Fast-Track Initiative submitted proposals for consideration. The Fast-Track partners commit- ted, upon verification of the estimated financing gaps, to ensure that these gaps are filled for the next three years, contingent on the countries' continued progress in exe- cuting the accelerated program and improving sector functioning in line with their indicative framework targets. The partners also agreed to meet every six months to consider additional country proposals, review implementation, and harmonize their education assistance to Fast-Track countries. Intensified collaboration among donor representatives at the client country level is a key part of this process. For their part, Fast-Track recipient countries are committed to annual monitor- ing of their progress against indicative framework targets. Key outcomes such as the net intake rate into first grade for girls and boys, the primary completion rate for girls and boys, and student learning achievement will also be monitored, although it is understood that these outcome indicators can be slow to reflect progress. 124 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD . CONCLUSION Without a substantial acceleration of progress in as many as 86 developing coun- tries, the Millennium Development Goal of universal primary education by the year 2015 will not be met. The good news is that many of these countries could meet the goal, if they could achieve and sustain the same rate of primary comple- tion progress averaged by the best-performing developing countries over the 1990s. We know from country experience that it can be done. We also know, from country experience, the key building blocks for a healthy system of primary education. We know what constitutes a broadly adequate level of resources, and how to balance two key elements of the resource mix: (a) spend- ing on teachers and (b) the equally crucial spending on complementary inputs, supervision, and support needed to make teachers effective in the classroom. Domestic commitment to universalizing primary education is the first key. As countries from the Republic of Korea in the 1960s to Zimbabwe in the 1980s to Brazil in the 1990s have demonstrated, when political will is mustered, primary completion progress can accelerate quickly. But for the 48 low-income countries with the world's lowest completion rates, even maximum commitments of domestic resources for primary education plus steady progress in reforming service delivery will not be enough to ensure achieve- ment of the education MDG by 2015. A resource gap of at least $2.5 billion per year in these countries will threaten their achievement of the goal in the absence of external assistance. Mobilizing this amount of increased aid should not be an insurmountable challenge for the international community, but it will require very significant changes in donor priorities and the mechanisms through which aid is channeled. Accompanying increased financing with support for capacity building and imple- mentation is also important to ensure that donor and national resources effectively produce the desired results. The new compact for accelerating progress embodied in the EFA Fast-Track Initiative-if launched successfully and expanded steadily to include all of the at- risk developing countries-could provide a framework in which countries' steady progress in improving core indicators of education system functioning and pro- gressive improvement in outcomes would be the quid pro quo for sustained and predictable budgetary support from international partners, channeled in new and more flexible ways. Few global goals have been as consistendy and deeply sup- ported as the notion that every child in the world should have the chance to com- plete primary school. With global effort, it could become a reality. CHAPTER FIVE * IMPLICATIONS FOR COUNTRIES AND DONORS 125 BIBLIOGRAPHY: Azariadis, Costas, and Allan Drazen. 1990. "Threshold Externalities in Economic Development." QuarterlyJournalofEconomics 105 (2): 501-26. Barro, Robert J. 1999a. "Determinants of Democracy." Journal of Political Econ- omy 107 (6): 158-83. . 1999b. "Human Capital and Growth in Cross-Country Regressions." Swedish Economic Policy Review 6 (2): 237-77. Barro, Robert J., and Jong-Wha Lee. 2000. "International Data on Educational Attainment Updates and Implications." Working Paper 7911. National Bureau of Economic Research, Cambridge, Mass. Birdsall, Nancy, and Juan Luis Londofio. 1998. "No Tradeoff: Efficient Growth Via More Equal Human Capital Accumulation in Latin America." In Nancy Birdsall, Carol Graham, and Richard Sabot, eds., Beyond Trade-Offi: Market Reform and Equitable Growth in Latin America . Washington, D.C.: Brookings Institution and Inter-American Development Bank. Brossard, Mathieu, and Luc-Charles Gacougnolle. 2001. "Financing Primary Edu- cation for All: Yesterday, Today and Tomorrow." United Nations Educational, Scientific, and Cultural Organization, Paris. Processed. Colclough, Christopher, and Samer Al-Samarrai. 2000. "Achieving Schooling for All: Budgetary Expenditures on Education in Sub-Saharan Africa and South Asia." World Development 28 (11): 1927-44. Colclough, Christopher, and Keith Lewin. 1993. Educating All the Children: Strategies for Primary Schooling in the South. New York: Oxford University Press. De, Anuradha, and Jean Dreze. 1999. Public Report on Basic Education in India (PROBE). New Delhi: Oxford University Press. de la Fuente, Angel, and Rafael Domenech. 2000. "Human Capital in Growth Regressions: How Much Difference Does Data Quality Make?" Paper 2466. Cen- tre for Economic Policy Research Discussion, London. Delamonica, Enrique, Santosh Mehrotra, and Jan Vandemoortele. 2001. "Is EFA Affordable? Estimating the Global Minimum Cost of Education for All." Inno- centi Working Paper 87. United Nations Children's Fund, New York. Devarajan, Shantayanan, Margaret J. Miller, and Eric V. Swanson. 2002. "Goals for Development: History, Prospects and Costs." Policy Research Working Paper 2819. World Bank, Washington, D.C. 122 di Gropello, Emanuela, Ashutosh Dubey, and Donald Winkler. 2002. "Reaching Millennium Development and Regional Goals for Education in Latin America and the Caribbean-Second Draft." World Bank, Latin America and Caribbean Region, Human Development Department, Washington, D.C. Processed. Filmer, Deon. 2000. "The Structure of Social Disparities in Education: Gender and Wealth." Policy Research Working Paper 2268. World Bank, Development Research Group, Washington, D.C. Underlying data available at http://www. worldbank.org/research/projects/edattain/edattain.htm Filmer, Deon, and Lant Pritchett. 1999. "The Effect of Household Wealth on Educational Attainment: Evidence from 35 Countries." Population and Develop- ment Review 25 (1): 85-120. Underlying data available at http://www.worldbank. org/research/projects/edattain/edattain.htm Godoy, Ricardo, and Manuel Contreras. 2001. "A Comparative Study of Educa- tion and Tropical Deforestation armong Lowland Bolivian Amerindians: Forest Val- ues, Environmental Externality, and School Subsidies." Economic Development and Cultural Change 49 (3): 555-74. Gregson, S., H. Waddell, and S. Chandiwana. 2001. "School Education and HIV Control in Sub-Saharan Africa: From Discord to Harmony?" Journal of Interna- tionalDevelopment 13: 467-85. Hanushek, Eric A., and Dennis D. Kimko. 2000. "Schooling, Labor Force Qual- ity, and the Growth of Nations." American Economic Review 90 (5): 1184-1208. Kelly, M. J. 2000. "Planning for Education in the Context of HIV/AIDS." UNESCO, International Institute for Educational Planning, Paris. Kitaev, Igor. 1999. "Private Education in Sub-Saharan Africa: A Re-examination of Theories and Concepts Related to Its Development and Finance." UNESCO, International Institute for Educational Planning, Paris. Lockheed, Marlaine E., Adriaan M. Verspoor, and associates. 1991. Improving Primary Education in Developing Countries. New York: Oxford University Press for the World Bank. Mehrotra, Santosh. 1998. "Education for All: Policy Lessons from High-Achieving Countries." Staff Working Paper, Evaluation, Policy and Planning Series, No. EPP- EVL-98-005. United Nations Children's Fund, New York. Mehrotra, Santosh, and Peter Buckland. 1998. "Managing Teacher Costs for Access and Quality." Staff Working Paper, Evaluation, Policy and Planning Series, No. EPP-EVL-98-004. United Nations Children's Fund, New York. Mingat, Alain. 2001. "Teacher Salary Issues in African Countries." World Bank, Africa Region, Human Development Analysis and Policy Development Support Team. Washington, D.C. Processed. 128 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD Mingat, Alain, Ramahatra Rakotomalala, and Jee-Peng Tan. 2002. "Financing Education for All by 2015 in Africa: Simulation Results for 33 Countries." Human Development Working Paper. World Bank, Africa Region, Washington, D.C. O'Connell, Lesley, and Nancy Birdsall. 2001. "Race, Human Capital Inequal- ity, and Income Distribution in South Africa, Brazil and the United States." Discussion Paper 4. Carnegie Endowment for International Peace, Washington, D.C. OECD (Organisation for Economic Co-operation and Development). 2000. Investing in Education: Analysis of the 1999 World Education Indicators. Paris. OECD and Statistics Canada. 2000. Literacy in the Information Age: Final Report of the InternationalAdult Literacy Survey. Paris: OECD. Oxfam International Global Campaign for Education. 2002. "Broken Promises? Why Donors Must Deliver on the EFA Action Plan." Global Campaign for Edu- cation Briefing Paper. Oxfam International, Oxford, U.K. Porter, Michael. 1998. "Microeconomic Foundations of Competitiveness: The Role of Education." In The Global Competitiveness Report 1998. Cambridge, Mass.: Harvard University, Center for International Development, World Economic Forum. Psacharopoulos, George. 1995. "The Profitability of Investment in Education: Concept and Methods." Human Capital Development and Operations Policy Working Paper. World Bank, Washington, D.C. Psacharopoulos, George, and Harry Anthony Patrinos. 2002. "Returns to Invest- ment in Education: A Further Update." Policy Research Paper 2881. World Bank, Education Department, Latin America and Caribbean Region, Washington, D.C. Sen, Amartya. 1999. Development as Freedom. New York: Knopf. Smith, Lisa C., and Lawrence Haddad. 2000. "Explaining Child Malnutrition in Developing Countries: A Cross-Country Analysis." Food Consumption and Nutrition Division Discussion Paper 60. International Food Policy Research Insti- tute, Washington, D.C. Subbarao, K., Angel Mattimore, and Kathrin Plangemann. 2001. "Social Protec- tion of Africa's Orphans and Other Vulnerable Children." Human Development Working Paper. World Bank, Africa Region, Washington, D.C. Vandemoortele, J., and E. Delamonica. 2000. "Education 'Vaccine' Against HIV/ AIDS." Current Issues in Comparative Education 3 (1). Voluntary Services Organization. 2002. "What Makes Teachers Tick? A Policy Research Report on Teachers' Motivation in Developing Countries." London. Processed. BIBLIOGRAPHY 129 Watkins, Kevin. 1999. Education Now: Break the Cycle of Poverty. Oxford, U.K.: Oxfam International. World Bank. 2001. Engendering Development: Through Gender Equality in Rights, Resources, and Voice. World Bank Policy Research Report 21776. New York: Oxford University Press. _ . 2002a. "Development Effectiveness and Scaling Up: Lessons and Chal- lenges from Case Studies." Synthesis paper DC2002-0018, prepared for the Sep- tember 28, 2002 Development Committee meeting. Washington, D.C. . 2002b. Education and HIV/AIDS: A Window of Hope. Washington, D.C. . 2002c. "EFA: The Lessons of Experience: The Impact of Policy in 20 Case Studies." World Bank, Education Department, Human Development Network, Washington, D.C. Processed. _ . 2002d. "India; Karnataka: Financing Education in the Context of Eco- nomic Restructuring." Report 24207-IN. South Asia Region, Human Develop- ment Department, Washington, D.C. ___. 2002e. "School Enrollment and Attainment in India: Progress during the 1990s." South Asia Region, Washington, D.C. Processed. 130 ACHIEVING UNIVERSAL PRIMARY EDUCATION BY 2015: A CHANCE FOR EVERY CHILD TECHNICAL ANNEXES: Annex A. Background Data and Aggregate Simulation Results 141 About the Simulation Model 141 Figure A.1. Stylized Dynamics of Selected Indicators of Student Flow, 2000-2015 142 Figure A.2. Stylized Dynamics of Primary School Enrollments, 2000-2015 142 Table A.1. Countries Included in Financing Simulation 144 Table A.2. Core Education Parameters for 55 Low-Income Countries 146 Table A.3. Key Education System Parameters for Countries, Grouped by Relative EFA Success (Adjusted Sample) 149 Table A.4. Group 1 Parameters with Six Eastern European and Central Asian Countries Included 152 Table A.5. Demographic Burden and Resource Mobilization, by Country and Region, 2000 and 2015 153 Table A.6. Actual and Projected Expenditures on Primary Education, Domestic Financing, and Financing Gap, by Country, under Scenario C2 157 Table A.7. Domestic Resources Mobilized and Financing Gap, by Country, under Scenarios Cl and C3 160 Table A.8. Estimated Annual Financing Gap, by Country, 2001-2015 (Scenario C2) 162 Annex B. Primary Completion Rate Estimates and Projections 165 Figure B.1. Global Progress in Primary Completion, 1990-2000 and Projected Trends, Country-Weighted 165 Figure B.2. Global Progress in Primary Completion, 1990-2000 and Projected Trends, Population-Weighted 165 Figure B.3. Primary Completion Progress by Region, 1990-2000 and Projected Trends, Country-Weighted 166 Figure B.4. Primary Completion Progress by Region, 1990-2000 and Projected Trends, Population-Weighted 166 Figure B.5. Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted 167 131 Figure B.6. Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Population-Weighted 168 Table B.1. Primary Completion Progress by Region, 1990-2000, Country-Weighted 169 Table B.2. Primary Completion Progress by Region, 1990-2000, Population-Weighted 170 Table B.3. Best Performers (IDA Countries) in Improving Primary Completion Rate, 1990 to Most Recent Year 171 Table B.4. Best Performers (IBRD Countries) in Improving Primary Completion Rate, 1990 to Most Recent Year 172 Table B.5. IDA Countries with Declining Primary Completion Rate, 1990 to Most Recent Year 173 Table B.6. IBRD Countries with Declining Primary Completion Rate, 1990 to Most Recent Year 174 Table B.7. IDA Countries That Have Achieved Universal Primary Completion 174 Table B.8. IBRD Countries That Have Achieved Universal Primary Completion 175 Table B.9. IDA Countries "On Track" to Achieve Universal Primary Completion by 2015 176 Table B.1O. IBRD Countries "On Track" to Achieve Universal Primary Completion by 2015 177 Table B.11. IDA Countries "Off Track" to Achieve Universal Primary Completion by 2015 178 Table B.12. IBRD Countries "Off Track" to Achieve Universal Primary Completion by 2015 179 Table B.13. IDA Countries "Seriously Off Track" to Achieve Universal Primary Completion by 2015 180 Table B.14. IBRD Countries "Seriously Off Track" to Achieve Universal Primary Completion by 2015 181 Table B.15. Countries with No Data Available 181 Annex C. Country Simulation Results 183 Table C.1. All 47 Countries: MDG-2015 Financing Gap under Alternative Policy Measures 183 132 TECHNICAL ANNEXES Table C.2. All 47 Countries: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 183 Table C.3. World, except Africa: Financing Gap under Alternative Policy Measures 184 Table C.4. World, except Africa: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 184 Table C.5. Africa: MDG-2015 Financing Gap under Alternative Policy Measures 185 Table C.6. Africa: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 185 Table C.7. Armenia: MDG-2015 Financing Gap under Alternative Policy Measures 186 Table C.8. Armenia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 186 Table C.9. Bangladesh: MDG-2015 Financing Gap under Alternative Policy Measures 187 Table C.10. Bangladesh: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 187 Table C.1l. Cambodia: MDG-2015 Financing Gap under Alternative Policy Measures 188 Table C.12. Cambodia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 188 Table C.13. Georgia: MDG-2015 Financing Gap under Alternative Policy Measures 189 Table C.14. Georgia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 189 Table C.15. Haiti: MDG-2015 Financing Gap under Alternative Policy Measures 190 ANNEX CONTENTS 133 Table C.16. Haiti: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 190 Table C.7. Honduras: MDG-2015 Financing Gap under Alternative Policy Measures 191 Table C.18. Honduras: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 191 Table C.19. India: MDG-2015 Financing Gap under Alternative Policy Measures 192 Table C.20. India: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 192 Table C.21. Lao PDR: MDG-2015 Financing Gap under Alternative Policy Measures 193 Table C.22. Lao PDR: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 193 Table C.23. Moldova: MDG-2015 Financing Gap under Alternative Policy Measures 194 Table C.24. Moldova: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 194 Table C.25. Mongolia: MDG-2015 Financing Gap under Alternative Policy Measures 195 Table C.26. Mongolia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 195 Table C.27. Nepal: MDG-2015 Financing Gap under Alternative Policy Measures 196 Table C.28. Nepal: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 196 Table C.29. Nicaragua: MDG-2015 Financing Gap under Alternative Policy Measures 197 Table C.30. Nicaragua: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 197 134 TECHNICAL ANNEXES Table C.31. Pakistan: MDG-2015 Financing Gap under Alternative Policy Measures 198 Table C.32. Pakistan: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 198 Table C.33. Yemen: MDG-2015 Financing Gap under Alternative Policy Measures 199 Table C.34. Yemen: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 199 Table C.35. Angola: MDG-2015 Financing Gap under Alternative Policy Measures 200 Table C.36. Angola: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 200 Table C.37. Benin: MDG-2015 Financing Gap under Alternative Policy Measures 201 Table C.38. Benin: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 201 Table C.39. Burkina Faso: MDG-2015 Financing Gap under Alternative Policy Measures 202 Table C.40. Burkina Faso: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 202 Table C.41. Burundi: MDG-2015 Financing Gap under Alternative Policy Measures 203 Table C.42. Burundi: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 203 Table C.43. Cameroon: MDG-2015 Financing Gap under Alternative Policy Measures 204 Table C.44. Cameroon: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 204 Table C.45. Central African Republic: MDG-2015 Financing Gap under Alternative Policy Measures 205 ANNEX CONTENTS 135 Table C.46. Central African Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 205 Table C.47. Chad: MDG-2015 Financing Gap under Alternative Policy Measures 206 Table C.48. Chad: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 206 Table C.49. Congo, Democratic Republic: MDG-2015 Financing Gap under Alternative Policy Measures 207 Table C.50. Congo, Democratic Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 207 Table C.51. Congo, Republic: MDG-2015 Financing Gap under Alternative Policy Measures 208 Table C.52. Congo, Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 208 Table C.53. C6te d'Ivoire: MDG-2015 Financing Gap under Alternative Policy Measures 209 Table C.54. Cote d'Ivoire: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 209 Table C.55. Eritrea: MDG-2015 Financing Gap under Alternative Policy Measures 210 Table C.56. Eritrea: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 210 Table C.57. Ethiopia: MDG-2015 Financing Gap under Alternative Policy Measures 211 Table C.58. Ethiopia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 211 Table C.59. The Gambia: MDG-2015 Financing Gap under Alternative Policy Measures 212 Table C.60. The Gambia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 212 136 TECHNICAL ANNEXES Table C.61. Ghana: MDG-20 15 Financing Gap under Alternative Policy Measures 213 Table C.62. Ghana: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 213 Table C.63. Guinea: MDG-20 15 Financing Gap under Alternative Policy Measures 214 Table C.64. Guinea: MDG-2015 Cost Estimates and Sources of Fina-ncing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 214 Table C.65. Guinea-Bissau: MDG-2015 Financing Gap under Alternative Policy Measures 215 Table C.66. Guinea-Bissau: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 215 Table C.67. Kenya: MDG-2015 Financing Gap under Alternative Policy Measures 216 Table C.68. Kenya: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 216 Table C.69. Lesotho: MDG-20 15 Financing Gap under Alternative Policy Measures 217 Table C.70. Lesotho: MDG-20 15 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 217 Table C.71. Madagascar: MDG-2015 Financing Gap under Alternative Policy Measures 218 Table C.72. Madagascar: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 218 Table C.73. Malawi: MDG-20 15 Financing Gap under Alternative Policy Measures 219 Table C.74. Malawi: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 219 Table C.75. Mali: MDG-201 5 Financing Gap under Alternative Policy Measures 220 ANNEX CONTENTS 137 Table C.76. Mali: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 220 Table C.77. Mauritania: MDG-2015 Financing Gap under Alternative Policy Measures 221 Table C.78. Mauritania: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 221 Table C.79. Mozambique: MDG-2015 Financing Gap under Alternative Policy Measures 222 Table C.80. Mozambique: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 222 Table C.81. Niger: MDG-2015 Financing Gap under Alternative Policy Measures 223 Table C.82. Niger: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 223 Table C.83. Nigeria: MDG-2015 Financing Gap under Alternative Policy Measures 224 Table C.84. Nigeria: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 224 Table C.85. Rwanda: MDG-2015 Financing Gap under Alternative Policy Measures 225 Table C.86. Rwanda: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 225 Table C.87. Senegal: MDG-2015 Financing Gap under Alternative Policy Measures 226 Table C.88. Senegal: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 226 Table C.89 Sierra Leone: MDG-2015 Financing Gap under Alternative Policy Measures 227 Table C.90. Sierra Leone: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 227 138 TECHNICAL ANNEXES Table C.91. Sudan: MDG-2015 Financing Gap under Alternative Policy Measures 228 Table C.92. Sudan: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 228 Table C.93. Tanzania: MDG-2015 Financing Gap under Alternative Policy Measures 229 Table C.94. Tanzania: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 229 Table C.95. Togo: MDG-2015 Financing Gap under Alternative Policy Measures 230 Table C.96. Togo: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 230 Table C.97. Uganda: MDG-2015 Financing Gap under Alternative Policy Measures 231 Table C.98. Uganda: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 231 Table C.99. Zambia: MDG-2015 Financing Gap under Alternative Policy Measures 232 Table C.100. Zambia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 232 Annex D. Aid for Primary Education 233 Table D.1. Development Assistance Committee and Multilateral Official Commitments for Education and Basic Education, 1997-2000 233 Table 0.2. Official Commitments for Basic Education as a Percentage of Total Education Commitments, 1997-2000 234 Table D.3. Multilateral Official Commitments for Education and Basic Education, by Donor and Region, 1997-2000 235 Table D.4. Bilateral Official Commitments for Education and Basic Education, by Donor and Region, 1997-2000 236 Table D.5. Total Official Commitments for Education and Basic Education, by Region, 1997-2000 237 ANNEX CONTENTS 139 Table D.6. Official Development Assistance to Education by Region, All Sources, 1998-2000 238 Table D.7. Official Development Assistance to Basic Education in Sub-Saharan Africa, by Donor, 1998-2000 240 Table D.8. IDA Disbursements for Primary Education, by Expenditure Type and Region, Fiscal Years 1999-2002 241 Table D.9. Proportion of IDA Primary Education Disbursements for Capital and Recurrent Expenditures, by Region, Fiscal Years 1999-2002 241 140 TECHNICAL ANNEXES AN N EX Background Data and Aggregate Simulation Results A ABOUT THE SIMULATION MODEL A discussion of the key assumptions underlying the target variables used for the 2000- 2015 simulations may be found in chapter 4. This section describes the mechanics and key building blocks of the relatively simple simulation model we used. The stylized dynamics of some key variables. Achieving the MDG by 2015 implies that all children in an age cohort enter the system and remain in it until they have completed the primary cycle (or six years of the cycle if it lasts longer than that). Figure A. 1 shows how the student flow profile in a hypothetical coun- try would shift over time as it attains the MDG. In panel A, the profiles at four points in time are shown. In the initial year (2000), 75 percent of the school-age population enters first grade, and 40 percent of that population completes sixth grade. By 2005, the dotted line profile shows these rates rising to 90 percent and 65 percent, respectively. By 2010 the entry rate reaches 100 percent, while the completion rate climbs to 85 percent. By 2015, both the entry and completion rates reach 100 percent, signifying attainment of universal primary completion. In panel B of the figure, the same information is presented differently to emphasize that in order to attain the MDG by 2015, countries must necessarily have achieved universal entry to first grade by 2010, in the case of a five-year pri- mary system, or 2009, for a six-year cycle. Panel B also shows that the repetition rate would need to decline over the period (if it is high in the initial year), given that extensive grade repetition is correlated with dropout and is therefore incom- patible with the MDG goal. The student flow profiles in figure A. 1 are applied to projections of the school- age population in order to obtain estimates of total primary school enrollments over the course of the simulation period. Figure A.2 shows in panel A typical pat- terns in the size of the school-age population, as well as the corresponding number of primary school pupils and non-repeaters among them. Notice that as the year approaches 2015, the gap between total enrollments and the number of non- repeaters narrows, reflecting the assumed decline in repetition rates. Panel B of the figure shows the distribution of enrollments between publicly and privately financed schools. The share of pupils in the public sector is expected to grow because as schooling extends to the whole population of school-age children, the new entrants would increasingly be children from poorer families or otherwise deprived circumstances (rural children, girls, orphans). Our thesis is that the families of these children are neither willing nor able to afford fee-charging schools in the private sector. Hence, the simulation model assumes that the share of enrollments in the privately financed sector would fall as countries approach the 141 FIGURE A. 1 Stylized Dynamics of Selected Indicators of Student Flow, 2000-2015 Panel A Panel B Enrollment rate in age cohort (percent) Percent 100 - 2015 100 - 80 - - 00 80 - 6 2005 Intake 60 - 60 - Completion 40 - 2000 40- Repetition 20 - 20 - 0- I I 0- I I 0 1 2 3 4 5 6 1995 2000 2005 2010 2015 Grade FIGURE A.2 Stylized Dynamics of Primary School Enrollments, 2000-2015 Panel A Panel B Number of children Number of children School-age population Pupils enrolled Pupils in ,,, ~~~~~~~~~~All pupils schoolic Non-repeaters - Pupils in private schools 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 MDG by 2015. As noted in chapter 4, however, this assumption does not preclude growing enrollments in schools that are publicly funded but privately managed. Overview of the simulation model. The model has a simple and parsimonious structure that seeks to incorporate key policy variables relevant to projecting the EFA external financing gap while minimizing data requirements. For each country, estimation of the external financing gap involves four steps. The first step is to estimate the recurrent cost of achieving the MDG by 2015. Two sets of variables are relevant to the calculation. The first set pertains to the pro- jected number of children served by publicly funded schools. As indicated earlier, this number depends on: (a) the projected size of the school-age population over 142 TECHNICAL ANNEXES the simulation period; (b) the intake rate to first grade; (c) the share of first graders who complete the primary cycle; (d) the prevalence of grade repetition (which affects the efficiency of student flows through the system); and (e) the proportion of pupils enrolled in publicly funded schools. The second set of variables relates to the cost of service delivery in the publicly funded sector. The estimate depends on the following factors: (a) the ratio of pupils to teachers in publicly funded schools (which allows the number of teachers in such schools to be estimated); (b) the aver- age salaries of public sector teachers (which is combined with the estimate of teacher numbers to obtain the aggregate salary bill); and (c) the cost of comple- mentary school inputs to promote student learning as well as to finance targeted subsidies to the most disadvantaged children who might otherwise not attend school (such as those from the poorest families or hard-to-reach communities). Because of the high prevalence of HIV/AIDS in many African countries, a sepa- rate calculation is made to estimate the epidemic's impact on the cost of delivering primary education services. The epidemic's impact on the size of the school-age cohort is already incorporated into the population projections referred to in the pre- ceding paragraph. Beyond this effect, the simulation model allows for two additional effects: (a) the costs associated with increased teacher absenteeism, which effectively translates into an increase in the number of teachers needed to ensure service delivery beyond the number determined solely by the average pupil-teacher ratio; and (b) the cost to support the growing cohorts of orphans created by the epidemic (see box 4.1 in chapter 4 for more details on the AIDS-related part of the simulation model). The second step in the simulation involves estimation of the capital costs of build- ing and furnishing classrooms to accommodate the growing number of children enrolled in the publidy funded sector. The calculation involves determining the number of new dassrooms implied by the growth in enrollments, assuming that the pupil-dass- room ratio is the same as the pupil-teacher ratio (which is equivalent to assuming that teachers are assigned to teach all subjects to the same group of pupils). The unit cost to construct and furnish dassrooms ranges widely across countries. In these simulations, we set it at what regional experts consider to be a "good practice" level for each country. The third step in the simulation involves estimating the volume of resources that the country itself will be able to mobilize to reach the MDG. The amount mobilized depends on the behavior of the following factors over the course of 2001-2015: (a) the projected growth rate of the country's GDP; (b) the size of government revenues relative to the GDP; (c) the share of government revenues devoted to education; and (d) the share of public spending on education that is set aside specifically for primary education. We used a GDP growth projection of 5 percent per year for all countries across the period. The fourth and final step in the calculation is to determine the resource gap for external financing in order to achieve the MDG by 2015. The calculation simply involves subtracting the results of steps 1 and 2 from those of step 3. The algorithm is arranged to reflect the assumption that domestic resources would first be used to finance the recurrent costs of service delivery, and any surplus would be applied toward the capital costs. In countries where domestic resources are insufficient to cover the recurrent costs, the assumption is that external resources would be avail- able to help finance the shortfall in financing for both the recurrent and capital costs. ANNEX A * BACKGROUND DATA AND AGGREGATE SIMULATION RESULTS 143 Table A. 1 .................................................. Countries Included in Financing Simulation I. Countries eligible for IDA lending in 2001 (GDP per capita less than $885) II. Countries with populations less than 1 million Ill. Countries for which adequate data could not be obtained IV. Full sample of 55 countries V. Countries omitted from adjusted sample (see chapter 3, footnote 1, for rationale for omitting these) VI. Countries not modeled (estimated to have achieved or to be close to achieving UPC, defined as 90 percent completion through grade 6) VIl. Countries for which simulations were run (47) 1. IDA countries: 79 Afghanistan Chad . Haiti Mongolia Solomon Islands Albania Comoros Honduras Mozambique Somalia Angola Congo, Dem. Rep. of India Myanmar . SriLanka Armenia Congo, Rep. of Indonesia Nepal Sudan Azerbaijan Cote d'Ivoire Kenya Nicaragua Tajikistan Bangladesh Djibouti . Kiribati Niger . Tanzania Benin Dominica Kyrgyz Republic Nigeria Togo Bhutan * Eritrea . Lao PDR Pakistan Tonga Bolivia Ethiopia . Lesotho Rwanda Uganda Bosnia and Gambia, The Liberia St. Lucia Uzbekistan Herzegovina Burkina Faso Georgia Madagascar St. Vincent Vanuatu Burundi Ghana Malawi Sao Tome and Vietnam Principe Cambodia . Grenada Maldives Samoa Yemen, Rep. of Cameroon Guinea Mali Senegal Zambia Cape Verde Guinea-Bissau Mauritania Serbia and Montenegro Zimbabwe Central African Guyana * Moldova Sierra Leone Republic II. Countries with populations less than 1 million: 16 Bhutan Dominica Maldives St. Lucia Cape Verde Grenada Samoa . St. Vincent Comoros Guyana Sao Tome and Principe Tongo Djibouti Kiribati Solomon Islands . Vanuatu 144 TECHNICAL ANNEXES 111. Countries with inadequate data: 8 Afghanistana Liberia Serbia and Montenegro SriLanka Bosnia and Herzegovina Myanmar Somalia Tajikistan a. Included in alternative simulation. IV. Full sample: 55 Albania Central African Republic Guinea-Bissau Mali Senegal Angola Chad Haiti Mauritania Sierra Leone Armenia Congo, Dem. Rep. of Honduras Moldova Sudan Azerbaijan Congo, Rep. of India . Mongolia Tanzania Bangladesh Cote d'Ivoire Indonesia Mozambique Togo Benin Eritrea Kenya Nepal Uganda Bolivia Ethiopia Kyrgyz Republic Nicaragua Uzbekistan Burkina Faso Gambia, The Lao PDR Niger Vietnam Burundi . Georgia Lesotho Nigeria Yemen, Rep. of Cambodia Ghana Madagascar Pakistan Zambia Cameroon Guinea Malawi Rwanda Zimbabwe V. Countries omitted from adjusted sample: 6 Albania Armenia Azerbaijan Georgia Kyrgyz Republic Uzbekistan VI. Countries not modeled (PCR through grade 5 or 6 over 90 percent): 8 Albania *Bolivia . Kyrgyz Republic .Vietnam Azerbaijan Indonesia Uzbekistan Zimbabwe VIl. Final adjusted sample for which simulations were run: 47 Angola Congo, Rep. of India Nepal Uganda Armenia C6te d'Ivoire Kenya Nicaragua Yemen, Rep. of Bangladesh . Eritrea . Lao PDR Niger . Zambia Benin Ethiopia . Lesotho Nigeria Burkina Faso Gambia, The Madagascar Pakistan Burundi Georgia . Malawi Rwanda Cambodia Ghana Mali Senegal Cameroon Guinea Mauritania Sierra Leone Central African Republic Guinea-Bissau Moldova Sudan Chad Haiti Mongolia Tanzania Congo, Dem. Rep. of Honduras Mozambique Togo ANNEX A * BACKGROUND DATA AND AGGREGATE SIMULATION RESULTS 145 Table A.2 Core Education Parameters for 55 Lori-income Countries Z-DUCATION PRM Y EgDUCATION * . ~~~~~RECURRENT * * . * SPENDING ~~~RECURRENT SPENDINGENINGSpending on __________ _____ _________ - .Average inputs other * ~~~~~~~~~Government :annual than * ~~~~~~~~~current As % of :Unit teacher teachers Year* ~~~~~~~revenues government: cost salary (as (as % of (excluding : current As % of as % multiple primary Private ofa Primary .Primary grants) revenues, As % recurrent As % of per of per education Pupils enrollments Average of GER completion as % ecuig o edatn of capita capita recurrent per (as % of repetition Country data (% rate (% of GOP grants GOP spending GOP GOP GOP) spending) teacher: total) rate (% Albania 1998 I 103 91 19.5 12.6 2.5 41.3 1.0 7.7 1.4 17.5 22.7 0.6 0.0 Angola 2000 80 29 55.7 4.3 2.4 41.6 1.0 7.8 1.5 19.0 24.4 6.0 25.0 Armenia :2000 87 70 15.8 15.1 2.4 51.3 1.2 12.6 0.6 52.9 12.8 0.0 0.1 Azerbaijan 2000 107 99 20.8 18.4 3.8 19.7 0.8 5.9 0.9 15.8 17.6 1.0 0.4 Bangladesh 2000 112 70 12.8 14.7 1.9 49.9 0.9 6.6 2.7 25.0 55.2 12.5 15.0 Benin 1998 86 I 39 15.3 16.5 2.5 62.6 1.6 11.6 4.6 26.4 54.0 10.8 25.0 Bolivia 2000 113 97 21.2 25.2 5.3 47.5 2.5 15.9 2.2 19.4 20.6 9.4 3.7 Burkina Faso 1998 45 25 14.7 17.1 2.5 64.0 1.6 23.6 8.0 30.7 48.9 10.8 17.7 Buirundi 1998 60 43 17.4 20.4 3.6 35.5 1.3 12.4 5.3 22.1 55.1 0.0 27.5 Cambodia 200 14 6 15 7 51.0 0.9 3.6 15 20.0 53.3 06 Cameroon .1999 82 43 15.5 10.8 1.7 66.3 1.1 9.5 3.4 32.5 64.6 1 90 25.9 Central African Republic 2000 45 19 9.6 12.5 1.2 52.4 0.6 8.7 4.9 28.5 78.9 3.3 32.8 Chad 2000 71 19 8.0 20.9 1.7 65.5 1.1 10.1 4.8 34.2 72.0 8.8 24.6 Congo, Dem. Rep.of 2000 60 40 10.6 3.2 0.3 65.1 0.2 2.4 0.9 10.3 42.2 10.0 15.0 Congo, Rep. of: 2000 84 4426.7 8.6 23 36.6 0.8 7.0 3.4 20.3 601231.1 Cote d'lvoire * 1999 77 40 16.5 21.5 3.5 49.0 1.7 16.0 5.7 22.5 46.0 11.6 24.7 Eritrea 1999 53 35 34.6 8.0 2.8 53.6 1.5 22.2 7.7 29.6 49.2 10.1 19.4 Ethiopia 1999 55 24 17.8 15.0 2.7 46.2 1.2 14.0 6.8 20.5 61.3 5.0 , 8.0 Gambia,The 2000 88 70 18.5 16.6 3.1 51.7 1.6 13.2 3.7 24.9 37.0 8.5 10.6 Georgia 1999. 86 79 13.7 9.3 1.3 26.0 0.3 4.2 0.6 16.0 16.5 1.0 0.5 Ghana 1999' 79 * 64 * 21.8 . 17.6 3.8 37.2 1.4 12.7 3.6 17.7 34.1 18.0 5.0 Guinea 2000 62 34 11.1 18.1 2.0 37.2 0.8 8.4 2.7 34.7 48.9 16.1 23.3 Guinea-Bissau 2000 70 31 19.6 9.8 1.9 35.0 0.7 6.7 1.6 34.3 37.4 8.5 27.1 Haiti 1997 112 40 9.3 20.6 1.9 38.7 0.7 16.2 6.8 10.0 46.3 76.6 17.0 Honduras 2000 105 67 18.5 30.1 5.6 49.6 2.8 17.8 5.0 12.0 31.7 6.9 8.0 India 1999 101 76 21.2 12.4 2.6 38.5 1.0 8.4 3.4 23.2 51.8 12.5 20.0 Indonesia 2000 106 91 18.4 9.7 1.8 59.3 1.1 10.3 1.6 19.9 19.9 15.7 5.9 Kenya 1999 91 64 24.2 26.2 6.3 44.2 2.8 17.6 5.3 4.2 31.4 2.2 14.2 Kyrgyz Rep. 1999 100 99 18.7 22.0 4.1 36.0 1.5 10.3 1.2 21.8 15.2 0.0 0.0 Lao PDR 1999 121 67 38.1 1.4 0.5 58.6 0.3 1.5 0.4 19.6 30.6 1.9 22.6 Lesotho .2000 103 70 35.9 * 22.2 8.0 40.2 * 3.2 20.8 6.6 29.9 45.3 0.0 18.3 Madagascar 1998 90 26 10.6 18.8 2.0 54.7 1.1 10.8 3.3 42.4 53.7 22.0 33.0 Malawi 1999 117 65 18.1 19.8 3.6 49.2 1.8 8.8 4.0 14.0 52.8 2.0 14.7 Mali 1998 49 23 16.8 13.7 2.3 42.1 1.0 14.3 6.1 31.1 61.0 21.2 17.9 Mauritania *. 1998 88 46 26.5 13.7 3.6 49.0 1.8 13.1 5.1 18.2 48.0 1.8 160 Moldova 1999 81 79 29.8 18.5 5.5 25.5 1.4 15.8 1.1 67.8 21.4 0.0 0.9 Mongolia 1999 92 67 29.2 24.6 7.2 33.6 2.4 15.1 3.9 15.0 30.7 0.5 0.9 Mozamnbique 1998 I 79 36 11.3 18.1 2.0 46.4 1.0 7.9 3.2 26.1 54.4 0.0 23.7 Nepal 1998 113 56 10.4 18.4 1.9 63.8 1.2 7.1 2.0 20.0 35.9 8.1 29.9 Nicaragua 2000 101 65 25.3 16.8 4.3 48.7 2.1 13.2 3.2 32.7 35.9 6.8 12.0 Niger 1998 31 20 9.1 31.5 2.9 62.0 1.8 35.5 9.6 25.9 36.5 4.0 13.0 Nigeria 2000 85 67 46.1 9.9 4.6 41.0 1.9 13.8 4.9 9.1 39.0 1.0 1.0 Pakistan 2000 65 59 16.7 10.2 1.7 51.8 0.9 14.0 3.6 19.3 32.1 29.4 6.2 (Continued) ub Table A.2 (C o ntin uied ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ EDUCATIDNCPR IMAR RECURRENT EDCTO SPENDI ~~RECURRENT Spending on SPENDING ~~Average inputs other Government annual than current As /o Or Unit teacher teachers revenues government: cost salary (as (as % of (excluding current As /0 of as%/ multiple primary Private Year Primary Primary grants) revenues, As % recurrent As %*: of per of per education Pupils enrollments Average of GER completion as%/ excluding of education of capita capita recurrent per (as % of repetition Countr-y data (/0 rate (% of GDP grants GDP spending GDP GDP GOP) spending teacher total) rate (% Rwanda 2000 101 28 9.8 32.6 3.2 44.7 1.4 91i 4.0 8.6 47.7 0.8 36.1 Senegal 2000 70 41 18.1 18.6 3.4 43.9 1.5 14.2 4.9 36.6 54.7 10.7 13.6 Sierra Leone 2000 64 37 11.4 30.4 3.5 51.3 1.8 16.4 4.3 33.1 39.6 0.0 9.3 Sudan 2000 61 35 11.1 16.2 1.8 50.5 0.9 10.3 2.2 22.5 27.5 0.0 1.2 Tanzania 1999 66 59 10.9 16.4 1.8 63.0 1.1 10.0 3.6 11.2 40.0 0.0 3.2 Togo 1999 115 63 14.9 25.6 3.8 48.3 1.8 13.2 4.5 25.2 45.5 35.6 27.0 Uga-nda 2000 102 82 10.8 30.1 3.2 53.2 1.7 9.8 2.9 26.2 40.9 2.0 9.8 Uzbekistan 1999 100 100 32.4 22.3 7.2 41.2 3.0 18.7 2.9 27.0 21.0 0.0 0.1 Vietnam 1998 110 99 22.3 10.9 2.4 46.5 1.1 7.1 1.2 45.0 30.4 0.0 3.5 Yemen, Rep. of 1998 67 58 35.2 15.8 5.6 48.0 2.7 18.5 3.4 26.7 25.0 1.4 7.0 Zambia 1998 85 72 18.8 12.3 2.3 43.2 1.0 6.9 2.7 21.7 50.0 1.6 6.2 Zimbabwe 1997 112 103 27.4 28.3 7.1 46.1 3.3 19.4 6.1 25.0 39.0 11.0 2.0 Average, all countries 86 57 19.7 17.3 3.2 47.3 1.4 12.2 3.7 24.5 40.9 8.4 14.1 Average, Afirican countries 77 45 19.0 17.8 3.0 49.2 1.4 12.9 4.5 24.1 47.7 8.2 17.7 Average, best performers ___ 103 83 20.7 18.2 3.8 47.6 1.7 11.8 3.3 26.0 39.0 7.3 9.5 Niote: Primary GERs and PCRs presented here have been adjusted to an equivalent basis for purposes of comparison. For countries with five-year or six-year systems, the values presented here correspond to the official cycle. But for countries whose primary cycle is shorter or longer than this, we have estimated a six-year equivalent GER and PCR. For this reason, data in this and subsequent tables may differfrmm completion-rate data in chapter 2. Table A.3 ..................................................................................................................... Key Education System Parameters for Countries, Grouped by Relative EFA Success (Adjusted Sample) * * ~~~~~~~~PRIMARY EDUCATION RECURRENT EDUCATION , , .S ENDIN RECURRENT .EDUCATION.SD GSpending on . . * * SPENDING SPENDING Average inputs other .Government annual than * . . current As % of Unit teacher teachers revenues government cost salary (as (as%of , * : (excluding current As % of . as % multiple primary Private Primary Primary grants) revenues, As% recurrent As % of per of per education Pupils enrollments Average GER . completion . as% excluding of education, of capRta capita recurrent per (as%of repetition Country (%) rate (%) of GDP grants GDP spending GDP GDP GDP) spending) teacher total) rate (%) Group 1: High enrollment (GER 85% or above) and high completion (PCR 70% or above) Bangladesh . 112 70 12.8 14.7 1.9 49.9 0.9 6.6 2.7 25.0 55.2 12.5 15.0 Bolivia 113 97 21.2 25.2 5.3 47.5 2.5 15.9 2.2 19.4 20.6 9.4 3.7 Gamnbia,The 88 70 18.5 16.6 3.1 51.7 1.6 13.2 3.7 25.0 37.0 8.0 10.6 India 101 76 21.2 12.4 2.6 38.5 1.0 8.4 3.4 23.2 51.8 12.5 20.0 Indonesia 106 91 18.4 9.7 1.8 59.3 1.1 10.3 1.6 19.9 19.9 15.7 5.9 Lesotho 103 70 35.9 22.2 8.0 40.2 3.2 20.8 6.6 30.0 45.3 0.0 18.3 Uganda 102 82 10.8 30.1 3.2 53.2 1.7 9.8 2.9 26.0 40.9 2.0 9.8 Vietnam 110 99 22.3 10.9 2.4 46.5 1.1 7.1 1.2 45.0 30.4 0.0 3.5 Zambia 85 72 18.8 12.3 2.3 43.2 1.0 6.9 2.7 22.0 50.0 2.0 6.2 Zimbabwe 112 103 27.4 28.3 7.1 46.1 3.3 19.4 6.1 25.0 39.0 11.0 2.0 Average 103 83 20.7 18.2 3.8 47.6 1.7 11.8 3.3 26.1 39.0 7.3 9.5 Group 2: High enrollment (GER 80% or above) but low completion (PCR 60% or lower) Angola 80 29 55.7 4.3 2.4 41.6 1.0. 7.8 . 1.5 19.0 *24.4 6.0 25.0 Benin 86 39 15.3 , 16.5 .2.5 62.6 1.6 11.6 . 4.6 26.0 54.0 11.0 25.0 Cameroon 82 43 15.5 10.8 1.7 66.3 1.1 9.5 3.4 33.0 64.6 19.0 25.9 Congo, Rep. of: 84 44 26.7 8.6 2.3 36.6 0.8 7.0 3.4 20.0 61.0 15.0 31.1 Madagascar 90 26 10.6 18.8 2.0 54.7 1.1 10.8 3.3 42.0 53.7 22.0 . 33.0 Mauritania . 88 * 46 26.5 13.7 3.6 49.0 1.8 13.1 5.1 18.0 48.0 2.0 16.0 Nepal 113 56 10.4 18.4 1.9 63.8 1.2 7.1 2.0 20.0 35.9 8.1 29.9 Rwanda 101 28 9.8 32.6 3.2 44.7 1.4 9.1 * 4.0 9.0 . 47.7 1.0 36.1 Average ,91 39 21.3 15.5 2.5 52.4 *1.3 9.5 3.4 23.4 48.7 10.5 27.8 (Continued) Table A.3 .............. (Continued) ~OUCATIOPRIMARY EDUCATIONTIO * '*'RECURRENT ' E'JCRREKINT Spendingon * . SPEENDIN G * SPENDING Average inputs other *Government : annual than : : : current As % of Unit * teacher teachers revenues government cost salary (as (as%of . * * (excluding current As % of . as % multiple primary Private Primary Primary grants) revenues, As % recurrent As % ofiper of per * education Pupils enrollments Average GER .completion as% excluding of education of capita capita recurrent per (as % of repetition Country (%) rate (%) of GOP grants GOP spending GOP GOP GOP) spending) teacher total) rate (%) Group 3: Low enrollment and low completion (GER and PCR both 60% or lower) Burkina Faso 45 25 14.7 17.1 2.5 64.0 1.6 23.6 8.0 31.0 48.9 11.0 17.7 Burundi 60 43 17.4 20.4 .3.6, 35.5 .1.3 12.4 5.3 * 22.0 55.1 0.0 27.5 Central African 45 19 9.6 12.5 1.2 52.4 0.6 8.7 4.9 29.0 .78.9 * 3.0 32.8 Republic Eritrea 53 35 34.6 8.0 2.8 53.6 1.5 22.2 7.7 30.0 49.2 10.0 19.4 Ethiopia * 55 24 178 15.0 2.7 46.2 1.2 14.0 6.8 21.0 61.3 5.0 8.0 Mali 49 23 16.8 13.7 2.3 42.1 1.0 14.3 6.1 31.0 61.0 21.0 17.9 Niger 31 20 9.1 31.5 2.9 62.0 1.8 35.5 9.6 26.0 36.5 4.0 13.0 Average 48 27 17.1 16.9 2.6 50.8 1.3 18.7 6.9 27.1 55.8 7.7 19.5 Group 4: Remaining countries Cambodia . 134 60 11.5 15.0 1.7 : 51.0 0.9 3.6 1.5 20.0 53.3 0.0 16.6 Chad 71 19 8.0 20.9 .1.7 65.5 ,1.1 10.1 * 4.8 34.0 . 72.0 9.0 , 24.6 Congo, Dem. 60 40 10.6 3.2 0.3 65.1 0.2 2.4 0.9 10.0 * 42.2 10.0 15.0 Rep. of Cote d'lvoire 77 40 16.5 21.5 3.5 49.0 1.7 16.0 5.7 . 22.0 . 46.0 12.0 24.7 Ghana 79 64 21.8 * 17.6 .3.8 37.2 1.4 12.7 3.6 18.0 ,34.1 18.0 5.0 Guinea 62 34 11.1 18.1 .2.0 37.2 .0.8 8.4 2.7 35.0 48.9 16.0 23.3 Guinea-Bissau 70 31 19.6 9.8 1.9 35.0 0.7 6.7 1.6 34.0 37.4 9.0 27.1 Haiti 112 40 9.3 206 1.9 38.7 0.7 16.2 6.8 10.0 46.3 76.6 17.0 Honduras 105 67 18.5 30.1 5.6 49.6 * 2.8 17.8 5.0 12.0 31.7 6.9 8.0 Kenya 91 64 24.2 26.2 6.3 44.2 2.8 17.6 5.3 4.0 31.4 2.0 14.2 Lao PDR 121 67 38.1 1.4 0.5 58.6 0.3 1.5 0.4 19.6 30.6 1.9 22.6 Malawi 117 65 18.1 19.8 3.6 49.2 * 1.8 8.8 4.0 14.0 52.8 2.0 14.7 Moldova * 81 * 79 29.8 18.5 5.5 25.5 1.4 15.8 1.1 67.8 21.4. 0.0 0.9 Mongolia 92 67 29.2 24.6 7.2 33.6 2.4 15.1 3.9 15.0 30.7 0.5 0.9 Mozambique 79 39 11.3 18.1 2.0 46.4 1.0 7.9 3.2 26.0 54.4 00 23.7 Nicaragua 101 65 25.3 16.8 4.3 48.7 2.1 13.2 3.2 327 35.9 68 12.0 Nigeria 85 67 46.1 9.9 4.6 41.0 1.9 13.8 4.9 9.0 39.0 1.0 1.0 Pakistan 65 59 16.7 10.2 1.7 51.8 0.9 14.0 3.6 19.3 32.1 29.4 6.2 Senegal 70 41 18.1 18.6 3.4 43.9 1.5 14.2 4.9 37.0 54.7 11.0 13.6 Sierra Leone 64 37 11.4 30.4 3.5 51.3 1.8 164 4.3 33.0 39.6 0.0 9.3 Sudan 61 46 11.1 16.2 1.8 50.5 0.9 10.3 2.2 23.0 27.5 0.0 1.2 Tanzania 66 59 10.9 16.4 1.8 63.0 1.1 10.0 3.6 11.0 40.0 0.0 3.2 Togo 115 63 14.9 25.6 3.8 48.3 1.8 13.2 4.5 25.0 45.5 36.0 27.0 Yemen, Rep. of 67 58 35.2 15.8 5.6 48.0 2.7 18.5 3.4 26.7 25.0 1.4 7.0 Average 85 53 19.5 17.7 3.3 47.2 1.4 11.8 3.5 23.3 40.5 10.4 13.3 Average 84 53 19.7 17.3 3.1 48.6 1.5 12.4 4.0 24.4 43.7 9.4 15.8 Adjusted Sample . . . *. . *. . .__ .._'_ Note: Primary GERs and PCRs presented here have been adjusted to an equivalent basis for purposes of comparison. For countries with five-year or six-year systems, the values presented here correspond to the official cycle. But for countries whose primary cycle is shorter or longer than this, we have estimated a six-year equivalent GER and PCR. I' PA Table A.4 .................................................................................................. Group 1 Parameters with Six Eastern European and Central Asian Countries Included * EDUCATION * PRIMARY * RECURRENT EDUCATION SPENDING RECURRENT Spendingon * * * ON SPENDING SPENDING . Average inputs other Government : annual: than current As % of Unit teacher teachers revenues government . cost salary (as (as % of (excluding current As % of as % multiple primary Private Primary Primary grants) revenues, As% recurrent As % of per of per education Pupils enrollments Average G GER completion as% excluding of education of capita . capita recurrent per (as % of repetition Country (%) rate (%) of GOP grants GDP . spending GOP gdp gdp) spending) ::teacher total) rate (%) Albania 104 91 19.5 12.6 2.5 41.3 1.0 7.7 1.4 17.5 22.7 0.6 0.0 Armenia 87 70 15.8 15.1 2.4 51.3 1.2 12.6 0.6 52.9 12.8 0.0 0.1 Azerbaijan 107 99 20.8 18.4 3.8 19.7 0.8 * 5.9 0.9 15.8 17.6 1.0 0.4 Bangladesh 112 70 12.8 14.7 1.9 49.9 0.9 6.6 2.7 25.0 55.2 12.5 15.0 Bolivia 113 97 21.2 25.2 5.3 47.5 2.5 15.9 2.2 19.4 20.6 9.4 3.7 Gambia,The 88 70 18.5 16.6 3.1 51.7 1.6 13.2 3.7 25.0 37.0 8.0 10.6 Georgia . 86 79 13.7 9.3 1.3 26.0 0.3 4.2 o06 16.0 16.5 1.0 0.5 India 101 76 21.2 12.4 2.6 38.5 1.0 8.4 3.4 23.2 51.8 12.5 20.0 Indonesia 106 91 18.4 9.7 1.8 59.3 1.1 10.3 1.6 19.9 19.9 15.7 5.9 KyrgyzRep. 100 99 18.7 22.0 4.1 36.0 1.5 10.3 1.2 21.8 15.2 0.0 0.0 Lesorho 103 70 35.9 22.2 8.0 40.2 3.2 20.8 6.6 30.0 45.3 0.0 18.3 Uganda 102 82 10.8 30.1 3.2 53.2 1.7 9.8 2.9 26.0 40.9 2.0 9.8 Uzbekistan 100 100 32.4 22.3 7.2 41.2 3.0 18.7 2.9 27.0 21.0 0.0 0.1 Zambia 85 72 18.8 12.3 2.3 43.2 1.0 6.9 2.7 22.0 50.0 2.0 6.2 Zimbabwe 112 103 27.4 28.3 7.1 46.1 3.3 19,4 6.1 25.0 39.0 11.0 2.0 Vietnam 110i 99 22.3 10.9 2.4 46.5 1.1 7.1 1.2 45.0 30.4 0.0 35 Average 101 86 21.0 17.6 3.7 43.2 16 11.1 2.5 25.7 31.0 4.7 6.0 Table A.5 ..................................................................................................... Demographic Burden and Resource Mobilization, by Country and Region, 2000 and 2015 Total Number of Recurrent available * *.Number of *public Per-pupil resources . . . . . students * spending . spending on for GDP per enrolled in on primary public primary capita *School-age, Total * primary school education education education (2000 . n . *. populatiion.- .population Demographic * (5 or 6 years, (millions of (PCGDP . (millions of constant Country Year (thousands) (thousands) burdena thousands) dollars) unit) dollars) dollars) Angola 2000 2,182 12,717 0.17 1,728 88 0.08 88 695 . 2015 3,180 18,961 0.17 3,498 400 0.13 331 969 Armenia 2000 432 3,827 0.11 375 24 0.11 24 500 ' 2015 282 4,061 0.07 282 33 0.13 64 980 Bangladesh 2000 15,748 129,754 0.16 17,668 . 285 0.07 285 281 * 2015 * 14,941 166,142 0.11 16,435 * 884 0.13 890 * 456 Benin 2000 1,050 5,948 0.18 907 31 0.12 31 326 . 2015 1,411 8,972 0.16 1,552 100 0.14 71 496 Burkina Faso 2000 1,805 10,730 0.18 816 33 0.24 33 192 * 2015 * 2,567 15,340 0.17 2,824 * 123 0.16 66 309 Burundi 2000 1,123 * 6,548 0.18 671 7 0.12 7 83 : 2015 : 1,460 8,796 0.17 1,606 30 : 0.15 : 18 : 142 Cambodia 2000 1,803 12,021 0.15 2,408 24 0.04 24 282 2015 1,807 14,757 0.12 1,988 112 0.13 . 99 478 Cameroon 2000 2,451 14,691 0.17 2,010 84 0.10 84 518 2015 3,299 19,264 0.17 3,629 380 0.13 266 863 Cenrral African* 2000 590 3,597 0.16 267 6 0.09 6 267 Republic 2015 675 4,461 0.15 743 39 0.13 28 448 Chad 2000 1,283 7,694 0.20 914 15 0.10 15 183 2015 1,956 11,661 0.17 2,152 71 0.15 43 252 Congo, Dem. 2000 8,748 51,390 0.17 5,270 9 0.02 9 75 Rep. of 2015 12,779 74,952 0.17 14,057 178 0.13 112 107 Congo, Rep. of* 2000 497 2,936 0-17 419 27 0.07 27 1,082 : 2015 705 4,326 0.16 775 * 140 0.13 119 1,526 C6te d'lvoire 2000 2,471 15,545 0.16 1,911 189 0.16 189 700 2015 2,828 20,250 0.14 3,111 526 0.16 460 1,174 (Continued) Table A.5 .............. (Continued) Recurrent via3 C1Um>z: * . ;161s * [er-PuRil * OrSD?qces stuE3e6s spending spendingc an Mar . per .nrolWJ in on pr6mary public .primal capita 7 '. * r * (3 cr$ 3 v7rs, ( .riioE1ns . (o T MG5 . (milliionS '~ coWIstaM8l 1 IL, !. CIIrI * l1sa . (Ohousands) i ( -F. sf z- _). r-dsN , :.dzE-_'*CXiars) . un>t) t EC:127S) . 3a;Svs) Eritrea 2000 559 3,991 0.17 296 8 0.22 8 140 * 2015 744 5,616 0.16 819 26 0.16 14 217 Ethiopia 2000 10,719 62,782 0.18 5,847 74 0.14 74 95 . 2015 14,173 87,895 0.16 15,307 364 0.15 212 173 Gambia, The 2000 192 1,286 0.15 168 . 7 0.13 . 7 322 *. 2015 258 1,769 0.15 284 17 0.13 12 487 Georgia 2000 510 5,452 0.09 438 11 0.04 11 612 . 2015 371 5,349. 0.07 373 60 0.13 117 1,362 Ghana 2000 3,264 18,785 0.17 2,578 54 0.13 54 201 . 2015 3,851 24,349 0.16 4,044 163 0.13 148 338 Guinea 2000 1,270 7,415 0.17 790 23 0.08 23 421 * 2015 1,516 9,845 0.15 1,668 130 0.13 91 659 Guinea-Bissau 2000 191 1,207 0.16 133 I 0.07 I 179 2015 235 1,586 0.15 259 9 0.13 6 283 Haiti 2000 1,306 7,492 0.17 1,459 18 ' 0.16 18 325 2015 1,352 10,039 0.13 1,487 151 0.15 121 751 Honduras 2000 1,064 6,653 016 1,109 173 0.18 173 944 2015 1,154 8,890 0.13 1,246 216 0.13 235 1,469 India 2000 112,970 997,515 0.14 113,610 3,666 0.08 3,666 437 2015 97,346 1,221,862 0.10 107,081 9,840 0.13 12,775 778 Kenya 2000 5,248 29,410 0.18 4,792 272 0.18 272 330 2015 5,752 37,185 0.16 6,327 579 0.18 445 570 LaoPDR 2000 710 5,097 0 17 859 3 0.02 3 256 2015 918 7,214 0.15 1,010 47 0.13 38 395 Lesotho .2000 324 2,154 0.15 333 31 0.21 31 454 * 2015 371 2,450 0.15 408 59 0.19 41 : 830 Madagascar 2000 2,114 . 14,592 . 0.18 1,893 33 0.11 33 * 207 . 2015 2,838 22,247 0.15 3,122 * 115 0.13 81 311 Malawi 2000 1,879 10,788 0.18 2,198 24 0.09 24 124 . 2015 2,397 14,229 0.17 2,636 68 0.14 41 206 Mali :2000 1,766 10,334 0.18 863 21 0.14 21 * 209 . 2015 2,813 14,876 0.19 3,094 131 0.14 69 * 332 Mauritania * 2000 . 399 2,529 0.17 353 . 14 0.13 14 * 315 .2015 561 3,675 0.15 618 42 0.15 29 497 Moldova * 2000 469 4,281 0.11 378 . 14 0.16 14 229 :2015 329 4,178 0.08 332 20 0.13 34 513 Mongolia 2000 413 2,378 0.17 380 20 0.15 20 * 341 .2015 328 3,019 . 0.11 331 26 0.15 28 587 Mozambique 2000 2,396 . 15,705 . 0.18 1,881 38 * 0.08 . 38 . 252 . 2015 2,768 . 22,614 0.15 3,045 170 0.13 125 471 Nepal . 2000 3,167 . 22,851 0.17 3,588 43 * 0.07 43 185 :________ 2015 :3,736 32,546 0.14 4,110 164 0.13 129 338 Nicaragua 2000 841 5,044 0.16 848 49 0.13 49 . 469 2015 897 6930 0.13 987 83 0.13 . 79 709 Niger 2000 1,700 * 10,146 0.18 530 31 0.36 31 170 . 2015 2,940 16,691 0.18 3,234 110 0.16 72 237 Nigeria 2000 20,457 126,910 0.16 17,411 772 0.14 772 324 2015 29,585 169,441 0.17 29,881 2,100 0.15 1,537 504 Pakistan 2000 17,417 138,080 0.16 12,453 526 0.14 526 428 . 2015 22,217 192,948 0.13 23,594 1,776 0.13 1,653 637 Rwanda 2000 1,336 8,508 0.16 1,351 26 0.09 26 211 * 2015 1,978 11,149 0.18 2,176 90 0.14 68 334 Senegal 2000 1,596 9,530 0.17 1,109 65 0.14 65 460 . 2015 2,192 13,342 0.16 2,411 220 0.15 146 683 Sierra Leone 2000 858 5,031 0.17 548 11 0.16 11 126 . 2015 1,007 6,745 015 1,101 28 0.14 24 195 Sudan 2000 4,277 29,677 0.14 2,612 104 0.10 104 384 . 2015 5,950 40,551 0.15 6,021 416 0.13 332 584 Tanzania * 2000 5,589 32 923 0.17 3,710 89 0.10 89 239 11 . 2015 6,984 43,556 0.16 7,210 339 0.13 241 395 (Continued) Table A.5 (Continued) m~er o~ eRecurrenmt avaiiable ' Nunelber ol * quNlic Per-pupiD resources . studevs spending spending on for . per * . * * ^ eavolerde on pIlwDJry Rtlic pnmarv c29RS *. * .Zc -ao . .ri11try schoiol e ducation educatVo * education (2083 B * 19 fL: .populetion .Demographc (5 or 6 years, (*tDDons of (PiGf (milHDons of constant CoUntry __s____s (tlzS2'S) (tiLwswnds) .wdn ODD.- sznds) d.aDtrs) Togo 2000 829 4,894 0.16 954 23 0.13 23 252 *. 2015 1,118 , 6,706 0.16 1,230 60 014 48 401 Uganda 2000 3,885 22,063 0.18 3,963 107 0.10 107 281 . 2015 5,053 . 31,437 .16 5,549 269 0.13 235 410 Yemen, Rep. 2000 3,141 18,046 0.18 2,644 247 0.18 247 514 of .2015 4,617 27,276 0.14 4,940 6 624 . 0.13 309 1,069 Zambia 2000 1,667 9,666 0.18 1,416 19 0.07 19 200 2015 1,980 . 11,982 0.17 2,103 . 92 0.13 62 * 371 AFR 2000 94,716 * 572,121 0.17 70,642 2,334 0.13 2,335 304 * 2015 127,925 786,919 * 0.16 136,492 7,583 0.14 5,595 478 EAP * 2000 6,093 42,347 * 0.16 7,234 * 90 0.07 90 * 266 * 2015 * 6,789 57,536 0.13 7,438 349 0.14 295 450 ECA . 2000 1,411 13,560 0.10 1,191 48 0.10 48 447 * 2015 982 13,588 0.07 987 113 0.13 215 952 LCR . 2000 * 3,211 . 19,189 o .I6 3,416 * 240 O.I6 240 579 * 2015 3,403 25,859 . 0.13 3,720 450 0.14 434 976 SAR . 2000 * 146,135 1,265,349 0.15 143,731 4,477 0.10 4,477 382 . 2015 134,504 1,580,952 0.11 147,110 12,501 0.13 15,318 624 MNA . 2000 17,417 138,080 0 .16 12,453 526 0.14 526 428 * 2015 22,217 192,948 0.13 * 23,594 . 1,776 0.13 . 1,653 637 Total 2000 254,707 1,930,612 n.a. 228,858 7,438 n.a. 7,438 n.a. * 2015 278,220 2,492,130 n.a. 300,688 21,620 n.a.. 22,165 * n.a. Average 2000 n.a. * n.a. 0.16 n.a. . n.a. * 0.12 n.a. . 337 , 2015 n.a. * n.a. . 0.15 n.a. * n.a. 0.14 * n.a. 559 n.a. Not applicable. a. Population aged 7-12 as a share of the total population. Table A. 6 Actual and Projected Expenditures on Primary Education, Domestic Financing, and Financing Gap, by Country, under Scenario C2 (millions of U.S. dollars per year) PROJECTED UNDER SCENARIO C2 (ANNUAL AVERAGE FOR PERIOD 2001-201 5) * * * RECURRENT * RECURRENT : Actual EXPENDITURE GAP expenditures Domestic Capital Total Country (circa 2000) financing Operation AIOS expenditure expenditure Operation AIDS Capital Total gap Angola 88 253 289 *2 5 296 * 36 2 5 44 Armenia 24 33 47 *0 0 * 47 15 0 0 15 Bangladesh * 285 * 556 . 568 .0 71 . 640 12 0 71 84 Benin 37 . 52 62 * I 10 73 11 10 21 Burkina Faso 40 51 76 11 23 , 1 25 11 23 60 Burundi 11 12 18 5 11 34 6 5 11 21 Central African 6 15 20 2 7 29 5 2 7 14 Republic Cambodia * 24 . 57 76 0 4 80 19 0 4 23 Cameroon 102 175 192 11 28 231 18 11 28 56 Chad 15 . 28 * 33 2 19 55 6 2 19 27 Congo, Dem. 9 54 97 *17 103 * 217 43 :17 103 163 Rep. of : : : - : : : : : Congo, Rep. of 27 . 70 72 .I 6 80 3 .I 6 10 C6te d'lvoire 218 307 347 16 16 379 40 .16 16 72 Eritrea 10 13 17 0 6 23 3 0 6 10 Ethiopia * 80 141 196 .30 125 351 55 :30 125 210 Gambia, The 7 9 11 : 0 1 13 . 2 :0: 1 * 4 Geri . _ __ :_ _ :O : : 13 :0 : 0 Georgia * 11 . 40 . 53 0 0 53 13 0 0 13 (Continued) u' ni C=5 Table A.6 .............. (Continued) * * PROJECTED UNDER SCENARIO C2 (ANNUAL AVERAGE FOR PERIOD 2001-2015) * . * RECURRENT : * RECURRENT * Actual EXPENDITURE i GAP expenditures Domestic I . Capital Total Country (circa 2000) financing Operation AIDS expenditure expenditure Operation AIDS Capital Total gap Ghana 111 100 107 6 14 127 7 6 14 27 Guinea . 23 53 76 1 12 * 89 23 1 12 37 Guinea-Bissau . I . 4 6 :O: 1 7 2 0 1 . 4 Haiti * 23 65 76 :0 23 99 * 11 *0 23 34 Honduras . 173 . 203 . 170 :0 37 * 208 . 0 0 5 * 5 India . 3,827 7,685 7,317 0 435 7,752 0 0 67 67 Kenya * 295 316 429 21 0 450 113 21 0 134 Lao PDR 6 26 37 0 2 40 11 0 2 13 Lesotho * 31 36 44 2 1 46 8 2 I 11 Madagascar * 41 58 70 0 21 91 12 0 21 33 Malawi 32 33 42 10 9 62 9 10. 9 29 Mali 25 45 68 2 28 98 23 2 28 53 Mauritania * 18 23 27 0 3 30 4 0 3 8 Moldova 16 21 26 0 0 26 5 0 0 5 Mongolia 20 21 21 0 0 21 0 0 0 0 Mozambique 38 80 103 9 14 126 23 9 14 46 Nepal 46 89 119 0 14 133 30 0 14 44 Nicaragua 49 65 67 0 6 73 3 0 6 8 Niger 37 46 72 1 27 100 26 I 27 54 Nigeria 772 1,275 1,367 58 122 1,547 92 58 122 272 Pakistan 526 1,061 1,117 0 148 1,265 55 0 148 204 Rwanda 26 39 50 7 11 68 11 7 11 29 Senegal 65 103 122 :2 19 142 19 *2 19 40 Sierra Leone . 1 15 19 *2 6 * 27 . 4 :2: 6 12 Sudan 104 205 290 1 20 311 85 1 20 106 Tanzania 96 164 216 22 28 265 52 22 28 102 Togo 26 31 39 4 7 51 9 4 7 20 Uganda 107 146 195 24 14 233 49 24 14 87 Yemen, 247 295 316 0 49 365 21 0 49 70 Rep. of Zambia 32 41 58 15 9 81 16 15 9 40 Total 7,818 14,210 14,840 285 1,515 16,645 1,035 285 1,115 2,441 a. Excluding projected financial surpluses in three countries, this would total $17,751 million. Table A. 7 ................................................................................................. Domestic Resources Mobilized and Financing Gap, by Country, under Scenarios Cl and C3 (millions of U.S. dollars per year) * SCENARIO Cl SCENARIO C3 RECURRENT RECURRENT:; Domestic GAP Capital Total Domestic GAP Capital Total Country .resources- Operation: AIDS gap gap .resources: Operation. AIDS: gap gap Angola 253 37 2 5 44 479 0 0 0 0 Armenia 33 15 0 0 15 33 15 0 0 15 Bangladesh 556 12 0 71 84 556 12 0 71 84 Benin 52 11 1 10 21 52 11 1 10 21 Burkina Faso 51 25 11 23 60 52 23 11 23 58 Burundi 13 5 5 11 21 14 4 5 11 20 CentralAfrican: 15 5 2 7 14 15 5 2 7 14 Republic Cambodia . 57 19 .0 4 .23 57 . 19 .0 4 ,23 Cameroon 175 18 *11 28 *56 175 18 *11 28 56 Chad 29 5 2 19 26 28 5 2 19 27 Congo, Dem. 54 43 17 103 163 54 43 17 103 163 Rep. of Congo, Rep. of 70 3 1 6 10 89 0 0 0 0 C6te d' Ivoire 321 26 16 16 58 307 40 16 16 72 Eritrea 13 3 0 6 10 22 0 0 1 2 Ethiopia 141 55 30 125 210 164 32 30 125 187 Gambia, The 9 2 0 1 4 11 0 0 1 2 Georgia 40 13 :0 0 .13 40 13 0 : 0 13 Ghana 100 7 .6. 14 27 113 0 6 8 14 Guinea 53 23 1 12 .37 53 23 .1. 12 37 Guinea-Bissau. 4 2 0 *1 4 5 1 *0 1: 3 Haiti 66 10 0 23 33 65 11 0 23 34 Honduras 208 0 0 0 0 204 0 0 3 3 India 7,685 0 0 67 67 7,724 0 0 28 28 Kenya 368 61 21 0 82 397 32 21 0 53 Lao PDR 26 11 0 2 13 33 4 0 2 6 Lesotho 38 6 2 1 9 51 0 0 0 0 Madagascar 58 12 0 21 33 58 12 0 21 33 Malawi 33 9 10 9 29 38 4 10 9 23 Mali 45 23 2 28 53 51 17 2 28 47 Mauritania 23 4 0 3 8 31 0 0 0 0 Table A. 7 .............. (Continued) SCENARIO C 1 SCENARIO C3 RECURRENT RECURRENT; Domestic 3 GAP Capital . Total Domestic GAP Capital Total Country resources Operation: AIDS gap gap :resources: Operation: AIDS: gap gap Moldova 21 5 0 0 5 22 4 0 0 4 Mongolia 21 0 0 0 0 21 0 0 0 0 Mozambique 80 23 9 14 46 80 23 9 14 46 Nepal 89 30 0 14 44 89 30 0 14 44 Nicaragua 65 3 0 6 8 71 0 0 2 2 Niger 53 19 1 27 47 46 26 I 27 54 Nigeria 1,275 92 58 122 272 2,131. 0 0 0 0 Pakistan 1,061 55 0 148 204 1,092 25 0 148 173 Rwanda 45 5 7 11 23 39 11 7 11 a29 Senegal * 103 19 2 19 .40* 111 11 2 19 32 Sierra Leone 17 2 2 6 9 15 4 2 6 :12 Sudan 205 85 I 20 .106 205 85 I 20 106 Tanzania 164 52 .22 28 102 164 52 .22 28 102 Togo 35 4 4 7 15 32 7 4 7 19 Uganda 168 27 24 14 65 146 . 49 24 14 87 Yemen, Rep. of. 295 21 0 49 70 358 0 0 6 6 Zambia * 41 . 16 15 9 *40 50 8 15 9 31 Total * 14,327 * 923 285 *1,110 :2,323: 15,643 679 222 879 .1,785 ANNEX A * BACKGROUND DATA AND AGGREGATE SIMULATION RESULTS 161 m~Table A. 8 ....................................... NEstimated Annual Financing Gap, by Country, 2001-2015 (Scenario C2) CUMULATIVE ANNUAL COUURY REGION 2001 2002 i2003 2004 2005 2006 2001 2008 2009 2010 2011 2012 2013 2014 2015 ETOTAL AVERAGE Africa Angola AF 0 6 7 8 6 3 4 1 10 1 2 3 9 7 649 43 Benin APR 6 9 11 12 13 15 17 19 21 25 28 32 36 41 32 318 21 Burkina Faso :APR 26 29 33 38 43 48 5 1 9 6 75 82 90 95 78 890 60 Burundi AF 4 15 1 1 8 19 2 2 2 3 5 2 9 3 1 322 21 Cameroon APR 5 8 I11 1 5 19 25 32 40 50 65 8 1 100 122 148 127 848 56 Central African AFR~ 6 7 7 8 9 1 11 1 3 14 1 5 ~17 20 ~23 25 17 203 1 Republic _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Chad APR 9 1 11 3 15 1 8 20 23 27 33 38 45 53 52 33 40 27 Congo, Dem. APR 125 132 139 146 154 162 171 181 191 171 180 189 199 180 122 2,440 163 Rep. of__ __ _ _ Congo, Rep. of APR 5 4 4 4 4 4 5 5 6 9 12 16 21 22 21 144 10 G6te d'lvoire APR 27 31 36 41 46 52 58 65 72 82 93 104 116 130 127 1,080 72 Eritrea APR 6 6 6 7 7 8 9 10 1 1 I11 12 14 16 13 10 146 10 Ethiopia APR 106 116 128 141 156 173 191 212 236 239 264 293 326 343 226 3,150 210 Gambia, The APR 1 2 2 2 2 3 3 3 4 4 4 5 4 4 3 47 3 Gha-na AR 1 19 2 22 2 25 27 29 3 1 29 30 32 34 32 30 403 27 Guinea APR 20 22 25 27 29 32 35 37 40 42 45 48 51 52 44 548 37 Guinea-Bissau APR 3 3 3 3 3 3 3 3 3 4 3 3 3 4 3 5 1 3 Kenya APR 34 44 56 68 80 94 109 124 141 158 178 198 220 245 268 2,017 134 Lesotho ~APR 2 :3 :3 4 i5 6 8 9 I10 13 1 5 i17 20 26 28 169 I11 Madagascar ~APR 10 17 19 21 22 24 27 29 31 41 46 E52 58 62 37 495 33 Malawi AFR 12 13 0 19 22:24:27:29:32:37:41:46:50:42:35: 430 29 Mali AF 9 2 5 2 4 3 4 5 9 6 0 8 1 10 7 797:53 Maurita-nia APR 1 ~3 ~3 4 ~4 5 6 7 8 9 10 12 13 15 14 113 8 Mozambique : AFR 28 30 31 33 3 37 40 42 44 52 56 61 66 75 60 690 46 Niger : AFR~ 31 34 37 41 45 50 55 60 66 58 63 68 73 71 63 814 54 Nigeria ~ AFR 107 101 98 100 108: 123_ 147_ _ _ 224_ 272_ 341_ 427_ 53:29 401 272 Rwanda ~ AFR~ 5 5 7 15:17 19 E21 24 27 36 42 E48:55:63 52 437 29 Senegal jAFRE 8 10 E14 17 21 26 31 37 44:48:55 63 72:82 69 599:40 Sierra Leone :AyR 6 8 9 10 10 I11 12 13 14 1 2 13 14 15 16 15 179 12 Sudan :AFR :92 96 100 104 107 111 115 118 122 111 110 108 105 95 91 1,584 106 Tanzania APFR E59 64 69 75 82 88 96 :104 :113 112 120 :128 137 141 :137 1,524 102 Togo EAPR 5 6 1 1 0 1 1 1 3 1 5 1 8 20 25 28 3 7 3 E 5296 20 Uganda :AFR 52 56 59 63 68 72 77 83 89 97 104 112 121 128 130 1,311 87 Zambia EAPR 25 26 28 30 32 34 37 40 43 44 47 50 53 58 56 603 40 Subtotal AFR 971 1,037 1,084 1,197 1,295 1,410 1,547 1,708 1,896 2,017 2,262 2,547 2,878 3,111 2,832 27,791 1,853 East Asia and the Pacific Cambodia EEAP 30 29 E27 E25 :23:22 2 1 20 1 8 23 24 24 25:22 1 8 351 23 Lao PDR EP 22 20 18 15 15 13 12 I1I 10 10 10 10 1 1 12 9 201 13 Mongolia EEAPE 0 0: 0 0 0 0 0 : 0 0 0 0 0 0 0 Subtotal EEAP 52 49 45 41 39 35 33 31 28 34 34 35 36 34 27 552 37 Europe and Central Asia Armenia EECAE 67 52 39 28 19 I11 4 0 0 0 0 0 0 0 0 221 1 5 Georgia EECA 48 43 37 30 22 15 7 1 0 00 0 0 0 0 202 13 Moldova EECAE 23 19 15 I11 7 4 2 0 0 0 0 0 0 0 0 80 5 Subtotal ECA 139 113 91 69 48 30 13 1I 0 0 0 0 503 34 Latin America and the Caribbean Haiti ELCR 14 15 16 18 22 24 27 30 34 40 45 5 1 58 63 54 513 34 Honduras LCR 0 0 0 0 0 0 0 0 0 0 3 6 10 14 37 71 5 Nicaragua ELCRE 13 I11 9 7 6 5 4 3 2 7 8 9 I11 13 16 123 8 Subtotal LCR I27 26 26 25 28 29 31 33 36 47 56 67 79 90 96 696 46 (Continued) Table A. 8 (Continued) CU1 ULATIVE A1CIUAL COUniTRY REGIOui 26D01 2002 2033 2004 2005 2f006 2007 2008 2009 2QJ10 2011 2012 2M 3 2014 2OM TOTAL AVERAGE Middle East and North Africa Yemen, Rep. of MNA: 32 29 25 25 20 27 44 61 83 87 107 118 134 150 110 1,052 70 Subtotal MNA: 32 29 25 25 20 27 44 61 83 87 107 118 134 150 110 1,052 70 South Asia Bangladesh SAR 7 1 70 78 96 85 85 85 86 87 92 96 102 93 70 60 1,255 84 India SAR 457 335 206 4 0 0 0 0 0 0 0 0 0 0 0 1,001 67 Nepal SAR 37 34 32 29 35 42 44 46 47 49 52 54 57 60 49 667 44 Pakista-n SAR 106 III 120 145 155 172 193 217 246 232 243 266 297 294 256 3,053 204 Subtotal SAR 671 551 435 273 275 300 322 348 380 372 391 413 441 I475 358 6,005 400 TOtaW__ 1,892 1,805 1,705 E1,629 1,704 1,832 1,990 E2,181 i2,423 2,557 2,850 3,180 3,568 3,858 3,416 36,591 2,439 Note: Gross financing gaps are presented (not adjusted for any current or estimated ODA inflows). Financing gaps for African countries include the impact of HIW/AIDS. Totals differ slightly from totals presented elsewhere due to rounding. A N N EX Primary Completion Rate p Estimates and Projections * FIGURE B.1 Global Progress in Primary Completion, 1990-2000 and Projected Trends, Country-Weighted Primary completion rate (percent) 100 - 80- 60 - X l l l l 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG - Current trend FIGURE B.2 i Global Progress in Primary Completion, 1990-2000 and Projected Trends, Population-Weighted Primary completion rate (percent) 10- ,' 90 - -- , . - 80 - 70 - 60- I l I 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG Current trend 165 FIGURE B.3* Primary Completion Progress by Region, 1990-2000 and Projected Trends, Country-Weighted Primary completion rate (percent) 100 - ECA 90 _ EAP 80 - CR 70 _ 60 - SAR 50 _ AFR 40- l l l I 1990 2000 2010 2020 2030 2040 2050 - Primary completion progress Projected trend FIGURE B.4: Primary Completion Progress by Region, 1990-2000 and Projected Trends, Population-Weighted Primary completion rate (percent) 100 - 80- 60 - AFR 40- l l 1 l I 1990 2000 2010 2020 2030 2040 2050 - Primary completion progress - Projected trend TECHNICAL ANNEXES FIGURE B.5 Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Country-Weighted Regions off-track Regions on-track Primary completion rate (percent) Primary completion rate (percent) 100 ~~~~~~~~~100 - 1 00 - ,t l 1 90- ^ ECA 90 - 9I 80 - MA80 - 70- 70- 60 - 60 - 50 - 50 - 40- l l l 40- I l l 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 2015 Primary completion rate (percent) Primary completion rate (percent) 100 - ,, 100 - .,, 8-EAP 70 - ~~~~~~~~~70 - 60 . 60- 50- 50- 40 l l l 40- l l l 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 2015 Primary completion rate (percent) Primary completion rate (percent) 100- 100- 90 - 90 80- " - 70 - 70 - __________________________ 60 - 60 50- 5 X l0- 40 I I 40 I 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 2015 ---- Required trend to achieve MDG - Current trend ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 162 FIGURE B.6 Primary Completion Progress in Africa, Middle East and North Africa, and South Asia Regions, 1990-2015, Population-Weighted Regions off-track Regions on-track Primary completion rate (percent) Primary completion rate (percent) 100 100 90- 9 80 MNA 80_- 60 - 60 - 50- 50- 40- l 40 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 201 Primary completion rate (percent) Primary completion rate (percent) 100- __ v 100- 90 - . 90- EAP 80 - 80- 70 SAR 70 - 60- 60- 50- 50- 40- l I 40- 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 201 Primary completion rate (percent) Primary completion rate (percent) 100- -( 100- 80- 80- 70- ,, 70- L 60- 60- AFR 50- 50- 40- 40- 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 201 ---- Required trend to achieve MDG - Current trend gee TECHNICAL ANNEXES Table B. I ................................................................................... Primary Completion Progress by Region, 1990-2000, Country-Weighted 1990 MOST RECENT YEAR" Mean Median Minimum Maximum Mean Median Minimum Maximum Africa 49 42 11 135 55 45 19 117 East Asia and 78 89 39 99 84 90 54 108 the Pacific Europe and 89 90 67 100 92 93 77 109 Central Asia Latin America and: 76 86 28 112 85 89 40 110 the Caribbean Middle East and 75 75 32 102 74 76 30 104 North Africa South Asia 64 50 22 111 70 67 8 112 All developing * 72 81 11 135 77 83 8 117 countries IDA 50 45 11 112 * 62 64 8 117 IBRD 84 89 43 135 87 92 44 111 a. Usually 1999/2000. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 169 Table B.2 Primary Completion Progress by Region, 1990-2000l, Population-Weighted 1990 MOST RECENT YEARa Girls Boys Both Girls Boys Both Africa 43 57 50 46 56 51 East Asia and the Pacific 92 97 96 98 98 97 Europe and Central Asia 85 95 90 93 95 93 Latin America and the Caribbean 71 64 69 85 81 83 Middle East and North Africa 71 84 78 78 86 83 South Asia 59 77 68 63 84 74 All developing 65 79 73 76 85 81 countries Note: For population-weighted estimates, all PCRs over 100 were set to equal 100, in order to avoid distortions. a. Usually 1999/2000. 170 TECHNICAL ANNEXES Table B.3 Best Performers (IDA Countries) in Improving Primary Completion Rate, 1990 to Most Recent Year PRIMARY COMPLETION RATE :Years in Primary Trend Rate PCR Country Cycle Region Level in 1990 Level in MRY * (1990 to MRY)b 2015 Cambodia 6 EAP 39 70 7.63 100 Malawi 8 AFR 30 50 4.10 100 Gambia, The 6 AFR 40 70 3.35 100 Togo 6 AFR 41 63 2.51 100 Lao PDR 5 EAP 56 69 2.50 100 Serbia and 8 ECA 72 96 2.40 100 Montenegro Uganda 7 AFR 39 65 2.36 98 Zimbabwe 7 AFR 97 113 2.29 100 Nicaragua 6 LCR 45 65 2.03 96 Benin 6 AFR 23 39 1.96 72 Bangladesh 5 SAR 50 70 1.96 99 Guinea 6 AMR 16 34 1.75 60 Haiti 6 LCR 28 40 1.74 71 Eritrea 5 AFR 22 * 35 1.72 63 Bolivia 8 LCR 55 72 1.70 97 Tanzania 7 AFR 46 59 1.67 89 Moldova 4 * ECA 67 79 1.51 * 100 Mali 6 AFR 11 23 1.51 49 Mauritania 6 AFR 34 46 1.47 * 71 Pakistan 5 SAR 44 59 1.39 80 a. Usually 1999/2000. b. Trend rate is average annual percentage point change in primary completion rate from 1990 to most recent year. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 171 Table B.4 Best Performers (IBRD Countries) in Improving Primary Completion Rate, 1990 to Most Recent Year PRIMARY COMPLETION RATE :Years in:: :Primary:: Trend Rate * PCR Country . Cycle . Region Level in 1990 Level in MRYa (1990 to MRY)b 2015 Czech Republic 4 ECA 89 109 6.55 100 South Africa * 7 . AFR 76 98 4.32 100 Egypt, Arab Rep. of. 5 MNA 77 99 3.60 100 Namibia 7 AFR 70 , 90 2.85 100 Brazil . 8 LCR 48 72 2.64 100 Tunisia 6 MNA 75 91 2.63 100 Latvia 4 ECA 76 86 2.39 100 Gabon 6 AFR 71 80 2.12 100 Lithuania 4 ECA 88 95 1.97 100 El Salvador 6 LCR 61 80 1.77 100 Morocco 6 MNA 47 55 1.65 87 Costa Rica . 6 LAC . 73 89 1.65 100 China .5 EAP 99 108 1.57 100 Hungary * 4 ECA 93 102 1.48 100 Swaziland , 7 APR . 71 81 1.46 100 Algeria :6 MNA 82 91 1.37 100 Colombia . 5 LCR 72 85 1.36 100 Paraguay 6 LCR . 65 78 1.34 98 Mexico :6 LCR 89 100 1.15 100 Croatia . 8 ECA * 86 96 1.11 100 a. Usually 1999/2000. b. Trend rate is average annual percentage point change in primary completion rate from 1990 to most recent year. D22 TECHNICAL ANNEXES Table B.5 .........................................................................I.................... IDA Countries with Declining Primary Completion Rate,j 990 to Most Recent Year 4. . ,PRIMARY COMPLETION RATE Years in Primary Trend Rate . PCR Country Cycle Region . Level in 1990 Level in MRYa (1990 to MRY)b 2015 Zarmbia 7 AFR 97 83 -2.10 41 Congo, Rep. of 6 AFR 61 44 -1.73 18 Albania 8 ECA 97 89 - 1.55 58 Cameroon 6 AFR 57 43 -1.48 20 Afghanistan 6 . SAR * 22 8 -1.39 . 0 Vanuatu 6 EAP * 89 86 -1.38 . 55 Comoros 6 AFR 35 33 -1.22 6 Kenya 8 . AFR 63 58 -1.10 36 Madagascar 5 . AFR 34 . 26 -0.93 11 Central African 6 AFR 28 19 -0.87 6 Republic Congo, Dem. 6 AFR 48 . 40 -0.79 28 Rep. of Nigeria * 6 * AFR 72 . 67 * -0.56 * 58 Rwanda 1 61 AFR 34 . 28 . -0.54 20 Senegal 6 AFR . 45 41 * -0.36 36 C6te d'Ivoire * 6 . AFR * 44 40 * -0.35 * 35 a. Usually 1999/2000. b. Trend rate is average annual percentage point change in primary completion rate from 1990 to most recent year. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 123 Table B. 6 ........... .................................................................................. IBRD Countries with Declining Primary Completion Rate, 1990 to Pvlost Recent Year * . 4 PRIMARY COMPLETION RATE Years in * Primary . Trend Rate PCR Country Cycle Region Level in 1990 Level in MRY * (1990 to MRY)b 2015 Qatar . 6 MNA 74 44 -6.10 0 UnitedArab 6 MNA 94 80 -2.22 38 Emirates Estonia 6 ECA 93 88 - 1.78 52 Bahrain 6 MNA 101 91 -1.55 62 Venezuela, RB 5 LCR 91 78 -1.52 54 Syrian Arab Rep. 6 MNA 98 90 -1.25 66 Iraq 6 MNA 63 57 -1.16 34 Belarus 4 ECA 97 93 -0.87 77 Belize 6 LCR 90 82 -0.87 68 Rwanda 6 AFR 34 28 -0.54 20 Iran, Islamic 5 MNA 94 92 -.41 84 Rep. of . * : Thailand . 6 EAP . 93 90 -0.26 86 Malaysia 6 EAP 91 90 -0.24 85 a. Usually 1999/2000. b. Trend rate is average annual percentage point change in primary completion rate from 1990 to most recent year. Table B.7 .......................................................................... IDA Countries That Have Achieved Universal Primary Completion V*o rs in * iDA/ o . Primary . P .M racliio . Pcn Country I.1ROD R1egion cycle 1'S * vear , l1 . [15.11 Cape Verde IDA AFR 6 117 1997 Zimbabwe IDA AFR 7 97 1990 113 1997 Samoa IDA EAP 8 99 1997 Vietnam IDA . EA 5 101 2001 Azerbaijan IDA ECA 4 100 1998 Serbia and IDA ECA 8 72 1990 96 2000 Montenegro Dominica IDA LCR 6 103 2000 Grenada IDA LCR 7 - - 106 2001 St. Lucia IDA LCR 7 112 1990 . 106 2001 Maldives IDA SAR 5 111 1992 112 1993 Sri Lanka IDA SAR 5 . 100 * 1990 111 2001 - Not available. Note: Countries have ACHIEVED universal primary completion if the completion rate in the most recent year is 95 percent or higher. awlc TECHNICAL ANNEXES Table B.8 ............................................................................ IBRD Countries That Have Achieved Universal Primary Completion * * ' Years in IDA/ Primary PCR Baseline PCR Country IBRO Region Cycle 1990 Year MRY MRY Borswana IBRD AFR 7 114 1990 102 1996 Mauritius IBRD AFR 6 135 1990 111 1997 SouthAfrica *IBRD AFR 7 76 . 1990 98 1995 China IBRD EAP 5 99 1990 108 1996 Fiji .IBRD EAP 6 - 95 1992 Korea, Rep. of *IBRD EAP 6 96 1990 96 2000 Croatia IBRD ECA 8 86 1992 96 2001 Czech Republic IBRD ECA 4 89 1992 109 1995 Hungary IBRD ECA 4 93 1989 102 1995 Lithuania IBRD ECA 4 88 1992 95 1996 Poland IBRD ECA 8 100 1990 96 1995 Romania IBRD ECA 4 96 1989 98 1996 Russia IBRD ECA 3 96 2001 Slovak Republic IBRD ECA 4 96 1992 97 1996 Slovenia IBRD ECA 4 99 1992 Antigua and IBRD LCR . 7 - - 95-1OOa 2000 Barbuda Argentina .IBRD LCR 7 - - 96 2000 Chile .IBRD LCR 6 94 1990 99 2000 Cuba n. a. LCR 6 95-100a 2001 Ecuador IBRD LCR 6 99 1992 96 1999 Mexico IBRD LCR 6 89 1990 100 2000 Peru IBRD LCR 6 85 1988 98 2000 St. Kitts and Nevis: IBRD LCR 6 110 2001 Uruguay IBRD LCR 6 95 1990 98 2000 Egypt, Arab IBRD MNA 5 77 1990 99 1996 Rep. of Jordan IBRD MNA 6 102 1990 104 2000 n. a. Not applicable. - Not available. Note: Countries have ACHIEVED universal primary completion if the completion rate in the most recent year is 95 percent or higher. a. Staff estimate. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 175 Table B.9 ....................................................................................... IDA Countries "On Track" to Achieve Universal Primary Completion by 2015 Years in * PCR IDA/ Primary PCR . Baseline PCR Annual PC.1 Country l* D * Region Cycle I9D *. Year .MRY , PRV : Increase 2015 : ~~ ~ . * : : Gambia, The IDA AFR 6 40 1991 70 *2000 3.35 100 Malawi .IDA AFR . 8 .30 1990. 50 .1995 4.10 100 Togo IDA AFR 6 41 1990 63 1999 2.51 100 Uganda IDA AFR 7 39 1990 65 2001 2.36 98 Cambodia IDA EAP 6 39 1997 70 2001 7.63 100 Lao PDR IDA EAP 5 56 1995 69 2000 2.50 100 Bosnia and IDA ECA 4 88 1999 - *95-100a Herzegovina Moldova IDA ECA 4 67 1991 79 1999 1.51 100 Bolivia IDA LCR 8 55 1990 72 2000 1.70 97 Nicaragua IDA LCR 6 45 1990 65 2000 2.03 96 Bangladesh IDA SAR 5 50 1990 70 2000 1.96 99 - Not available. Note: Countries are ON TRACK if a projection of the observed trend results in a completion rate of 95 percent or higher by 2015. If a country does not have two data points, it is ON TRACK if the most recent year completion rate is 85-94 percent, inclusive. a. Staff estimate. SEX3 TECHNICAL ANNEXES Table B.1O .................. ... ...................................I........................ IBRD Countries "On Track" to Achieve Universal Primary Completion by 2015 Yearsin PCR . IDA/ * '. Primary PCR Baseline PCR Annual PCR Country IBRO Region Cycle 1990 Year MRY MRY :Increase 2015 Gabon .IBRD AFR 6 71 1991 80 1995 2.12 100 Namibia *IBRD AFR 7 :70 *1990 90 *1997 2.85 100 Swaziland IBRD AFR 7 71 1990 81 1997 1.46 100 Philippines IBRD EAP 6 89 1989 92 1996 0.38 99 Bulgaria IBRD ECA 4 90 1990 92 1996 0.30 97 Latvia IBRD ECA 4 76 1992 86 1996 2.39 100 Macedonia, FYR IBRD ECA 8 89 1992 91 1996 0.44 99 Turkey IBRD ECA 5 90 1990 92 1994 0.52 100 Ukraine IBRD ECA . 3 - * 94a 2002 - :95-100a Brazil I IBRD LCR 8 . 48 1990 72 1999 2.64 100 Colombia IBRD LCR 5 72 a1990 85 i2000 1.36 100 Costa Rica IBRD LCR 6 * 73 1990 89 2000 1.65 100 El Salvador IBRD LCR 6 61 1989 80 2000 1.77 100 Jamaica IBRD LCR 6 90 1990 94 2000 0.40 100 Panama IBRD LCR 6 87 1990 94 2000 0.65 100 Paraguay IBRD LCR 6 65 1990 78 2000 1.34 98 Algeria IBRD MNA 6 82 1990 91 1996 1.37 100 Kuwait . n. a. MNA 4 56 1991 70 1996 2.74 100 Oman n. a. MNA 6 67 1989 76 1996 1.31 100 Saudi Arabia n. a. MNA 6 60 1990 69 1996 1.50 97 Tunisia IBRD MNA 6 75 1990 91 1996 2.63 100 n. a. Not applicable. - Not available. Note: Countries are ON TRACK if a projection of the observed trend results in a completion rate of 95 percent or higher by 2015. If a country does not have two data points, it is ON TRACK if the most recent year completion rate is 85-94 percent, inclusive. a. Staff estimate. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 177 Table B. II ....................................................................................... IDA Countries "Off Track" to Achieve Universal Primary Completion by 2015 D * , Years In * PCR IDna/ Primary PCGR Baseline PCR1 , Annual * I Country 'tlR Region Cycle 199O Vear ° V M 'MRV ,Increase 2Di5 Benin IDA *AFR 6 23 1990 39 1998 1.96 72 Eritrea .IDA AFR * 5 .22 1991 35 1999 1.72 63 Ghana IDA *AFR 6 63 1990 64 1999 . 0.11 65 Guinea IDA AFR 6 16 1990 34 2000 1.75 60 Lesotho IDA AFR 7 64 1990 69 1996 0.86. 85 Mauritania IDA AFR 6 34 1990 46 1998 1.47 71 Mozambique IDA AFR 5 30 1990 36 1998 0.82 50 Nigeria IDA .AFR 6 72 1990 67 .2000 -0.56 58 Sao Tome and IDA AFR 4 84 2001 50-84a Principe Tanzania IDA AFR 7 46 1989 59 1997 1.67 89 Indonesia IDA EAP 6 92 1990 91 2000 0.10 89 Mongolia IDA L-4AP 4 82 1998 50-84a Solomon Islands IDA EAP 6 65 1990 * 66 1994 0.19 70 Timor-Leste, IDA EAP 6 - - 54 2001 50-84a Dem. Rep. Vanuatu IDA EAP 6 89 1990. 86 1992 1.38 55 Albania IDA ECA 8 .97 .1990 89 .1995 -1.55 58 Armenia IDA ECA 4 - - 82 1996 50-84a Georgia IDA ECA 4 82 1998 50-84a Tajikistan IDA ECA 4 77 1996 50-84a Guyana IDA LCR 6 92 1990 89 2000 0.24 86 Haiti IDA LCR 6 28 1990 40 1997 1.74 71 Honduras IDA LCR 6 66 1991 67 2000 0.21 71 St. Vincent IDA LCR 6 84 2001 50-84a Yemen, Rep. of IDA MNA 6 58 2000 50-84a Bhutan IDA SAR 7 59 2001 50-84a India IDA SAR 5 70 1992 76 1999 0.93 91 Nepal IDA SAR 5 49 1988 65 2000 1.33 85 Pakistan IDA SAR 5 44 1989 59 2000 1.39 80 -Not available. SIote: Countries are OFF TRACK if a projection of the observed trend results in a completion rate in 2015 of 50-94 percent, inclusive. If a country does not have two data points, it is OFF TRACK if the most recent year estimate is 50-84 percent, inclusive. a. Staff estimate. AwS TECHNICAL ANNEXES Table B. 12 ..........................................................................I............... IBRO Countries "Off Track" to Achieve Universal Primary Completion by 2015 Years in PCR IDA/ Primary PCR Baseline PCR Annual PCR Country IBRO Region Cycle 1990 Year . MRY MRY ::Increase 2015 Malaysia IBRD EAP 6 91 1990 90 1994 *-0.24. 85 Papua New IBRD EAP 6 53 1990 59 1995 1.04 79 Guinea Thailand IBRD EAP 6 93 1990 90 2000 -0.26 86 Belarus IBRD ECA 4 97 1992 93 1996 -0.87 77 Estonia IBRD ECA 6 93 1992 88 1995 -1.78 52 Belize IBRD LCR 6 90 1990 82 1999 -0.87 68 Dominican IBRD LCR 8 62 2000 50-84a Republic Guatemala *IBRD LCR 6 43 1991 52 *2000 1.01* 67 Trinidad and * IBRD LCR 5 94 ,1990 94 2000 .-0.01. 94 Tobago Venezuela, RB . IBRD . LCR * 5 91 . 1990 . 78 * 1999 .-1.52 * 54 Bahrain * n. a. * MNA 6 . 101 1990 91 1996 -1.55 62 Iran, Islamic , IBRD . MNA . 5 . 94 .1990 92 1996 -0.41 . 84 Rep. of Lebanon ,IBRD MNA. 5 7- 1- .70 1996.- .50-84a Morocco IBRD . MNA 6 47 '1991 55 1996 1.65 87 Syrian Arab Rep. IBRD MNA * 6 * 98 * 1990 * 90 i 1996 *-1.25 . 66 n. a. Not applicable. - Not available. Note: Countries are OFF TRACK if a projection of the observed trend results in a completion rate in 2015 of 50-94 percent, inclusive. If a country does not have two data points, it is OFF TRACK if the most recent year estimate is 50-84 percent, inclusive. a. Staff estimate. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 179 Table B. 13 ................................................................................................... IDA Countries "Seriously Off Track" to Achieve Universal Primary Completion by 2015 Yearsin . PCR IDA/ Primary PCIR Baseline PCR Annual PM Country IBRD Region Cycle 190* Year IRY IRV :Increase 2Q15 Angola ' IDA .AFR 4 39 .1990 .less than 50a Burkina Faso IDA AFR 6 19 1990 25 1998 0.73 38 Burundi IDA AFR 6 46 1990 43 1998 -0.35 37 Cameroon IDA AFR 6 57 1990 43 1999 1.48 20 Central African IDA AFR 6 28 1990 19 2000 -0.87 6 Rep. Chad IDA AFR 6 19 1990 19 2000 -0.04 18 Comoros IDA AFR 6 35 1991 33 1993 *-1.22 6 Congo, Dem. IDA AFR 6 48 1990 40 2000 -0.79 28 Rep. of Congo, Rep. of IDA AFR 6 61 1990 44 2000 -1.73 18 C6te d'Ivoire IDA AFR 6 44 1990 40 1999 -0.35 35 Ethiopia IDA AFR 6 22 1990 24 1999 0.28 29 Guinea-Bissau IDA AFR 6 16 1988 31 2000 1.23 49 Kenya IDA AFR 8 63 1990 58 1995 -1.10 36 Madagascar IDA AFR 5 34 1990 26 1998 -0.93 11 Mali IDA AFR 6 11 1990 23 1998 1.51 49 Niger IDA AFR 6 18 1990 20 1998 0.25 25 Rwanda IDA AFR 6 34 1990 28 *2000 -0.54 20 Senegal IDA AFR 6 45 1989 41 2000 -0.36 36 Sierra Leone IDA AFR 7 32 2000 l- ess than 50a Sudan IDA AFR 8 35 1996 less than 50a Zambia IDA AFR 7 97 1988 83 1995 -2.10 41 Djibouti IDA MNA 6 32 .1990 30 1999 .-0.26. 26 Afghanistan IDA SAR 6 22 1989* 8 *1999 *-1.00* 0 -Not available. Note: Countries are SERIOUSLY OFF TRACK if a projection based on the observed trend results in a completion rate BELOW 50 percent in 2015. If a country does not have two data points, it is classed as SERIOUSLY OFF TRACK if the most recent year estimate is below 50 percent. a. Staff estimate. neXeD TECHNICAL ANNEXES Table B.14 .................................................................................................... IBRO Countries "Seriously Off Track" to Achieve Universal Primary Completion by 2015 Years in PCR IDA/ Primary PCR Baseline PCR Annual PCR Country IBRD Region Cycle 1990 Year MRY MRY :Increase 2015 Equatorial Guinea IBRD- AFR 5 46 1993 less than 50- Iraq IBRD. MNA 6 63 1990 57 1995 1.16 34 Qatar n. a. MNA 6 74 1990 44 1995 -6.10 0 UnitedArab n. a. MNA 6 94 1990 80 1996 -2.22 38 Emirates n. a. Not applicable. - Not available. Note: Countries are SERIOUSLY OFF TRACK if a projection based on the observed trend results in a completion rate BELOW 50 percent in 2015. If a country does not have two data points, it is classed as SERIOUSLY OFF TRACK if the most recent year estimate is below 50 percent. a. Staff estimate. Table B. 15 ...................................... Countries with No Data Available Years in IDA/ Primary Country IBRO Region Cycle Korea, DPR n.a. LEAP 4 Libya n.a. MNA 9 Seychelles IBRD AFR 6 Marshall Islands IBRD EAP 6 Micronesia, Fed. States of IBRD EAP 6 Palau IBRD EAP 8 Kazakhstan IBRD ECA 4 Turkmenistan IBRD ECA 4 Liberia IDA . AFR 6 Somalia IDA AFR 8 Kiribati IDA EAP 7 Myanmar IDA EAP 5 Tonga IDA EAP 6 Kyrgyz Republic IDA ECA 4 Uzbekistan IDA ECA 4 West Bank/Gaza IDA MNA 10 n.a. Not applicable. Note: Primary completion rates cannot be calculated for the countries listed because either enrollment or population data are not available. ANNEX B * PRIMARY COMPLETION RATE ESTIMATES AND PROJECTIONS 181 AN NE X Country Simulation Results Table Cl1 All 41 Countries: MDG-2015 Financing Gap under Alternative Policy Measures B: EFIFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary 1: Prtvate Spending on Teacher Salary (as iEducation Enrollments Annual Pupils Per Inputs Other mtultiple of per Average As % at % for iRecurrent (as % of Financing Policy Scenario' Teacher than Teachors' capita GDP) Rpeptitiuni Rate GOP Education Spendine4 total) Gap' Stattu quo 13 -79 0.1i - 45 0.6 -9.6 0 -360% 8 -56 1.4 -32.6 26 -66 0 -77 A only 36 33.7% 4.5 : 7,489 A+B 40 33.3% 3.8 8.2% 4,348 "Best :CI i 40 33.3% 3.7 8.2% 15.2% 21.1% 48.6% 10.0% 2,033 Pmcti' C2 ~~~~~~~~~~~~15.206 20.0% 2,151 A+ + C3 _________ ____120.3% 20.0% i1,563 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3. The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In milUons of 2Q00 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.2 All41 C"o'un'trie"s: M"DG-2'0'1"5C"o'st E'st'i'mat'es a'n"d Sou'rc'e"s o'fFi'nan'ci'ng u'nd'er ""Be'st P"r'a'ctic"e"'P"oli'cie's an'd Alfte'rn'ati've, Resource Mobilization Scenarios (millions of 2000 constant U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL DainoStCic COST OF MDG-2015 DomESTIC RESOURCES FINANCING coot Resources. . lItem Period Scenario, Mobhilized 1Recurrent Capital Total Recurrent capital I Total Recurrent ICapital Total 8 ~~~~Cl 214,897 222,665 22,728 245,393 208,811 6,085 214,896 13,853 16,643 30,496 Cunnulative, i C 213,124 222,665 22,728 245.393 207,109 6,014 I213,123 15,556 16,714 32,270 __________ C3 234,632 222,665 22,728 1245,393 2i2,430 9,521 221,951 10,234 13,208 23,442 Cl 14,326 14,844 1,515 16,360 391 406 436 924 110 23 Anul C2 14,208 14,844 1,515 16,360 13,807 401 14,208 1,037 111 4 2,151 C3 15,642 14,844 1,515 16,360 14,162 635 14,797 682 881 1,563 G ~~~~~~ci, 286 286 0 0 286 286 Arnnual C2 286 286 0 0 285 286 C3 _____ 286 ____ 286 63 ___ 3 223 _ _ _ 286 4 ___ __ ___~~~~~~~~~~~~~~~~~~~6 63_ _ __ _ Cl ~~~~~~15,130 1,515 16,646 13,921 406 14,326 1,210 1,110 2,319 Annual C2 15,130 1,515 16,646 13,807 401 14,208 1,323 1,114 2,437 C3 r 15,130 1,515 16,646 14,226 635 14,860 905 881 I 785 Note: "Best practice" policies refer to the combination of scenarios A+B+C, Shaded cells denote no change from values directly above. 183 Table C.3 Wo~rld, E'xc'ep't A'fric'a': Fi'na"ncin"g Ga"p u'n"d'er Alte"r'na'tive ..P'o i'cy ..Me' a's'u'r'e's B: EFFICIivNCY A: QUALITY MEACURES MEASUItEO C: FINANCING MOASURES Average Annual Govermment Revenues' Primary :ZPrivate Speeding on Teacher Salary (as Edcto nolet nneal Pupils Per Inputs Oilher multIple at per Average As % af % for Recurrent (as % of Financing Policy Scenario' Teacher than Toachers0 capita GDP) Repetition Rate GOP Education Spending' total) Gap' Statooquoo 34 25.7% 2.7 11.2% 20.5% 15.7% 42.8% 12.0% 538 A only 31 33.3% 3.9 3,967 A+ B 40 33.3% 3.7 6.7% 2,428 'Best C 40 33.3% _ 3.6 6.7% 15.7% 20.8% I47.1% 10.0% 579 practice': C2 15.7% 20.0% 585 A + B+ C3 20.8% 20.0% 438 Niote: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.4 World, Except Africa: WlDG-201 5 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCOES CAP rFOR ErtTERNAL ilemesiC COOT O~ MDG-2015 DOMUSTIC RESOURCES FINANCING cent Resources Item Perind ScertarIo Mohilized Recurrent Capital Total Recurrentt Capital natal Recurrent Capital Total 8 ~~~~Cl 153,340 150,180 11,849 162,029 147,254 16,085 153,339 2,927 5,764 8,690 E Cula6tive, C2 153,247 150,180 11,849 162,029 147,233 6,014 153,247 297 5,835 8,782 ~ . 2001 2015208 ___ 656 C . _________ ~C3 155,464 150,180 11,849 162,029 148,162 7,301 155,463 2,018_ 4,54~8 656 * . Ci 10,223 10,012 790 .10.802 :9,817 406 10,223 195 384 579 Annual C2 10,216 10,012 790 10,802 9,816 401 10,216 196 389 585 _________ ~C3 10,364 10,012 790 10,802 9.877 487 10,364 135 303 i438 Cl 0 0 0 Annua C2 0 0 0 : 0 ______ ~~~C3 _ __ 0 0 0 0 _ 0 __ _ 0 ci 10012 790 10,802 9,817 406 10,223 195 384 579 Annual C2 110,012 790 10,802 9,816 401 10,216 197 389 586 C3 *10,012 790 10,802 9,877 487 10,364 135 303 438 tlote "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. llt~36 TECHNICAL ANNEXES Table C.5 Africa: MD6-2015 Financing Gap under Alternative Policy Measures 1B: EFFIcIENCY A; QUALITY MEASURES MEASURES . c FINANCING MEASURES Average Annual Government Revenuoe0 Primary Private Spending on iTeacher Salary (as Educatlon Enrollments Annual Puplis Per Inputs Other muliple of per Average As % of % for Recurrent (as % of Flnancing Policy Scenario * Teacher than Teachersb capita GOP) Repetition Rate GOP Education Spendingd total) Gap' Staiw quo 48 24.1% 4.4 18.2% 18.7% 17.5% 49.3% 8.1% 1,700 A only 38 33.8% 4.7 i_i_j ij_ 3,522 A + B 40 33.3% 3.8 8-9% 1,920 . : f I : 2 .I + 'Best Cl 40 33.3% 3.8 8.9% 15.0% 21.2% 49.3% 10.0% 1,454 practice C2 150.% 20.0% 1,566 A + B + , C3 20.1% 20.0% 1,125 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.6 Africa: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FI.NANCING SOURCES' , . 7 . I EGAP FOR. EXTERNAL Domestic COST OF MDG-2015 DOMESTIC ESOURCE FNANCNG Cost Resources - : Tc_ RESOURCE Item Period Sconario Mobiilzed Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 61,558 72,484 . 10,880 . 83,364 61, 558 0 61,558 10,927 10,880 21,806 7 Cumu21 ve C2 59,876 72,484 10,880 83,364 59,876 0 59,876 12,608 10,880 23,488 x 72001 5 C3 79,168 772,484 7 10,880 83,364 I 64,268 2,220 66,488 8,216 8,660 7 16,876 t?s 3 i 7 CI 4,104 4,832 725 5,558 4,104 0 04 728 725 1454 Annual C2 3,992 4,832 725 5,558 3,992 0 3,992 841 725 1,566 ._______ 7 C3 5,278 4,832 725 5558 , 4285 148 4,433 548 577 1125 Ann~~~~~~~l (,0 2 &I a C .nnl 2 266 28u 0 . 28) 266 2___ ______ C3 -: 286 - . 286 63 . 63 222 286 . , Cl ; :.9, ^--c 5,118 725 5,844 4,104 0 4,104 1,015 725 1,740 Annual C2 5,118 725 5,844 3,992 0 3,992 1,127 725 1,852 so. 7 C3 5,118 725 5,844 4,348 148 4,496 770 577 1,348 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 18S Table C.7 ................................................................................ Armenia: GD6-2015 Financing Gap under Alternative Policy Measures B- EFFICIENCY A: QUALITY MEASSURES MEASURES C: FINANCING MEASURES . * . Average Annual Government Revenuese Prmary Private , * Spendingon TeacherSalary( as Educaton Enroliments Annual Pupils Per Inputs Other muftiple of per Average As % of % for Recurrent (as % of Flnancing Policy Scenarlo' Teacher *than Teachers' capita GOP) RepetitIon Rate GOP Educaton Sponding* total) cap, Starwquo 13 52.9% 0.6 0.1% 15.8% 15.1% 51.3% 0.0% 0 A only 13 33.3% 3.5 61 A + B 40 33.3% 3.5 0.1% 12 'Best Cl 40 33.3% 3.5 0.1% 16.0% 20.0% 50.0% 10.0% 15 practice": C2 160% 200% 15 A + B + . C3 16.0% 20.0% 15 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three altemative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered 'best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.8 Armenia: MD6-2015 Cost Estimates and Sources of Financing under bBest Practice" Policies and Alternative Resource Mobilization Scenarios 195 (millions of 2000 U.S. dollars) FINANCINC3 SOURCES GAP FOR EJXTERNAL Domestic COST OF MDG-2015 DO.MESTIC RECOURCES FINANCING Cost Resources . . - Item Pe*od scenado *Bobilized Recurrent g Capital Total Recorrent Capital Total Recurrent Capital ToSal 4. : : ~ : . 4 8 * * Cl 491 712 0 712 491 0 491 221 0 j221 E Cuulatve, C2 491 712 0 712 491 0 491 221 i 0 ,221 Q g j 2001-2015 i C3 491 712 0 712 491 0 491 221 i 0 j221 Cl 33 47 0 47 33 0 33 15 0 15 ' Annual C2 33 47 0 47 33 0 33 15 0 15 ____i __ i_ C3 33 47 0 47 33 0 33 15 0 15 ci 0 0 0 ,0 0 0 t S8 iAnnual C2 0 0 0 0 0 . 0 .__ *__ _ _ C3 ' _ 0 . 0 0 . _ 0_ 0 0 Cl, 47 0 47 33 0 33 15 0 15 0 rE i Annual C2 47 0 47 33 0 33 IS 0 15 i ,i C3 . . 47 0 47 33 0 33 15 0 15 Note: "Best practice policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ass TECHNICAL ANNEXES Table C.9 Bangladesh: MDG-201 5 Financing Gap under Alternative PoIcy . Measures S ~~~~~~B: E F FIC IE NC Y * A: QUALITY MEASURIES MEASURES C: FINANCING MEASURES Average Annual Govemnment Revenues' Primary Private Spending on .Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % hr Recurrent (as % ot Financing Pollcy Scenario Teacher lthan Teachers' capita GOP) Ropetiion Rate GOP Education Spendingd total) cap' Staw quo 55 25.0% 2.7 15.0% 12.8% 14.7% 41.6% 12.5% 0 A only 40 33.3% 3.5 , 224 A+B 40 33.3% 3.5 100% 195 "Best Cl i 40 33.3% 3-5 10.0% 14.0% 20.0% 42.0% 10.0% 84 practice" C2 14.0% 20.0%. 84 A +B+ C3 14.0% 20.0% 84 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three altemative resource mobilization scenanos (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.IO Bangladesh: MD6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 197 (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic . COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost . Resources . . : : _ Item Period Sconaria Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital: Total Cl 8,342 8,527 1,070 9,597 8,342 0 8,342 185 1,070 1,255 j. Cumulative, .. i C2 8,342 8,527 1,070 9,597 8,342 0 8,342 185 j 1,070 1,255 it j2001 2015. .. X C3 i 8,342 8,527 i,070, 9,597 8,342 0 8,342 185 1,070 1,255 Cl 556 568 71 640 556 0 556 12 71 84 Annual. . C2 556 568 71 640 556 0 556 12 71 84 _____i__i_ C3 556 568 71 640 556 0 556 12 71 84 Ci 0 0 0 0 0 0 * Annual C2 0 0 0 0 0 0 _ _ _ m ji j C3 0 0 0 0 0 ____ Ci 568 71 640 556 0 556 12 71 84 'o rY j Annual C2 568 71 640 556 0 556 12 71 84 C3 568 71 640 556 0 556 12 71 84 Note: "Best practice" policies refer to the combination of scenaros A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 187 Table C.ll ........................................................................... Cambodia: MDG-2015 Financing Gap under Alternative Policy Measures 3: EFFICICNCY A: QUALITY MEA£URES MEASURQ£ C. FINANCING MOASURES Average Annual .Government Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputt Other multiple of per Average As % of % tor Recurrent (as % of Financing Policy Scenarli . Teacher than Teachers' capitabSGDP) Repetition Rate GSDP Education Spondine total) Gap' S'anwquo 53 20.0% 1.5 16.6% 11.5% 12.2% 510.% 10.0% 0 A orny 40 33.3% 3.5 45 A + B 40 33.3% 3.5 10.0% 40 'Best Cl 40 33.3% 3.5 10.0% 14.0% 20.0% 500.% 10.0% 23 praccice": C2 14.0% 20.0% 23 A + B + C3 14.0% 20.0% 23 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenanos A+B+C is considered best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.12 Cambodia: MD6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 199 (millions of 2000 U.S. dollars) * . ~~~~~~~~~~~~~~~~~~~~~~~FINANCING SOURCES . . GAP FOR EXTOsRNAL Domestic COST OP MDG-201S DOMESTIC RESOURCoQ FINANCING Cost Resources .__..__ Item Period Scenario .Mobilized Recurrenti Capital Total Recurrent Capital Total Recurrent Capital Total Cl 852 1,144 58 1,202 852 0 852 292 58 350 . j Citnuiacn'c i C2 852 152 0 852 292 58 350 2001 2015 : _ N i 5 C3 852 1,144 58 1,202 852 0 852 292 58 350 0. - i , Cl 57 76 4 80 57 0 57 19 4 23 z j Annual C2 57 76 4 80 57 0 57 19 4 23 i____ i_________ C3 57 76 4 80 57 0 57 19 4 23 o 4~~~~~~~~~~~ ci 0 0 0 0 0 0 a ": Annual C2 0 0 0 0 0 0 8 g i C3 i _o_i0 0 0 o to 0 Cli 76 4 80 57 0 57 19 4 23 Annual C2 76 4 80 57 0 57 19 4 23 C3 76 4 80 57 0 57 19 4 23 Note: "Best practice policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. aims TECHNICAL ANNEXES Table C.13 Georgia: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES , . Average Annual Government Revenues- Primary Private Spending on Teacher Salary (as Educatlon Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (es % af Financing Policy Scenario , Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spending' total) Gap' Stanu qao 17 16.0% 06 05% 137% 93% 260% 1_0% 0 Aonly 17 33.3% 35 ,_,_,_129 A + B 40 33.3% 3 5 0.5% 39 'Best C1 40 33.3% 3.5 0.5% 16.0% 20.0% 50.0% 10.0% 13 prac:ce": C2 ii 16.0% 20.0% 13 A + B + , C3 16.0% 20.0% 13 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.14 ....... ................................................................................................................................ Georgia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) ; , FINANCING SOURCES GAP FOR E XTERNAL DOmestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Reseurces * Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 599 801 0 801 599 0 599 202 0 202 j. Cumuiadve, C2 599 801 0 801 599 0 599 202 0 202 2 001 20012015 ____ , ;C3 599 801 0 801 599 0 599 202 0 202 c:i XCi 40 53 0 53 40 0 40 13 0 13 Annua" C2 40 53 0 53 40 0 40 13 0 13 __________ C3 40 53 0 53 40 0 40 13 0 13 AS 8 Annual C2 2i 0 0 0 0 0 0 ___ C3 i_ - 0 0 0 i 0 0 0 Cl i - - 53 0 53 40 0 40 13 0 13 -51 '' Annuai C2 53 0 53 40 0 40 13 0 13 C3 53 0 53 40 0 40 13 0 13 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 189 Table C.15 H;aiti: MDG-2015 Financing Gap under Alternative Policy Measures ; ~B: EFrICIINCY A. QUALITY MEACURES MEASURES C: FINANCING MEASURES Average Annual Govemment Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per *Inputu Other multiple ef per . Average As % ol, % for Recurrent (as % of FIrnancIng Policy Scenarlo' Teacher than Teachers. capIta GP) Repetition Rate GOP Education Spendingr total) Gap' Srarw quo 46 10.0% 6.8 17.0% 9.3% I 20.6% 387% 76.6% 0 A only 40 33.3% 6.8 7 A+1B 40 33.3% . 5.3 10.0% 0 'Bcst Cl 40 33.3% 4.0 10.0% 16.0% 20.6% 50.0% 10.0% 33 praCtic": .C2 16.0% 20.0% 34 A +B8 + . C3 r . * : 16.0% 20.0% 34 Nlote: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency impmvement, and three altemative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.16 Haiti: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL .omesutic COST OF MOG 2015 DOMCSTiC RESOURCCE FINANCING cost Resources - - : -_-- Item Perlod Scenario .rOtulized Recurrent Capital Total Recurrent Capltal Total Recurrent Capital Total * . Cl 993 1,142 344 1,486 993 0 993 149 344 .493 Cumul2ative, C2 972 1,142 344 1,486 972 0 972 170 344 513 . 2001-2015 __ _ 5 * * .C3 972 1,142 344 1,486 972 0 972 170 .344 513 Cl 66 76 23 99 66 0 66 10 23 33 : . Annual , C2 65 76 23 99 65 0 65 11 23 34 _____ .________ . C3 65 76 23 99 65 0 65 I I 23 34 , I '' Cl ,' *J O . a O '' O * t' 0 0 0 0 0 0 C o,5* Annual C2 0 0 0 0 0 0 ___ ____ ; C3 0 0 0 0 0 0 Cl 76 23 99 66 0 66 10 23 33 Anual C2 76 23 99 65 0 65 11 23 34 C3 76 23 99 65 0 65 ii 23 34 Note: Best practice policies refer to the combination of scenaros A+B+C. Shaded cells denote no change from values directly above. nee TECHNICAL ANNEXES Table C.17 Honduras: MD6-201 5 Financing Gap under Alternative Policy Measures B FFC IENCY A: QUAL,TY MEASURES MEASURES C. FINANCING MEASURES Average Annual Government Revenue Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other murtiple of per Average As % of % tor Recurrent (as % of Financing Policy Scenario' Teacher lthan Teachers' capita OP) Repetition Rate GDP Education Spendingd ntotl) Gap' Statwquo 32 12.1% 5.0 8.0% 18.5% 30.1% 49.6% 6.9% 17 Aonly 32 33.3% 5.0 , . , .* - 169 A+B 40 33.3% 3.5 8.0% 0 'Best ' Cl 40 33.3% 3.5 8.0% 18.0% 30.1% 50.0% 10.0% 0 practcole: C2 18.0% 20.0% 5 A + B + C3 18.5% 20.0% 3 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three altemative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.18 .... .... .... ... ........ ....... .... .... ... .... .... ... ............... .... .... ... .... .... ... .... ....... Honduras: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) F,NA.c-o SOUACEr . GAP COR EI*ERNA . Domestc COST OF MDG 2015 DCo.E.-mC RESOURCES F.NANC,I,G Coal Resources Item Perlod j Scenario Mobliized Recurrent Capital Total Recurrent Capital Total Recurreat Capital Total 8 i Cl ! 3,114 2553 561 3,114 2553 561 3,114 0 o O sa , Cumulanivew , C2 3,043 2,553 561 3,114 2,553 490 3,043 0 71 71 li g i 2001 2015 ; C3 3,063 2,553 561 3,114 2,553 510 3,063 0 51 51 Cl 208 170 37 208 170 37 208 0 0 O = j Annua. C2 203 170 37 208 170 33 203 0 5 5 i'. C3 204 i 170 37 208 170 34 204 . 0 3 3 S ~~~~~~~i i 0 0 R , Annual C2 0 O 0 _ _ _ _ _ ~~~C3 0 0 0 ____ . . Cl ,g.2 170 ', 37 208 170 37 208 0 0 0 Annual C2 170 37 208 170 33 203 0 5 5 C3 170 37 208 170 34 204 0 3 3 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 191 Table C.19 India: MD6-2015 Financing 6ap under Alternative Policy Measures B3 EFFICIENoCY A: QUALITY MEASuRrco MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (as . Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of FInancing Policy Scenario' Teacher than Teachers' capita GDP) Repetition Rate GOP Education Spendineg total) Gap, Sranrsquo 52 23.2% 3.4 20.0% 21.2% 12.4% 32.1% 12.5% 146 A only 40 33.3% 3.5 2,470 A + B 40 33.3% 3.5 10.0% 1,782 Best , C1 40 33.3% 3.5 10,0% 16.0% 20.0% 42.0% 10.0% 67 practice": C2 16.0% 20.0% 67 A + B + *. C3 212% 20.0% 28 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.20 ..................................................................................................................................... India: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCINCG SOURClES GAP FOR EXTERNAL Domestic CoOr op MOG-2015 DOMESTIC RESOURCES FINANCING. Cost Resourcmsn : : : : ! : Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 115,278 109,754 6,525 116,279 109,754 5,524 115,278 0 1,001 1,001 -E . Cum0'lul2ative,X . C2 115,278 109,754 6,525 116,279 109,754 i 5,524 115,278 0 1,001 1,001 it~, 2001 2015 X ______ j_ C3 115,858 109,754 6,525 116,279 109,754 6,104 115,858 0 422 422 Cl 7,685 7,317 435 7,752 7,317 368 7,685 O 67 67 : , Annual] C2 7,685 7,317 435 7,752 7,317 368 7,685 0 67 67 C3 7,724 7,317 435 7,752 7,317 407 7,724 0 28 28 c 0 0 0 0 0 0 C0 8 iAnnual C2 0 0 0 0 0 0 ; _ j C3 j_ ;;_O_*. j e ! ° i I 0 0 0 0 0 ° Cl 7,317 435 7,752 7,317 368 7,685 0 67 67 Annual C2 7,317 435 7,752 7,317 368 7,685 0 67 67 C3 7,317 435 7,752 7,317 407 7,724 0 28 28 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. aga TECHNICAL ANNEXES Table C.21 ................. .............................................................. Lao PDR: MDG-2015 Financing Gap under Alternative Policy Measures I * B: EPFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES I ._______________________ . - . I . . _____________ Average Annual Government Revenues' * Primary Private Spending on Teacher Salary (as Educaton Enrollments Annual Pupils Per oInpuas Olher multIple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) Repetition Rate GDP Education Spending' total) Gap' Statfwquo 31 19.6% o04 22.6% 38.1% 1.4% 48.8% 1.9% 2 A only 31 33.3% 3.5 49 A + B 40 33.3% 3.5 10.0% , 33 "Besr Cl 40 33.3% 3.5 10.0% 16.0% 20.0% 42.0% 10.0% 13 practice': C2 160% 200% 13 A + B + . C3 t ' * *, 38.1% 20.0% 6 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.22 Lao PDR: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Oomestic COST OP MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources ._._: . . i Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital i Total Recurrent Capital Total Cl 393 558 36 594 393 0 393 165 36 j201 E j C20umulu2ative, i C2 393 558 36 594 393 0 393 165 36 201 r N C3 . 500 * 558 36 594 500 0 500 57 36 94 Cl 26 37 2 40 26 0 26 I I 2 13 : j Annual C2 26 37 2 40 26 0 26 I I 2 13 _________i_ C3 33 37 2 40 33 0 33 4 2 6 0 0 ~ ~ ~ ~~0 0 0 0 | | I, 9 '. , , i , : 0 Annual C2 00 0 0 0 ______ lt___j__ti_ C3 t _-_-i_ ° 0 0 0 0 o 0 C I 37 2 40 26 0 26 1 1 2 1 3 Annual C2 37 2 40 26 0 26 1 2 13 C3 37 2 40 33 0 33 4 2 6 Note: "Best practice' policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 193 Table C.23 Moidova: MDG-2015 Financing Gap under Alternative Policy Measures B EFFICIENCY A: QUALITY MEADURES MEASURES C: FiNAPICING MrASURES Average Annual Government Revenuesf Prmary Prtate Spending on Teacher Salary (as Education Erollnrments Annual Pupils Per Inputs Other multple 01 per Average As % of % for jRecurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) iORepetonRate SDP Educaton Spendiogi tetall Gap Satswquo 21 67.8% 1.1 0.9% I 29.8% 18.5% 25.5% 0.0% 0 A only 21 33.3% 3.5 17 A + B 40 33.3% 3.5 0.9% 5 Becs Ci 40 33.3% 3.5 0.9% 16.0% 20.0% 50.0% 10.0% 5 practice: C2 16.0% 20.0% 5 A + B + C3 29.8% 20.0% 4 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three altemative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.24 MJoldova: M~D6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCINCI SOURCEO GAP FOR EXTERNAL Domestic COST OF MDG-201 5i DOMESTIC RESOURCES FINANCING ceto Resources .. Itemn Penod Scenario jMobilized Recurrent Capital Total Recurrent Capital; Total Recurrent Capital Total Cl 309 388 0 388 309 0 309 79 : 0 79 i Cumnulative, i C2 309 388 0 388 309 ,0 i309 79 0 o 79 *2001-2015 . c ____ >ti 01215i C3 327 388 0 388 327 0 327 61 i 0 £61 Cl 21 26 0 26 21 0 21 5 0 5 Annual C2 21 26 0 26 21 0 21 5 0 5 _____ i____ ,i___ C3 22 26 0 26 22 0 22 4 0 4 ci 0 0 0 0 0 0 Annual C2 0 0 0 o 0 0 C3 i_ * 0 0 0 0 0 0 Cl1 + 26 0 26 21 0 21 5 0 5 -4 Annual C2 1 26 0 26 21 0 21 5 0 5 C3 26 0 26 22 0 22 4 0 4 Nlote: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 11049 TECHNICAL ANNEXES Table C.25 ................................................................................. Mongolia: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on TeacherSalary(as . Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of * Recurrent (asI of *inancing Policy Scenario' Teacher than Teachersj capita GDP) j Repetition Rate GDP Education Spendingd total) cap, Starur.quo 31 15.0% 39 09% 29-2% 24.6% 33.6% 05% 0 A only 31 33.3% 3.9 0 A + B 40 33.3% 3.9 0.9% 'Best Cl 40 33,3% 3.9 0i9% 16.0% 20.0% 50.0% 10.0% 0 practice: C2 16.0% 20.0% 0 A + B + C3 16.0% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios am: A for quality improvement B for efficiency improvement and three alternative resource mobilization scenarios (Cl C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.26 Mongolia: MD6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 319 318 0 . 318 318 . 0 318 0 0 0 C l ¢ 2001 2015' j C2 319 318 0 318 318 0 318 0 0 0 2001 2015 . . ___ _ g ; C3 319 318 0 318 318 , 0 318 0 0 0 0 _ _ _ _ _ _ _ _ _ _ Cl 21 21 0 21 21 0 21 0 O 0 v; . Annual C2 21 21 . 21 21 0 21 0 0 0 ________._ C3 21 21 0 21 21 0 21 0 0 0 Cl 0 0 0 0 0 0 0a e Annual C2 0 0 0 0 0 . 0 ______ e; _______ ; C3 0 0 0 0 . 0 _ j 0 Cl 21 0 21 21 0 21 0 0 0 O U ,i Annual ,i C2 21 0 21 21 0 21 0 0 O C3 . 21 0 21 21 0 21 0 0 0 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 195 Table C.27 Nepal: MDG-2015 Financing Gap under Alternative Policy Measures ,: EFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues0 Primary Private Spending on . Teacher Salary (as * Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % os % for Recurrent (as % of Financing Policy Scenarlso Teacher than Teachers' capita GOP) . Repetition Rate GOP Education Spendinge total) Gap' Starus quo 36 20.0% 2.0 29.9% 10.4% 18.4% 53.2% 8.1% 5 A only 36 33.3% 3.5 97 A + B 40 33.3% 3.5 10.0% . 54 "Best Cl 40 33.3% 3.5 100% 14.0% 20.0% 42.0% 10.0% 44 praccice": .C2 14.0% 20.0% 44 A + B + C3 14.0% 20.0% 44 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.28 Nepal: MD6 2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domeslic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING cost Resources . - i Item Period Scenario * Mobilized Recurrent Capital Total Recurrent CapRtal Total Recurrent Capital Total Cl 1,333 1,787 213 2,000 1333 0 1,333 454 213 667 Cumulacive, C2 1,333 1,787 213 2,000 1,333 0 , 1,333 454 213 667 e-, _ _ g. * C3 1,333 1,787 213 2,000 1,333, 0 1,333 454 .213 667 Cl 89 119 14 133 89 0 89 30 14 44 = * Annual C2 89 119 14 133 89 0 89 30 14 44 ___i__; C3 ; 89 ; 119 14 133 89 0 89 30 14 44 '- , . Cl 030 0 0 0 0 v Annual. C2 0 0 0 0 0 0 ________ j_ C3 _____ _ ; 0 0 0 0 0 ' 0 Cl 119 14 133 89 0 89 30 14 44 Annual C2 119 14 133 89 0 89 30 14 44 C3 119 14 133 89 0 89 30 14 44 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above, age TECHNICAL ANNEXES Table C.29 Nicaragua: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY * A: QUALITY MEASUJRES MEASURES C: FINANCING MEASURES * ~~~~~~~~~Average Annual Governeunt Revenvuee Primary Private *Spendinig on iTeacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % af % far Recurrent (ns % af Financing Policy Scenario' Teacher thn Teachers' capita GOP) Repetition Rate GDP Education Spendingd total) Gap' Status quo 36 32.7% 3.2 12.0% 25.3% 16.8% 48.7% 6.8% 0 A only 36 33.3% 3.5 19 A +B 40 33.3% 3.5 10.0% "Bevt Ci 40 33.3% 3.5 10.0% 16.0% 20.0% 50.0% 10.0% 8 prcie C2 *16.0% 20.0% 8 A + B+ C3 25.3% 20.0% 2 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3) The combination of scenarios A+B+C is considered "best practice. Ii. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.30 Nicaragua: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP ao. E.r- Domestic COST OF MDG-201 5 DomESTIc ResOURCES F.-.Nc-c.G Cost Resources .- 4 Item Period Scenana Mohilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Ci 974 1,012:85: 1,097 974 0 :974 38 05 123 Cumulauve C2 974 ~~~~~1,012 85 1.097 974 0 I 974 38 85 123 U i!. 2001 2015 _____ ~~~~C3 1,062 1,012 85 1,097 1,012 50 11,0621 0 35 ~35 Ci 65 67 6 73 65 0 65 3 6 8 AnnualI C2 65 67 6 73 65 0 65 3 6 8 _________ ~~C3 71 67 6 73 67 3 71 0 2 2 - ~~~~~~ ~~ci 0 0 0 0 0 Annu.al C2 0 0 0 0 0 0 _ _ _ _ _ C3 _ _ __3 0 0 0 0 0 0 CI 67 6 73 65 0 65 3 6 8 Arnnual C2 67 6 73 65 0 65 3 6 8 C3 67 6 73 67 3 71 0 2 2 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C - COUNTRY SIMULATION RESULTS 197 Table C.31 ................................................................................. Pakistan: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (asi Education Enrollments Annual Pupils Per i fnputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GDP) Repetition Rate GOP Education Spending' total) Gap' Sautisguo 32 19.3% 3.6 6.2% 16.7% 10.2% 518% 294% 285 A only 32 33.3% 3-5 450 A + B 40 33.3% 3.5 6.2% 261 'Best Ci 40 33.3% 3.5 6.2% 16.0% 20.0% 42.0% 10.0% 204 practiCce jC2 16.0% 20-0% 204 A + B + C3 16.7% . 20.0% 173 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3Q. The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.32 Pakistan: MDG-2015 5 Cost Estimates and Sources of Financing under "B, est Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) I , r'INANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Rosources . ! : Item Period Scenario .obilized Recurrent Capital Total Recurrent *Captal Total Recurrent Capital Total 8 i Cl . 15,919, 16,748 * 2,224 18,972 15,919 15,919 829 2,224 3,053 2 cumlative, C2 15,919 16,748 2,224 18,972 15,919 0 15,919 829 2,224 3,053 a 2 0 20 ; C3 16,373 16,748 2,224 18,972 16,373 0 16,373 375 2,224 2,599 s t 5 ci 1,061 1,117 148 i 1,061 0 1,061 55 148 204 = . Annual C2 1061 1,117 148 1,265 1,061 0 1,061 55 148 204 ____ *______ ;C3 1,092 1,117 148 1,265 1,092 0 1,092 25 148 173 ,.t, * . ~~~~~Cl . 0 O 0o * O 6 i CI 0 0 ~ ~ ~~ ~~~~~~~~0 0 0 0 Annual C2 0 0 0 0 0 0 '__ _____ _ j C3 j0 j 0 0 0 0 * , 0 cI 1117 148 1,265 1,061 0 1,061 55 148 204 oE* Annual C2 1,117 148 1,265 1,061 0 1,061 55 148 204 C3 1,117 148 1,265 1,092 0 1,092 25 148 173 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. M@S TECHNICAL ANNEXES Table C.33 Yemen: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other * multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) RepetitIon Rate GOP Education Spendingd total) Gap Sratnvquo 25 26.7% 34 7.0% 35.2% 15.8% 48.0% 1.4% 82 A only 25 33.3% 3.5 231 A+B 40 33.3% 3.5 7.0% 6 'Best . Cl * 40 33.3% 3.5 7.0% 16.0% 20.0% 50.0% 10.0% 70 practice: C2 16.0% 20.0% 70 A +B+ C3 35.2% 20.0% 6 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.34 ....................................................................................................................................... Yemen: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL DOmeotit Cosr oF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources . . . . Item Period Scenarlo Mobilized Recurrent CapHtal Tntal Recurrent Captal Total Recurrent Capital Total Ci 4423 4736 732 .5468 4423 0 4,423 . 313 . 732 1,045 Ci-unulative C2 4423 4736 732 5468 4423 * 0 4,423 313 732 1,045 y . 2001 2015 637 5. 7 5 9 C3 5373 4736 732. 5468 4736163715373 0 .95 .95 u 3 * j Cl 295 316 49 365 295 . 295 21 49 70 5 Ann.ua C2 295 316 49 365 295 0 295 21 49 70 ______ ._________ . C3 358 316 49 365 316 42 358 0 6 6 ci 0 0 0 0 0 0 : E 8 Annual C2 0 0 0 0 0 0 ___ i ____ _ ; C3 ; _ 0 O 0 0 0 0 0 Cl 316 49 365 295 0 295 21 49 70 Annual C2 316 49 365 295 0 295 21 49 70 C3 , 316 49 365 316 42 358 0 6 6 Note "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 199 Table C.35 .............................................................................. Angola: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A; QUALITY MEASURES MEASURES C: FINANCING MEASURES I . . _ : ~~~~~~~~~~~~~.. .4 Average Annual Government Revenues' Primary Private . Spending on Teacher Salary (as . Education Enrollments Annual Pupils Per Inputs Other multiple ot per Average As % of i % for j Recurrent j (as % of Financing Policy Scenario' Teacher than Teachers' capita GDP) Repetition Rate GOP Education Spending' total) Gjap Srarw quo 24 19.0% 1.5 25.0% 55.7% 4.3% 41.6% 6.0% 87 A only 24 33.3% 3.5 354 A + B 40 333% 3.5 10.0% 161 "Besc CI 40 33.3% 3.5 10.0% 18.0% 20.0% 50i0% 10.0% 41 practice:. C2 18.0% 20.0% 41 A+B+ *C3, 55.7% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.36 .......................................................................................................I................................ Angola: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG 2015 DOMESTIC RESOURCES FINANCING cost Resources - i trem Period Scenario Mobilized Recurrent C CapHal Total Recurrent . Capital * Total Recurrent Capital . Total Cl 3,793 4,337 75 4,412 3,793 3 0 3,793 544 75 619 C Cumularive, C2 3,793 4,337 75 4412 3793 i3,793 544 75 619 ~ . 2001 2015 ____* C3 i 7,191 4,337 75 4,412 4,337 75 i4,412 . i 0 i Z', i i Cl * 253 289 5 294 253 0 253 36 5 41 : Annual C2 253 289 5 294 253 0 253 36 5 4i 3 ___i___j__ C3 479 289 5 294 289 5 294 0 i 0 i Ct 2 2 0 0 2 2 a t Annual C2 2 2 0 0 2 2 C3 2 2 2 2 0 2 Ci 292 5 296 253 0 253 39 5 44 0 Annual C2 292 5 296 253 0 253 39 5 44 C3 292 5 296 292 5 296 o o o Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 2©O TECHNICAL ANNEXES Table C.37 ............................................................................. Benin: MDG-2015 Financing Gap under Alternative Policy Measures B* EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Spendig on .Average Annual Covernment Revenue' Primary Private .Spending an Teacher Salary (as . Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % fur Recurrent (as % of Financing Policy Scenario Teacher thanTeachers' capita GOP) Repetition Rate GOP Education Spending' total) j op' Stats quo 54 26.4% 4.6 25.0% 15.3% 16.5% 62.6% 10.8% 16 A only 40 33.3% 4.6 39 A + B . 40 33.3% 3.8 10.0% 21 'Best Cl . 40 33.3% 3.8 10.0% 16.0% 20.0% 50.0% 10.0% 20 practiCe: C2 16.0% 20.0% 20 A + B + C3 16.0% 20.0% 20 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.38 ...................................................................................................................................... Benin: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources . . ' . .__ Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent CapHal Total 8 * ', ive Cl 778 936 145 1,081 778 0 778 158 145 303 ; Cuulativ i C2 i 778 936 i145 1,081 778 0 778 158 145 303 2001 2015. N 30 _ ! _ ; C3 778 936 1,45 1,081 778 0 778 158 145 303 ci 52 62 10 72 52 11 10 20 ' I Annual C2 52 62 10 72 52 0 52 11 10 20 ____i __ i_ C3 52 62 10 72 52 0 52 11 10 20 , 0 . j Cl j j I 1 I I 0 0 I I S Annual C2 0I I j I I *_____ .__________ C3 I I 0 0 I 1 c i 63 10 73 52 0 52 i 11i 10 21 0 , Annual C2 63 10 73 52 0 52 11i 10 21 C3 63 10 73 52 0 52 II 10 21 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 201 Table C.39 ...................................................................................... Burkina Faso: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MLASURES Average Annual Gevernment Revenue Primary . Private Spending an Teacher Salary (as . Education Enrollments Annual Pupils Per Inputs Other *multple ot per Average As % os % for Recurrent (as % of Financing Policy Scenario Teacher than Teachersb capita GOP) Repetition Rate GOP Education Spendinge total) Gap Stats quo 49 30.7% 8.0 17.7% 14.7% 17.1% 64.0% 10.8% 79 A only 40 33.3% 8.0 106 A + B 40 33.3% 4.2 10.0% 44 "Besr Cl 40 33.3% 4.2 10.0% 14.0% 20.0% 50.0% I 10.0% 48 practice: . C2 14.0% 20.0% 48 A + B + C3 14.7% 20.0% 47 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.40 ................................................................................................................................... Burkina Faso: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCINO SOURcEs GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING cost Resources - . . . . i Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital . Total Cl 764 1,138 351 1,489 764 0 764 374 351 725 Cumulative, C2 764 1,138 351 1,489 764 0 764 . 374 , 351 . 725 ~ . 2001 2015 _ _ > j i C3 787 1,138 1351 1,489 787 ' 0 787 351 .351 703 Cl 51 76 23 99 51 0 51 25 23 48 0 Annual C2 51 76 23 99 51 0 51 25 23 48 *._______ * C3 52 76 23 99 52 0 52 23 23 47 Cl 11 11 0 0 11 I I C) uj 0. Annual C2 I I 11 0 O 11 11 6 ________ C3 II 0 0 II C3 ;' Cl 87 23 111 51 0 51 36 23 60 e Annual C2 87 23 111 51 0 51 36 23 60 C3 87 23 111 52 0 52 35 23 58 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 212 TECHNICAL ANNEXES Table C.41 .... ............................................................................ Burundi: MDG-2015 Financing Gap under Alternative Policy Measures 1 .B: EFFICIENCY A: QUALITY MEASURES MEASURES C FINANCING MEASURES Average Annual Government Revenues' Prlmary Private Spending on Teacher Salary (as Educaton Enrollments Annual Pupils Per *iputs Other muliple of per A Average As % ot % tor Recurrent (es % of Flnancing Policy Scenarto Teacher than Teachersb capita GOP) Repetition Rate GOP Education Spendingd total) Gap, Statusquo 55 22.1% 5.3 27.5% 17.4% 20.4% 35.5% 0.0% 18 A only 40 33.3% 5-3 33 A + B . 40 33.3% 3.9 10.0% 17 'Best Cl 40 33.3% 3.9 10.0% 14.0% 20.4% 50.0% 10.0% 16 praatice: ^C2 14.0% 20.0% 16 A + B + C3 17.4% 20.0% 15 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality imprmvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.42 ........................................................................................................................................ Burundi: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR Ext'ERNAL .omestIc COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources : _ Item Period Scenario Mobilized Recurrent Capibl Total Recurrent Capibl Total !Recurrent Capital Total Cl 109 270 163 433 189 0 189 81 163 !244 Cumulative, C2 187 270 163 433 187 187 84 ~ . 2001 2015. c, ______ j C3 214 270 163, 433 214 0 214 56 1 163 219 Cl 13 18 I I 29 13 0 13 5 II 16 Annual C2 12 18 I I 29 12 0 12 6 I I 16 j_______ i C3 14 18 I I 29 14 0 14 4 I I 15 clI 5 5 0 0 5 5 Annual C2 5 5 0 5 5 ______ __________ C3 __ -_ __ 5 5 0 0 5 _' ' _ 5 cl 23 11 34 13 0 13 10 11 21 i B j Annual. C2 23 11 34 12 0 12 11 11 21 C3 23 I I 34 14 0 14 9 I I 20 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 203 Table C.43 Cameroon: MD6-2015 Financing Gap under Alternative Policy Measures ; e~~~~~~B EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEA5URES Average Annual Government Revenues, Primary Private . Spending on Teacher Salary (as . : i Education Enrollments Annual Pupils Per Inputs Other multiple of per i Average @ As % of % for Recurrent (as % of Financing Policy Scenario Teacher ttan Teachersb capita G6P)I Repetition Rate GDP oEducation Spending total) Gap Stas quo 65 32.5% 3.4 25.9% 15.5% 10.8% 66.3% 19.0% 24 Aonly 40 33.3% 3.5 100 A + B 40 33.3% 3.5 10.0% 89 'Best Cl 40 33.3% 3.5 10.0% 16.0% 20.0% 50.0% 10.0% 46 practice': C2 160% 200% 46 A + B + C3 16.0% 20.0% 46 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery underthe specific policy scenario and the total resources for primary education mobilized domestically. Table C.44 ..................................I.............................................................................................. Cameroon: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DomEsTIc RESOURCES FINANCING Cost Resounrces : - - : Item Period Scenario Mobilized Recurrent Capital Total Recurrent . Capital Total Recurrent Capital Total ' i Cl 2,625 2,887 420 3,308 2,625 0 2,625 263 420 683 E Cumulative, , C2 2,625 2,887 420 3,308 2,625 0 2,625 263 420 683 i . 2001 2015 _________ C3 2,625 2,887 420 3,308 2,625 0 2,625 263 * 420 683 Cl 175 192 28 221 175 0 175 18 28 46 Annuai. C2 175 192 28 221 175 0 175 18 28 46 C3 175 192 28 221 175 0 175 18 . 28 46 Cl ii 11 . 0 0 11 11 Ct Annual C2 i5 11 0 0 1, C3 _ _ , _11 ii 0 0 11 C i 203 28 231 175 0 175 28 28 56 d uO i Annual C2 203 28 231 175 0 175 28 28 56 C3 203 28 231 175 0 175 28 28 56 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 2016 TECHNICAL ANNEXES Ta6be C.45 ................................................................................................... Central African Republic: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues ! Primary Private Spending on Teacher Salary (as Educaton Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of %lor Recurrent (as % ef Flnancing Policy Scenario Teacher than Teachers' capita GOP) Repetition Rate GDP EducaAon Spending' total) Gap Status quo 79 28.5% 4.9 32.8% 9.6% 12.5% 52.4% 3.3% 14 A only 40 33.3% 4.9 32 A + B 40 33.3% 3.7 10.0% 18 'Best . Cl 40 33.3% 3.7 10.0% 14.0% 20.0% 50.0% 10.0% 12 prucrice": * C2 14.0% 20.0% 02 A + B + . C3 14.0% 20.0% 12 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.46 ................................................................................................................................... Central African Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURcEs GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources * Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capual. Total 8 , ' Cl * 232 302 103 405 232 0 232 70 103 173 Cumnulantic, C2 232 302 103 405 232 0 232 70 103 173 ~ ..2001 2015 _____. __. C3 232 302 103 405 232 0 232 70 103 173 A Cl] 15 20 7 27 15 0 15 5 7 12 ' . Annual . C2 15 20 7 27 15 0 15 5 7 12 _____ ;________ ;C3 15 20 7 27 15 0 15 5 7 12 Ci 2 2 0 0 2 2 Annual C2 2 2 0 0 2 2 E__ ; __ ;_ C3 2 2 ; 0 0 ; 2 . -. 2 Cl ' 22 7 29 , 15 0 15 7 7 14 Annua. C2 22 7 2 15 2 0 15 7 7 14 C3 22 . 7 29 . 15 0 15 . 7 ' 7 *14 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 205 Table C.47 ................................................ ............................. Chad: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES . MEASURES C: FINANCING MCASURES Average Annual Government Revenues' Primary Prniate Spending on ITeacher Salary (as Educatiuon Enrollments Annual Pupils Per Inputs Other multiple of per Average As % ef % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GDP) Repetition Rate GOP Education Spendingd etoal) Gap° Status quo 72 34.2% 4.8 24.6% 8.0% 20.9% 65.5% * 8.8% 21 A only 40 34.2% 4.8 48 A + B 40 33.3% 3.6 10.0% 29 'Best Cl 40 33.3% 3.6 10.0% 14.0% 20.9% 50.0% 10.0% 24 practice: C2 14.0% 20.0% 25 A + B + C3 14.0% 20.0% 25 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.48 ..................................................................................................................................... Chad: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR ExTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources . Item Perioda Scenaio Mohilized . Recurrent Capital . Total Recurrent . Capital Tutal * Recurrent t Capital . Total Cl 430 498 .290 788 430 0 430 69 .290 358 Cumulative, . C2 i 418 498 1290 788 418 0 418 80 290 370 e C;0 2 5 . C3 418 498 290 788 418 0 418 . 80 290 370 Cl 29 33 19 53 29 0 29 5 19 24 Annual C2 28 33 19 53 28 0 28 5 19 25 ______ ;_________ ;C3 28 33 19 53 28 0 28 5 19 . 25 iA~~~~~~~ ~~Cl 2 2 0 0 2 2 Annual C2 2 2 0 0 2 2 ;_____ ;___________ C3 2 2 0 0 2 2 Cl 36 19 55 29 0 29 7 19 26 Annual C2 . 36 19 55 28 0 28 8 19 27 C3 36 19 55 28 0 28 8 19 27 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ace TECHNICAL ANNEXES Table C.49 ....................................................................................................... Congo, Democratic Republic: MDG-2015 Financing Gap under Alternative Policy Measures B: EPFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Anneal Government Revenues' Primary Private . Spending on Teacher Salary(as Ia Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) RepetitienRate G6P Educafion Spending' total) Gap' Staruzquo 42 10.3% 0.9 15.0% 10.6% 3.2% 65.1% 10.0% 109 A only 40 33.3% 3.5 198 A + B 40 33.3% 3.5 10.0% 187 'Best C I 40 33.3% 3.5 10.0% 14.0% 20.0% 500% 100% 146 praCtice: C2 14.0% 20.0% 146 A + B + C3 14.0% 20.0% 146 Note-^ Shaded cells denote no change from Values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3. The combination of scenarios A+B+C is considered "best practice." b. As a share of pnmary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.50 .................................................................................................................................. Congo, Democratic Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL .Demestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources ' . . Item Period Scenario Mobilized Recurrent * Capital Total Recurrent Capital Total Recurrent Capial Total 8 Cl 814 1,456 1,542 2,998 814 0 814 642 1,542 2,185 E *COumulnU2atixve'' * C2 814 1,456 1,542 2,998 814 O 814 642 1,542 2,185 ~.2001 2015v c _ _ g* . C3 814 1,456 1,542 2,998 814 0 814 642 1,542 2,185 = b e ~~~~~Cl * 54 * 97 *103 , 200 * 54 , 0 54 . 43 ,103 .146 = Annual C2 54 97 103 200 54 0 54 43 103 146 ____;_____ C3 54 97 103 200 54 0 54 43 103 146 Cl 1 7 0 0 17 17 a~ 8.Annual C2 17 17 0 0 17 17 . j C3 17 . , 17 0 0 17 _ . 17 . . Cl , : . 114 103 217 54 0 54 60 103 163 Annual , C2 114 103 217 54 0 54 60 103 163 C3 114 103 217 54 0 54 60 103 163 Note: "Best practice' policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 207 Table C.51 ......................................................................................... Congo, Republic: MDG-2015 Financing Gap under Alternative Policy Measures .* EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES , * * Average Annual Government Revenues' Primary Private *. * Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spendinge total) Gap, Status quo 61 20.3% 3.4 31.1% 26.7% 8.6% 36.6% 15.2% 17 A only 40 33.3% 3.5 47 A + B 40 33.3% 3.5 10.0% 38 "Bcst Cl 40 33.3% 3.5 10.0% 18.0% 20.0% 50.0% 10.0% 9 ptScri1C . C2 18.0% 20.0% 9 A + B + . C3 26.7% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.52 Congo, Republic: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and "Aiternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-20 1 5 DOMESTIC RESOURCES , FINANCING Cost Resources . * Item Period Scenario Mobilized Recurrent Capital Total *Recurrent Capital Total Recurrent Capital Total u ', , Cl .1i,046 * 1,086 * 89 1,175 1,046 . 0 1,046 . 40 , 09, 129 ; Cumulatuve, C2 1,046 1,086 89 1,175 1,046 0 1,046 . 40 , 89 , 129 ~ . 2001 2015 _____ g ' C3 1,342 1,086 89 1,175 1,086 89 1,175 0 0 0 ',t-, '. . Cl 70 72 6 78 70 0 70 3 6 9 ' , Annual C2 70 72 6 78 70 0 70 3 6 9 i________ . C3 89 72 6 78 72 6 78 0 0 0 cil 0 1 C e O, Annual C2 *1 0 0 1 I __________ ' C3 * _I ' I ,_I I 1 1 0 I Ci 74 6 80 70 0 70 4 6 10 0E Annual . C2 74 6 80 70 0 70 4 6 10 C3 74 6 80 74 6 80 0 0 0 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above, 203 TECHNICAL ANNEXES Table C.53 ..................................................................................... C6te d'lvoire: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C FINANCING MEASURES Average Annual Government Revenues Primary Private . Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple ot per Average As% ao % for Recurrent (as % af Financing Policy Scenario . Teacher than Teachers' capita GOP) Repetition Rate GDP Education Spending' total) Gap' Skanwquo 46 22.5% 5.7 24.7% 16.5% 21.5% 49.0% 11.6% 102 Aonly 40 33.3% 5.7 187 A + B 40 33.3% 4.3 10.0% 63 "Best Cl 40 33.3% 4.3 10.0% 18.0% 21.5% 50.0% 10.0% 42 practice': C2 18.0% 20.0% 56 A + B + , C3 18.0% 20.0% 56 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.54 ................................................................................................................................... C6te d'lvoire: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic 1 COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources ; Item Pernod Scenario Mobilizedi Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 4,815 5,205 237 5,442 4,815 0 4,815 391 237 628 Cumulativc. C2 , 4,603 5,205 237 5,442 4 4,603 603 0 237 840 u . 2001 2015. .. c_.°°.___ C3 4,603 5,205 237 5,442 4,603 0 4,603 603 237 840 Cl 321 347 16 363 321 0 321 26 16 42 S AnnualA.. . . C2 307 347 16 363 307 0 307 40 16 56 '_'._______ C3 307 347 16 363 307 0 307 40 16 56 Cl 16 1 0 16 16 Annual . C2 16 16 0 0 16 16 _____. _,. C3 _, 16 . . 16 0 0 16 '6 Cl 363 16 379 321 0 321 42 16 58 Annual C2 363 16 379 307 0 307 56 16 72 C3 363 16 379 307 0 307 56 16 72 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 209 Table C.55 ............................................................................... Eritrea: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES Z C: FINANCING NMASURES I ... . :.-* n Average Annual u6vemnment Revenues, Primary Private Spendingan Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multple ot per Average As % of % for Recurrent (as % of Financing Policy Scenario Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spendingd total) Gap SZaw quo 49 29.6% 7.7 19.4% 34.6% 8.0% 53.6% 10.1% 16 A only 40 33.3% 7.7 22 A + B 40 33.3%0 . 4.3 10.0% 10 "Best Cl i 40 33.3% 4.3 10.0% 14.0% 20.0% 42.0% 10.0% 10 practice: C2 14.0% 20.0% 10 A + B + C3 34.6% 20.0% Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.56 Eritrea: MD6-20 1 5 Cost Estimates and Sources of Financing under "Best Pracctice" Policies and Alternat'ive Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES CAP FOR EXTERNAL DomeStlC I COST oF MOG.2015 DOGESTIC RESOURCeS FINANCIN0 Cast Resources - - . ___ Item Perlod Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total * ~ ~ ~ ~ ~ ~ ~ 1 *____ . *_*,__*__:-_;:_ Ci 198 250 94 344 198 0 198 z 52 94 z146 Cumuisove C2 198 250 94 344 198 0 198 52 94 146 . . 2001 2015 c _ _ z *.C3 I 325 250 94 344 250 75 325 o 19 19 Cl 13 z 6 23 13 0 13 3 6 10 *. Annual C2 13 17 6 23 13 0 1 3 6 10 ______ z_____ z__ C3 *. 22 17 6 23 17 5 22 0 1 I ci 0 0 0 0 0 0 a " 8. Annual C2 0 0 0 0 0 0 ,_ _ j_____ C3 j _ z_ 0 . 0 0 0 0 _-_. 0 Ci 17 6 23 13 0 13 4 6 10 Annual C2 17 6 23 13 0 13 4 6 10 C3 17 6 23 17 5 22 0 1 2 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 2BO A(TECHNICAL ANNEXES Table C.57 ..i:....................................................................... I...... Ethiopia: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES . MEASURES C: FINANCING MEASURES Average Annual Govemment Revenueoo Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per *nputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scnnaria Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spondine. total) Gap' Staws quo 61 205% 6.8 8.0% 17.8% 15.0% 46.2% 5.0% 146 A only 40 33.3% 6.8 312 A + B 40 33.3% 4.1 j 8.0%4. 192 'Best Cl 40 33.3% 4.i 8.0% 14.0% 20.0% 50.0% 10.0% 180 practice: C2 14.0% 20.0% 180 A + B + C3 17.8% 20.0% 157 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.58 Ethiopia: MD6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) -F.N.NCLNG SouRCES . . . % G-a pop EXs ERrA, Domestic COST OF MDG-201 5 DoMEST.C RESOURCES F.NSNC,NO Cost: Rerources L. ; Item Pennd Scenario iMohilized Recurrent Capital Total Recurrent Capital Total Recurrent Cat Total * ~ ~ ~ ~ ~ ~ ~ ____ , -:-I -I.t | Cl 2,113 2,940 1,872 4,811 2,113 0 2,113 j 827 1,872 j 2699 T Cumulativei C2 2,113 2.940 1.872 4,811 2,113 0 2,113 827 1,872 2,699 > g j 2001 2015 i C3 2454 2,940 1,872 4,811 2.454 0 2,454 486 1,872 2,358 Cl 141 196 125 321 141 0 141 55 125 180 Annual C2 141 196 125 321 141 0 141 55 125 180 ____i__j C3 i 164 196 125 321 164 0 164 32 125 157 0 . I ,I -j - , ,0 , : .6 ii 351 .i : 141 8 * 351 : 012 210 E Annual C ; 6 125 351 141 i * 141 * j 12i 'IU C3 226 125 351 j 164 0 164 62 125 187 Note: "Best practice' policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 211 Table C.59 The Gambia: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES M MEASURES C: FINANCING MEASURES * Average Annual Government Revenues' Primary Private . Spending on Teacher Salary (as : Education Enrollments Anntal Pupils Per *Inputs Other multipleatper Average As % of % for *Recurrent (as % of Financing Policy Scenario Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spendinge total) Gap' Staru quo 37 24.9% . 3.7 10.6% 18.5% . 16.6% 51.7% 8.5% 3 A only 37 33.3% 3.7 4 A + B 40 33.3% 3.6 10.0% 3 "Best . Cl 40 33.3% 3.6 10.0% 14.0% 20.0% 50.0% 10.0% 3 practice: : C2 14.0% 20.0% 3 A +B+ B C3 18.5% . 20.0% 2 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.60 The Gambia: MDG-2015 Cost Estimates and Sources OT Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources ' : : . : : ! : : Item Penod Scenario .iMbilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 140 170 17 187 140 0 140 31 17 47 Cumulative, C2 140 170 17 187 140 0 140 31 17 47 t ..200120 .5 C3 164 170 17 187 164 ' 0 164 7 17 23 Cl 9 1 12 9 O 9 2 1 3 = . Annual C2 9 11 I 12 9 0 9 2 1I 3 C3 i 11 i 11 I 12 i11 11 0 i 2 ,0 ~~~~~~c . . Cl '. O , O * 9 8 Annual C2 0 0 0 0 0 E Cumulative, C2 1,499 . 207 1,812 , 499 . 0 1,499 105 207 . 312 y . 2001-2015 12 c; __ 5_ i C3 1,692 1,604 207 1,812 1,604 88 1,692 0 120 120 C; ,, . g Cl 100 107 14 121 100 0 .00 7 14 21 Annual C2 100 107 14 121 100 0 iOo 7 14 21 _____. __* C3 113 107 14 121 107 6 113 0 8 8 Cl 6 6 0 0 6 6 a ~ 2,-, Annual. C2 6 6 0 0 6 6 E__ i __', _i C3 i ____ 6 , 6 0 i 0 6 * , 6 Cl 113 14 127 100 0 100 13 14 27 oO E jAnnual C2 113 i14 127 i100 0 100 13 14 27 C3 113 14 127 107 6 113 6 8 14 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 213 Table C.63 Guinea: MDG-2015 Financing Gap under 'Alternative Policy Measures B: EFFiCiENCY A: QUALITY MEAGURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues Primary Private Spending on Teacher Salary (as Educaton Enrollments Annual Pupls Per Inputs Other multiple of per Average As % co . % for Recurrent (as % of Financing Polcy Scenario , Teacher than Teachers' capita GOP) Repet'iton Rate GDP Education Spending' total) Gap' Sw quo 49 34.7% 2.7 23.3% 11.1% 18.1% 37.2% 161% 28 A only 40 34.7% 3.5 58 A + B 40 33.3% 3.5 100% 53 'Best Cl 40 33.3% 3.5 10.0% 14.0% 20.0% 50.0% 10.0% 35 practice": C2 14.0% 20.0% 35 A + B+ C3 14.0% 20.0% 35 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.64 Guinea: MDG-2015 Cost Estimates and Sources o.f Financing under "Best Practice" Policies and Alteernative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic I COST OF MDG-2015 DOMaSTIC RESOURCES FINANCING cost Resources - : item Perlod Scenario . obilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 790 1,141 181 1,322 790 0 790 351 181 .532 Ctniulat've, C2 790 1,141 181 1,322 790 0 790 351 181 532 u >.2 51 *. C3 790 1,141 *181* 1,322 790 0 790 351 181 *532 o,_. t I Cl 53 76 12 88 53 0 53 23 12 35 -i Annual C2 53 76 12 88 53 0 53 23 12 35 ______ i_________ i C3 i 53 76 12 88 53 0 53 23 12 35 ci I 0 0 1 I :- S i Annual C2 1 0 0 1 I _ _ _ ;__ _ - :I:C3 I i0 i : i 1 Ci 77 12 89 53 0 53 24 12 37 Annual C2 77 12 89 53 0 53 24 12 37 C3 77 12 89 53 0 53 24 12 37 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 2ge 9~TECHNICAL ANNEXES Table C.65 6uinea-Bissau: MDG-201 5 Financing Gap under Alternative Policy Measures .: EPFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES I. . . * , I * * * Average Annual Government Revenues- Primary Private i i Spending on i TeacherSalary (as Education Enrollments Annual . PupilsPer .Inputs Other mahtpleaper j Average As % a i % for Recurrent (as % *t Financing Policy Scenartar j Teacher than Teachers' capita GOP) Repetition Rate- GOP Education Spending' total) Gap' Sratw quo 37 34.3% 1.6 27.1% 19.6% 9.8% 35.0% 8.5% 3 A only 37 34,3% 3,5 7 A + B 40 33.3% , 3 5 10.0% , j ii 5 seSt j Cl 40 333% 35 100% 14.0% 20.0% 50.0% 10.0% 3 praccice: iC2 14.0% 20.0% 3 A + B+ C3 19.6% 20.0% 2 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.66 . .................................................................................................................................. 6uinea-Bissau: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FiNANCINrC SoUsc S' . OA. . :GF FO. EXTERNA. Domestic COST OF MDG-2015 DOMESTIC RESOURCES F.NANCINC Cost Resources . - Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total j Recurrent Capital Total Cl 55 87 18 105 55 0 55 . 32 18 50 Cumulative. . C2 55 87 18 105 55 0 55 32 18 50 *2001 2015 C _ _ i C3 68 87 18 105 68 0 68 19 18 37 Cl 4 6 1 7 4 0 4 2 I 3 = j Annual C2 4 6 7 4 0 4 2 3 ____'___;__ C3 5 6 7 5 0 5 I I 2 CI 0 0 0 0 0 Annual C2 0 0 0 0 ___i- __j _; C3 0 0 0 0 0 0 CI - : 6 7 4 0 4 2 I 4 '50 B iAnnual C2 6 7 4 0 4 2 I 4 C3 6 1 7 5 5 I 3 Note :Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 215 Table C.67 Kenya: MDG-201 5 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MeASURES C: FINANCING MEASURES Average Annual Government Revenuese pnmary Private iSpendirgoen Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple at per i Average As % oft % tar Recurrent (as % af Financing Policy Scenario' Teacher than Teachers' capita GOIP) Repetition Rate GDP Eduocation Spendingd totail) Gap' Sranni quo 31 4.2% 5.3 14.2% 24.2% 26.2% 44.2% 2.2% 53 *Aonly 31 33.3% 5.3 178 *A+B 40 33.3% 4.8 10.0% . 3 'Beost ClI 40 33.3% 4.8 100o% 16.o% 26.2% 50.0% 10.0% 61 practice': C216.0% 20.0% 113 A+ + IC3 24.2% 20.0% 32 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3) The combination of scenarios A+B+C is considered "best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C. 68 e~n-y'a: 5D-2 1 5 os t'E Es'ti m'a'te's 'a'n'd S o'u'r c'e's 'o'f 'Fin'a'n'cin'g 'u n'de r ""Be s't' 'Pr'a c'ti'ce"" 'P'olic'i'es 'a'n'd 'A'lt'e'rn'a'tiv'e ..R'e s o'u r c'e Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDOG2015 DomESTIC RESOURCES FINANCING EaSt Resourceeas ____ Item Peniod scenaniaI Mobilized Recurrenit Capital Total Recurrent Capital Total Recurrentt Capital Total Cl 5,517 6,436 0 :6,436 5,517 0 5,517 919 : 919 Cumulstive, C2 4,734 6,436 0 46 4,734 0 4,734 1,702 0 1,702 j ~. 2001 2015 _____ ~~~C3 5,959 64 6 0 636 5 9 0 5,959~ 477 0 1477 Cl 368 429 0 429 368 0 368 61 0 61 Annual C2 316 49 0 29 36 0 16 113 0 113 ______ __________ ~C3 397 429 0 429 397 0 397 32 0 32 cl 1 21 0 0 21 21 Annual C2 21 21 0 0 21 21 ___ _____ ~~~ ~~C3 __ 2 _ _ 21 0 0 21 21 cl 450 0 450 368 0 368 82 0 82 o E Annual C2 450 0 450 316 0 316 134 0 134 C3 450 0 450 397 0 397 53 0 53 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. a a (B TECHNICAL ANNEXES Table C.69 ................................................................................ Lesotho: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY . A: QUALITY MEASURES MEASURES C: FINANCING MEASURES I .. ' ~ . : 4 * Average Annual Government Revenues, Primary Private . . Spending on Teacher Salary (as Education Enrollments Annual . Pupils Per Inputs Other multiple of per Average . As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers5 capita GOP) CRepetitionRate . DP Education Spending* total) . ap, Start quo 45 29.9% 6.6 18.3% 35.9%0 . 22.2% . 40.2% 0.0% 5 A only 40 33.3% 6.6 12 A + B 40 33.3% , 5.2 10.0% - 2 'Best Cl 40 33.3% 5.2 10.0% 18.0% 22.2% 50.0% 10.0% 7 prarolce : C2 * 18.0% 20.0% 9 A + B + C3 . 35.9% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement B for efficiency improvement, and three alternative resource mobilization scenarios (C1 C2 and C3). The combination of scenarios A+B+C is considered best practice. b. As a share of primary education recurrent spending. c. Current revenues excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.70 Lesotho: MDG-Z015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCINO cost Resources: Item Penod Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 565 659 15 674 565 0 565 94 15 108 Cumulativ- C2 534 659 t1 674 534 0 5 1 y - 2001 2015. c _i>_. _5_ . C3 763 659 I 15 674 659 15 , 674 0 - 0 0 Cl 38 44 1 45 38 0 38 6 I 7 : Annual. C2 36 44 I 45 36 0 36 8 I 9 .____ __________ C3 51 44 I 45 44 I 45 0 0 0 Ci 2 2 0 0 2 2 a H O Annual C2 2 2 0 2 2 ____ * ____ _ . C3 _ 2 2 2 2 0 2 Cl 45 1 46 38 0 38 8 I 9 ; E AnAnual C2 45 I 46 36 0 36 10 I t1 C3 45 I 46 45 I 46 0 0 0 Note: Best practice policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 217 Table C.71 Madagascar: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES * i ,i Average Annual Government Revenues. Primayi Private ; : Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other i multiple of per Average As % ot % for Recurrent (as % of Financing Policy Scenaria . Teacher thanTeachers5 capita GDP) Repetition Rate Gi P Educauon Spendineg total) Gap' Statsw quo 54 42.4% 3.3 33.0% 10.6% 18.8% i54,7% 22.0% 23 A only 40 42.4% 3.5 50 A + B 40 33.3% 3.5 10.0% 36 'Best . Cl 40 33.3% 3.5 10.0% 14.0% 20.0% 42.0% 10.0% 33 praccicet : C2 14.0% 20.0% 33 A +B+ C3 14.0% 20.0% ; 33 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.72 Madagascar: MDGN-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL . Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING cost Resources -. - : - _ Item Period Scenarlo Mobilized Recurrent Capntal Total Recurrent CapRtal Total Recurrent CapRtal Total 8 i j Cl 874 1,053 j314 1,367 874 , 0 874 180 i 314 494 Cumulatlve, C2 874 1,053 314 1,367 874 0 874 180 314 494 a k. 2001 2015 C3 874 1,053 314 1,367 874 0 874 180 314 494 Cl 58 70 21 91 58 0 58 12 21 33 = Annual C2 58 70 21 91 58 0 58 12 21 33 _____ji __i_ C3 58 70 21 91 58 0 58 12 21 33 Cl 0 0 0 0 0 0"" 8 j Annual.. C2 0 0 0 0 0 0 E __3 ___ C3 0 0 0 Cl 70 21 91 58 0 58 12 21 33 Annual C2 70 21 91 58 0 58 12 21 33 C3 70 21 91 58 0 58 12 21 33 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. age TECHNICAL ANNEXES Table C.73 Malawi: MD6-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private . Spending on Teacher Salary (as Educatlon Enrollments Annual Pupils Per Inputs Other multiple ot per Average As % of % for Recurrent (as % of Flnancing Policy Scenarol Teacher than Teachers' capita GDP) Repetition Rate GDP Education Spending' total) Gap, Statnsw qu 53 14.0% 4.0 14.7% 18.1% 19.8% 49.2% 2.0% 2 A only 40 33.3% 4.0 24 A + B 40 33.3% 3.7 10.0% 14 "Besr Cl 40 33.3% 3.7 10.0% I 14.0% 20.0% 50.0% 10.0% 19 pructice':, C2 14.0% 20.0% 19 A + B + . C3 t X * , 18.1% 20.0% 13 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.74 Malawi: MO6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources . . . . _ Item Period Scenario Mobilized Recurrent Capital * Total Recurrent Capital Total * Recurrntt Capitai Total Cl 491 633 .136, 769 491 0 491 142 136 278 i, . CumulaslYc,. C2 491 633 136 769 491 0 491 142 136 278 ~ . 2001 2015. . . e _ _ g. . C3 - 571 633 136 769 571 0 571 62 136 198 Cl 33 42 9 51 33 0 33 9 9 19 Annual C2 33 42 9 51 33 0 33 9 9 19 ';___ i_________ C3 38 42 9 51 38 0 38 4 9 13 Ci 10 10 0 0 l0 10 O . i 8 jAnnual C2 10 10 0 0 10 10 .___ .___ _ C3 _ ._ 10 i i 10 . 0 i . 0 . 10 ili 10 Ci 52 9 62 33 0 33 20 9 29 Annuaj C2 52 9 62 33 0 33 20 9 29 C3 52 9 62 38 0 3 14 9 23 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 219 Table C.75 Mali: MDG-201 5 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES I MEASURES C: FINANCING MEASURES Average Annual Government Revenues' . Primary Private Spending on .Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher thasTeachers' capita GDP) Repetition Rate GDP Education Spending' total) Gap Suatusquo 61 31.1% 6.1 17.9% 16.8% 13.7% 42.1% 21.2% 54 Aonly 40 33.3% 6.1 91 A + B 40 33.3% 3.8 10.0% 62 'Best Cl i 40 33.3% 3.8 10.0% * 14.0% 20.0% 50.0% 10.0% 51 practice: C2 14.0% 20.0% 51 A +B+ C3 16.8% 20.0% 45 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.76 .............................. ............................................................................................ Mali: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG.201 5 DOMESTiC RESOURCES FINANCING Cost .Resources . . . Item Period Scenano Mlobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 676 1,014 427 1,442 676 0 676 339 427 766 . Cumulative, C2 676 1,014 427 1,442 676 0 676 339 427 766 L 2001 2015 c____5__; C3 762 1,014 427 . 1,442 762 0 762 252 427 679 Cl 45 68 28 96 45 0 45 23 28 51 Annual C2 45 68 28 96 45 0 45 23 28 51 Xi___ ,________ . C3 51 68 28 96 51 0 51 17 28 . 45 Cl 2 2 0 0 2 2 Annual C2 2 2 0 0 2 2 ;_____ i_____ _ C3 ;i __ ;_2_; 22 0 ; e 0 2 ,i , 2 C I 69 28 98 45 0 45 24 28 53 Q k * Annual C2 * i 69 * 28 * 98 45 0 45 24 28 53 C3 69 28 98 51 0 51 19 28 47 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 2 20 TECHNICAL ANNEXES Table C. 77 ..................................................................................... 'e"a's'"r'e's Mauritania: MDG-201 5 Financing Gap under Alternative Policy Measures .E EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per . Average As % of % for Recurrent (as% of Firnancing Policy Scenario' Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spending' total) Gap, Status quo 48 1 8.2% 5.1 16.0% 26.5% 13.7% 49.0% 1.8% 6 Aonly 40 33.3% 5.1 16 A + B 40 33.3% 4.0 10.0% 7 'Best Cl i 40 33.3% 4.0 10.0% 16.0% 20.0% 50.0% 10.0% 7 practice": C2 16.0% 20.0% 7 A + B + C3 26.5% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (C1, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.78 ................................................................................................................................ Mauritania: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources ' . . , . . Item Period Scenaro Mobilized Recurrent Capital j Total Recurrent Capital . Total Recurrent Capital * Total . . . . , , z , , : I ~ 4 4. Cu 343 406 . 48 454 343 0 343 63 48 .111 t *. Cumulative. *. C2 343 406 * 48 454 343 * 0 343 ~. 2001 2015 . e _ _ ¢* 5 * C3 466 406 48 . 454 406 .48 454 0 0 0 Ir .5 g Cl 23 27 3 30 23 0 23 4 3 7 Annua" C2 23 27 3 30 23 0 23 4 3 7 ______ __________ i C3 31 27 3 30 27 3 30 0 0 O ci 0 0 0 0 Annual. C2 0 0 0 0 0 O ; ; ~~~~~C3 i . 0 ______ ___________ C3 ________ 0 0 ~~~~ ~ ~~0 0 0 _ _ _ 0 Cl 27 3 30 23 0 23 4 3 8 Annual C2 27 3 30 23 0 23 4 3 8 C3 27 3 30 27 3 30 0 0 0 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 221 Table C.79 Mozambique: MDG-2015 Financing Gap under Alternative Policy Measures 13B EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual PGovernment Revenues Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario Teacher than Teachers' capita GDP) Repetition Rate GlP Education Spendinge total) Gap, Statu quo 54 26.1% 3.2 23.7% 11.3% 18.1% 46.4% 0.0% 27 A only 40 33.3% 3.5 75 A + B 40 33.3% 3.5 10.0% 48 'Best Cl 40 333% 3.5 10.0% 14.0% 20.0% 42.0% 10.0% 37 practice: C2 14.0%, 20.0% 37 A +B+ B C3 14.0% 20.0% 37 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.80 Mozambique: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FoR EXTERNAL Domestic COST OF MIIG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources ! Item Period Scenario *MobDized Recurrent Capital Total Returrent Capial Total . Recurrent Capital Total Cl 1,206 1,552 209 1,761 1,206 0 j 1,206 346 209 . 555 Cumulative, * C2 1,206 1,552 209 1,761 1,206 0 1,206 346 209 555 c ¢ , 2001-2015 ; C3 1,206 1,552 209 . 1,761 ; 1,206 * 0 , 1,206 346 209 555 -i A.nual Cl 80 103 14 117 80 0 80 * 23 14 37 a . Annual C2 80 103 14 117 80 0 80 23 14 37 ____,___ ; C3 80 103 14 117 80 0 80 23 14 37 C1 9 9 0 0 9 9 - 8'. Annual C2 9 9 0 0 9 9 _ _, __ ; C3 *; 9 9 0 0 9 _ ; 9 Cl 112 14 126 80 0 80 32 14 46 Annual C2 112 14 126 80 0 80 32 14 46 C3 112 14 126 80 0 80 32 14 46 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 222 TECHNICAL ANNEXES Table C.81 ............................................................................. Niger: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per Average As % af % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachersb capita GDP) Ropetition Rate GDP Education Spendinge total) Gap Statusquo 37 25.9% 9.6 13.0% 9.1% 31.5% 62.0% 4.0% 135 Aonly 37 33.3% 9.6 146 A + B 40 33.3% 4.3 10.0% 48 'Best Cl 40 33.3% 4.3 10.0% 14.0% 26.0% 50.0% 10.0% 46 practice": C2 14.0% 20.0% 53 A +B+ C3 14.0% 20.0% 53 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.82 ..................................................................................................................................... Niger: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING Cost Resources : * :! Item Period Scenardo Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 796 1,078 403 1,481 796 0 796 282 403 685 Cumulative. C2 684 1078 403 1,481 684 0 684 394 403 797 Iiu 2001 2015. .. C _ C3 684 1,078 403 1,481 684 0 684 394 403 797 Cl 53 72 27 99 53 0 53 19 27 46 = i Annual . C2 46 72 27 99 46 0 46 26 27 53 ____i___;_ C3 46 72 27 i 99 46 0 46 26 27 53 Ci I I 0 0 1 aO.O Annual C2 tI 0 O I ___ ____ * C3 15--_; . I * I 0 0 I iI Cl 73 27 100 53 0 53 20 27 47 Amual C2 73 27 100 46 0 46 28 27 54 C3 73 27 100 46 0 46 28 27 54 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 223 Table C.83 ... ............................................................................ Nigeria: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple ot per . Average As % at % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GDP) Repetitien Rate GOP Education Spending total) Gap' strat.quo 39 9.1% 4.9 1.0% 46.1% 9.9% 41.0% 1.0% 352 A only 39 33.3% 4.9 654 A + B 40 33.3% 4.1 1.0% 323 'Best Cl 40 33.3% 4.1 1.0% 18.0% 20.0% 50.0% 10.0% 214 practice": C2 18.0% 20.0% 214 A + B + C3 46.1% 20.0% 0 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.84 ........................................................................................................................................ Nigeria: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL j Onmestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources . ._._.__ Item Period Scenario i lobilized Recurrent CapRtal Total Recurrent Capital Total Recurrent Capial Total C T 19,123 20,502 1,831 22,333 19,123 0 19,123 1,380 1,831 3,210 t i Cumulativce. C2 19,123 20,502 1,831 22,333 19,123 0 19,123 1,380 1,831 3,210 2001 2015 c , ________ 200 j C3 31,966 20,502 1,831 22,333 20,502 1,831 22,333 0 0 0 i Annual Cl 1,275 1,367 122 1,489 1,275 0 1,275 92 122 214 C2 1,275 1,367 122 1,489 1,275 0 1,275 92 122 214 C3 2,131 1,367 122 1,489 1,367 122 1,489 0 0 j 0 Annual Cl 58 58 0 0 58 58 : ^ j i C2 i i 58 * 58 0 0 58 58 __e_8; _ . C3 _____ 58 58 58 j i 58 0 58 j Annual j Cl j 1,425 122 1,547 1,275 0 1275 150 122 272 0O Ec j j C2 1,425 122 1,547 1,275 0 1,275 150 122 272 C3 1,425 122 1,547 1,425 122 1,547 0 0 0 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 22d§ TECHNICAL ANNEXES Table C.85 ................................................................................ Rwanda: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES * MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spendingono Teacher Salary (as Education Enrollments Annual Pupils Per inputs Other multiple of per * Average As % oa % for Recurrent (as % of Financing Policy Scenario' Teacher thanTeachers' capita GOP) Repetition Rate GOP Education Spendingr total) Gap' Starv quo 48 8.6% 4.0 36.1% 9.8% 32.6% 44.7% 0.8% 24 A only 40 33.3% 4.0 48 A + B 40 33.3% 3.7 10.0% 23 'Best C I 40 33.3% 3.7 10.0% 14.0% 26.0% 50.0% 10.0% 16 practice: C2 14.0% 20.0% 22 A + B+ C3 14.0% 20.0% 22 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.86 ........................................................................................................................................ Rwanda: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL *omestic COST OF MDG-201 5 DOMESTIC RESOURCES FINANCING Cost Resources I * * * _ - Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capial Total Cl 682 756 .166 922 682 0 682 75 166 240 c 2C0umnlu2aOvce C2 591 756 .166 922 591 0 591 165 166 331 2001 201592 51 o __. _ i C3 591 756 *166_ 922 591 * 0 ', 591 165 166 331 .0 o ' Annual Cl 45 50 I I 61 45 0 45 . 11 16 C2 39 50 I i 61 39 0 39 1 1 I I 22 l ___i __ i_ C3 39 50 I I 61 39 0 39 I I I I 22 , _ .~ Annual Cl 7 7 0 0 7 7 C2 7 7 0 0 7 7 . __________ C3 . ____ _ i 7 . ' 7 0 0 7 , . 7 Annual Cl 57 I I 68 45 0 45 12 11 23 +0 E , , C2 57 I I 68 39 0 39 18 11 29 C3 57 I I 68 39 0 39 18 I I 29 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 225 Table C.87 Senegal: MDG-201 5 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues Primary Private Spending on Teacher Salary (as Educatlon Enrollments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GIP) Repetition Rate GDP Education Spendingd total) G6ap- Srarus quo 55 36.6% 4.9 13.6% 18.1% 18.6% 43.9% 10.7% 48 A only 40 36.6% 4.9 85 A + B 40 33.3% 3.7 10.0% 43 "Best Cl 40 333% 3.7 10.0% 16.0% 20.0% 500% 10.0% 38 practice: C2 16.0% 20.0% 38 A +B+ C3 18.1% 20.0% 30 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.88 ........................................................................................................................................ Senegal: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios 275 (millions of 2000 U.S. dollars) FINANCING SOURcEs GAP FOR EXTERNAL Domestic COST OF MOG-201 5 DoMESTIC RESOURCES FINANCING Cost Resources . . . Item Period Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total 8 , * ClI * 1,543 ' 1,830 i 282:. 2,112 1,543 * 0 1,543 288 . 282:569 Cumula[ive, C2 1,543 1,830 . 282 * 2,112 1,543 , 0 ° 1,543 . 288 282 . 569 c C3 1658 1830 ' 282 ' 2112 1,658 * 0 . 1,658 172 282 . 454 o t. . . ' I. ' , _ _ - ; Cl 103 122 19 141 103 0 103 19 19 38 Annual C2 103 122 19 * 141 103 0 103 19 19 38 . C3 111 122 19 141 111 0 111 I I 19 30 Cl 2 2 0 0 2 2 8 = S' Anlnual * C2 2 2 0 0 2 2 C3 2 2 0 0 2 2 Cl 124 19 142 103 0 103 21 19 40 +0 Annual C2 124 19 142 103 0 103 21 19 40 C3 124 19 142 111 o 111 13 19 32 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 226 TECHNICAL ANNEXES Table C.89 Sier'r"a Leon'e': MD"G'-20"15F Fin"an"ci'n'g Gap u"n'de"r Alt"e'r'nati'v"e P'oli'cy ..M'e"a'sur'es B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenuene Pnmary Private Spending on Teacher Salary (as Education Enrolimenta Annual Pupils Per Inputs Other imultiple of per Average As % of % fer Recurrent (as % of Financing Policy Scenaria' Teacher than Teachers' capIta GOIP) ReelnRt LP Euatn Sedn4 ttl a Statusquo 40 33.1% 4.3 9.3% 11.4% 30.4% 51.3% 0.0% 13 A only 40 33.3% 4.3 13 *A+B 40 33.3% 3.8 9%8 'Best Cl 40 33.3% 3.8 9.3% 14.0% 26.0% 50.0% 10.0% 8 practice: C2 14.0% 20.0% 10 A+ + C3 14.0% 20.0% 10 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3. The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.90 Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Donmestic COST OF MDG-2015 DomESTIC RESOURCES FINANCING Cost Reosnurcens .* - Item Period Scenario Moblized Recurrent Capital Total Recuffren capital Total Recurrernt Capital Total 8 Ci 261 288 88 376 261 0 261 27 88:115 Curnulati~e, C 227 288 08 376 227 0 2271 62 88 150 2001 2015 .. __ __ C3 *227 288 88 376 227 i 0 227 62 88 150 Cl 17 19 6 25 17 0 17 2 6 8 Annual C2 15 19 6 25 15 0 15 4 6 10 __________ ~C3 15 19 6 25 15 0 15 4 6 10 ' g Annua Cl 2 2 0 0 2 2 C2 ~ ~ ~ ~ ~~~2 2 0 0 22 ______ __________ ~C3 __ _ _ _ 2 2 0 0 2 2 ci 21 6 27 17 0 17 4 6 9 E. Annual C2 21 6 27 15 0 15 6 6 12 C3 21 6 27 IS 0 15 6 6 12 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 227 Table C.91 Sudan: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Goveruiment Revenues, Primary Private Spending on Teacher Salary (aSn EducatUon EnrOllments Annual Pupils Per Inputs Other multiple of per Average As % of % fOr Recurrent (as % of Flnancing Policy Scenario, Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spendingd total) Gap- Statusquo 28 22.5% 2.2 1.2% 11.1% 16.2% 50.5% 0.0% 131 A only 28 33.3% 3.5 301 A + B 40 33,3% 3.5 1.2% 153 "Bst ; Cl 40 33,3% 3.5 1.2% 14.0% 20.0% 50.0% 10.0% 105 PraC;Cice ' C2 14.0% 20.0% 105 A + B + C3 14.0% 20.0% 105 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.92 Sudan: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DOMESTIC RrSOURCES FINANCING COSt, ; e0rCS: : E Resources: Item PeriOd Scenario MubiliZed Recurrent Capital; TOeal ReCUrrent Captalo TOtal ReCUrrent CaPial; Tntal Cl 3,078 4,349 297 4.646 3,078 ; 0 3,078 1,271 297 1,568 . CUmUIaUiVe, * C2 3,078 4,349 297 4,646 3,078 0 3,078 1,271 297 1,568 2001 205 ;C3 3,078 4,349 297 4,646 3,078 0 3,078 1,271 297 1,568 '.-U . * Cl 205 290 20 . 310 . 205 0 * 205 85 20 105 ' . Annual ,; C2 .; 205 . 290 , 20 ; 310 205 0 205 85 20 105 ______ ______;;__ C3 205 290 20 310 205 0 205 85 20 105 Cl 1 1 ; 0 I 1 - XX 8; Annual C2 I 0 0 I I ____ ;______ C3 ; ; I 0 0 1 1 Cl 291 20 311 205 0 205 86 20 106 Annual C2 291 20 311 205 0 205 86 20 106 C3 291 20 311 205 0 205 86 20 106 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 228 TECHNICAL ANNEXES Table C.93 ................................................................................. Tanzania: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY . A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Private Spending on .Teacher Salary (as Education Enrolments Annual Pupils Per Inputs Other multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachersb capita GOP) Repetition Rate GOP Education Spending' total) G.ap Statuusquo 40 11.2% 3.6 3.2% 10.9% 16.4% 63.0% 0.0% 87 A only 40 33.3% 3.6 128 A + B 40 33.3% 3.5 3.2% 102 "Best Cl 40 33.3% 3.5 3.2% 14.0% 20.0% 50.0% 10.0% 80 practice": C2 14.0% 20.0% 80 A + B + C3 A 14.0% 20.0% 80 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.94 ........... ki........................................................................................................................ Tanzania: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domostic COST OF MDG-201 5 DOMESTIc RESOURCES FINANCING cost Resources . , ._i Item Period Scenario Mobilized Recurrent C Capital Total Recurrent Capital Total Recurrent t Capital * Total Cl 2,454 3,234 . 413 3,646 2,454 5 0 2,454 780 413 1,193 : Cumulative, . C2 2,454 3,234 . 413 3,646 2,454 * 0 2,454 780 * 413 1,193 ''. 2001 2015 . . e _____. _ . C3 2,454 3,234 413 3,646 2,454 ' 0 2,454 780 . 413 ' 1,193 8 Annu l Cs 164 216 28 243 * 164 0 164 52 28 80 = Arnua C2 164 216 28 243 164 0 164 52 28 80 Si ; ;C3 164 216 28 243 ; 164 0 164 52 28 80 Cl 22 22 0 0 22 22 o Annual C2 22 22 I 0 0 22 22 _ _ 8 * * C3 22 22 . 0 0 22 22 Cl 238 28 265 * 164 I 0 164 74 28 102 -6 E, Annual C2 238 28 265 164 0 164 74 28 102 C3 238 28 265 . 164 0 164 74 28 102 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 229 Table C95 Togo: MDG-2015 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual Government Revenues' Primary Priwate Spending on Teacher Salary (as * Education Enrollments Annual Pupils Per Inputs Other multiple ot per Average As % oi % tor Recurrent (as % of Financing Policy Scenario Teacher than Teachers' capita GOP) RepettiUon Rate GOP Education Spending' total) Gap' status quo 45 25.2% 4.5 27.0% 14.9% 25.6% 48.3% 35.6% 3 A only 40 33.3% 4.5 10 A + B 40 33.3% 3.9 10.0% 12 "Best Cli 40 33.3% 3.9 10.0% 14.0% 25.6% 50.0% 10.0% 11 pracicac: *C2 * ' i , 14.0% 20.0% 16 A + B+ C3 14.9% 20.0% 15 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement B for efficiency improvement and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.96 Togo: MD6-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL jDomstic COST OF MDG-2015 DOMESTIC RESOURCES FINANCING cost Resources: . : Item Scenario Mobilized Recurrent Capital Total Recurrent Capital Total Recurrent Capital Total Cl 531 589 j107 696 531 0 531 58 107 165 k Cumulative, C2 460 589 107 696 460 0 460 129 107 i 236 ~ . 2001 2015 c _j2121 , C3 478 589 107 696 478 0 478 112 107 219 CI 35 39 7 46 35 0 35 4 7 11 Anual C2 31 39 7 46 31 0 31 9 7 16 ____i___i__ C3 32 39 7 46 32 0 32 7 7 15 ci 4 4 0 0 4 4 a O i Annual C2 4 4 0 0 4 4 _______ C3 4 4 0 0 4 4 Cli 43 7 51 35 0 35 8 7 15 00- E g Annual j C2 43 7 51 31 0 31 13 7 20 C3 43 7 51 32 0 32 12 7 19 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 230 TECHNICAL ANNEXES Table C.97 Uganda: MDG-201 5 Financing Gap under Alternative Policy Measures B: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES Average Annual oenetRvne' Pninary Private Spending on Teacher Salary (as iEducation Enrollments Annuial Pupils Per Inpuits lOtther multiple of per Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) Repetitino Rate GDP Education Spenidingu total) Gap' Status quo 41 26.2% 2.9 9.8% 10.8% 30.1% 53.2% 2.0% 14 *Aonly 40 33.3% 3.5 ___63 *A+B 40 33.3% 3.5 9.8% 48 'Best C I 40 33.3% 3.5 9.8% 14.0% 26.0% 50.0% 10.0% 41 prcie: C2 14.0% 20.0% 63 A+B+ C3 ~~~~~ ~~~~~~~~~~~~~~~~~~~~14.0% 20.0% 63 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice. b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specif ic policy scenario and the total resources for primary education mobilized domestically. Table C.98 Uganda: MDG-2D015 Cost15 .. Co Estimatesi andat Sourcesn of Financinge under "Best Practice" Policies and Alternative Resourcee"' P'ol'ic''e'' a' n'' '' 't'' r'na'tive .. 'e's''o'' ' ''e Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MDG-2015 DoMESTIC RESOURCES FINAN~CING cost Resourtes - -I Item Period Scenario Mobilized Recurrent capital Total Recurrent Capital Total iRecurrent Capital iTatal Cl 2,523 2,930 210 3,040 2,523 I 0 2,523 407 :210 607 ve, C2 2,189 2,930 210 3,140 2,189 0 2,189 741 210 951 ~ , 2001 2015. .. . _________ ~C3 2,189 2,930 210 3,140 208 I 01,89 741 210 95 '~Annual Cl 168 195 14 209 168 0 168 27 14 41 C2 146 195 14 209 146 0 046 49 14 63 ______ ___________ C3 146 195 14 209 146 0 146 4 4 6 Annual Ci 24 4 00 2 24 C ~~~~~~~~C2 24 24 0 0 24 24 ___ _______ ~~~C3 24 24 0 0 24 24 Annual Ci 219 14 233 168 0 168 5 1 14 65 C2 219 14 233 146 0 146 73 14 87 C3 219 14 233 146 0 146 73 14 87 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. ANNEX C * COUNTRY SIMULATION RESULTS 231 Table C.99 Zambia: MDG-2015 Financing Gap under Alternative Policy Measures ,: EFFICIENCY A: QUALITY MEASURES MEASURES C: FINANCING MEASURES I. . - Average Annual Government Revenues' Primary Private Spendingon Teacher Salary (as Education Enrollments Annual Pupils Per Inputs Other multiple of per . Average As % of % for Recurrent (as % of Financing Policy Scenario' Teacher than Teachers' capita GOP) Repetition Rate GOP Education Spendinge total) Gap' Status quo 50 21.7% 2.7 6.2% 18.8% 12.3% 43.2% 1.6% 10 A only 40 33.3%0 ' 3.5 , 40 A + B 40 33.3% 3.5 6.2% . . i34 "Best Cl 40 33.3% 3.5 6.2% * 14.0% 20.0% 50.0% 1 10.0% 25 practice": *C2 @ ' * , 14.0% 20.0% 25 A + B + . C '. ' 18.8% 20.0% 17 Note: Shaded cells denote no change from values directly above. a. Policy scenarios are: A for quality improvement, B for efficiency improvement, and three alternative resource mobilization scenarios (Cl, C2, and C3). The combination of scenarios A+B+C is considered "best practice." b. As a share of primary education recurrent spending. c. Current revenues, excluding grants. d. As a share of total education recurrent spending. e. In millions of 2000 U.S. dollars. Calculated as the difference between the total cost of service delivery under the specific policy scenario and the total resources for primary education mobilized domestically. Table C.100 ........................................................................................................................................ Zambia: MDG-2015 Cost Estimates and Sources of Financing under "Best Practice" Policies and Alternative Resource Mobilization Scenarios (millions of 2000 U.S. dollars) FINANCING SOURCES GAP FOR EXTERNAL Domestic COST OF MIOG-201 5 DOMESTIC RESOURCES FINANCING Cast Resources - : . - Item Period Scenario Mobillzed Recurrent Capital . Total Recurrent Capital Total Recurrent Capital Total Cl 619 866 .131 996 619 0 : 619 247 131 378 Cumulative, C2 619 866 131 996 619 0 619 247 131 378 2001 2015 __u_. _5 C3 747 866 131 996 747 0 747 119 131 249 Cl 41 58 9 66 41 0 41 16 9 25 a . Annual *. C2 41 58 9 66 41 0 41 16 9 25 ______ *_________ C3 50 58 9 66 50 0 50 8 9 17 Cl 15 15 0 0 15 15 __ t4 Anus) C2 15 15 0 0 15 15 ____ i______ C3 i ;_ 15 * 15 i 0 15 15 Cl 72 9 81 41 0 41 31 9 40 a0 E ' Annual C2 72 9 81 41 0 41 31 9 40 C3 72 9 81 50 0 50 23 9 31 Note: "Best practice" policies refer to the combination of scenarios A+B+C. Shaded cells denote no change from values directly above. 232 TECHNICAL ANNEXES AN N EX Aid for Primary Education Table D. I Development Assistance Committee and Multilateral Official Commitments for Education and Basic Education, 1997-2000 (millions of current U.S. dollars) Annual Donor 1997 1998 1999 2000 Average IBRD, total education 739.3 I 1,927.8 799.4 215.3 920.5 o/w basic education 163.2 612.7 363.5 76.6 304.0 IDA, total education 255.1 1,201.5 534.8 468.7 615.0 o/w basic education * 96.3 1,018.8 283.4 298.7 424.3 World Bank, total education 994.4 3,129.3 1,334.2 684.0 1,535.5 o/w basic education 259.5 1,631.5 646.9 375.3 . 728.3 Other multilateral, total education 486.1 1,274.7 773.7 1,335.5 967.5 o/w basic education * 184.7 181.0 184.1 199.7 187.4 All MDBs, total education 1,480.5 4,404.0 2,107.9 2,019.5 2,503.0 o/w basic education 444.2 1,812.5 831.0 575.0 915.7 G-7, total education 3,838.2 3,568.5 4,093.9 2,639.7 3,535.1 o/w basic education 356.5 218.0 345.9 432.5 338.2 EU members, total education 3,278.3 3,229.1 3,066.0 2,155.0 2,932.1 o/w basic education 241.5 306.9 293.8 379.1 305.3 All DAC countries, total education 4,804.3 4,459.2 5,014.3 3,541.7 4,454.9 o/w basic education 534.0 434.3 599.7 684.4 563.1 Grand total, all education donors 6,284.8 8,863.2 7,122.2 5,561.2 6,957.8 o/w basic education 978.3 2,246.3 1,430.6 1,259.4 1,478.7 Sources: World Bank Business Warehouse and OECD DAC Database. 233 Table D.2 Official Commitments for Basic Education as a Percentage of Total Education Commitments, 1997-2000 40or > r- Donor -;-- _9 -499 - 8 t-1993ib0 0O IBRD 22 32 45 36 34 IDA 38 85 53 64 60 World Bank 26 52 48 55 45 Other multilateral 38 14 24 15 23 All MDBs 30 41 39 28 35 G-7 9 6 8 16 10 EU members 7 10 10 18 11 ARlDACcountries 11 10 12 19 13 Total, all sources 16 25 20 23 21 234 TECHNICAL ANNEXES Table D.3 ......................................................................................................... Multilateral Official Commitments for Education and Basic Education, by Donor and Region, 1997-2000 (millions of current U.S. dollars) * . * . Annual Donor 1997 1998 . 1999 2000 Average World Bank, total education 994.4 3,129.3 .1,334.2 684.0 1,535.5 o/w basic education * 259.5 1,631.5 646.9 375.3 728.3 AFR education 75.1 372.3 194.1 159.7 200.3 o/wAFR basic education * 21.3 218.3 . 131.0 58.8 107.4 LCR education 61.5 1,199.9 393.6 77.5 433.1 o/w LCR basic education . 33.0 387.7 243.2 72.5 184.1 EAP education . 645.0 103.5 557.2 5.0 327.7 o/w EAP basic education 113.4 138.4 5.0 85.6 ECA education 137.8 592.4 36.1 22.6 197.2 o/w ECA basic education 16.8 321.0 36.1 124.6 SAR education 0.0 718.2 98.2 200.0 254.1 o/w SAR basic education 0.0 . 704.5 * 98.2 * 182.4 246.3 MNA education . 75.0 143.0 50.0 219.2 121.8 o/w MNA basic education * 75.0 . .. 56.6 65.8 Other multilateral, total education 486.1 1,274.7 773.7 1,335.5 967.5 o/w basic education * 184.7 181.0 : 184.1 199.7 . 187.4 AFR education * 161.2 868.9 309.5 1,041.3 595.2 o/wAFR basic education 33.3 28.2 110.7 116.2 72.1 LCR education . 150.0 165.9 89.8 64.1 117.5 o/w LCR basic education . 110.0 * 72.6 73.6 11.1 66.8 EAP education * 150.4 203.5 144.7 90.3 147.2 o/w EAP basic education . 85.0 29.0 36.1 6.2 . 39.1 ECA education * 21.0 0.0 0.1 85.7 26.7 o/w ECA basic education . . .. . 0.1 0.1 SAR education * 46.6 45.7 38.0 54.0 46.1 o/w SAR basic education . 20.2 17.9 12.8 * 10.0 15.2 MNA education . . 50.6 321.5 186.0 o/w MNA basic education . .. 30.4 56.1 43.3 All MDBs, total education 1,480.5 4,404.0 2,107.9 2,019.5 2,503.0 o/w basic education * 444.2 1,812.5 831.0 575.0 915.7 AFR education . 236.3 1,241.2 503.6 1,201.0 795.5 o/wAFRbasiceducation * 54.6 246.5 241.7 175.0 179.4 LCR education 211.5 1,365.8 . 483.4 141.6. 550.6 o/w LCR basic education * 143.0 460.3 316.8 83.6 250.9 EAP education 795.4 307.0 701.9 95.3 474.9 o/w EAP basic education . 198.4 29.0 174.5 11.2 103.3 ECA education 158.8 592.4 36.2 108.3 223.9 o/w ECA basic education 16.8 321.0 36.1 0.1 93.5 SAR education 46.6 763.9 136.2 254.0 300.2 o/w SAR basic education . 20.2 * 722.4 111.0 192.4 261.5 MNA education 75.0 143.0 100.6 540.7 214.8 o/w MNA basic education 75.0 30.4 112.7 72 7 Negligible. Sources: World Bank Business Warehouse and OECD DAC Database. ANNEX D * AID FOR PRIMARY EDUCATION 235 Table D.4 ........................... ...................................................................................... Bilateral Official Commitments for Education and Basic Education, by Donor and Region, 1997-2000 Annual Donor 1997 1998 1999 2000 Average All DAC countries, total education 4,804.3 4,459.2 5,014.3 3,541.7 4,454.9 o/w basic education 534.0 434.3 599.7 684.4 563.1 AFR education 1,781.8 2,328.4 1,259.2 1,405.6 1,693.7 o/wAFR basic education 210.2 422.2 326.1 378.8 334.3 LCR education 880.1 590.4 571.2 517.1 639.7 o/wLCRbasiceducation 74.5 69.0 106.9 116.3 91.7 EAP education 771.8 722.1 1,846.7 502.7 960.8 o/w EAP basic education 128.0 48.6 76.0 64.4 79.2 ECA education 139.3 47.8 253.6 230.5 167.8 o/w ECA basic education 1.7 5.4 1.9 12.7 5.4 SAR education 701.0 428.9 466.8 339.3 484.0 o/wSARbasiceducation 92.0 113.4 136.6 102.9 111.2 MNA education 530.4 341.7 616.9 546.4 508.9 o/w MNA basic education 52.2 61.2 44.7 73.8 58.0 G-7, total education 3,838.2 3,568.5 4,093.9 2,639.7 3,535.1 o/w basic education 356.5 218.0 345.9 432.5 338.2 AFR education 1,584.1 2,099.5 912.3 1,050.0 1,411.5 o/wAFRbasiceducation 154.6 339.4 224.1 270.6 247.2 LCR education 741.9 384.3 391.9 319.1 459.3 o/w LCR basic education 61.0 47.7 80.7 90.2 69.9 EAP education 584.6 413.6 1,609.5 366.5 743.6 o/w EAP basic education 122.2 39.0 60.5 25.7 61.8 ECA education 42.9 32.1 196.2 161.3 108.1 o/w ECA basic education 0.5 0.5 0.4 7.5 2.2 SAR education 355.8 360.2 395.3 247.7 339.7 o/w SAR basic education 27.9 82.7 106.5 65.0 70.5 MNA education 528.9 278.8 588.8 495.2 472.9 o/w MNA basic education 7.5 . 42.0 35.0 60.8 36.3 EU members, total education 3,278.3 3,229.1 3,066.0 2,155.0 2,932.1 o/w basic education 241.5 306.9 293.8 379.1 305.3 AFR education 1,453.0 1,851.2 1,014.9 889.0 1,302.0 o/wAFRbasiceducation 77.9 260.0 185.7 247.9 192.9 LCR education 648.6 448.0 433.3 306.7 459.2 o/w LCR basic education 48.2 49.0 53.7 39.0 47.5 EAP education 165.4 301.8 564.2 242.9 318.6 o/w EAP basic education 9.1 24.5 32.1 37.1 25.7 ECAeducation 168.1 44.5 256.2 152.1 155.2 o/w ECA basic education 0.7 5.4 1.4 5.6 3.3 SAR education 495.0 287.4 285.2 203.5 317.8 o/w SAR basic education 62.5 82.9 84.4 72.9 75.7 MNA education 348.2 296.3 512.2 360.7 379.3 o/w MNA basic education 4.7 40.9 17.2 31.1 23.5 Sources: World Bank Business Warehouse and OECD DAC Database. 236 TECHNICAL ANNEXES Table D.5 .............. ..... .......... .... ........................................................... Total Official Commitments for Education and Basic Education, by Region, 1997-2000 * * . Annual Donor 1997 1998 1999 2000 Average Education, all sources 6,284.8 8,863.2 7,122.2 5,561.2 6,957.8 o/w basic education 978.3 2,246.8 1,430.6 1,259.4 1,478.8 AFR education 2,018.1 3,243.7 1,762.8 1,840.8 2,216.4 o/wAFR basic education 264.7 668.7 567.8 468.8 492.5 LCR education 1,091.6 . 1,894.0 . 1,054.6 611.6 1,162.9 o/w LCR basic education 217.5 529.3 423.6 199.8 342.6 EAPeducation 1,567.1 952.8 2,548.6 531.6 1,400.0 o/w EAP basic education 326.4 77.6 250.5 75.6 182.5 ECA education 298.1 640.2 289.8 275.8 376.0 o/w ECA basic education 18.5 326.4 38.0 12.7 98.9 SAR education* 747.5 1,175.7 603.1 553.6 770.0 o/w SAR basic education * 112.2 835.8 247.7 295.3 372.7 MNAeducation 605.4 484.7 717.5 850.7 664.6 o/w MNA basic education 127.2 61.2 75.1 186.6 112.5 Sources: World Bank Business Warehouse and OECD DAC Database. ANNEX D * AID FOR PRIMARY EDUCATION 237 Table D.6 ............................................................................................. Official Development Assistance to Education by Region, All Sources, 1998-2000 (commitments basis, millions of current U.S. dollars) Donor 1998 AFR EAP ECA LCR MNA SAR 1999 AFR IBRD, total education 1,927.8 799.4 o/w basic education 612.7 363.5 IDA, total education 1,201.5 534.8 o/w basic education 1,018.8 283.4 World Bank, total education 3,129.3 372.3 103.5 592.4 1,199.9 143.0 718.2 1,334.2 194.1 o/w basic education 1,631.5 218.3 321.0 387.7 704.5 646.9 131.0 Other multilateral, total 1,274.7 868.9 203.5 165.9 45.7 773.7 309.5 education o/w basic education 180.9 28.2 29.0 72.6 17.9 184.0 110.6 All MDBs, total education 4,404.0 1,241.2 307.0 592.4 1,365.8 143.0 763.9 2,107.9 503.6 o/w basic education 1,812.5 246.5 29.0 321.0 460.3 722.4 830.9 241.6 G-7, total education 3,568.4 2,099.5 413.6 32.1 384.3 278.8 360.2 4,093.9 912.3 o/w basic education 217.9 339.4 38.9 0.5 47.7 42.0 82.7 345.9 224.1 EU, total education 3,229.1 1,851.2 301.8 44.5 448.0 296.3 287.4 3,065.9 1,014.9 o/w basic education 306.9 260.0 24.5 5.4 49.0 40.8 82.9 293.8 185.7 All DAC countries, 4,459.2 2,328.4 722.0 47.8 590.4 341.7 428.9 5,014.3 1,259.2 total education o/wbasiceducation 434.3 422.2 48.6 5.4 68.9 61.2 113.4 599.7 326.1 Grand total, education 8,863.2 3,243.7 952.8 640.2 1,893.9 484.7 1,175.7 7,122.2 1,762.8 o/w basic education 2,246.8 668.7 77.6 326.4 529.3 61.2 835.8 1,430.6 567.8 Note: The World Bank (IBRD and IDA) reports flows to "primary education" whereas the DAC reports flows to "basic education." Regional breakdowns could not be obtained for some multilateral sources; hence, regional totals may be underreported. Regional classifications reported by the DAC are not consistent with those reported by the World Bank. Distribution across regions subject to authors' estimates. Grand total estimates include some flows that could not be attributed to individual regions. Sources: World Bank (IBRD and IDA) figures from World Bank Business Warehouse; all other figures from OECD DAC Database. 238 TECHNICAL ANNEXES EAP ECA LCR MNA SAR 2000 AFR EAP ECA LCR MNA SAR 215.3 76.6 468.7 298.7 557.2 36.1 393.6 50.0 98.2 684.0 159.7 5.0 22.6 77.5 219.2 200.0 138.4 36.1 243.2 98.2 375.3 58.8 5.0 72.5 56.6 182.4 144.7 0.1 89.8 50.6 38.0 1,335.5 1,041.3 90.3 85.6 64.1 321.4 53.9 36.1 73.6 30.4 12.8 199.7 116.2 6.2 0.1 11.1 56.1 10.0 701.9 36.2 483.4 100.6 136.2 2,019.5 1,200.9 95.3 108.2 141.6 540.6 253.9 174.5 36.2 316.8 30.4 111.0 575.0 175.0 11.2 0.1 83.6 112.7 192.4 1,609.4 196.2 391.8 588.8 395.3 2,639.7 1,049.9 366.5 161.3 319.1 495.2 247.7 60.5 0.4 80.7 34.9 106.5 432.5 270.6 25.7 7.5 90.2 60.8 65.0 564.2 256.2 433.3 512.2 285.2 2,154.9 889.0 242.9 152.1 306.7 360.7 203.5 32.1 1.4 53.7 17.2 84.4 379.0 247.9 37.1 5.6 38.9 31.1 72.9 1,846.7 253.6 571.2 616.9 466.8 3,541.6 1,405.6 502.7 230.5 517.1 546.4 339.3 76.0 1.9 106.8 44.7 136.6 684.4 378.8 64.4 12.7 116.3 73.8 102.9 2,548.6 289.8 1054.6 717.5 603.1 5,561.2 1,840.8 531.6 275.7 611.6 850.7 553.6 250.5 38.0 423.6 75.0 247.6 1,259.4 468.8 75.6 12.7 199.8 186.6 295.3 ANNEX D * AID FOR PRIMARY EDUCATION 239 Table D. 7 ............................................................I..........................I........................ Official Development Assistance to Basic Education in Sub-Saharan Africa, by Donor, 1998-2000 (commitment basis, millions of current U.S. dollars) Donor 1998 1999 2000 International Development Association Total education 1,201.5 534.8 468.7 Africa education 372.3 194.1 159.7 Africa basic education 218.3 131.0 58.8 Other multilateral development banks Total education 1,274.7 773.7 1,335.5 Africa education 868.9 309.5 1,041.3 Africa basic education 28.2 110.6 116.2 Development Assistance Committee countries Total education 4,459.2 5,014.3 3,541.6 Africa education * 2,328.4 1,259.2 1,405.6 Africa basic education 422.2 418.4 378.8 All donors Total education 6,935.4 6,322.8 5,345.8 Africa education 3,569.6 1,762.8 2,606.6 Africa basic education 668.7 660.0 553.8 Note: The World Bank (IBRD and IDA) reports flows to "primary education," whereas the DAC reports flows to "basic education." DAC and African Development Bank regional classifications cover continental Africa and not Sub-Saharan Africa, as reported by the World Bank. Therefore the regional totals are approximate. Sources: World Bank Business Warehouse; OECD DAC Database. 240 TECHNICAL ANNEXES Table D.8 ................................................................................................................... IDA Disbursements for Primary Education, by Expenditure Type and Region, Fiscal Years 1999-2002 (millions of current U.S. dollars) FY99 FYOO FYOI FY02 REGION Capital Recurrent Total Capital Recurrent Total Capital Recurrent Total Capital Cecurrent Total Africa 42.2 115.7 157.9 47.2 59.5 106.7 35.9 52.5 88.4 40.0 75.8 115.8 East Asia and the Pacific 3.8 1.6 5.4 2.1 0.9 3.0 0.5 0.3 0.8 0.6 0.4 1.0 Europe and Central Asia 1.2 2.7 3.9 2.3 3.5 5.8 2.4 4.7 7.1 1.8 4.7 6.5 Latin America and the 3.0 6.0 9.0 4.8 18.8 23.6 2.8 11.3 14.1 2.6 7.6 10.2 Caribbean Middle East and North 6.4 2.7 9.1 5.4 2.3 7.7 4.0 1.6 5.6 2.9 1.4 4.3 Africa South Asia 66.5 119.9 186.4 43.1 99.0 142.1 48.8 138.4 187.2 27.3 74.9 102.2 Total 123.0 248.6 371.7 104.9 184.0 288.9 94.4 208.8 303.2 75.3 164.7 240.0 Table D.9 ...................................................................................................................................... Proportion of IDA Primary Education Disbursements for Capital and Recurrent Expenditures, by Region, Fiscal Years 1999-2002 (percent) FY99 FY0 oo FYOI . FY02 Region Capital Recurrent Total Capital Recurrent Total Capital Cecurrent Total Capital Recurrent Total Africa *27 73 100 44 56 100 41 59 100 35 65 100 East Asia and the Pacific 70 30 100 70 30 100 62 38 100 60 40 100 Europe and Central Asia, 31 69 100 . 40 60 100 . 34 66 100 * 28 72 100 Latin America and the 33 67 100 20 80 100 20 80 100 25 75 100 Caribbean Middle East and North 27 73 100 70 30 100 : 71 29 100 * 67 33 100 Africa South Asia 36 64 100 30 70 100 26 74 100 27 73 100 Total :37 63 100 46 54 100 42 58 100 40 60 100 ANNEX 0 * AID FOR PRIMARY EDUCATION 241 0~~~~1 0 ~ ~ ~ 4~~ ~~~ '' (I 9 0 *~" If! fl. lv<(tlj *)r 'LI IJ[ asorrv ~, L..I vgal All PIND~~~~~~~~~~~~~~~~~ "OZAU~~~~~~~~~i~o t. hiWd~~'red and, eighynine'countries aveiu One hui`ri have committed themselves'to the Millennium .,,'', Development G soal,(MDGs),',aimed at eradicating extreme poverty and improving the ., , welfare of,people,,by' the,ye;ar 2915. Few..global goals have been as consistently and ....deeply'supported-as the second MDG,;.wfich states that, by 2015,."children.everywhere, boys and girls alike, will-be able to complete a full coLirse of primary schooling." -l Achievement,of this goal is crucial as.education is one of the most powerful instruments known for'reducing poverty andinequality and for lyitig the foundation for sustained ,, economic growth,, effective institutions, and sound governance. -. Achieving' Universal Primary Education. by 2015: A Chance for Every Child assesses . 'WIhether universal'primary ',du`acion can be achieved by 2015.'.The study focuses on. _;-A c £_ __ and homne to .about ', 75 percent of the .childreni oout- of schol globall.y. By an'alyzinrg education policies ;- a,nd financing: patterns.in relatively high,performi,ng countries, the study identifies 'a iw t . r,te.w, policy, a,,nd$!nancing.framevw.orkr fkr, faster global piogress,in prinary education. The auih6ors! use a simulatio, n `model to show how adoption of this framework could accelerate.progress in low-i'icomne countrines currendy at,r;isk'of not reachin,g the-educa-:... non MDG' .- : - \ , -'..- *., 'The study makes4it clear that.worldwide attainment of universal primnary education by ~2015-will-necessitate an -eVen 'stronger combination of political ,will, deep' and. 2015 will ger combina~~~~~Ctio ofianilefot sustained.'reform,ifaster dissemination of best pracices,'and intensified financial effort' g * lR""|Rgriq"warte!We The book provides practical tools and 'guidance on I the-policies and .strategies'that 'can help lows-income countries improve education , L _system performance-and,on the corresponding actions by international donors that.. : , containing a ."hands-on" versi6n' of the simulation-model dev em -. ' and all of the background data used. This book will be of grea - - ai policymakers, researchers, an'd others interested in international r r4n. ,;, - ~~~~~~~~~~~rs..; '. e ~,, --X''rs. inf terested- ,, S .; *- ' W'. "' . ' . ' .-' . ' I .U -- 15 34 5 9 780821 1353455