::: 0 i ' A - - . ' wi- A ^ s \g- -- - - - o , g-fW -le ~ ~ ~ ~ ~ ~ ~ ~ ~ - a 's3 . I , r J - i I ; :. "1~~~~~~~~~~~~~~~~~~~i : r ~~~~~~~I qBlF_Ja *~~~~~~~~~~~~~~I *s]H .f2............ Z,X '"<'' ~~~~~~~~~~~S -.U D E V E L O P M E N T I N P R A C T I C E Priorities and Strategies for Education Priorities and Strategies for Education A World Bank Review T H E W O R L D B A N K W A S H I N G T 0 N, D. C. © 1995 The International Bank for Reconstruction and Development / THE WORLD BANK All rights reserved Manufactured in the United States of America First printing August 1995 The Development it Practice series publishes reviews of the World Bank's activities in different regions and sectors. It lays particular emphasis on the progress that is being made and on the policies and practices that hold the most promise of success in the effort to reduce poverty in the developing world. This report is a study by the World Bank's staff, and the judgments made herein do not necessarily retlect the views of the Board of Executive Directors or of the governments they represent. Cover photo of schoolgirls in Fatehpur Sikri, India, by Maurice Asseo. Library of Congress Cataloging-in-Publication Dala Priorities and strategies in education: a World Bank review. p. cm. - (Development in practice) Includes bibliographical references. ISBN 0-8213-331 1-9 1. Education-Developing countries. 2. Education-Developing countries-Finance. 3. Education. Secondary-Developing countries. 4. Education and state-Developing countries. 5. Educational equalization-Developing countries. 6. Economic development-Effect of education on. 1. International Bank for Reconstruction and Development. 11. Series: Development in practice (Washington. D.C.) LC2605.P756 1995 370'.9172'4-dc2O 95-18770 CIP Contents FOREWORD xi ACKNOWLEDGMENTS xiii DEFINITIONS AND DATA NOTES xv SUMMARY 1 PART I THE RECORD OF EXPERIENCE AND THE TASKS AHEAD 17 1 Education and Development 19 Education and Economic Growth 19 Labor Market Linkages 24 Poverty Reduction 27 Fertility and Health 28 2 Achievements and Challenges 32 Access 33 Equity 43 Quality 46 Delays in Reforming Education 49 Appendix. The Poverty of Education Data 50 3 Public Finance for Efficiency and Equity 53 The Rationale for Public Finance 53 Misallocation among Education Subsectors 56 Misallocation within Education Subsectors 58 Inequitable Public Spending 62 v vi PRIORITIES AND STRATEGIES FOR EDUCATION The Potential for Increased Efficiency and Equity 67 Financing Education 69 4 Improving Quality 73 Setting Standards 73 Support for Effective Inputs 74 Flexibility in Providing Inputs 86 PART 11 SIX KEY REFORMS 89 5 A Higher Priority for Education 91 6 Attention to Outcomes 94 Using Outcomes to Set and Monitor Public Priorities 94 Setting Standards and Monitoring Performance 100 7 Public Investment Focused on Basic Education 103 Pricing Policy for Public Education 104 Priorities for Public Spending I1( Sustainability 112 8 Attention to Equity 113 Financial Measures 114 Special Measures 116 9 Household Involvement 120 School Governance 120 School Choice 122 Risks 124 10 Autonomous Institutions 126 Administrative Measures 126 Financial Measures 130 Risks 134 PART III IMPLEMENTING CHANGE 135 11 The Political and Social Context of Change 137 Appendix. Priorities for Educational Reform in Eastern and Central Europe 142 12 The World Bank and Education 145 Developments since 1980 145 Future Bank Support for Education 153 CONTENTS vii REFERENCES 157 TABLES I Challenges in Education and Key Reforms 9 1.1 Rates of Return to Investment in Education by Region and Level of Schooling 22 1.2 The Externalities of Investing in Girls' Education. Pakistan 31 2.1 Children Age 6-11 out of School. 1980-90 and Projections for 2000 and 2015 38 2.2 Countries with Primary Gross Enrollment Ratio below 90 Percent, 1990 39 3.1 Education Expenditure by Source of Funds. All Levels of Education Combined, Selected Countries, 1991 54 3.2 Public Recurrent Expenditure on Education by Level, 1990 56 3.3 Public Spending Per Student: Higher Education as a Multiple of Primary Education, 1980-90 58 3.4 Higher Education Students by Family Income 65 3.5 Allocation of Education Benefits in East Asia. 1985 67 3.6 Government and Household Expenditure on Education in Kenya by Level, 1992-93 72 5.1 Rates of Return to Investments in Different Sectors of the Economv 93 6.1 Compulsory Educationi, Enrollment Ratios, and Minimum Age for Employment, Selected Countries, 1990s 95 7.1 Fees for Public Higher Education as a Share of Unit Operating Expenditure, Selected Countries 108 7.2 Share of Secondary and Higher Education Budget Devoted to Student Welfare, Sahelian Countries. Sub-Saharan Africa. and Asia, about 1985 109 7.3 Savings in GNP Allocated to Education as a Result of Lower Fertility Rates in East Asia I]] 9.1 Relative Average Cost and Efficiency of Public and Private Schools. Early 1980s 124 10.1 Locus of Decisionnaking Authority in Primary Education Systems in Selected Developing Countries 128 10.2 Decisions Made at the School Level as a Share of All Decisions by Public Schools in OECD Countries, by Level of Education, 1991 129 12.1 External Support for Education, 1975-90 146 viii PRIORITIES AND STRATEGIES FOR EDUCATION FIGURES 1.1 Total Fertility Rate by Mother's Education and Region 29 1.2 Chances of Children Dying before Age Two by Mother's Education 30 2.1 Gross Enrollment Ratios by Region and Level of Education. 1980 and 1990 34 2.2 Expected Years of Schooling by Region. 1980 and 1990 35 2.3 Percentage Growth in Enrollments by Region and Level of Education, 1980-90 35 2.4 Growth of the Primary-School-Age (6-11) Population, 1990-2000 and 2000-2010 37 2.5 Primary School Enrollment and Retention by Region, about 1990 42 2.6 Gender Gaps in Expected Years of Schooling by Region, 1980 and 1990 44 2.7 Variation in Reading Achievement for 14-Year-Olds in Selected Countries, 1990-91 47 2.8 Difference in Reading Achievement between Urban and Rural Schools for 14-Year-Olds in Selected Countries, 1990-91 48 3.1 Change in Allocation of Public Recurrent Expenditure on Education by Region and Level, 1980-90 57 3.2 Primary and Secondary Student-Teacher Ratios, 1980 and 1990 60 3.3 Distribution of Subsidies for Education in Colombia, Indonesia, and Kenya, Selected Years 64 3.4 Public Education Expenditure as a Share of GNP and of Central Government Expenditure, 1980 and 1990 66 3.5 Relation between Public Spending on Education and Gross Enrollment Ratios for Population Ages 6-23, Selected Countries. 1990 68 3.6 Growth of Enrollment in and Public Expenditure oii Higher Education by Income Group, 1980-88 70 3.7 Relation between Gross Enrollment Ratios in Higher Education and Extent of Private Financing. Selected Asian Countries, about 1985 71 4.1 Determinants of Effective Learning at the Primary Level 82 12.1 World Bank Education Lending, Fiscal 1980-94 147 12.2 World Bank Education Lending by Subsector, Fiscal 1964-94 148 12.3 World Bank Education Lending by Region, Fiscal 1964-94 151 12.4 World Bank Education Lending by Expenditure Category, Fiscal 1964-94 152 CONTENTS ix BOXES 1.1 Rates of Return to Education 20 1.2 Education and Economic Growth in East Asia 23 2.1 AIDS and Education 40 2.2 Education in Eastern and Central Europe during the Political and Economic Transition 50 4.1 Gender Bias in Textbooks 79 4.2 A Promising Avenue: School Leadership 87 8.1 Reducing the Household Costs of Girls' Education in Bangladesh 115 9.1 School Development Boards in Sri Lanka 121 10.1 Involvement of NGOS in Education: The BRAC Story 130 11.1 Tradeoffs between Access and Quality: Lessons from Kenya and Thailand 139 11.2 Stakeholder Partnership in Mauritius 141 12.1 Massive Education Reforms in Mexico's Poor Southern States 149 Foreword EDUCATION produces knowledge, skills, values, and attitudes. It is essential for civic order and citizenship and for sustained economic growth and the reduction of poverty. Education is also about culture; it is the main instrument for disseminating the accomplishments of human civilization. These multiple purposes make education a key area of public policy in all countries. Its importance is recognized in several intemnational conventions and in many national constitutions. In 1990 it was the subject of a landmark international meeting: the World Conference on Education for All, held in Jomtien. Thai- land, under the joint sponsorship of the United Nations Development Progranmme (UNDP), the United Nations Educational, Scientific, and Cultural Organization (UNESCO), the United Nations Children's Fund (UNICEF), and the World Bank. The civic purpose of education-the sharing of values throughout soci- ety-is becoming more salient in light of the widespread political liberaliza- tion of the past decade. This trend, which is most notable in Easteni Europe and Central Asia, also includes the consolidation of civilian democratic rule in Latin America, the introduction of multiparty systems in Africa, and the devolution of political power to subnational levels of govemnment in many regions of the world. Research and experience have also led to a deeper understanding of how education contributes to economic growth, the reduction of poverty, and the good governance essential for implementing sound economic and social poli- cies. In line with these changing circumstances and perceptions. the World Bank's financing of education has grownl rapidly in the past fifteen years, and the Bank is now the single largest source of external finance for education in low- and middle-income countries. Projects to support primary and lower- secondary education-basic education-have become increasingly promi- nent in Bank lending for the sector. This emphasis is in harmony with the recommendations of the World Conference on Education for All. xi xii PRIORITIES AND STRATEGIES FOR EDUCATION The expansionl of World Bank lending for education has been accompa- nied by a series of studlies on education policy in developing countries: Educa- tionl in Sub-Saharan Af ica (1988). Primarv Ediucation (1990). V\ocational anzd Technical Education an7d Training (1991), and Higlher Education (1994). In addition, recent World Development Reports-Povery ( 1 990), The Challenge oJ Development (1991), Investing in Health (1993), and Workers in ani tlute- grating World (H 995)-have highlighted the importance of educatioll for de- velopmelnt. This Development in Practice book-the first overall review of education by the World Bank since the 1980 sector policy paper-synthesizes the find- ings of the publicationis issued in the interveninig years, adds a review of secondary education, reflecting the results of ongoing work in the Bank's Human Development Department, and extends these results into the areas of sectoral finance and management. It also draws heavily Oil UNESCO'S World Education Report (1993). The report outlines policy options that low- and middle-income counities can adopt to meet educational challenges as they move toward the twenty-fi-st century. It is designed to assist policymakers in these countries. especially those concernecd with the education system as a whole and with the allocation of public resour-ces to education. It is also intended for World Bank staff who work with client countries to support education policies and projects. The report discusses policy options in education, not the details of educa- tionI projects. It focuses on the formal edlucationl system and the role that governments can play throuoh sound financial and managerial policies that encourage the expansionl of the private sector and improvement in the func- tioning of public institutions. It does not pretend to be exhaustive: for ex- ample, it does not discuss in-depthi training (covered in the 1991 paper) or adult education, which is in the current work programii of the Human Develop- ment Departmllent. What the report does do is to treat the mainistr-eamii formal educationi sector as a whole. It focuses on the contribution of formal education to sustained economic growth and the reduction of poverty. It emphasizes approachies and ways of determining priorities and strategies. recognizinig that policies must be tailored to each country according to its stage of educational and economllic development and its historical and political context. Throughiout it keeps in view the World Bank's fundamental objective, whicih guides the Bank's wolk in educationi as in every sector: helping borrowers reduce poverty and improve living standards thrLough sustainable growth and investment in people. A4rneane M. Clxoksi Vice President Human Capital Development and Operations Polic y The World Bank Acknhowledgments THIS report was prepared by a team led by Nicholas Bumnett and consisting of Toni Eisemon, Kari Marble. and Harry Anthony Patrilos. ULinder the general direction of K. Y. Amoako and the immediate supervision of Peter R. Moock in the Education and Social Policy Department. Other niajor con- tributors to the report were Aruin Joshi. Marlaine Lockheed and Kin Bing Wu; material was also provided by Barbara Bruns, Sarbani Chakraborty. Helen Craig. Joy Del Rosso. Reed Garfield, Indennit Gill. Masooma Habib. Jane Hannaway, Ward Heneveld, Donald Holsinger, Theresa Moran, Christina Rawley, OriipoIii Regel. Rajendra Swamy and Stella Tamiavo. Helpful com- merits on earlier drafts were given by Arvil Van Adams, Jean-Claude Eicher, Vincent Greaniey, Lauritz Holm-Nielsen. Bruno Laporte. Jon Lauglo. Michael Mertaugih, John Middleton. Alain Mingat. Paud Murphy, Fran,ois Orivel. Janiil Salmi. Nate Scovronick, Lyvi Squire, Jee-Penig Tan, Zafiris Tzariiiatos. Michael Walton. Maureen Woodhall, and Adrian Ziderman. Members of a Banikwide advisory panel that provided valuable assistance were Mark Baird. Carl Dahlman, Birger Fredriksen, Wacdi Haddad. Ralph Harbison. Roslyn Hees. Stephen Heyneinani, Enimaniuel Jimenez, Homi Kharas. Jack Maas. Hiimelda Martinez, Philip Musgrove, George Psacharopoulos. Jurlian Schweitzer, Richard Skoliik. James Socknat and Donald Winkler. The paper was reviewed in September 1994 by an external panel consisting of ministers. senior officials, anid academics from Arnienia. Colombia, France. Guinea. India, Japan. Jordan. Mexico, Nigeria. Pakistan. the Philippines. the Russian Federation, the Slovak Republic. Thailand, Uganida. and the United Kingdomi. Discussions were also held with staff mnembers of the Organisation for Eco- nomic Co-operation and Developmient (OECD), UNESCO. UNICEF, the Canadian Intemnational Development Agency (CIDA); the Swedish Intemnational Devel- xiii xiv PRIORITIES AND STRATEGIES FOR EDUCATION opment Authority (SIDA), and the U.S. Agency for International Development (USAID); with the UNESCO Commission on Education for the Twenty-First Century, chaired by Jacques Delors; with donor agency representatives meet- ing under the auspices of the International Working Group on Education; with the Commonwealth Ministers of Education: with British academics and offi- cials at a meeting organized by the British Council; and with Education International. Jo Bischoff, Ian Conachy, Richard Crum, Kari Labrie. and Margot Verbeeck helped prepare drafts of the report. Definitions and Data Notes FOR ITS operational purposes, the World Bank groups low- and middle-income countries (as defined by the Bank's International Economics Department) into six regions: Sub-Saharan Africa, East Asia and the Pacific, Europe and Central Asia, Latin America and the Caribbean, the Middle East and North Africa, and South Asia. The analysis in this report uses these regions and, for some comparisons, two other groups: all low- and middle- income countries, and the members of the Organisation for Economic Co- operation and Development (OECD). Because of gaps in data availability, some regional averages for Europe and Central Asia do not include the countries of the fonner Soviet Union. Unless otherwise stated, dollar amounts are current U.S. dollars. A billion is a thousand million. xv Summary EDUCATION is critical for economic growth and poverty reduc- tion. Changing technology and economic reforms are creating dramatic shifts in the structure of economies, industries, and labor markets throughout the world. The rapid increase in knowledge and the pace of changing technology raise the possibility of sustained economic growth with more frequent job changes during individuals' lives. These developments have created two key priorities for education: it must meet economies' growing demands for adapt- able workers who can readily acquire new skills, and it must support the contin- ued expansion of knowledge. This paper synthesizes World Bank work on education since publication of the last sector policy paper, in 1980. and consid- ers options for the Bank's borrowing countries. The World Bank's strategy for reducing poverty focuses on promoting the productive use of labor-the main asset of the poor-and providing basic social services to the poor. Investment in education contributes to the accumu- lation of human capital. which is essential for higher incomes and sustained economic growth. Education-especially basic (primary and lower-secondary) education-helps reduce poverty by increasing the productivity of the poor. by reducing fertility and improving health, and by equipping people with the skills they need to participate fully in the economy and in society. More generally. education helps strengthen civil institutions and build national capacity and 2 PRIORITIES AND STRATEGIES FOR EDUCATION good governanice-critical elements in the implementation of sound economic and social policies. Basic education encompasses general skills such as language, science and mathematics, and communications that provide the foundation for further edu- cation and training. It also includes the development of attitudes necessary for the workplace. Academic and vocational skills are imparted at higher levels; on-the-job training and work-related continuing education update those skills. Progr-ess and Challenges The economies of low- and middle-ilcome countries have been growing at historically rapid rates. Progress in education-expanded enrollments and longer schooling-has contributed to this growth and so has helped to reduce poverty in developing countries. In 1990 a typical 6-year-old child in a developing country could expect to attend school for 8.5 years. up from 7.6 years in 1980. In Eastern Europe and Central Asia schooling for nine to ten years is the rule: in East Asia and in Latin Ainerica and the Caribbean primary education is almost universal. Countries in the Middle East and North Africa are making steady progress; so are those in South Asia. although they have a considerable distance to go. Sub-Saharan Africa is lagging: certain countries are making gains, but overall, primary enrollment ratios are actually declining. Yet despite these substantial achievements in the world as a whole, major challenges remain: to increase access to education in some countries, to im- prove equity, to improve quality, and, where needed, to speed educational reform. Access If the current high population growth rates in Africa, South Asia, and the Middle East and North Africa continue, the number of 6-to-I 1-year-old chil- dren not in school will increase to 162 million by 2015, from 129 million in 1990. To make matters worse, only two-thirds of children who start primary school complete it. As a result, adult illiteracy, which already affects over 900 million people, most of them women, is likely to remain a major problern. In most countries. more children wish to go to secondary school than are able to enroll, and the demand for higher education is in general increasing faster than the supply. The enrollment gap between the transition economies of Europe and Central Asia and the members of the Organisation for Economic Co-operation and Development (OECD) is also widening as enrollment ratios decline in the former and rise in the OECD countries. SUMMARY 3 Equity The issue of equity mainly affects several overlapping disadvantaged groups, including the poor, linguistic and ethnic minorities, nomads, refugees, and street and working children. The different access that boys and girls have to the education system in some parts of the world is also very important because it contributes to gender differences later in life. The gender gap in expected years of schooling is now very small in most countries in Europe and Central Asia and in Latin America. It remains large in the Middle East and North Africa and in South Asia, where it is not closing at all. Quality The quality of education is poor at all levels in low- and middle-income coun- tries. Students in developing countries have a mean level of achievement below that in industrial countries, and their performance shows a much greater varia- tion around the mean. Speeding up Refornm Delays in reforming education systems to keep pace with economic structures are most apparent in the transition economies of Eastern and Central Europe. Lags in reform can hinder growth; conversely, timely reform can pay off in terms of economic growth and poverty reduction, as evidenced by the East Asian countries that have generally invested heavily in basic human capital, both male and female. Fiiiance and(I Maana"eiiient Present systems for financing and managing education are often inappropriate for meeting the challenges discussed above. Public financing, moreover. is growing more difficult as enrollments expand. Public intervention in education can be justified on several counts: it can reduce inequality. open opportunities for the poor and disadvantaged, compen- sate for market failures in lending for education, and make information about the benefits and availability of education generally available. But public spend- ing on education is often inefficient and inequitable. It is inefficient when it is misallocated among uses; it is inequitable when qualified potential students are unable to enroll in institutions because educational opportunities are lacking or because of inability to pay. 4 PRIORITIES AND STRATEGIES FOR EDUCATION Basic education ought to be the priority for public spending on education in those countries that have yet to achieve near-universal enrolilinelit at the pri- mary and lower-secondary levels. Most countries are already allocating the highest share of public spending on education to primary education. Subsidies increase the demand for higher education. Although puLblic spending per stu- dent in higher education is falling in comparison with that per primary student, it remains very high. In Africa. for instance, spending per student in higher education is about forty-four times that per student in primary education, and the share of higher education in total public spending on education is now higiher- than in any other region of the world. Yet one-half of Africa's primary- school-age childr-en are not enrolled in school, and universities in the region are often of low qLiality. Inefficient mixes of inputs-for example, between staff and instrLuctional materials-can contribute to low learning achievement and high repetition and dropout rates. For effective learning, the input mix must vary from country to country and school to school according to local conditions. Intenilationial coin- parisonis and interschool comparisons can, however. provide broad guidance. Modest increases in student-teacher ratios tend to improve education when they permit resources to be reallocated to other critical inputs, such as text- books. School buildinigs can be built more cheaply than is usually the case at present. and they will last longer if adequate maintenance fundinig is ensured. Consolidation of small schools and the use of multigrade teaching andc multiple shifts offer other possibilities for using buildings more efficiently. Public spencding on primary education generally favors the poor, but public spending on education as a whole often favors the affluent because of the heavy subsidizationi of the upper-secondary and higher educationi levels, which USLI- ally have relatively few students from pooi families. Public sector spencding for higiler education is particularly iiecluitable because the subsidy per stuLdent is highier than that for basic education, even though higher edtIcatioln students come disproportionately from richer famiiilies. The PotentialJ('r Saving thlioglh Effic ienc v In 1990 public spending on educationi equLaled 5.2 percent of gross national product (GNP) in the Midcdle East and North Africa but only 3.4 percent in East Asia. Yet in both regions an average 6-year-old child could expect to complete more than nine years of school. Public spending oil education in Africa, which has the lowest enr ollment r atios of any region. represents a greater share ofGNP (4.2 percent) than in Latin America (3.7 percent) or East Asia, whichi have largely achieved uiliversal primary education. Some cotLntries that spend very little on education could dramatically improve results simply by increasing SUMMARY 5 public spending. In many countries, however, improved education could be achieved with the same or even less public spending by focusing public spend- ing on the lower levels of education and increasing internal efficiency, as has been done in East Asia. The Need for New Sources of Finance The inefficiencies and inequities described above, along with expanding en- r ollnents in public schools at all levels, have contributed to increasing the share of GNP devoted to public spending on education. The result is increasing pres- sure on public funds at the same time that many countries, especially in Eastern Europe and Africa. ar e experiencing general fiscal difficulties. During the 1980s public spending on education as a share of GNP remained stable or increased. and its share of total central govermilent spending increased, in nearly every region of the developing world. In Latin America, which experienced debt- induced recession. real public expenditure per primary student fell. In Africa real spending per student decreased at both the primary and the secondarv levels. Real spending per student in higher education fell in all regions. As enrollments increase. resources per student decline, and so will the quality of schooling, unless public spending becomes more efficient. Although measuLes to increase the efficiency of public spending on educa- tion can make existing funds more productive, such measures alone may not be enough. Sonie countries have chosen to reallocate public spending to education fromil other publicly fLunded activities, such as defense and iiefficient public enterprises that can be runi better by the private sector. Other countries have found a wa', withini their macroeconomic policies, to increase the revenues of government and thereby have more to spend on education. Yet others have sought to supplemlienit public funds for educationi with private funds. Private financing can be encouraged either to fund private institutions or to supplemiienit the income of publicly funded ilstitUtions. Some countries prohibit private schools and universities; others regulate them excessively. Since pri- vate schools are usually financed through fees. such restrictions crowd out private spending on education that would otherwise have occurred and so increase the pressure on publicly funlded schools. Another argumilent for private schools and ulliversities is that, even though they tend to draw their students from more advantaged socioeconomic backgrounds, they promote diversity and provide LIseful comilpetitioni for public institutionis, especially at the higher levels of education. Charging fees for students at publicly funded institutions raises difficult questions concerniig equity. access. and taxation. If all students attendiig public schools at all levels are charged. the poor will be hit particularly hard. 6 PRIORITIES AND STRATEGIES FOR EDUCATION discouraging their enrollment. Scholarship and other systems can be intro- duced to offset this problem. but they are complex to administer at lower levels of education. At the upper-secondary and higher levels, a much stronger case exists for payment of fees by students at public institutions. For these levels, the gap between the private and the social returns to education is generally much greater than in basic education. This inequlity can be overcome by charging the student either from current family income or from future earnings, through a loan scheme or through the tax system. Organizing Eclucationi for Effective Schooling Most education systems are directly managed by central or state governments, which put a great deal of effort into dealing with such issues as teacher salary negotiations, school construction programs, and curriculum reform. This cen- tral management. extending even to instructional inputs and the classroom environment, allows little room for the flexibility that leads to effective learning. The main ways in which governments can help improve the quality of education are setting standards. supporting inputs known to improve achieve- ment. adopting flexible strategies for the acquisition and use of inputs, and monitoring performance. Generally, however, these steps are not taken because of the weight of existing education spending and management practices and the vested interests associated with them. Standards Governments can help improve academnic achievement by setting clear and high performance standards in core subjects. Inputs Learning requires five inputs: the student's capacity and motivation to learn, the subject to be learned, a teacher who knows the subject and can teach it, time for learning, and tools for teaching and learning. Capacityfor learniing can be increased through high-quality preschool edu- cational programs and through preschool and school-based programs to rem- edy short-term hunger, protein-energy and micronutrient malnutrition, hearing and vision impainnents. highly prevalent health conditions such as parasitic infections, and inappropriate health and nutrition practices. Such programs are even more effective when combined with efforts to improve the physical envi- ronment in the school. SUMMARY 7 The curriculum defines the subjects to be taught and furnishes general guidance on the frequency and duration of instruction. Curricula and syllabi should be closely tied to performance standards and measures of outcome. No single curriculum is appropriate for all or most developing countries, but some generalizations can be made. At the primary level curricula are fairly standard. but there are often too many subjects, reducing the time available for teaching core skills. The most effective initial language of instruction is the child's native language. At the secondary level curricula vary considerably, particu- larly in science education and vocational education. Science education. be- cause of its importance for economic development, is increasingly integrated in curricula; specialized vocational and technical education, which yields much lower social returns than does general secondary education. is best conducted in the workplace, after a grounding in general education. Vocational education works best when the private sector is directly involved in its provision, financ- ing, and governance. At all levels. making the curriculum gender-sensitive is particularly important for encouraging girls' education. The most effective teachers appear to be those with good knowledge of the subject and a wide repertoire of teaching skills. The most effective strategy for ensuring that teachers have adequate subject knowledge is to recruit suitably educated teachers whose knowledge has been demonstrated through assessed performance. This strategy is followed for secondary and higher education teachers but is rare at the primary level. In-service training to improve teachers' subject knowledge and related pedagogical practices is niost effective when it is linked directly to classroom practice and provided by the head teacher. The amount of actual tinie devoted to learning is consistently related to achievement. Students in low- and middle-income countries receive fewer hours of classroom instruction than those in OECD countries-a consequence of a shorter official school year, unscheduled school closings, teacher and student absences. and miscellaneous disruptions. Instructional time can be increased by extendilng the official school year; by permitting flexible scheduling to accom- modate the demands of agricultural seasons, religious holidays, and children's domestic chores: and by assigning homework. The most effective instructional mnaterials are blackboards, chalk, and text- books. Supplementary reading material is particularly critical lor improving reading skills. Flexibility Flexibility in combining and managing inputs and monitoring performance s vital for effective schooling. Yet many education systems in developing coun- 8 PRIORITIES AND STRATEGIES FOR EDUCATION tries are still rigidly centralized; for instance, a central authority often selects and purchases textbooks and prescribes teaching methodology. School govern- ing bodies, principals, and teachers. with their intimate knowledge of local conditions, are best able to select the most appropriate package of inputs. Under the right circumstances, making schools and higher education institutions ac- countable to parents. communities, and students helps bring about more effec- tive learning and hence improves educational quality. Three conditions are necessary for this result: shared goals regarding the learning objectives of the school, professionalism on the part of teachers, and schools' autonomy in allocating instructional resources flexibly. A further promising approach is school-based leadership that ensures an effective climate for learning. Priorities {or -Refo-rm Reforming education finance and management means redefining the role of government in six key ways, with the order of priority depending on country circumstances. Table I illustrates how these measures contribute to the goals of improving access, equity. and quality and overcoming delays in reform. A Higher Priority for Education Education is more important than ever for economic development and poverty reduction, and its role in this effort is better understood. Education, for girls as well as for boys. therefore deserves a higher priority on government agendas- not just those of ministries of education. This imperative has long been recog- nized in East Asia and is increasingly coming to be understood elsewhere. particularly Latin America. Other countries still need to give education more attention. Education alone will not reduce poverty: complementary macroeconomic policies and physical investments are also needed. Attention to Outcomes Educational priorities should be set with reference to outcomes, using eco- nomic analysis, standard setting, and measurement of achievement through learning assessments. An approach that looks at the whole sector is key for setting priorities; attention to only one level of education is insufficient. Eco- nomic analysis usually compares benefits (in labor productivity, as measured by wages) with costs, for individuals and for society. It identifies as priorities for public investment those investments for which the social rate of return is highest and the level of public subsidization is lowest. Rates of return must be calculated for specific country circumstances and cannot be assumed. Because SUMMARY 9 TABLE 1 CHALLENGES IN EDUCATION AND KEY REFORMS Over-c omiing delays Stiategy Access Equity Qualify in reform Higher priority for education / / / / Attention to outcomes / / Emphasis on basic education in public investment v/ / / Attention to equity / Household involvement V/ / / Autonomous institutions / / of the difficulty of valuing external benefits not reflected in wages. cost-benefit analysis should be applied with caution. The high rates of return estimated for basic education in most developing countries strongly suggest that investments to improve enrollments and reten- tion in basic education should generally have the highest priority in countries that have not yet achieved universal basic education. Some improvements in educational efficiency or quality will often be possible through policy changes that require no specific investments. Decisions about public education priorities beyond basic education have to be taken within a broad sectoral approach and will vary according to country. Countries that have largely achieved universal primary and lower-secondary education are likely to consider upper-secondary and higher education the priorities. and they can often make informed decisions about these postcompulsory levels through the prudent use of economic analysis focused on labor market outcomes. It has been shown, for example, that the returns to general secondary education are much higher than those to highly specialized vocational secondary education, although analysis has not yet looked into the returns to investment in the more "general" type of vocational education that is now becoming prevalent in many OECD countries. Countries that have yet to achieve universal basic education will need to pay attention to all levels of education, using economic analysis to guide decisions about which invest- ments will have the greatest effect. Focusing on outcomes also entails the establishment of performance standards, particularly for primary and general secondary schools, and development of a system of assessments to monitor what students are learning. Standards, curricula. and monitoring are most effec- tive when they are directly linked through appropriate incentives. 10 PRIORITIES AND STRATEGIES FOR EDUCATION Em-Sphasis on Basic Education in Public Investment A more efficient, equitable. and sustainable allocation of new public invest- ment on education would do much to meet the challenges that education sys- tems face today. Efficiency is achieved by making public investments where they will yield the highest returns-usually, for education investments, in basic education. To achieve equity. the government needs to ensure that no qualified student is denied access to education because of inability to pay. Because the gap between private and social returns is larger for higher education than for basic education, students and parents may well be willing to bear part of the costs of higher education. Governments can also encourage private financing by taking oii some of the risk that makes financial institutions reluctant to lend for higher education. A policy package of fees and efficient expenditure in the public sector might consist of: * Free basic education, including cost-sharing with communities and targeted stipends for children from poor households * Selective charging of fees for upper-secondary education. combined with targeted scholarships * Fees for all public higher education, combined with loans, taxes, and other schemes to allow needy students to defer payment until they become income- earners, and a targeted scholarship scheme to overcome the reluctance of the poor to accumulate debt against uncertain fuiture earnings * Assurance of quality primary education for all children by making that level the top priority for public spending on education in all countries * Improved access to quality general secondary education (initially lower- secondary and later all levels of secondary) as the second priority, once all children are receiving good primary education * Efficient public spending at the school and institution level. Fiscal sustainability also requires the continuous projection of the implications of public expenditure and consistent efforts to ensure that financing plans and mechanisms are in place. Attention to Equit' Equity has two principal aspects: (a) everyone's right to a basic education-the basic knowledge and skills necessary to function effectively in society-and SUMMARY 11 (b) the government's obligation to ensure that qualified potential students are not denied education because they are poor or female, are from disadvantaged ethnic mniorities or geographically remote regions, or have special educational needs. At the lowest and compulsory levels of education, equity simply means ensuring that schools are available. Beyond that, it means having fair and valid ways of determining potential students' qualifications for entry. Achieving equity requires both financial and administrative measures. Fi- nancial measures, such as scholarships, are important at all levels to enable the poor to gain an education. Scholarships can cover fees and other direct costs. such as transport and uniforms and, when appropriate, can compensate families for the indirect costs of sending children to school-for example, loss of labor services for the household. Administrative measures can increase enrollments of the poor, females, linguistic minorities, and students with special educa- tional needs. Programs designed to demonstrate the importance of educating children can increase the demand for schooling among the poor. Measures to encourage schooling for girls can include setting aside all-girl classrooms and schools, locating schools within easy access of girls' homes, providing separate sanitary facilities, constructing boundary walls, increasing the number of fe- male teachers, providing childcare centers, and adjusting school hours to ac- commodate girls' work at home. For linguistic minorities, bilingual programs and schools offering a choice of language of instruction are important, espe- cially in primary education. Special programs. such as deworming and micro- nutrient supplementation, to improve the nutrition and health of schoolchildren can reduce the number of children with physical and learning disabilities. The (usually low) costs of educating children with minor impairments can often be shared with nongovernmental organizations. Househlold Involvement Around the world, parents and communities are becoming more involved in the governance of their children's schools. Effective involvement in school gover- nance does not come easily, however, and training is generally advisable. Several countries have a long tradition of parental choice. Increased experi- mentation with school choice is another recent educational reform. particularly in OECD countries. There is as yet no evidence that competition among schools either improves or worsens school performance. For choice to be effective, the student must have more than one possible school. The institutions should have some distinguishing characteristics-for example, in what aspects of the curriculum are emphasized, in teaching styles, and. at higher levels, in course offerings. Finally. institutions need to enjoy 12 PRIORITIES AND STRATEGIES FOR EDUCATION considerable autonomy in how they teach. The availability of a variety of institutions enables parents and students to exercise choice and thus gives institutions an incentive to adapt to demand. Increased household involvement carries major risks: * Implementation of systemwide education policies can be more difficult. * Enforcement of broader national objectives can be hampered. * Social segr-egation may increase if schools become polarized between elite academies and schools for the children of the poor and uneducated. * Equity may be reduced if schools and institutions accept students on the basis of their ability to pay rather than on academic entrance qualifications. * Parents may lack the information they need to make judgments about quality. The first four risks can be mitigated relatively easily through policies for the provision of public funding. Such funds can be made available only to schools that follow certain practices, can be higher per StuLdent for poor children. and can be accompanied by restrictions on fee levels. The fifth risk can be reduced through government efforts to provide open and independent information about school quality. Autoinomous Institutions The quality of education can benefit when schools have the autonomy to use instructional inputs according to local school and community conditions and are accountable to parents and communities. Fully autonomous institutions have authority to allocate their resources (not necessarily to raise them), and they are able to create an educational environment adapted to local conditions inside and outside the school. There is still little evidence about the impact of increased school-level flexibility on the overall quality of education systems in developing countries. As with school choice, therefore, some caution is called for as more countries experiment with increased school autonomy. Accountable autonomous institutions can be encouraged by both adminis- trative and financial means. Administrative measures include giving school management the authority to allocate resources-for example. the aLlthority to deploy personnel and to alter such things as the timing of the school day and year and the language of instruction to fit local conditions. Most critically. teachers need to have the authority to determine classroom practices, within limits set by a broad national curriculum. Financial measures to encourage school and institutional autonomy and accountability can include: SUMMARY 13 * The use of local and central government taxation * Cost-sharing with local communities * The allocation of block grants to communities and schools without restric- tions on the allocation of the funds * Fee charging at higher levels of education * Encouragement of revenue diversification * The use of financing mechanisms in whiich money follows students, such as capitation grants. vouchers and student loans * Fundinlg based on output and quality. Reliance oni local funding must be tempered with adjustments by higher levels of government to compensate for differing resource levels among localities. Local control of resources need not imply local raising of revenues. The goal of local financing of schools should be to improve learning, not to reduce overall resources. The main risks of school autonomy are the creation of inequalities in edu- cational opportunities and failure to adhere to the national standards and cur- riculuin. These risks can be largely mitigated by clearly distinguishing school- level management and resource allocation from exclusive reliance on local financinig and by handling performance standards. curricula, and learning as- sessnients at the national or regional level. There are few risks to increased institutional autonomy in higher education. Implementiing Chanige The relative priority accorded each reform depends on specific country circum- stances. A sectoral approach-one that considers the efficiency. equity. and quality of the education sector as a whole and gives due attention to the larger policy environment and institution building-is essential. In all countries entrenched ways of operating and vested interests will make change difficult. Education is intensely political: it affects the majority of citi- zens. involves all levels of government, almost always makes up the single largest component of public spending in developing countries, and involves public subsidies that are biased in favor of the elite. Prevailing systems of education spending and management often protect the interests of teachers' unions, university students, elites, and the central government rather than those of parents. communities, ancl the poor. There are, however, strategies that can 14 PRIORITIES AND STRATEGIES FOR EDUCATION ease change. Financing and management reforms are best introduced in paral- lel with the expansion of educational opportunities. Sometimes the change itself makes for expansion-for example. when prohibitions on the private sector are lifted. Increased cost-sharing in public higher education is politically most feasible when it is linked to expansion of opportunities for higher educa- tion. Building niational consensus involves stakeholders in the education sys- tern in national consultation mechanisms. Increasing the involvement of par- enits and communities by making schools autonomous and accountable can offset the power of vested interests: it is also critical for increasing flexibility and improving instructional quality. Careful design of refornm measures is nec- essary to avoid disrupting the vital links among education subsectors. An es- sential, although often neglected, step is the provision of appropriate resources and mechanisms to accompany policy changes. The transition economies of Eastern and Central Europe have high primary and secondary enrollment ratios but need to adjust the entire education system toward the needs of a market economy. It is particularly important for these countries to maintain funding levels for basic and upper-secondary edu- cation. to shift away from overspecialization at vocational. technical, and higher education institutions, and to reform the governance and financing of higher education. 'lihe Wor-ld Bank and Education The World Bank made its first loan for education in 1963, and the Bank is now the largest single source of external financing for education in developing countries. Since 1980 the total volume of lending for education has tripled, and its share in overall Bank lending has doubled. Primary and secondary education are increasingly important; in fiscal 1990-94 these levels represented half of all Bank lending for education. Early Bank lending for education concentrated on Africa. East Asia, and the Middle East. but today lending is significant in all regions. Girls' education is at the forefront. and increasing attention is being given to the educational needs of ethnic minorities and indigenous people. Today Bank funds are used less for buildings and more for other educational inputs. The narrow project focus of the past is increasingly giving way to a broad sectoral approach. The World Bank is strongly committed to continued support for education. However, even though Bank funding now accounts for about a quarter of all aid to education, this funding still represents only about 0.5 percent of developing countries' total spending on the sector. Thus, the World Bank's main contribu- tion must be advice, designed to help governments develop education policies suitable for the circumstances of their countries. Bank financinig will generally SUMMARY 15 be designed to leverage spending and policy change by national authorities. Future operations will therefore adopt an even more explicit sectorwide policy focus in order to support changes in educational financing and management. Because of the need to consult key stakeholders, this strategy may increase both the resources and the time needed to prepare projects. In increasingly decen- tralized contexts. the stakeholders will include not only central governments but also other levels of government, as well as communities, parents. teachers, and employers. Donor cooperation will extend to broad policy advice, as well as investment coordination. Bank programs will encourage governments to give a higher priority to education and educational reform, particularly as economic refonn takes hold as a permanent process. Projects will take more account of outcomes and their relation to inputs, making explicit use of cost-benefit analysis. participatory methods, leaniing assessments. and improved monitoring and evaluation. The share of basic education in total Bank lending for education is expected to continue to increase, especially in the poorest countries, which receive Interna- tional Development Association (IDA) funds. This emphasis will be embedded in a sectoral approach that recognizes the importance of the various parts of the education system. the interdependencies among these parts. and the need to base both the emphasis and the nature of Bank assistance on a determination of where the Bank can be most useful in the particular circumstances of each country. Bank-sUpported projects will pay greater attention to equity-especially education for girls, for disadvantaged ethnic minorities. and for the poor-and consequently to early childhood education. Projects will support household involvement in school govemnance and in school choice through an increased emphasis on the regulatory framework for education, on quality-enhancing mechanisms such as outcome monitoring and inspection, on recurrent cost financing, and on demand-side financing mechanisms such as targeted scholar- ships for the poor, stipends for girls, and student loan schemes for higher education. They will encourage flexible management of instructional resources. complemented by national assessment and examination systems to provide incentives. In all these areas, Bank-supported projects will focus more intently on institutional development-including strengthening educational adminis- tration-and appropriate financial mechanisms, and the Bank's staff will pay increased attention to implementation. Basic education will continue to receive the highest priority in the Bank's education lending to countries that have not yet achieved universal literacy and adequate access, equity, and quality at that level. The Bank's sectoral approach means that in countries that have yet to achieve universal literacy. its involvc- ment in higher education will continue to concentrate mainly on making the 16 PRIORITIES AND STRATEGIES FOR EDUCATION financing of this level more equitable and cost-effective so that primary and secondary education can receive increased attention at the margin. As the basic education system develops in coverage and effectiveness, more attention can be devoted to the upper-secondary and higher levels. Bank lending for higher education will support countries' efforts to adopt policy reforms that will allow the subsector to operate more efficiently and at lower public cost. Countries prepared to adopt a higher education policy framework that stresses a differentiated institutional structure and diversified resource base. with greater emphasis on private providers and private funding, will continue to receive priority. P A R T I The Record of Experience and the Tasks Ahead EDUCATION-in particular, primary and lower-secondary educa- tion-is critical for economic growth and reduction of poverty, especially at a time when, as a result of technological change and economic reform, labor market structures are shifting dramatically. The spread of education has re- duced poverty by helping developing countries' economies grow at histor-ically rapid rates. Yet major challenges remain: to expand access in some countries and, in many others, to increase equity, improve educational quality. and speed educational reform. The current systems of finance and management are fre- quently not well suited to meeting these challenges. Public spending on educa- tion is too often inefficient and inequitable. In view of the competition for and pressure on public funds, new sources of fiianicinig are needed. And changes may be needed in the organization and management of education systems to permit the flexibility and choice that contribute to better student achievement and prospects. The next four chapters examine how education contributes to economic development and what actions are needed to meet the demands of a changing world and workplace. 17 C H A P T E R O N E Educationi and Developmi-enit EDUCATION is a maijor instrulilenit for economic and social devel- opment. It is central to the World Bank's strategy for helpinig countries reduce poverty and improve living standards throughI sustainable growth and invest- menit in people. This twofold strategy calls for promotinig the productive use of labor-the principal asset of the poor-and providing basic social services to the poor (World Bank 1990b). Investiimenit in eclucation leads to the accumulation ot' human capital. which is key to sustained economic growth and iicreasing incomiles. Education, espe- cially basic (primiiary and lower-secondary) education, also contributes to pov- erty reduction by increasing the productivity of the poor's labor, by reducing fertility and iilprovinig healthi, and by equipping people to participate fully in the economiiy and in society. In addlition, education contributes to the strength- eninig of the institutions of civil society. to nationial capacity building, and to good governance, all of which are increasinigly recogilized as critical elements in the effective implemenitationi of sound economic and social policies. Educationl a31(1 Econiomiic ( I row tih Educationi contributes to economnic growth, but by itself it will not generate growth. The strongest growth comes about when investment in both human and physical capital takes place in economies with competitive markets for goods and factors of production. Such markets are the product of macroeconomic 19 20 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD stability, well-functioning labor markets, and an openness to international trade and flows of technology. Economic growth is explained only in part by stocks of labor and physical capital. A large component of growth stems from improvements in the quality of the labor force, including increased education and better health, together with technological progress and economies of scale (T. W. Schultz 1961: Denison 1967: World Bank 1991d). New theories of economic growth suggest that faster technological change increases the long-run economic growth rate. Tech- nological change increases faster. in turn, when workers are more highly edu- cated. Thus, the accumulationi of human capital, and specifically of knowledge. facilitates the development of new technologies and is a source of self-sustain- ing growth (Romer 1986: Lucas 1988: Azariadis and Drazen 1990; Barro 1991). Education contributes to economic growth both thiough the increased indi- vidual productivity brought about by the acquisition of skills and attitudes and through the accumulationi of knowledge. The contribution of education can be estimated by its impact on productivity, measured by comparing the difference BOX 1.1 RATES OF RETURN TO whether to "invest" in a four-year uni- EDUCATION versity degree. The prospective stu- dent has to compare the costs and The concept of the rate of return to benefits associated with going to col- investment in education is very simi- lege. The direct cost is $10,000 a year lar to that for any other investment for tuition and other expenses related project: it is a summary of the costs tostudy. Inaddition,thestudentwould and benefits of the investment that incur an indirect (or opportunity) cost apply at different points in time, and it because of not being able to work while is expressed in an annual (percent- attendingcollege.Thisisapproximated age) yield, similar to that quoted for by what 18-to-21 -year-olds with a sec- bank savings accounts or government ondary-school leaving certificate earn bonds. If the rate of return to educa- in the labor market-say, $20,000 a tion is 10 percent, this means that year. On the benefits side, the student when $100,000 is invested in educa- expects to be making, on average, ap- tion, there will be an annual benefit of proximately $15,000 a year more in $10,000 over the lifetime of the aver- the future than a secondary-school age graduate, over and above what graduate would over his or her life- the same person would have earned time. without the investment. A rough way of summarizing these Assume that an 18-year-old sec- costs and benefits is to divide the an- ondary-school graduate is driven only nual benefit of $15,000 by the lump- by monetary considerations on sum cost of $120,000, yielding a 12.5 EDUCATION AND DEVELOPMENT 21 in earnings over time of individuals with and without a particular course of education and the cost to the economy of producing that education. This mea- sure is known as the social rate of return to investing in education. although it does not capture all social benefits or all extemnal effects (box 1.1). Rates of return to education can be difficult to measure in some cases (Weale 1993)- though no more so than for other sectors such as agriculture and transport-but they have withstood the tests of more than three decades of careful scrutiny (T. P. Schultz 1994). And, as was concluded almost two decades ago. human capital theory has no genuine rival of equal breadth and rigor (Blaug 1976). Rates of return to education are very high in low- and middle-income countries (table 1.1). Country circumstances differ, but, in general, in econo- milies with less than universal basic education, rates of return are greatest for primary education, followed by secondary and then higher education. Interest- ingly. economies with universal primary education that have undergone rapid growth tend to show a higher rate of return to secondary than to primary education (Jain 1991; T. P. Schultz 1993. 1994). percent rate of return to investment in ideally include the externalities, or higher education. The logic of this cal- spillover benefits, associated with edu- culation is similar to that of buying a cation. One of the main arguments $120,000 bond with a annual coupon used to justify public subsidy of edu- of $15,000. The yield of the bond is cation has to do with externalities, 12.5 percent. which affect society as a whole but are The example given above refers to not captured by the individual. Since a private rate of return, where the costs the social benefits of education exceed are what the individual actually pays the private benefits, governments sub- in order to receive an education. A sidize education to prevent under- social rate of return calculation includes investment. on the cost side the full resource cost As in most other sectors, externali- of one's education-that is, not only ties are extremely difficult to measure what the individual pays but also what and are not reflected in earnings. So- it really costs society to educate one cial rates of return, as conventionally person. Since in most countries edu- computed on the basis of monetary cation is heavily subsidized, the social earnings and costs, thus underesti- cost is much higher than the private mate the true social returns to invest- cost. Hence a social rate of return is ment in education. If one could include typically less than the corresponding externalities, social rates of return private rate of return. might well be higher than private rates Beyond these monetary adjust- of return to education. ments, the social rate of return should 22 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD TABLE 1.1 RATES OF RETURN TO INVESTMENT IN EDUCATION BY REGION AND LEVEL OF SCHOOLING .sucial Pirivate Re.'io,n Primar, Se( {hary Highjer Priman- ySe o) V Higholr Low- and middle-income countries Sub-Saharan Africa 24.3 18.2 11.2 41.3 26.6 27.8 Asia 19.9 13.3 11.7 39.0 18.9 19.9 Europe. Middle East. anid North Africa 15.5 11.2 10.6 17.4 15.9 21.7 Latin America and the Caribbean 17.9 12.8 12.3 26.2 16.8 19.7 OECD countries n.a. 10(.2 8.7 n.a. 1 2.4 12.3 n.a. Not applicable. S rt, 1 c: Psacharopoults 19'4. In almost all countries. rates of return to investmenit in all levels of educa- tion exceed the long-run opportuniity cost of capital (usually estimated at 8-10 percent in real tenrs). making education an excellent investment. Caution. it should be noted, is necessary when looking at rates of return. They can be misleadinig when, for instance, labor markets are heavily regulated and earn- ings do not reflect marginial productivity. Recent studies confirm the importance of education. and especially primarv education. for growth. Cross-country studies suggest the possibility of a thresh- old level of human capital accumulation beyond which a countr-y's growth tnav accelerate (Azariadis and Drazen 1990; Lau, Jamison. and Louat 1991). This cooncept is essenitially a reconfirmation of the original hypothesis that formial- ized a threshold-type relationship between huluanl capital and economic growth (Bowman and Anderson 1963; Easterlin 1981). Primary education is the single largest contr-ibutor to growth in both the cross-counitry and cross-regional com11- parisons and the within-country analyses calTied out to explain the East Asiani iitiracle" of developmnent (box 1.2). Differences il the educational level of the labor force explain about 20 percent of the differences in growthi across states tll Brazil. They suggest a threshold level of average education somewhere be- tween thiee atid four years of schooling (Lau and others 1993). a result con- firmed for Brazil usinig individual-level inforimiationl (Griffini and Cox-Edwards 1993) and corroborated for Guatetmala (World Bank 1994d). The East Asian high-growth countries invested heavily in both primary and secondary eciucationi in an effort to enhanice the quality of labor. This effort was EDUCATION AND DEVELOPMENT 23 BOX1.2 EDUCATIONANDECONOMIC The high-performing Asian econo- GROWTH IN EAST ASIA mies show a significantly higher rate of growth attributable to education than Primary education is the largest single all the other economies. When East contributor to the economic growth Asia and Latin America are compared, rates of the high-performing Asian 34 percent of the predicted difference economies (World Bank 1993a). In- in growth rates can be attributed to vestment in physical capital is second, higher investment levels and 38 per- followed by secondary school enroll- cent to higher enrollment rates. Simi- ments and population growth. These larly, the major difference between results are based on a 1 13-nation cross- East Asia and Sub-Saharan Africa is country regression that estimates the due to variations in primary school en- relationship between the rate of real rollment rates. Investment in physical per capita income growth, the share of capital accounts for only 20 percent of investment in gross domestic product the difference. (GDP). and educational attainment. complemented on the demand sicle by a pattern of growth that made productive use of labor and by complementary investment in physical capital. Substantial spending on education increased growth. For example, if in 1960 the Republic of Korea had had the same low school enrollment rate as Pakistan, its GDP pel capita by 1985 would have been 40 percent lower than it actually was (World Bank 1993a). Higher education also contributes to self-sustaining growth through the impact of graduates on the spread of knowledge (Becker 1964). Institutions of higher education have the main responsibility for equipping individuals with the advanced knowledge and skills required for positions of responsibility in government, business. ancl the professions. These institutions produce new scientific and technical knowledge through research and advanced training and serve as conduits for the transfer. adaptation, and dissemination of knowledge generated elsewhere in the world. Estimateci social rates of return of 10 percent or more in many low- and middle-income countries indicate that investments in higiler education contribute to increases in labor productivity and to higher long-term growth (World Bank 1994e). Not all of the extei-nal effects of higher education-such as the benefits from basic research and from technology development and transfer-are lully reflectedl in the earnings used in calculating these rates of return. The returns to higher education, as to basic education, are thus greater than those measured usilng earnings, and it is very possible that the contribution of higher education 24 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD to growth may increase with levels of technology and as countries achieve universal primary and secondary education. The external effect of education is important for economic growth and is suggested both by the possibility of a threshold effect at the primary level and by the likely spread of knowledge facilitated by higher education. The new theories of economic growth. like the older theories, show the complementarity of humanl and physical capital: a higher stock of human capital eiihances the rental value of machines; an increasing stock of physical capital boosts the efficiency of educational investment; and general investment plays a weak role in economic growth when not supported by education (Lucas 1988: Becker 1964). Empirical experience in East Asia demonstrates this complementarity. as well as the importance of sound macroeconomic policies in a broadly coin- petitive economy. The latter finding is reinforced by the experience of the former Soviet Union. There, rapid and sustained physical and human capital investnent led at first to rapid growth. However, excessive state intervenition in the economy. low capital-labor substitution, the nature of economic planning. and-perhaps most importantly-failure to allow the substantial investments in human capital to flourish and to stimulate qualitative improvements led to a lack of productivity growth and, in the long run, to stagnation (Easterly and Fischer 1994). Labor Market Linkages The dramatic recent shifts in labor markets brought about by economic re- forms. the integration of the world economy, technological change (especially in information technology), and migration have important consequences for education. International trade and the deregulation of economies and labor markets have not only contributed to growth but have also led to changes in the employment structures of advanced. transition, and developing countries. The rate of accumulation of new knowledge and the pace of technological change raise the possibility of sustainied growth and more frequent job changes during individuals' lives. Work tasks are becoming more abstract and more removed from the actual physical processes of production, which require less and less manual involvement. These developments have two important implications for education sys- tems. First, education must be designed to meet economies' iicreasing de- mands for adaptable workers who can readily acquire new skills rather than for workers with a fixed set of technical skills that are used throughout their work- ing lives. This need increases the importance of the basic competencies learned in primary and general secondary schools. Second, education systems-prima- EDUCATION AND DEVELOPMENT 25 rily at the higher and postgraduate levels-must support the continued expan- sion of the stock of knowledge. Major shifts in labor markets occurred during the 1980s. beginninig with a reversal of the trend during the 1970s of declining rewards to higher education in advanced market economies. Overwhelming empirical evidence suggests that the rewards to higher education are now increasing in many advanced countries (see, for example, Davis 1992). This trend set in at a time when earnings ineqLiality was growing at unprecedentecd rates and the average level of schooling in the labor force was very high. The improvement in the position of the more educated in advanced countries, even though their numbers in- creased. suggests that the demand for more educated workers has increased over time, causing an increase in the eamings premium associated with more schooling. Although schooling and earnings inequality are related, the earnings premium may increase despite an increase in the average level of schooling (or a decrease in the variance of schooling) if the demand for schooling has also increased. Recent techniological changes have involved both the deskilling of manv jobs that previously required some skills and a greater demand for workers to fill more highly skilled positions (Blackburn. Bloom. and Freeman 1990: Blackburn 1990). A decreased demand by employers for manual dexterity, physical strength, and traditional craftsmanship has increased the demand for educatecl over less educated workers, resulting in relative wage increases favor- ing more educated workers. Thus, technological change leads to increasing inequality (Bound and Johnson 1992). More educated workers can deal more effectively with a rapidly changing envi-onmenit (T. W. Schultz 1975: Mincer 1989: World Bank 1991cl). Highly educated workers are more likely thani unskilled workers to be found in new technology indListries. wher-e they are relatively better paid than in traditional industries: this comparative advantage holds true in high-, middle-, and low- income countries (Bartel and Lichtenberg 1987: Loh 1992; Gill and Riboud 1993). Schooling raises productivity in the market and in the houselholdl by en- hanicing information acquisition; it improves the ability to learn. But if returns trom investments in schooling are to be realized, the scope for productive learning has to be expanded thioughi technical innovation and changes in malr- ket and political regimes. The introdtictionl of new technologies can raise the returins to schooling if the new techinology increases rather thani decr-eases the nee(d for learning or the scope for misuse of inputs. For examliple. the "green revolutioln" in agricuilture led to an increased premium for acquisition of infor- mattion. The new high-vielding imported seed varieties that were the enginies of 26 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD growth of the green revolution were signilicanitly more sensitive to the use of such inputs as water and fertilizer. Famiers formerly engaged in "traditional" farming practices faced the challenge of mastering the appropriate allocation of inputs in order to realize large potential payoffs, and the continluing introduc- tion of new seeds every few years may have raised the returns to skills in informationi decoding (Rosenzweig 1995). Relative poverty is generally reduced as the labor force becomes mor-e educated. Evidence for some low- and middle-income countries indicates that equality in schooling corresponds to equality in earnings over the period of the 1980s, in contrast to the situation prevailing in industrial countries (see Patrinos 1994). An increase in the number of educated workers leads to decreased eainings differentials between them and the less educated. This eftect is re- flected in declining education-earninlgs premiums as education expands (Psacharopoulos 1989) and in declining wage differentials in the 1970s and 1980s in such countries as Brazil. Colombia, Indonesia, Korea, and Venezuela (Davis 1992; McMahon and Boediono 1992). Not only is the level of education important in adapting to rapidly changing labor markets: so is its content. It is often suggested, particularly at times of growing youth unemployment, that the school curriculum should be vocationialized or that technical skills should be taught in secondal-y school in order to equip school leavers for work in the modern sector. Skills training can indeed increase labor market productivity and earnings, but only when the skills are actually used in employment. International experience suggests that vocational and technical educationi and trainilng are most effective when they follow a sound general education and are job-related. In practice, many coun- tries. especially East Asian countries and members of the OECD, are moving to increase the technology content of general education and the general content of vocational education at the upper-secondary level and to provide many more course options. The effect on employment and earnings of this convergence of the two upper-secondary cunicula has yet to be evaluated. Comparative evalu- ations of earlier, more differentiated, general and vocational secondaly cur- ricula indicate clearly, however, that the rate of return to investment has been much higher in general than in vocational secondary education (Psacharopoulos 1987). The roles o'f the different educational levels are thus becoming clearer with these labor market changes. Primary and secondary schools focus on basic general competencies-language, science and mathematics, and, increasingly, communications skills, as well as the development of attitudes necessary for the workplace. These conipetencies provide the foundationi for subsequent edu- cation and training: even vocational secondary education is becoming more general. Further education and training tlieni consist of the acquisition of aca- EDUCATION AND DEVELOPMENT 27 demic and vocational skills, in institutions of higher education and in special- ized on-the-job training, with periodic updating through work-related continu- ing education (OECD 1994a). IPo%erti Redtietlion The low eanings of the poor are partly the result of their relatively lower human capital endowments and partly of labor market discrimination. Educa- tion can help with the first, but other steps are necessary to deal with the second. The differential in earnings between women and men in Latin America. for instance. is little explained by differences in human capital (Psacharopoulos and Tzannatos 1992). By contrast, human capital endowments explain most of the overall earnings differential between minority and majority male indig- enous workers in Bolivia and between monolingual Guarani speakers and Span- ish speakers in Paraguay. If the relatively poorer Guarani speakers of Paraguay had the same schooling level as Spanish speakers. the earnings differential would disappear. Education can therefore make a significant contribution to the reduction of poverty. It confers skills, knowledge, and attitudes that increase the productiv- ity of the poor's labor by increasing their output as farmers and. whenl discrimi- nation is absent, their access to jobs in both the fonmial and the infomial sectors. Studies have found that a farmer with four years of complete schooling has a much higher productivity than one with no education (Lockheed. Jainison, and Lau 1980: Moock 1994). Education also makes workers in industry more pro- ductive (Haddad and others 1990) and can contribute to entreprelieliurship (World Bank 199'1d). The creation of humani capital is the creation and distribution of new wealth. It contributes to the reductioll of both absolute and relative poverty, but it can take a whole generation to have an effect-in contrast to the more rapid effects of redistributinig existing capital. through. for example, tax reform and land reform. Resources invested today in education may lead to less poverty only after several years. when the poor whose human capital has been enhanced start to benefit from increased earnings, greater ability in self-employment. and improved efficiency in the use of household resources (T. W. Schultz 1982). In many developing countries the lInk between the labor market and the education system that is most important for the poor is the urban informal sector. In Sub-Saharan Africa during the 1980s. for example. about 15 million jobs were created in the informal sector. compared with only I million in the urbani modern sector. Since the poor often find it difficult to obtain employmenlt in the modein sector. increasing the productivity of workers in the infolriial sector is an effective way to reduce poverty (Moock. Musgrove. and Stelcner 28 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD 1990). In these circumstances, as for the modern sector, a sound general educa- tion may be more effective, and far more cost-effective, than providing specific vocational and technical skills, as it equips work-ers to acquire skills on the job. Studies on the determinants of earnings show that the early home environ- ment plays an imporlant role in the development of a child's intellectual ability. For example, pre-school-age children from lower socioeconomic groups per- form substantially worse on tests of cognitive development than do children from higher-income groups (Selowsky 1983). These differences can be attrib- uted to malnutritioni, lack of sanitation and health facilities, lack of parental stimulation, and other environmental deficits surrounding children living in poverty. It has also been shown that early childhood interventions (such as the provision of health care, education, and nutrition) can have a positive impact on the lives of children from poor backgrounds (Halpern 1986). Various attempts to equalize the opportunities of children from disadvantaged backgrounds have been made, but they often stalt too late. Research shows that by age three or four, children have already been conditioned by their family environment (Selowsky 1980: Young 1994). There is a need, therefore, to invest more in early childhood programs designed to enhance the growth and development of children (Myers 1992) and in subsequent programs to sustain the advantages provided by early interventions. Fertilitv and Health The more educated a woman, the lower her fertility (figure 1. 1; see also World Bank 199 Id, 1993t0. Education influences fertility through higher age at mar- riage for women and increased contraceptive use. For example, age at marriage has been rising steadily in North African countries, largely as a result of school attendance (Westoff 1992). In Honduras, Indonesia. Kenya, and Mexico schooled women desire fewer children, and they express this desire through a higher rate of contraceptive use. The more educated the parents. particularly the mother, the lower is mater- nal mortality and the healthier is the child. Parental education is significantly associated with the health status of'children (defined by a reduction in mortality or an improvement in chances of survival), even after controlling for socioeco- nomic status and access to health services (Rodriguez and Cleland 1980; United Nations 1986; Cleland and Wilson 1987; Hobcraft 1993). Rising levels of maternal education reduce the odds of the child's dying before age twvo (figure 1.2). in both urban and rural settings. On average, child mortality seems to fall by about 8 percent for each additional year of parental schooling, for at least the first eight to ten years of schooling (that is, includinig secondary as well as primary education). EDUCATION AND DEVELOPMENT 29 FIGURE 1.1 TOTAL FERTILITY RATE BY MOTHER'S EDUCATION AND REGION Total fertility rate 8 7 6 5 4 3 No schooling 1-3 years 4-6 years 7+ years Mother's educational level Africa (9 countries) Asia and Oceania (7-8 countries) Latin America (13 countries) Note Data are from dennographic surveys taken duiring the 1970(s and 198(s. The total fenility rate represenlt the iiunuber of children that would be boot to a woman if she were to live to the end of her childbearing years and hear children at each age in accordance with prevailinig age-specific fertility rates. Si,tgrc(c: United Nationis 1987. Parental education influences child mortality through the use of medical services (such as prenatal care and clinic visits) anid changes in household health behavior (such as washing hands and boiling water). These behavioral changes may result from perceptual and attitudinal changes and from the ability of the educated (whose incomes are higher than those of the uneducated) to 30 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 1.2 CHANCES OF CHILDREN DYING BEFORE AGE TWO BY MOTHER'S EDUCATION Percentage' 80 60 40 10(1 1-3 years 4-6 years 7+ \ears Mother's educational level Now. Datrn are trom : saimnple of iweiitv-five counitrie' ini Africa. Asia. ani Latini Amlerica.. .i. Chaincer of clitildren dying before a,ge two comilpared wit i chanices for children of iolihers witl no sowhooligi (reprCseltedl y tI 0) percnic oh seilica aXis). So'uicc: H-bcAfit 1993. afford better nutritioni and better health services for their childieni (Caldwell 1979: Lindenbaum. Chakrabortv. and Elias 1989: LeVine and others 1991). Even before taking accounlt of these etfects, the returils to investment in women s education exceed those to men's educationi for women who obtain employmilent (Psacharopoulos 1994). When the health and fertility externalities are addedl. the case for educating girls becomes even stronger. The benefit-cost ratio of these healthi and fertility externalities in Pakistan. for instance. has been estimated at about 3:1 (table 1.2). EDUCATION AND DEVELOPMENT 31 TABLE 1.2. THE EXTERNALITIES OF INVESTING IN GIRLS' EDUCATION, PAKISTAN Co.t 01' henef it hetm Ca/h Il/wrlOII (U.S. dofills ) Recurrent cost of one year of education for 1.000 womeni 30.000 Benefits Reductioni in child mortality Total deathis averted 60 Set cost (U.S. dollars) 80( V\alue of averted deaths 48.000 Reductioll in fertility Births averted 500 Set cost (U.S. dollars) 65 Value of births averted 37.500 Reduction in maternial mortality Total maternal deaths averted 3 Set cost (U.S. dollars) 2,500 Value of averted maternal deaths 7.500 S,w, e. Stiuimiimer. 1992. C H A P T E R T W O Achievements and Challenges THE EDUCATION systems of developing countries have made unprecedented progress in recent years. However, the future holds major chal- lenges for countries at all stages of educational and economic development. Some challenges are of crisis proportions. Enrollments are falliig in Africa. and there are still more than a billion illiterate adults in the world. The gender gap between boys' and girls' enrolliments is still very wide in the Middle East and in South Asia (where it has not closed at all in the past decade). In low- and middle-iniconie countr-ies the quality of education is poor. compared with OECD countries. Finally. as the pace of technological change quickens. there is a worrying lag between the reform of economic structLires and that of education systems, notably in the countries that have embarked on the transition from command to market economies. This chapter identifies these challenges; chap- ters 3 and 4 describe the ways in which cunenit patteins of educationi finance anid education manlagemenit are not fully appropriate for meeting them. East Asia's recor-d shows. however, that the challeniges can be met if the lessons of successful experience are adopted. The analysis is mainily regional. Each of the six regions definied by the World Bank for its operational purposes conltains a wide range of country conditionls, and the findings therefore do not apply to every couLntry in a region. (The regionis are describecd in the Definitionis and Data Notes at the front of this book.) The analysis has beenl severely hamilperecl by the poor availability and quality of data on educatioll and educationi fiianice (see the appendix to this 32 ACHIEVEMENTS AND CHALLENGES 33 chapter). In the graphics, the numilber of couLntries included in each region varies depending on data availability. Quantitative conclusions thus represent orders of magnitude and directions of trends rather than precise indications. Access The average level of education in developing countries is increasinlg. For the first time in world history, most cihildren at least start school. By 1990. 76 percent of the 538 million 6-to- 11-year-olds in developing countries were in school, up from 48 percent in 1960 and 69 percent in 1980 (UNESCO 1993a). These numbers reflect an increasing ratio of enrollments to the primary-school- age population during the 1980s in every region except Africa. At the second- ary level. 46 percent of 12-to-I 7-year-olds attended school in 1990( the propor- tion havinig increased during the 1980s in every region. At the tertiary level. enrollment ratios also increased during the 1980s in every region (figure 2.1). As a result of these gains, an average 6-year-old in one of the low- and middle-income countries in 1990 coulcd expect to complete 8.5 years of school, tIp from 7.6 years in 1980; the numiiber of years rose in every region except Afiica (figure 2.2). This impressive increase does not, of course. reveal any- thiing about the quality of education. The achievements in eniollmenlts are all the more remarkable when consid- ered in absolute terms (figure .3). as they occurred at a time of general fiscal restiainit and. in many regions, of rapid population growth. In Eastern Europe and Central Asia nine or ten vears of schooling is the norimi. In East Asia and in Latini America and the Caribbean primary education is ahlmost universal. Coun- tries in South Asia and in the Middle East and North Africa are also makine steady progress. althlough those in South Asia still have a conisiderable distance to go. Sub-Saharani Africa is not doing as well. With the easing of demographic pressuIre. coupledl with past success in increasinig access, especially at the primary level, prospects appear encourag- ing. The trends outlinied her-e give no reason for comiiplacency. however: * The absolute nulilber of cliildrenl in the wor[cl who receive no education at all is likely to increase in the next twenty years. * Only two-thirds of primary school students coniplete the primialry cycle. * Adult litera-cy appear-s likely to reimiainl a nijor probleim especially for women. * In part because of past success at the Plrinmry level. the demiand for second- ary and tertiary education is growinig faster thantl ivianly education systems caln accnll illodate. 34 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 2.1 GROSS ENROLLMENT RATIOS BY REGION AND LEVEL OF EDUCATION, 1980 AND 1990 Percent 12() L = Primiaarv 8(1 -l -.11 60 40- 0 12(: _ IOl {= 1'!9 Secondar' 60- 12() - 19811 1)9() Tertiary 800 8() _ 6() 61t) - 401 2 V AIriL LI A,i i nid Aiilliei; Eaqi aind Asizu IriLille - ioiltrics Iid lile (ellll andI tle Nonlih ie.nie l i3 ) Ar1i'./ The er&ivs eirllrNd iioedlt in D Io hle vetrh A ilrtnil edlrllio il)4. reeaiitlers t t .ic. ill 19 i9 lL3 level it ed.u;nli 1 e1 ti viilat)iizIiI ;Ige eiliilp 01 ildeiiilrle1eiii liilie illticill :DSeliiIg ne tiili levell dii edEiOIe I ill i NIl .Siiiiii- BrXeed riji ilnn iii Doripi-r iii Al ir;lil Edul. ljtiiii 19'4 ziidlEtl I 993n)1'i);. I 9931 ACHIEVEMENTS AND CHALLENGES 35 FIGURE 2.2 EXPECTED YEARS OF SCHOOLING BY REGION, 1980 AND 1990 Years I 5 (1 2 9- 6- Sub- Sah.aian Eam Eu rope Latiin Miiddle SouIth LIu- aind OECD Airi,a Ai nil Aerica. E.,'lmIdnl Aa,.i 1irarldle- oOLir1irlea lld [li e entrial aid tirc Noriih inl.Ilclle P:Rclii. AX,i Caii hheibbii AC11"I 0ccor111 _ 19)8) E= 1g)) S,,,,, B.r,rd in d,. uI.rn. D ,, , [ Irroi AI IO ,10 - EL 'rcrrr l rc 0t1ir 4 a I,r,Li INLNr r I l)Q A,i. I IN FIGURE 2.3 PERCENTAGE GROWTH IN ENROLLMENTS BY REGION AND LEVEL OF EDUCATION, 1980-90 Percenit I 5(1 20 --3) \iib-S.iirr iii Liar LLir,iaqi Lairi \!,rddi, Stair Lo-idr - (rl IN Dr SIrLLI anal ArI I,, t o ld Atm -V rlial,di - 0cIMiri 'Ca "' ' i iii ' a111t1 .err l.1'1t, 1i .rad n N ribh r i -11c \I ii I, b1,5 Ahla 11L-IMl.C Nyu. or Kiacil ii air), .i,,r, .1- Nr 153 irir)b 36 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD * The educational gap between the OECD countries and the transition econo- mics of Eastern Europe and Central Asia is widening. Demographic Pr essul-e Demographic pressure on enrollments will remain strong for the next decade but will start to ease in the next century as the rate of increase falls. The primary-school-age population in developing countries will increase by about 89 million children between 1990 and 2000 but by only 22 million between 2000 and 2010. Depending on when they started the demographic transition, some countries now face absolute declines in their school-age populations. This decline is already occurring in Eastern Europe and Central Asia and will occur in the first decade of the twenty-first century in East Asia and Latin America (figure 2.4). Many countries in these regions-for example. Colom- bia, Indonesia, and Korea-will experience declining school-age populations even before the region as a whole. This explains the fall in absolute enrollments at the primary and secondary levels in East Asia, as seen in figure 2.3. An opposite trend occurred in Afirica: there absolute primary enrollments increased, but not as much as the school-age population, so the gross enrollment ratio fell. In Africa, South Asia, and the Middle East and North Africa the school-age population will continue to increase. but more slowly in the first decade of the 2000s than in the 1990s. The main demographic pressure on enrollments will continiue to be in the three regions with the lowest enrollment ratios for girls and the highest fertility levels: Africa, South Asia, and the Middle East and North Africa. Between 1990 and 2010 Africa's primary-school-age population is projected to increase by 59 million, South Asia's by 28 million, and the Middle East and North Africa region's by 16 million. The Out-of-School Poplulation In 1990 about 130 million primary-school-age children-60 percent of them girls-were not enrolled in school. (The number had been 160 million in 1980.) The three regions with the greatest demographic pressure account for about two tlhirds of all children not enrolled in school. In Afirica 50 percent of all primary- age children are in this category, in South Asia, 27 percent. and in the Middle East and North Africa, 24 percent. The largest absolute numbers are in South Asia because of its large population (table 2. 1). School-age populations are growing in all three regions but almost twice as fast in Africa as elsewhere (see figure 2.4). Enrollnent ratios in Africa are low and are decreasing on average: only 46 percent of primary-age girls are in school. By contrast, primary enroll- ACHIEVEMENTS AND CHALLENGES 37 FIGURE 2.4 GROWTH OF THE PRIMARY-SCHOOL-AGE (6-11) POPULATION, 1990-2000 AND 2000-2010 Percent per year 3.0 2.5- 2.0- 1.5- 1.0 0.5- 0.0 -0.5 - 1.0 Sub-Saharan East Europe Latin Middle Soutli OECD Africa Asia and America East and Asia countries and tie Central and the North Pacifc Asia Caribbean Africa 1990-2n00 2000-2010 Sozwr C: World Bank projeclmons. ment ratios are increasing for both boys and girls in South Asia and the Middle East, although they are still very low. Unless the pace of enrollment accelerates. the absolute number of children not attending school at all is likely to increase in the next two decades-for the first time since 1960-reaching 145 million in 2000 and 162 million in 2015 (see table 2.1). This outcome would be brought about by continued high population growth rates, combined with falling enroll- ment ratios in some coulntries. Despite overall success, at least forty-two low- and middle-income countries have gross primary enrollment ratios below 90 percent (table 2.2). These countries are concentrated in Africa and South Asia, whichi contain all twelve countries with gross ratios below 50 percent and twenty-one of the thirty countries with ratios between 50 and 90 percent. These two regions also have the highest growtil r ates of school-age population. Gross enrollment ratios include overage students, but they do indicate trends, if not the absolute change. Net enrollment ratios-the proportion of children of primary TABLE 2.1 CHILDREN AGE 6-11 OUT OF SCHOOL, 1960-90 AND PROJECTIONS FOR 2000 AND 2015 (millions) /9h( /9,S() I 90 200091 20115 Re )wi,i Twal 11e1/fuah,b' noal Femau,le Tarll tenIclel Total Femauule 1l,,/ Fcuurle All develO)piLo CountlieS 65 96 158 94 12 9 77 145 85 162 92 (52) (62) (31) (38) (4) (29) (2) (27) (23) (2 7) SLil Sarharan Africa' 25 14 26 15 41 Ps 59 32 83 45 (75) (X2) (43) (49) (50) (54) (51) (55) (51) (55) Middle Eat, 9 5 9 6 9 5 1)) 6 1( 7 ((1) (72) (33) (43) (24) (31) (21) (27) (2 1) (26) Latini Amerwia and the C rihhean 15 7 9 5 8 4 7 4 7 4 (42) (43) (17) (18) (13) (13) (11) (12) (11) (11) Last Ahia 67 39 55 32 26 14 27 1-5 2 II (47) (56) (25) (30) (14) (16) (13) (14) (12) (12) Soutih Asia 49 30) 59 381 48 32 47 31 46 29 (56) (71) (40) (53) (27) (28) (23) (32) (2))) (27) \, u1,iii p.ilii-. il cI.I i il-.I I 'JI-c)luhildil a,ull . hlc I clll (lIhildl un I o ll(II ICeiuC lli(uiui. Regi,ml 'l1 l dol 1 u1,) i0t1,ll I)II IIICd01il cl ' llr u(IC' Cllluiell 1c-10l1 HC Iluu1dcLIL. 1)Ie<. I Iigilc :H i(It ;,.l t!.I II r lilrulll \le le phC ilIl.II! c LILICrdeiiii 1rgii1i .11 t ce 7. \iioiNo - I N I i i mll,,, cuIiI ie, Ji1e MCileiC II i IN ll[l SI h-S,llhero M IAi;l u 11i 1 Hdli ItiId IC Fzt. %,.1f,~ ~~IQ ; 1S } Ij.Ti ACHIEVEMENTS AND CHALLENGES 39 TABLE 2.2 COUNTRIES WITH PRIMARY GROSS ENROLLMENT RATIO BELOW 90 PERCENT, 1990 Giov Rfirlewl cmI,IUJ1oln10 Ss,, >s2/ir Pt,1( llt \ '" (')!()IflCIl Re f/l vi~ t'ijl/ioc 5(-90 persewut Sub-Saharani Africa Middle East and North Africa Beniin 67 Dem. Yemeni 88 Burunidi 73 Yemeni Arab Rep.' 76 Central African Rep. 68 Morocco 65 Chad 64 Saudi Arabia 77 Comiior-os 75 Latin America and the Caribbean C6te d'lvoire 69 Bolivia 85 Gambia. The 64 El Salvador 79 Ghania 77 GGuatemiiala 79 Guiniea-Bissan 60 Haiiti 56 Malawi 66 Mauritania 5 1 Be/ow 50 perl(e Mozambique 64 Stib-Saharan Africa Nigeria 72 Burkina Faso 37 Rwanda 71 Djibotiti 44 Senegal 58 Ethiopia 39 Sudan 50 Guinea 37 Tainizanlia 69 Liberia 30 UIganda 80 Mali 24 Zaire 76 Niger 29 Siena Leone 48 East Asia and Pacific Somalia 10 Papua New Guinea 72 Southi Asia South Asia Afghanistan 24 Bangladesh 77 Bhutan 25 Nepal 82 Pakistan 42 a BetortinI it It/clmn. Si, ,t.t Dounors to African Eduication 1t94: UNESCO 1991h 40 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD school age who are actually enrolled in primary school-would be a better measure of those not in school, but they are not readily available. Particularly alarming are Africa's falling primary enrollment ratios and, in some countries, falling absolute enrollments. Ratios have not fallen in all Afri- can countries. Of the thirty-five countries for wilich reasonably good data exist. gross primary ratios increased in twenty between 1980 and 1990. In fourteen others, however, including the most populous, the ratios fell, often by a large amount. Thus. the regional average ratio (calculated as the average of coulitry ratios) unweighted by population, fell only from 79 to 78 percent. This may seem minor, but Africa was the only region in the world to register a declining ratio. Because many of the individual countries in which the ratio fell have large populations. the weighted regional average fell from 80 to 69 percent. Clear evidence is lacking on the reasons for falling enrollments in many African countries. Civil disruption and war explain both declining absolute numbers and declining ratios in several countries, such as Angola and Mozambique. Populations are growing rapidly, despite the higih prevalence of HIV/AIDS (box 2.1), and in many countries the supply of education has been unable to keep up with demand, resulting in declining enrollment ratios. A reduction in overage studenits enrolled does not explain the fall: in six of the seven countries with declining gross ratios for which net ratios are available., the pattern of decline is confirmed. Where absolute enrollments have fallen, however, demand has also fallen due to low quality, poor employment pros- pects. the need for children's help with hotisehold work, and difficulties in paying fees and other school-related expenses (World Bank 1988). Whatever the explanation, even arresting the decline in enrollnent ratios will not be BOX 2.1 AIDS AND EDUCATION impact on the demand for education. Children who lose their parents to AIDS Human immunodeficiency virus (HIV), are often forced to drop out of school the virus that causes AIDS, is on the to survive. In Tanzania, for instance, increase. The World Health Organiza- the widespread prevalence of HIV/AIDS tion projects that by 2000 as many as is associated with the withdrawal of 26 million people could carry the virus girls from school and with marriage at andthat1.8millionwilldieofAIDseach an early age, eroding much of the year. Most victims are young, in their progress made in female education years of prime working productivity. (Ainsworth, Over, and Rwegarulira These deaths could have a profound 1992; Shaeffer 1993). ACHIEVEMENTS AND CHALLENGES 41 sufficient to prevent an absolute increase in the number of African children who do not attend school. Simply put, the rate of increase of the school-age population is higher than that of enrollments. Lou PrimaiaX Completion Rates About 30 percent of the children in developing countries who enroll in primary school do not complete it. More thani half the countries in East Asia and in the Middle East have completion rates above 80 percent, as do all countries in Europe and Central Asia. By comparison, only one-third of the countries in Latin America and only one-fifth of the countries in Africa and in South Asia have completion rates above 80 percent (UNESCO 1993b). Low completion rates mean that the proportion of children reaching grade 5 is roughly the same in Africa, South Asia. and South America despite very different enrolilment ratios in first grade (figure 2.5). Low primary completion rates result from high repetition and dropout rates. Repetition and dropout are closely linked: the first often leads to the second. although their causes are usually different. On the supply side, the low comple- tion rates may reflect problems with the quality of instruction. On the demand side. families may need children to work (for instance, in agricultural produc- tion) and may withdraw children, especially girls, from school temporarily- leading to grade repetition-or even pennanently. Dropout clearly affects learn- ing outcomes, but this may not be the case for repetition if students leamn more by repeating a grade (Eisemon, Schwille. and Prouty 1992: Psacharopoulos and Velez 1993). Repetition is. of course, costly to the system. And when a student repeats a grade more than once, repetition frequently leacis to dropout. Adult Illiteraev The combinationi of an increasing absolute number of children out of school and low primary completion rates means that the formial education system in the poorest countries is likely to continue to be inadequate as a mechanism for overcoming illiteracy. Overall illiteracy rates declined from about 55 percent of all adults in low- and middle-income countries in 1970 to about 35 percent in 1990, but this percentage still represents over 900 million illiterates, up from 840 million in 1970. Many more of the illiterate adults are women than men- a striking aspect of the gender disparities that still characterize many countries. Moreover. although illiteracy rates are declining, they remained at about 50 percent in Africa, the Middle East. and South Asia in 1990 and will not fall much below 40 percent in these regionis by 2000 (UNESCO 1990) without new interventions. 42 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 2.5 PRIMARY SCHOOL ENROLLMENT AND RETENTION BY REGION, ABOUT 1990 Percent 10( 80 60K 40 20 0 Sub-Salharain Middle South East SoUth Central Africa East Asia Asia America America and the Caribbean Percentade enerting grade I Perc.etlLge rea,hing grmde 5 Note. Data do not iicltide overage chilIdren anid are projected using reconstructed cohort aalsdysis. Reeions are those used by L NIC EF. souf)rce t[NICEF 1993. GrowinUin t Unmet Demiand tor Secondary and Telrtiary Education In most low- and middle-income coulitries substantially more students seek entrance to secondary and higher-level institutions than there are places avail- able. and the proportion of applicants to successful entrants is increasing. (For detailed evidence on Asia. for instance, see Tan and Mingat 1992.) At the tertiary level this gap partly reflects the provision of free or heavily subsiclizecl public education. At the University of the Punjab in Pakistan, 94 percent of those applying in 1 986 were not acimitted, up fi-om 91 percent five years before (Butt and Sheikh 1988). In maniy countries. such as Korea and Thailand. par- ents often pay for private tutoring outside regular school hours to increase their children's chances of admissioni. Repetition of the final year of a level-a form of queuinge for admission to the next level-is also common. In Mauritius more than 40 percent of secondary students repeat at least one grade to improve their chances of admission to higher education: in Burundi more than 70 percent of primary students repeat the final primaary grade. ACHIEVEMENTS AND CHALLENGES 43 The growing gap between demand and supply at the secondary level re- flects population growth, the increasing proportion of students completing pri- mary school, governments' difficulties in financing an expanded public sys- tem, poor parents' difficulties in paying school fees. and restrictions on private schooling. There is strong evidence around the developing world that many 12- to-17-year-olds at the secondary level are not in school because of a lack of places rather than a lack of interest (Holsinger and Baker 1993). In Tanzania. for instance, successful applicants to public secondary schools represented 11 percenit of primary school leavers in 1970 but only I percent in 1984 because Tanzania neither permitted private secondary schools nor expanded public ones. Since Tanzania started to license private schools in the mid- I 980s. enroll- ments have mushroomed and nlow exceed those in public secondary schools. This leap demonstrates the previously unmet demand for secondary education. A comparison with Kenya shows clearly that encouraging private schools can help accommodate the demanld for secondary education (Knight and Sabot 1990). The same pheniomilenlonl is seen in higher education. In Romania second- ary school graduates increased niore than 20 percent a year during thle 1980s. This increase stimulatecd an enormous pent-up demand for higher education that has lecl to the appearance of more than sixty private ulivelsities since such institutions became legal (Woricl Bank 1991b). The Widen ing Gapv between OECD Colunitries anditl the Transition Economies A wide gap in years of schooling separates OECD members and the transitional economilies of Easterin and Central Europe. Average "expected years in school," defined as the nuniber of years of schooling a child of six can be expected to complete. are conisiderably lower iil tile transitionial economilies thall ill OECD coulitries (see figure 2.2). Moreover. this average is a movin-g target: the ex- pected years an average 6-year-old child ini al OECD country will spend ill school rose from 13.4 in 1980 to 14.3 in 1990 and will continiue to increase in the 1990s. In Eastern and Central Europe expected years in school iicreased durinig the 1980s. but initial indicationis are that the level is falling il the transitionl economies in the 1990s. As the level of schooling iicr-eases in the OECD and falls in the tranisitioni economilies. the gap is widening. Equiity Girls, the rural poor. children from litnguistic and ethnic miniorities, nomads. refugees, street ancd workinig children. and childienl with special needs go to school less than others. In part. this reflects limited access, in part, lower de- mand. Despite an overall increase in the proportion of girls enrolled in school, 44 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD boys are still more likely to be enrolled. In 1990 an average 6-year-old girl in a low- or middle-income country could expect to attend school for 7.7 years, up from 6.7 years in 1980: an average 6-year-old boy could expect 9.3 years of education. The gap between boys and girls is widest in South Asia, where in 1990 a girl could expect 6.0 years of schooling and a boy 8.9 years, and in the Middle East, where a girl could expect 8.6 years and a boy 10.7 years. The gender gap is now very small in Eastemn and Central Europe and in Latin America, although such regional generalizations mask country exceptions sucIl as Turkey. In all regions except South Asia, the gender gap is closing (figure 2.6). The gender gap in school enrollments is. of course, not just a matter of access. In addition to a shortage of school spaces for girls. in many countries parents' demand for education for their daughters is low, reflecting both cul- tural nonns and girls' work in and around the home. Literate parents are more likely than illiterate ones to enroll their daughters in school, and the regions with the highest proportions of illiterate aclults are therefore those with the widest gender gaps. Overcoming the gender gap will require not only providing FIGURE 2.6 GENDER GAPS IN EXPECTED YEARS OF SCHOOLING BY REGION, 1980 AND 1990 Years of schooling 3.5 3.0 2.5 2).0 1.5 1.0 0.5 0.0 Suib-Sahatzan East Europe Latini Middle Southl Low- and OECD Africa Asia and Anierica East attd Asia ntiddlle- countrie, and [he Cenitral attd the Northi incorne PaciF-ic Asia Caribbeain Africa counitries 0 19_( 1 I I - - Soures: Based on dlata in Donors to African Education 1994 and UNESCO 993a. 1 993b. ACHIEVEMENTS AND CHALLENGES 45 school places for girls but also overcoming many parents' igniorance of the gains that will result from enrolling their female children. Rural populations are less educated than urban populations. Only 3 percent of Indonesia's urban population in 1980 had received no schooling, as against 10 percent for the rural population. In Venezuela in 1991, 95 percent of urban I 0-to- 14-year olds, but only 86 percent of rurlal children of the same age, were enrolled (World Bank 1 993e). Gender differences are particularly acute when disaggregated by urban-rural residence. In Pakistan in 1991 the proportion of girls and boys aged 7 to 14 years who ever attended school was 73 and 83 percent in urban areas, but 40 and 74 percent in rural areas (Sathar and Lloyd 1993). In Egypt 35 percent of rural people are literate. compared with 61 percent of the urban population (World Bank 199 Id). Approximately 60 per- cent of urban students in Colombia complete primary education, but only 20 percenit in rural areas do (World Bank 1990b). Relatively lower enrollments among the poor are most pronounced in higher education, largely as a consequence of inequities at the primary and secondary levels. For example, in the late 1980s, 63 percent of higher-educationl students in Chile caine from the top income quartile of households. and 92 percent of students in Indonesia and 77 percent in Venezuela came from the top income quintile (Tilak 1989: World Bank 1993c, 1993e). Linguistic minorities also suffer from relatively lower enrollments because they are often poor and because of language policies. Most countries are multi- lingual. either officially or in practice. More than 5,000 languages and dialects are spoken around the world, including more than 200 in Mexico and more than 400 in lndia and in Nigeria. Linguistic diversity reflects ethnic diversity and is often associated with high levels of illiteracy. In GLuatemala. for instance, 80 percent of the rural indigenous population is illiterate, and indigenous males in the labor force average only 1.8 years of schooling. In rural Peru, where the majority of the population is indigenous, 70 percent of Quechua-speaking people over the age of five have never been to school. compared with only 40 percent of nonindigenous Peruvians (Psacharopoulos and Patrinos 1994). Others who have difficulty going to school are nomads, refugees, street and working children, and children with learning and physical impairments. Refu- gees suffer from the unwilliigness of host governments to fund expenses for temporary immigrants. Street children suLffer from lack of parental guidanice, and working children because of the need to contribute to faamily income. Disease and malinitrition in developing countries result in a high proportion of children with leamning and physical impairments, estimated at 10 to 12 percent of all those under age 15. Most of tihese children with impairments come from poor familfies, and most lack access to schooling. Official estimates from devel- 46 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD oping coLintries suggest that of every hundred children with special needs. only one receives any form of schooling (Mittler. Brouillette, and Harris 1993). Quality in education is difficult to define and measure. An adequate defilition must include student outcomes. Most eclucators would also include in the defi- nitioni the nature of the educational experiences that help produce those out- comes-the learning environmenit (see Ross and Mhlack 1990). On both counts. the quality of education at all levels in low- ancd middle-income countries is not of the same standard as in OECD countries. although the lack of time-ser-ies data on outcomes makes it impossible to discern trenIs. Furthierimiore, students in low- and middle-inicome countries drop out and repeat more iltani those in high- incomtle countries. An important indicator of the Lquality of educationi is the value addecd of schoolinig-a measure of outcomes (Bridge. Judd, and Moock 1979: Lockheed and flanushek 1988). The value added consists ot leamning gain and the in- creased probability of income-earning activity. (The value addecd of higher- education also incluLdes research productivity) Cognitive learinig gain can be measured by achievement tests. Measurement of changes in the probability of income-earnine activity is extremely difficult, since it is affected by chanlges in the demanid for labor in an economiy. For example. a university graduate could receive a quality edlucationi in ancienlt languages, but there might be no clemanid flo such skills. Recent interilationial comparisonis of acilievenient have been malde of 9- year-old andl 14-year-old students in readintg and in mathiemiiatics and science. Althoughi most of the countries included in the comparisons are OECD mem- bers, enough developing countries were included to shiow thait test scores in developing couLItr-ies are lower-in some cases by miore than one standaid deviation-than the inter-national mean for all countries compared. The readinlg results for 14-year-olds in Botswana. the Philippinies, Thailanid, Trinidad and Tobago. Venezuela, and Zimibabwe shown in figule 2.7 illustrate this finding. In Burkina Faso and other Sahelian counitries, meani achievenienit scores some- times approachi ranidomniess, suggeSting that students are learning very little (Jarousse aid Min-at 1993). As strikinig as the lower mean score in developing countr-ies is the greater variation arounLd the meani, hoth of studenlt scores and of school scores. Some Venezuelan studenits, for instance. test as hiigh on reading as the internationial mean: others test in the bottom decile. In the Philippines 15 percelit of scilools scored higher than tile median for all countries in a test of achievement in general science (Lockheed, Fonacier, and Bianchi 1989). The variance in read- ing achievemenit in (levelopilng couLitries appears to be related to differences between urbani and rural schools. which are many times more pronounced than ACHIEVEMENTS AND CHALLENGES 47 FIGURE 2.7 VARIATION IN READING ACHIEVEMENT FOR 14-YEAR-OLDS IN SELECTED COUNTRIES, 1990-91 International icin FinlandI Firance I I I Sweden II New ZeakIand Switzerlanicd Singapore Lu niteLl States,, Iceland Slovenia German Deni. Rep.' ` I I Denmark Portugal Germ-1aniy, Fed. Rep.' I I Canada ( B( ) I I Norway I I I Ital1y I I Netliel l.laric Irelanldc Greece Sysi~~~~~~~~~~~i BCIYprUI11 I I I Spain Trio dad t)li(i OblgBil I I I Thanl.and Phihppines Venezuela I I I Zimbabwe Botswan I I I I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 51) 300) 35)) 401() 411 51))0 55)) 6)00 650 70)0 Reading achievement core Distribution of sc1hool merian achievement score Distribulton0 ot student aclhicvemen, score Ni,tl Fulr c.ich ciliiiiilv. lie vcertit.) eclici lline ill cacti hlil iiilic.iiee Ilie ciiie(l. the huIes p.ainilel il Ilie ceililer line repieselil + 1.9h ild -1 ."h sQ(I idlitadl errTors 01 NIIsiiliiig: Ilie enId p(lili (of IlIe ai' Illndicate lie 25th and TSih iiiieTlIlle pJlillel: ;1r-teilie e[tL - Olc lionzo lille ind xiendate10iIICI1nliile tIc' l r1ait 1)mii1 1,1 Ile i II d 0111 ITe111C aclhieiellieili scole. R etore tin) 1ic.it c1.ii b. Brilislt Cilninbi.i. c. Freichpl-eaking tiarea Sitlie IEA 1994 48 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 2.8 DIFFERENCE IN READING ACHIEVEMENT BETWEEN URBAN AND RURAL SCHOOLS FOR 14-YEAR-OLDS IN SELECTED COUNTRIES, 1990-91 Greece Philippirtes Zimbabwe Cyprus Hungary Trinidad and Tobago Indonesia Spain Thailand Slovenia Botswana Singapore Ireland German Dem. Rep.3 Germany, Fed. Rep.a' Venezuela Canada (Bc)t Portugal Norway United States Switzerland Denimark Iceland France Finlanid Belginlll' Italy S weden Netherlands New Zealand -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 Effect size No'le: The effect size ineasitres the difterenice betwee [Ilte inean scores for sitdents in urban il[d ruiral areas in relation to the pooled stanidar-d deviation. Anl tidex value less thani 0.2 can be cons idered insignificant. a differenice betweeit 0.2 anid 15. smaIll. anid ai difference greater than ().5, moderale to large. a. Before uniiication. b. British Columbia. c. Frenich-speakinl area. Solflr(: IEA 1994. ACHIEVEMENTS AND CHALLENGES 49 in advanced countries (figure 2.8). Raising quality will thus imply not only increasing average performance but also reducing variation across students and schools by making the greatest improvements in the leamning environmelits and performance at the worst schools. Delays in Reforming Education A more general and very disturbing issue is the lag between reform of coun- tries' economic systems and that of their education systems. As chapter I showed. technological progress is accelerating, and along with it the pace of change of economic structures. In these circumstances, delays in reforming the educationi system to keep pace with the economic system can imply lower growth and more poverty than would otherwise occur. This dynamic is particu- larly pronounced in the fornler socialist economies of Eastern and Central Europe, where many of the impressive educational legacies of the communist period are now threatened by austerity, uncertainty, and too slow a response by the education system to political and economic changes (box 2.2). The education systems that Eastern and Central Europe inherited from the socialist period were designed for a centrally planned economy that required labor with specialized professional, technical, and vocational skills. The result was a proliferation of narrow training programs. Because resource allocations weere determined by politically established plan objectives. there was little need for well-trained managers, for a skillecl labor force, or lor citizens capable of showing individual initiative. Studies in the applied social sciences and hu- maniities were discouraged. Teaching and learning practices allowed relatively little scope for independenit studies or for development of critical thinking skills. Nevertheless, the educational legacies of socialism are impressive. They imiclude almost universal adult literacy; universal access to primary and lower- secondar-y education: high average levels of educational attainment; significant reduction of unequal access associated with gender. ethnicity, rural residence, and socioeconomic status: the provision of high-quality compulsory eduication: the establishment of a large network of preschools; and international excel- lence in many fields of advanced scientific training and research. These accom- plishments are now threatened by austerity, by political and economiiic uncer- tainty, and especially by the slow responlse of education systems throughout the region to the emergence of participatory political systems and market econo- miiies and the consequent demand for new kinds of skills. The failure to adjust education systems is just as serious, if not as visible, in coLuntries in other regions in the f:ace of increasing global economic competi- tion and more open markets. These changes underline the need for a labor force with ever higher average levels of skills and knowledge and for more even 50 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD BOX 2.2 EDUCATION IN EASTERN AND In Poland, as in Russia, educational CENTRAL EUROPE DURING THE POLITICAL expenditures have declined as a pro- AND ECONOMIC TRANSITION portion of a shrinking GDP, even though Poland's economy is now growing. In Despite the paucity of reliable time- Bulgaria, Hungary, Romania, and sev- series data, there is evidence of de- eral other Eastern and Central Euro- cline in important educational indica- pean countries, education budgets tors in the transition economies during have been better protected, but public the 1990s. In Russia, for example, expenditures have declined in real enrollment in higher education con- terms. In Romania per student expen- tracted by 5 percent and enrollment in ditures in public higher education de- technical and vocational institutions by creased 36 percent between fiscal 9 and 7 percent, respectively. Pre- 1990 and 1993, although enrollments school enrollments dropped by 22 per- increased by about 44 percent. cent between 1991 and 1993. Between A vibrant private education sector 1992 and 1993 alone, total educational has developed in Bulgaria, Estonia, expenditures fell 29 percent in real Romania, Russia, and elsewhere as terms. In a country where state control an alternative to state-provided com- produced a high degree of uniformity pulsory and higher education. The cur- in educational financing, variations in ricula of many of these institutions educational expenditures by rich and emphasize foreign language training, poor localities are increasing. management, and other market- distribution of these skills within the population. The East Asian countries., which have generally invested heavily in basic human capital for both men and women, are outstanding examples of what can be achieved when the education system is reformed along with the economic system. Appelwfix. 'I'he Povertv of' Edticationi Data Data and research on education are generally insuifficient for monitoring. policymaking, and resource allocation. In Syria. for instance. 50 percent more students recently completed secondary school than had been estimated, with enonnous repercussions for the higher education system. Estimates by Uganda's Ministry of Finance showed 85.000 primary-level teachers in the system in 1992, while the Ministry of Education counted 140.000 (Puryear 1995). In Mauritius the rationality of the refonn of basic education in the 1990s was undeimined by the poor quality and analysis of data on education (Bhowon and Chinapah 1993). ACHIEVEMENTS AND CHALLENGES 51 oriented skills. Nevertheless, legisla- embraced, either formally through leg- tion to accredit private institutions islation or informally by the introduc- and recognize the qualifications they tion of new fees. award has been introduced only re- In most countries in Eastern and cently in a few countries, such as Ro- Central Europe. however, structures mania, which by 1994 had sixty-six for administering and allocating public private universities. funding for education have remained Reform efforts during the first years largely unchanged, notwithstanding a of transition focused on depoliticizing plethora of reform proposals. One con- school curricula and management, re- sequence of the lack of comprehen- establishing the political autonomy of sive reform is the increasing reliance universities, redefining the educational on decrees and regulations for man- rights of ethnic and linguistic minori- aging education systems. In Romania, ties, and, especially in Russia, increas- which has not as yet passed an or- ing local control of schooling. Guaran- ganic law on education, the govern- tees of employment for graduates of ment has found it necessary to issue the education system were abrogated, more than 2,000 temporary decrees as were policies that mandated state- and regulations since 1990 to manage owned enterprises to provide and fi- the higher education subsector (Eise- nance various education and training mon and others forthcoming; Laporte activities. The principle of cost-shar- andSchweitzer1994:Vlasceanu1993; ing in noncompulsory education was World Bank 1994k, 19941). The problems arise mainly because: Existing education statistics are -euerally not reliable. * Statistics are often out of date and hence of limited use in informing policy decisions. i Statistics are often collected as a matter of course, with too little critical reflection on the underlyinig theoretical framework, the comparative perspec- tive, and the purposes for whiich the data are intended. * The infonnation collected focuses more on counting inputs than on assessing achievement and monitoring labor market outcomes. * Educational research is usually not available or is not used to complement statistics in monitorinig education systems. Efforts to improve the situation are under way in many coUlntr-ies. The OECD initiative to develop a limited set of comparable indicators of national educa- 52 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD tion systems is a large cooperative project designed to improve the reliability, timeliness. policy relevance. and comparability of a core set of essential statis- tics on education finance, expenditure, and student achievement (OECD 1993, Tuijnman and Bottani 1994). Similar initiatives are being launciled elsewhere. particularly in Asia. And some countries in Latin America have reasonably good data. Although laudable, these efforts do not go far enough because they do not address the principal causes of the problem in a global perspective. In most countries there is little incentive, and often little funding, for the collection and analysis of data, especially that needed for the assessment of learning, and the monitoring and evaluation of educational developments. In maniy countries fear of the potential political repercussions of reporting negative trends and weaknesses in the education system is an impediment. Intemnationally, global leadership is lacking. UNESCO. tor instance, compiles intemnational statistics supplied by its member countries but does not verify them. A major interna- tional cooperative effort, spearheaded by UNESCO and the World Bank. to improve education data and research in developing countries is nlOW beginning. This work is perhaps similar to the efforts in the 1950s that led to intemnational consensus on the value of using the United Nations system of national eco- nomic accounts-a practice that continues to the present. C H A P T E R T H R E E Public Finance for Efficiency and Equity PUBLIC finance is the main instruLIment for implemeniiting public priorities, and there is a strong rationale for public interventioni in the financing of education. In general, public investment accounts for about two-thirds of all education spendinig, although the share varies from as much as 93 percent in Hungary to below 50 percent in Uganda (table 3. I). Public spending on edLuca- tion is often inefficient, however, when it is misallocated across levels and within levels, and it is inequitable when qualified potential students are unable to enroll in institutions because there are no educational opportunities available or because they are unable to pay or to obtain financing. The Rationale tor- Public Finance The high private rates of return to investments at all levels of eclucation justify large investments by individuals. They also justify self-financing bv families or students. through immediate or defelTed cost-sharing. Despite these highi pri- vate returns and the justification for private finance, there is also a strong case for public intervention, especially for basic education, for reasons of income distribution. capital market irnperfections. information asymmetries. and exter- nalities. In fact most governments are heavily involved in all levels of educa- tion-an activity which in many cases takes up a significant portion of public expenditLures. 53 54 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD TABLE 3.1 EDUCATION EXPENDITURE BY SOURCE OF FUNDS, ALL LEVELS OF EDUCATION COMBINED, SELECTED COUNTRIES, 1991 (percent) Gr(o,ip and il o04li1? V P7blitc sutces Prjis otet solrtc (es OECD c ur1tries Australia 85.0 15.0 Canada 90.1 9.9 Denmiark 99.4 0.6 Finland 92.3 7.7 Franice 89.7 10.3 Germiany 72.9 27.1 Ireland 93.4 6.6 Japan 73.9 26.1 Nethela-nds 98.( 2.0 Spatini 80.1 19.9 United States 78.6 21.4 Lovv- cilici widdle-i#tcomnie countrniess Haiti 2 0. 0 80.0 Hungary 93.1 6.9 Indiia 89.0 1 1.0 Indonesia' 62.8 37.2 Kenyab (1992/93) 62.2 37.8 Uganda (1989/90) 43.0 57.0 VeniezuLela (1 987) 73.0 27.0 a. Puhli.c itiittions oniv. Privste sources refer o hotusehotolds olr. h. Pri.laiv and secondary leve ls only. Private sources refer to household' onily. Si ,ur,s : Noss 1991: nEC[) 1993: Tilak 1993: World Banik 1993c, 1993e. 1994g. 199411i. Income Distribition Education can reduce income inequality by promotiiig productivity gains in agriculture and facilitating the absorption of labor into the modern industrial sector. Equality of distribution of education usually results in equality of distri- bution of income. Education opens new opportunities for the poor and so increases social mobility. Public spending on basic education defilitely helps the poor, for two reasons. First, because the poor tend to have large families, a larger subsidy accrues to a poor family than to a rich one. Second, the rich ma' opt out and buy private education. again increasing the amount of the subsidy that flows to the poor. PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 55 Not all groups in society can afford the direct and indirect costs associated with investing in education, and so the state has a role in promoting equality of opportunity. If education were provided under market conditions, only those who could afford to pay tuition fees could enroll. Not only would there be underinvestment from the social point of view, but income inequalities would be preserved from one generation to the next, since education is itself a deter- minant of lifetime income. Capital Mark-et hInpetfertions The private purchase of schooling. especially of higher education. is beyond the means of many poor families. Most credit markets do not provide an effec- tive solution because of strong imperfections that reduce participation, particu- larly by very poor people. In principle, the budget constraints can be overcome by borrowing, given the high private rates of return to education. However, there are high risks for both borrowers and lenders in educational financing, and banks do not accept the promise of future earnings as collateral. The failure of the capital market thus affects not only lower-income groups but also middle- income groups that cannot finance tertiary education without credit. Info)rmation Asvmmetries Parents with little education tend to be less informed than better-educated parents about the benefits or quality of education. In the United Kingdom working-class parents tend not to encourage their childr-en to aspire to an Llii- versity education (Barr 1993).The capital market for education is far from perfect. Students from poor households are understandably reluctant to saddle themselves with debt or to enter into fixed obligations because they do not kniow their future incomes. Furthermore, those from poor backgrounds tend to underestimate their prospects. Lenders hesitate to accept risks backed only by the uncertain future incomes of reluctant debtors (Arrow 1993). Externalities The benefits of education accrue not only to its direct recipients but also to society at large. In the absence of govenilmenlt provision, expenditures on edu- cation are smaller than would be desirable. According to an adaptation of new growth theory, a worker's productivity is affected by the average level of human capital, as well as by the worker's own human capital (Lucas 1988). Widespread public education at the basic level may be a thieshold for develop- ment. The optimal distribution of education for maximizing the spillover effects 56 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD associated with human capital and taking advantage of these potential thresh- old levels would appear to be a distribution that is equitable. External effects on health and fertility are not maximized on the basis of private spendilig alone but can be captured for society as a w hole through public spending. Misallocation amonng Education Subsectors In low- and niiddle-inconie coulitries the rates of return to investmelits il basic (primary and lower-secondary) education are generally gleater thanl those to higher education. Therefore basic education should usually be the priority for public spending on education in those countries that have yet to achieve near- universal enrollment in basic education. Indeed, in most countries the highest share of public spending on education goes to primary education (table 3 2) In all regions except South Asia, moreover. the share of public education spend- ing going to secondary education increased during the 1980s (figure 3.1). reflecting growing enrollments and the near achievement of universal primary education in several regions. Few low- and middle-income countries, except those of Europe and Central Asia arid some colintries in East Asia and the Middle East, have achieved near-Liniversal secondary education. Hence the increasing share of public spending going to higher education during thie 1980s in regions without high primary and secondary enrollment ratios is ulilikely to be efficient, as the rates of return to primary and secondary education are likely to be higher in most countries. Similarly, the declining share of public spending going to higher education in the Europe and Central Asia region may TABLE 3.2 PUBLIC RECURRENT EXPENDITURE ON EDUCATION BY LEVEL, 1990 (percent) Regiion P-i-m,vl SC'colmde v Tc,,licn v Low- and middle-income countries Sub-Saharani Africa (22) 42.9 29 8. t 9.7 East Asia and the Pacific (4) 41.3 30.5 14.8 Europe and Central Asia (5) 49.3 26.8 15.9 Latin Amilerica and the Caribbean (I 11) 39.4 28.5 18.4 Middle East and North Africa (3) 36.0 41.5 16.1 South Asia (3) 41.5 30.4 13.9 OECD countries (15) 30.7 39(0 20.6 N'or LITlrWeighted averages: figuires in pareritheses reter to Ihe Tlumilber of counurie, ii [he regionill saiple. Sour, es: DOTnors to AfriCIIIi Education I 1994: LI NESCo dalabase. PUBLIC FINANCE FOR EFFICIENCY AND EOUITY 57 FIGURE 3.1 CHANGE IN ALLOCATION OF PUBLIC RECURRENT EXPENDITURE ON EDUCATION BY REGION AND LEVEL, 1980-90 Percent 2( 10 5 0 -5 -10 Sub-Saharan East Europe Latihn Middle South OECD Atrica Asia and America East and Asia coun1tries (22) and the Central and the Norlh 4) (15) Pacitic Asia Caribbean Africa (4) (5) (l l) l 3) Primary Secondaryv Tertiary N,, Percenrtige, are unweighted a,erages. FigLLres in paienlhte,e ieter to tie niLmber ot colntriie, ii the regionial samrlple. 5,nm eN- >:,Donor to Atrican EductlolIT I J)Q4: LINESCO dolatiase. be inappropriate. depending on1 the returnls to investmnenit at different levels of education. Even thougil spencling per higher education stLidenit fell as a proportion of spending per primar-y student in all regiotis (table 3.3), subsidization of higher education is still very high. This subsi(dization increaises the demnand for higher education, even though edLIcationl at th,at level is generally less efficient for society as a whole in countries that have vet to achieve uilliversal priimiary and 58 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD TABLE 3.3 PUBLIC SPENDING PER STUDENT: HIGHER EDUCATION AS A MULTIPLE OF PRIMARY EDUCATION, 1980-90 Region 19&) 1990 Low- and middle-income countries Sub-Saharan Afiica (8) 65.3 44.1 East Asia and the Pacific & South Asia (4) 30.8 14.1 Latin America and the Caribbean (4) 8.( 7.4 Middle East and North Africa (2) 14.6 8.2 OECD countries (15) 3.0 2.5 Nowt: Unweighte(d averages: figures in parentheses refer to tile number ot couintries in the regionial sample. Source: LNESCO database. secondary education. The subsidization of higher education is most acute in Africa. Although private rates of return to higher education are 2.5 times higher than social rates (see table 1. I ), public spending per student in higher education in Africa is about 44 times spending per student in primary school. The share of tertiary education in public spending on education is higher in Africa than in any other region. at the samne level as in OECD countries, according to the lUNESCO database. Misallocation within Ecdltcationi Subsectors Inefficiencies, which are prevalent within all education subsectors, reflect an inefficient mix of inputs, such as staffand instructional materials. They can also result from high repetition and dropout rates. For effective learning, the input mix inevitably varies from country to country and institution to institution. according to local conditions. Important broad guidance about the internal efficiency of education systems can come, however, from international coin- parisons and interschool comparisons, especially with regard to student-teacher ratios and school buildings. The student-teacher ratio is one overall measure of staff efficiency, al- though it excludes nonteaching staff and shows systemwide averages, not ac- tual class size. To take one example, China's student-teacher ratio is 25:1 at the primiary level and 17:1 at the secondary level, compared with an average in Asia of 34:1 and 23:1. Chinese teachers teach for only 12-18 hours a week. compared with 20-25 hours a week in other countries (Tsang 1993). Schools in low- and middle-income countries could save costs and improve learnmig by PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 59 increasing student-teacher ratios. They would thereby use fewer teachers and would be able to allocate resources from teachers to other inputs that improve achievement, such as textbooks and in-service teacher training, as discussed in chapter 4. (In practice. such savings are rarely allocated to other inputs.) Yet in all regions except South Asia primary and secondary student-teacher ratios are decreasing (figure 3.2). In Africa the number of teachers increased by 24 per- cent between 1985 and 1990, while the enrollment ratio declined by 3 percent (Donors to African Education 1994). The scope for improving efficiency through modest increases in student- teacher ratios is enormous because teacher costs typically account for about two-thirds of total spending on education (UNESCO 1993b). In Botswana a gain of one year of learning per grade could be achieved in junior secondary school by reducing class size (requiring more teachers) at a cost of $9,414 per grade, or by introducing supplementary reading materials at a cost of $727. or by in- creasing in-service teacher training, at a cost of $328 (Fuller, Hua, and Snyder 1994). Some countries, such as Bangladesh, Malawi, and Namibia, where the first two grades often have more than sixty students per teacher. would benefit greatly from significantly reducing rather than increasing class size. School buildings are not entirely necessary for obtaining desired academic outcomes. Indeed, the first "academe" in Europe was a public grove of trees where Plato taught. Even today, learning occurs despite the absence of build- ings in many countries, includinig parts of rural India. Hlowever, school build- ings are everywhere acknowledged as conventional locations for teaching and learning. There are many opportunities for increasing efficiency in construct- ing and Using buildings, thus saving resources for other purposes. Many coun- tries. particularly those in Afi-ica with colonial legacies, adopt expensive design standards and use imported construction materials. This is evident in compara- tive construction costs for World Bank postsecondary education projects in Africa and Asia in the early 1980s: the estinated total construction cost per boarding place in nonuniversity eclucation in West Africa was nearly double that in South Asia and 50 percent higher than that in East Asia and the Pacific (Singii 1990). In some African coLintries the annualized capital costs of new school facilities are as much as 80 percent of annual recunenlt costs (World Bank 1988). Constructioni costs can be cut by simplifying designs and by using appropri- ate materials and comnmuniity labor, supervised by trained engineers to maintain safety standar-ds (for example. earthquake-proof construction for certain regions). Cost-sharinig with communities in school construction is coiimmon, especially at the primary level. These approaches have brought down costs in World Bank projects in India. Mexico. and Senegal by as much as 50 percent. Flexible tloor plans can also improve space utilization by accommodating changes in enroll- 60 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 3.2 PRIMARY AND SECONDARY STUDENT-TEACHER RATIOS, 1980 AND 1990 Number of students per teacherf 50 Primary level 40- 30 - 20 _ 1980 1 199( 50 Secondary level 40 30- 20- SLub-Saliarani East ELlrope Latiti Middle Souttl OECD Africa Asia and Americat East and Asia coulitries and the Central and tlie North Pacific Asia Caribbean Africa =3 1980 C 199( a. Liweighted averages. .S,,,e. Based on data in U1NESCO) 19'J1b. PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 61 inents. For example. multipurpose school buildings in Bangladeshl have renmov- able partitions to permit differenit class sizes and accommodate large school and commnuniity gatherings. There are, however, tradeoffs between flexibility and instructional effectiveness. It is also critical to maintaini a building's physical plant and equipment. yet frequently sufficient funds are not allocated. This oversight has been a particu- lar problem in Africa. where responsibility for school maintenance often rests with the central government rather than the local level (World Bank 1988). More intensive use of existing school buildings can reduce the need for new school construction. In Jordan a systematic program of school consolidation has led to the closinc of about 1.000 schools. In Thailand lower-secondary classes now share with primary classes buildings that were previously used only as primary schools. Anotlher way sclhools can be used more intensively is to adopt multiple shifts, which reduce per student capital costs. For example. the per student costs of school construction and furnishing in Jamaica are J$1,500 for single-shift schools, J$1,139 for overlapping double shifts, and J$1 ,027 for end-on clouble shifts, in which the second shift starts after the first- shift pupils leave (Bray 1990; Leo-Rhynie 1981). Project preparation estimates in Zambia show that maxilumll use of multiple shifts in primary grades I throLIgh 7 can reduce per student construction costs by half (Bray 1990: Kelly and others 1986). Multiple shifts can reduce achievement. however, if they lead to reduced instructioll time per student. Multiple-shift schools therefore often increase the niumber- of days of school attendance per year to compensate for a shorter nuImber of hours per day. Such techniques have been used extensively and successfully in Korea, Malaysia, and other East Asian countr-ies. If annual instructional houL-s are maintainied, there is no significant loss in quality com- paredl with single-shift schools (Bray 1990; Leo-Rhynie 1981). Multigrade teaching, in which one teacher instructs more than one grade. can be cost-effective in rural areas, where teachers are often scarce and classes are often small because few childreni are in a particular grade. A successful example is the Escuela Nueva programii in Colombia (Thomas and Shaw 1992). Multigrade teaching can reduce the costs of repetition and dropout if it enables students to repeat only the parts of the curriculum they found difficult. The overall costs of multigrade teachinlg are highler than for single-grade schemes, however, because of the need for special teacher training, study guides, and teaching materials. In Colombia these needs raise unit costs by 5 to 10 percent comparecd with costs for single-grade teaching, largely because teacher training is thr-ee times as costly'. Since learinig achievement in language and mathiemat- ics is higher, however, the extra costs are justifiedl by the benefit-cost calcula- 62 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD tion (Psacharopoulos. Rojas. and Velez 1993). In Colombia and other Latin Americani countries multigrade schools tend to perforn better thani single- grade schools (Velez. Schiefelbein. and Valenzuela 1993). Multigrade teach- ing is not successful where single-grade techniques are simply applied without adaptation to a multigrade class, as in Pakistan. At the higher level. buildings can sometimes be dispensed with altogether. Open universities for motivated higher education students, for instance, are much less costly than conventional universities. In Korea the unit costs for distance learning are only 10 percent those of residential students. Similar results were found in Thailand (14 percent), Pakistan (22 percent), and China (50 percent) (Lockheed, Middleton. and Nettleton 1991). The lower costs result from the much higher student-staff ratios. Dropout rates are very high in dis- tance higher education courses-typically 50 percenit or more-and so costs per graduate are double the costs per student. In China, where dropout rates in distance higher education are 69 percent, the unit cost per graduate is higlher than in a conventional university. A proper comparison of distance and conven- tional higher education would involve calculating the returns to each. This comparison has not generally been made, largely because of the unavailability of data on graduate earnings according to type of higher education institution. It has. however, been calculated for Thailand, where the cost per studenit at open universities is only one-fiftieth that at conventional ones but the earnings of open uLniversity graduates are only 2 percent less, on average. than those of' private university graduates (Tan 1991). Repetition and dropout are a further result of inefficiency. although their causes are complex and some repetition can even improve achievement. They are most pronounced in Africa and Latin America but are declining in both regions. The simple solution to repetition is automatic promotion. However. this solution is ftrequently not possible, when repetition is a form of queuing for entry to the next level of schooling, or not desirable. when repetition occurs because students have ftailed to master certain skills. Improving access andc quality are, in general. more appropriate solutions to the problems of r epetition amid dropout. Although public spending on primary education generally benefits the poor. total public spending on education in low- anid middle-income countries often favors the affluent, largely because relatively fewer poor childien attend sec- ondary and highler education institutions. In developing countries as a whole. 71 percent of school-age children share only 22 percent of overall public re- PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 63 sources for education, whereas the 6 percent who receive higher education get 39 percent of public resources (Mingat and Tan 1985). Various criteria can be used to assess the equity impact of public spending. A weak technical criterion is whether the poor receive a share of the public financing subsidy larger than their share in national income. If so. the distribu- tion of the subsidy improves the distribution of real income, and the relative size of the per capita stibsidy expressed as a proportion of per capita income is larger for the poor than for the better-off. A stronger technical criterion is whether the poor receive a share of the subsidy that is larger than their share in the population, meaning that the absolute size of the per capita subsidy is larger for the poor. An even stronger. and better, criterion is whether public expendi- ture, including loan guarantees, is arranged so that no qualified student is unable to enroll in education at any level because of inability to pay. There is no simple measure of this criterion, which hinges on the existence of an appropri- ate selection mechanism at the postcompulsory level that can be used to define .qualified." In the absence of such a mechanism, the weaker technical measure can be used as a minlimal criterion. In Indonesia in 1989, Kenya in 1992. and Colombia in 1974, Lorenz curves for the distribution of the education subsidy compared with the distribution of personal income show simi lar patterns (figure 3.3). The total education subsidy is more evenly distributed than personal income: its Lorenz curve lies above the income distribution curve. Overall, the education subsidy only weakly favors the poor, however, because it lies below the 450 diagonal indicating equal shares of the total subsidy. The only line above the diagonal in all three cases is the curve for primary education, which shows a strong pro-poor distribution: lower-income individuals receive a larger share of the primary school subsidy than their share in the overall population. The secondary and tertiary education subsidies do not even remotely favor the poor because so few poor children are enr-olled at these levels. Not only does public spending on secondary and ter- tiary education favor the better-off in absolute tenns (their Lorenz curves lie below the 450 diagonal); it is less equal than even the distribution of personal income. Rich households receive a larger share of postprimary education sub- sidies than their share of total income. Given these results for the relatively weak technical criteria, it is clear that education spending by the public sector is very inequitable in terms of the more intuitive criterion that no qualified person should be unable to enroll in education because of inability to pay. It is possible to reallocate total public expenditure by encouraging the enrollmenit of the poor so that spending is no longer biased in favor of the affluent. During the 1970s and 1980s Colombia increased poor family enroll- ments and improved the targeting of its spending on secondary and tertiary 64 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 3.3 DISTRIBUTION OF SUBSIDIES FOR EDUCATION IN COLOMBIA, INDONESIA, AND KENYA, SELECTED YEARS Colombia, 1974 Colombia, 1992 Percenitage of income or subsidy Percentatge of incomiie or subsidy 80 - 0 8 60 -60 - 40 - --40 ~~O - -20 00 0 20 40 60 80 10n) 0 20 40 60 80 100 Percentage of houseliolds Percentage of houIsehiolds Indonesia, 1989 Kenya 1992-93 Percentage of expenditure or Subsidy Percentage of expenditure or subsidy 100 100 8(1 - 8( - 60 - 601 - 40 -> J 40 - S 20 - 2(1 2(0 40 6(0 80 1(1(1 (I 2() 41) 6(0 81) 1(1( Percenitage of population Percentagc of population Cumulative primary subsitly -- Cumulative seconclary CumLIlative upper secondary SLbsidy' subsidy Cumulative tertiar-y subsicdy C1.1 Cumulative subsidy foir all ------ CumLIlative incomile education or cumiulative expenditUre a. For hidoniesia. lower secodcary. tilpper secLILaIiry subsidy iS show ii f'o Indonlesia only. Som,e.r: FLo ColomLbiaW. 1974. SeIon sky 1979: tor Colombia. 1902. World Batnk 1 954a: for lihirdori:.a Wold Baniik 1993c: loI Keirva, Worldl Banik 19941. PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 65 education. As a result, education spending as a whole in 1992 benefited the poor. even thoughl spending on tertiary education continued to favor the upper socioeconomic classes (see figure 3.3). Public expenditure is also biased against the rural population. This pattern is consistenit with the bias againist the poor. as the incidence of poverty is usually greater in rural than in ulban areas. In Indonesia in 1989. for instance. the monthly subsidy for all education programs averaged Rp.l.520, but the average for the urban population was Rp. 1.894, compared with Rp. 1.366 for the rural population (World Bank 1993c). In China primary education in the rural areas that contain 70 percenlt of the population is financed mainly by parents and communities through cash and in-kind contributions to teachers' salaries and school construction. Urbani primary and secondary schools are financed by provincial, town, and district govenmments. Chinese uLiiversities charged no tuition before 1989 (Tsang 1993). Spending on higher edLication also illustrates the bias against the poor. Spending more public funds per higher education student than per primary student is inefficient in most countries because the social returnls are generally lower to higher educationi than to primary education, at least in countries with less than universal primary and secondary eniollimenits. It is also inequitable: those students who gain access to higher education receive a larger absolute subsidy thani those at lower levels, and higher educationi students come dispro- portionately from richer families (table 3.4), which are better able to pay for higher studies. Yet public higher edLIcation is free, or almost free, to students in most countries. In only twenty developing countries do tuition fees account for imore than IO( percent of recunent expenditures. There are importanit regional TABLE 3.4 HIGHER EDUCATION STUDENTS BY FAMILY INCOME (percentage oftotal enrollment) -ndn. JI om _o 2) /It e ,( Z/f couni,citi)I %vt oIho/ v^ehbold bv in, onle Chilc 1987 63h Colombia, t979 67 India. 1987 45 tndonesia. 1 989 9-2 Japan. 1987 46 MiavIy,Jiat. 1979 48 LJIniited States. 1987 37 Venezuela. 1986 77 zI Top 2s per,enil 1t hd byN HiLn0oe. Til,ik S). 1994: \W', ILI Bzi,ik 10")3c. 1'9'). 66 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 3.4. PUBLIC EDUCATION EXPENDITURE AS A SHARE OF GNP AND OF CENTRAL GOVERNMENT EXPENDITURE, 1980 AND 1990 As a share of GNP Pet-cent 6 5 4 3 2 Sub-Saharan East Europe Latin Middle South OECD Africa Asia anid Anenica East anid Asia counitries (25) and the Central aid the North (4) (19) Pacific Asia Caribbean Africa (6) (8) (15) (8) As a share of central government expenditure Percent 20 15 10 5 0 Sub-Siaharan East Europe Latin Middle Souith OECI) Africa Asia zilid America East anid Asia countries (6) and the Central and the North (3) (13) Pacific AsNia Caribbean Aftica (4) (3) (6) (4 1 1980 1 199( Note: Percentages are 1ri Wcighted ax rages. Figures in paretiilxeses refer to the imumi her otf cOun1tries in the regiontul sample. Soultes: IMF and t INESCiO databases. PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 67 TABLE 3.5 ALLOCATION OF EDUCATION BENEFITS IN EAST ASIA, 1985 Phiblic Slharle of Share tf Pulb/ hc eApendillo/n edluc ation education ('e. eiilitr e Ol ol p) izaiv and dllb ,l'et bii ,lf 't allocfated oil e'dliicitlOil st'collelrv C'cluication (c/loceated to) to pIiiillr V as a Share of or al Shell of hilgher and .'t and secon ay Economy ' GNP GNP edclitioln eduf-cation Hong Kong 2.8 1.9 25 69 Indonesia 2.3 2.0 9 89 Korea. Rep. of 3.0 2.5 10( 84 Malavsia 7.9 5.9 15 75 Singapore 5.0 3.2 31 65 Thailand 3.2 2.6 12 8 1 Socrce World Banik 199)3a. differences in the pattern of fee-charging. Countries in Afiica. the Middle East, and Eastern Europe and Central Asia have little or no tradition of cost recovery in higher education. In half of all Asian countries. however, and in a fifth of Latin American ones, cost recovery accounts for more than 10 percent of reculTenlt expenditures in public higher education (World Bank 1994e). lt Il t ilol'ill ti'le liWI ;Iced(l I:tt'iC-it'iclC1 andI [qtlil Increasinlg public spending on education is not necessary in many cases be- cause of the enormous potential of efficiency gainis at culTent levels. This can be seen through a simple comparison of regions. Public spending on education in Africa, which has the lowest enrollinent ratios of any region. represents a greater share of GNP (4.2 percent) thani in East Asia (3.4 percent) and Latin America (3.7 percent), where primary education is nearly universal (see the top panel of figure 3.4). By 1990 an average 6-year-old in East Asia or in the Middle East and North Africa region could expect to complete over nine years of school. Yet public spending on education by countries in the Middle East anid North Africa represented 5.9 percent of GNP, compared with only 3.4 percent in East Asia. Some, but not most, of the difference is accouLitedl for by demographic structure. There is no theoretically appropriate proportion of CiNP or public spending that shoLIld be devoted to education. In many countries, however, more educa- tional attainment couldl be achieved with the same or even less public spending, particularly by following the East Asian pattern of focusing public spending on the lower levels of education and increasing its internal efficiency (table 3.5) 68 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 3.5 RELATION BETWEEN PUBLIC SPENDING ON EDUCATION AND GROSS ENROLLMENT RATIOS FOR POPULATION AGES 6-23, SELECTED COUNTRIES, 1990 Gross enrollimenit ratio (percent) Jordani m 7(0 - Ecuador * Syrian Arab Rep. Egypt n Philippines * Tunisia 60 - Jamiiaica Paraulay' * Bolivi Colombia 50( - * Canieroon * Honduras F1l Salvaidor Thailand . 40 - 30 - GuaIternala naolorocco o *3(1 - Senegal n * Mauritania 0 1 2 3 4 5 6 7 Spending as a pericentage of GNP m- N tin[g;, aild rill, 1494. while relying more on private financing at the higher- levels. Figure 3.5 com- pares public spending on education as a percentage of GNP with gross enioll- ment tatios for a sample of countries. Wlhile net enrollment ratios would be a better meiasure, such art inteliiationial cotmparison using gross ratios does help identify countries wher-e public spending appears to stimullate relatively low levels of hutman capital formation. Mauritania and Morocco. for instance, show patticular-lv poor results despite high levels of public spending on education; their public spentding on edtication is very inefficient compared with that of Jamiaica. Jordan, and Tunisia. Similarly, Senegal anid Thailand are inefficient compared with Syria. SLIcI comparisons also show that public spending is very low il some countt-ies compared with international averages. In Paraguay, for instance, puib- lic spending appears to be relatively efficienit compared with spendilig in Colombia and( Thailanid. which have the same outcotnes but devote more than PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 69 twice the share OfGNP to public spendinig on educationl. Paraguay could probably increase educational attainment by spending more public funds on education. Finaniicing( [ducation The often-enicounitered iniefficiencies and inequities of public expenditure on education have combinied with expanding public sector enrollilenlts at all levels to increase the shar-e of GNP of public spending on education in many regions. particularly since Ullit costs are higher- for secondar-y and tertiary students than for primary students. (The exceptionis in shal-es spent are Latin America and the Middle East.) This trend has frequently put increasinig pressure on public fuiids at the same time as many countries, especially in Easterin Europe and Africa. have experienced general fiscal difficulties. Indeed some countr-ies' Ifiacroeconolinic clifficulties have in part been ciriven by fiscal problems result- ing from education spending. In Kenya. for instance, spending on education rose from 30 percent of the government budget in 1980 to almost 40 percenit by 1990. largely because admi.ssions to public Llliversities quacirupled. During the 1980s public spending on educationi iicr-eased or maintained its share of GNP and increased as a share of total central governmienit spending in every region of the developing world, except in Latin America with its debt- induced recessioni (see figure 3.4). Even more significant than total spending is spending per studeit, although such data are particularly limited in availability and reliability. Real public expenditure per student fell at the primary level not only in seven of the ninie Latin American countries for which data are available but also in thirteeni of twenty African countries. Real spending per student fell at the secondar-y level in Africa as well, by 1 8 percenlt. At the tertiary level. rapidly increasing elirollmenits in the 1980s combined with falling recurrent expenditLire resulted in a decline in real per student expenditure (figure 3.6). with a particularly acute fall in Africa of 34 percent. Amonig OECD countries. by contrast, all but one of the fourteenl for whicil data are available increased real expenditure per student at both the primary and seconclar-y levels during the 1980s, and half increased spending at the tertiary level (Donors to African Education 1994: LNESCO database). Measures to iicr-ease the efficiency of public spendinig on educatioll can irelease funds for more productive investmelIt in educatioll. The share of higher education in centr;al and state governlment spelidilig in India. for instance. fell from 21 to 19 percent between 1976 and 1991, although primary education remains underfunlded at 48 percent compared with secondary at 33 percent. SLIcI reallocationis may still not be enough. and other souLIces of funds may be required, paiticularly when overall public spendinig is falling. In BuL-kina Faso. for instance, the share of education spendinig devotedl to primaLy education 70 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 3.6 GROWTH OF ENROLLMENT IN AND PUBLIC EXPENDITURE ON HIGHER EDUCATION BY INCOME GROUP, 1980-88 Average annual change (percent) 9 0 -6- -9 __ -15 Low-income Lower-middle- Upper-middle- High-incoine countries income countries income countries countries _ Enrollmenits _ Recurrent expenditure CSw (c: Salnii 199 ". increased from 23 to 42 percent between 1980 and 1990 but dropped in abso- lute terms, as spending fell from 2.9 to 2.3 percent of GNP. In these circum- stances. some countries have chosen to switch further public expenditure into education and away from other publicly funded activities. such as defense and inefficient public enterprises that can be run better by the private sector. Other countries have decided that their macroeconomic policies have the scope to expand spending on education by increasing the revenues of government. Mili- tary spending in many developing countries exceeds that for education, having quintupled in constant dollars between 1960 and 1991-twice the rate of in- crease of per capita income-anid is only slightly less tihani the combined total for both education and health spending (McNamara 1992). Uganda reduced its clefense spending from 3.8 percent of GNP in 1989 to 1.5 percent by 1992, increasing education spending from 1.4 to 1.7 percent OfGNP and healtil spend- ina from 0.5 to 0.8 percent (World Bank 1994mn). Several Indian states in- creased spending on education from about 2.5 percent of state domestic prod- PUBLIC FINANCE FOR EFFICIENCY AND EQUITY 71 FIGURE 3.7 RELATION BETWEEN GROSS ENROLLMENT RATIOS IN HIGHER EDUCATION AND EXTENT OF PRIVATE FINANCING, SELECTED ASIAN COUNTRIES, ABOUT 1985 Gross enrollment ratio (percent) 40 Philippines * 35 - * Korea 30 _ 25- 20 - * Thailand 15- 10 _ India Bangladesh Sri Lanik-a Malaysia * * Indonesia 5 China * * * Nepal . Papua New Guinea 0 I II I I I II 0 10 20 30 40 50 60 70 80 90 100 Index of private financing (percent " a. Retlects the rate of cost reCOVerv across institutional types weighted by their share oftotal enrollitients. Source Tatn attd Mingat 1992. uct in the mid- 1970s to more than 4 percent in 1990. Ghana increased its share of education in public spending from 27 percent in 1984 to 36 percent in 1988. Not all countries will be able to reallocate resources from, for example, defense to education, or to raise revenues. Some countries have sought to supplement public funds for education by injecting private funds. Private funds can increase enrollments, whether they are used at private or at publicly funded institutions. In Asia the more that higher education costs are financed through student fees, the greater is the overall coverage of the education system (figure 3.7). The very existence of private schools and universities promotes diversity and provides useful competition for public institutions, especially at the higher levels of education. Some countries, however, prohibit private sclhools and universities. and others regulate them excessively. Since private schools are usually financed mainly by household payments of fees, such restrictions pre- vent private spending on education that could have substituted for public spend- 72 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD TABLE 3.6 GOVERNMENT AND HOUSEHOLD EXPENDITURE ON EDUCATION IN KENYA, BY LEVEL, 1992-93 (percentage of GDP) Go,'rnmelt Household Detliv ed t ,.t Level ll'. The early years of life are crucial in the formation and development of intelligence. personality, and social behav- ior, and targeted integrated programs of health, nutrition, ancl cognitive stimu- lation can give disadvantaged children a better start in school. Early childhood prograrns need to monitor the child's health status, supply health and nutrition supplements wheni needed. and provide age-appropriate curricula, activities, and materials to encourage cognitive development. Programs should be so structured that eveiy child receives attentioni every day. Parental and commu- nitv inivolvement should be promoted (Young 1994). World Bank-assisted projects are beginling to include such programs. In Colombia. for instance. a project helps women repair and renovate their homes so that they can offer 76 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD childcare facilities for the commllunity. A project in Bolivia helps expand home- based day care in poor urban and periurban areas, and one in Mexico supports a program of parental education targeted to rural and indigenous poor in the states with the lowest per capita incomes. In India health. nutrition, and early education services are provided to more than 12 million children between 6 nionths and 6 years of age. NLITRITION AND HEALTH PROGRAMS. Temporary hLinger-, chronic nialnutri- tion. micronutrient deficiencies. parasitic infections. and vision and hearing imipairments reduce a child's ability to achieve in school (Levinger 1992; Pollitt 1990). Most children with a history of malnutrition and poor health, it is now knowil, are capable of successful school performance if measures are taken to compensate for their deficiencies. (Severe health and nutrition deficiencies that inflict gross changes in the brain or irreversible physical damage cannot be so simply hanidled.) For many of the poor nutrition and health conditions affecting childrenl, there are effective. safe, and relatively inexpensive interventionls. Where costs are higher. interventionis can be targeted to the needy poor. Schoolchildren who suffer concurrent malnutrition and poor health per- form less well and attend less regularly, leading to grade repetition or dropout. School-based vitamin A. iron. and iodine supplemenitation and mass delivery of deworming drugs through schools are perhaps the most cost-effective ways of improving children's readliness to leani via imlproved nutrition and health. These remedies are inexpensive, and providers do not require medical training. although distribution and logistical infrastructure do need to be well developed. Per child per year, deworminig costs less than $1.50: vitamini A supplementa- tion, less than $0.50; ferrous sulfate tablets for iron deficiency, between $2.00 and $4.00; and oral iodine supplementation, less than $0.50 (Bundy and other-s 1990: World Bank 1994c). Integration of programs can reduce costs further. and education programs designed to change specific nutrition and health prac- tices or to increase knowledge among schoolchiildlren can complement ancd sustain these shorter-term interventions. Childlren with inpaired visioi arid hearing can be identified at negligible cost with the use of simple eye charts and whisper tests. Glasses and hearinig aids could then be provided, or teachers could at least seat the children near the front of the room or take other helpful measures. Temporary huniger affects the child's ability to pay attention and so has a detrimental effect on learning. Many gov\ernments support large and expensive school feeding programs. Such programs can be made mole cost-effective by targeting them to the poor, by providing breakfast or snacks before school rather than a larger meal later in the day, and by selecting fortified foods or foods high in essential micronut-ients. IMPROVING QUALITY 77 Increasingly, World Bank-assisted projects are being designed to iilprove the quality of primary education through school nuLItitioni and health interven- tions. A project in Brazil supports iilprovemenlts in the school feeding pro- gram, screeninit of schoolchildren for health and nutritioni status, the integra- tion of health and nutrition instruction into the curriculumil, and pilot programs for school-based vitamin A and ironi supplementation. A project in Guinea is developinig a national dewormliti andl iodine supplementation program through the schools. Another, in the Dominicani Republic. is assisting the implementa- tion of a school snack programii in poor urban areas. a national height census of first-graders, a micronutrient deficiency survey, and pilot school-based iron and vitamin A supplementation prograrms. CURRICULUM. The curTicultinu defines the subjects to be taught and fur- nishes general gui(danice regarding the firequency and dulationi of instruction. In some cases an accompanying syllabus specifies more precisely what is to be taught and what will be assessecd. Curricula and syllabi shoLIld be closely liniked to performance standaids and measures of outcome. The curriculum typically includes fewer subjects at lower levels and more subjects at higher ones. At the primary level there are broad international similarities in the relative emphasis placed on approximately eight major subjects; reading, writing, and mathemat- ics account for about 50 percent of curricular emphasis (Benavot and Kamens 1989). Within each subject, the coverage, sequencing, and pacing of topics may vary widlely between and within couLn1tries. At the secondary level countries differ with respect to the number of subjects taught, the balance between gen- eral and vocational subjects. the designation of mandatory and elective sub- jects. and the sequencinig of subjects. The range of national variation among relatively successful education sys- teins shows clearly that iio single curriculum is appropriate to all or most low- and middle-income countries. Intendced variation between and within countries results from differences in desired outcomes, in theories of instuLiction, and in local conditions. These differences can affect which subjects are taught, when they are intioduced, and how long they are taught. For example, schools il BuLundi offer relatively fewer hours of instructioni and more emphasis on lan- guage ancd mathematics thani do schiools in Kenya, where the school year is longer ancl the curriculum covers more subjects, with an emphasis on science (Eisemon and Schwille 1991; Eiseinon. Schwille. and Prouty 1989). In Japan finite mathemiatics. including statistics, is introcduced in grade 6 and precalcULIus and calculus in grades 7-9: in the United States these subjects are taught in grades I I and 12. Another type of cuLITicular variation is unintended: the discrepancy between the official curriculuim and the one actually implemented in schools and class- 78 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD rooms. This variation has two main causes. The first is technical: in many couLitries educational systems, schools, and classrooms are unable to deliver the key ingredients for learning: a teacher who can teach the subject. time for leamning. and the requisite tools for teaching and leaming. Second. competing incentives (or disincentives) can affect the number of hours teachers or students spend in school and the attention paid to specific subjects. Opportunity costs of student and teacher time are the most important disincentive to maintaining official norms for instructional time. Student dropout in Ghania was highest when the direct and indirect costs to the family for children's school participa- tion were highest-the direct costs were highest during (amine tLimes and the indirect costs durinig peak agricultural seasons. In Jamaica. studenit absenteeism is highest on market days. Teacher absenteeism can result when teachers hold multiple jobs or have to commute long distances to their assigned schools. Selection examinations provide strong incentives for ignorinig certain subjects in favor of the subjects that are tested. For example. in Jamaica students in grades 5 and 6 spend a disproportionate amount of time studying vocabulary lists and mathematics problems in preparation for the Common Entrance Ex- amination for secondary-level schoolinig, to the detrimilenit of subjects il the primary curricululil that are not tested on that examniation. Children's gender-role development is affected by maniy aspects of the school environiment, such as curricula and instructional materials. While most countries have a nationial curriculum meant to expose boys and girls to the same subjects, gender-differentiated courses are still offered in many schools. Boys tend to outnumber girls in mathematics, science. and meclianical courses. while the reverse is true for homemaking courses in many countries. Girls may con- sequently perform poorly on quanititative tests, discouraging expectations of academic achievement (Martin and Levy 1994). A heavy bias in the curricula toward low-paying skills for females. such as sewing, knitting. and secretarial work. can also strongly intlIuence girls' future employment opportunities by limiting women's access to better-paying jobs (Herz and others 1991). Textbooks and other educational materials in many countries have been found to have a pronounced gender bias that portrays women as admiring, passive, and powerless, suited for traditional roles only. In contrast. men are depicted as intelligent and capable of employment in a great number of exciting and profitable fields (box 4.1). These messages can reinforce negative stereo- types, discouraging girls from viewing themselves as good students, as intelli- g,ent, or as capable of pursuinig any occupations other thani a few traditionial ones (Herz and others 1991). Many primary curricula include too many subjects. reducilig the time avail- able for teaching core language and number skills. Furthermore, many cur- ricula require the teaching of several languages (the native language, the IMPROVING OUALITY 79 BOX 4.1. GENDER BIAS IN TEXTBOOKS females were usually shown doing domestic tasks and caring for children. Since the mid-1 970s parental pressure In one story a poor female street ven- on publishers has led to reduction of dor drops her basket of wares while gender bias in textbooks in industrial thinking of plans for the future. The countries, but in developing countries accompanying text-"What should the relatively little change has occurred woman have been doing instead of (Stromquist 1994). A study in Zambia imagining future possibilities?"-implies revealed that while textbooks system- negative consequences of women's atically treated the activities of men as imagination. Costa Rica has since in- admirable, women, if featured at all, troduced a new series of books in an were portrayed in domestic roles and effort to reduce gender bias (Gonzalez- depicted as stupid, ignorant, and pas- Suarez 1987; Lockheed, Verspoor, sive (Hyde 1989). An analysis of gov- and others 1991). Such discrepancies ernment and commercial textbooks in in the representation of men and Costa Rica in 1985 found that 75 per- women in textbooks have been found cent of the images were of males and across regions and cultures in, for ex- 25 percent of females. Males were ample, Colombia, Egypt, India, Kuwait, commonly shown as historical figures, Lebanon, Qatar, Saudi Arabia, Tuni- in pursuit of intellectual activity or work- sia, and Yemen (Lockheed, Verspoor, ing in agriculture and ranching, while and others 1991; Stromquist 1994). national or regional language of instruction, the metropolitan language, and so on). Learniny is more effective, and time is saved. if instruction in the first several grades is in the child's native language. This approach allows for inas- tery of the first lanigLiae an(d promotes the cognitive development needed for learning a second language (Dutcher 1994). Once solid skills in the first lan- guage have been acquired. a nationial, regional, or metropolitan language can be learned in the later primary grades to prepare for secondary education. Production of textbooks in native languages may. however, increase the costs of education. "Language capital--the ability to speak. read, or write one or more lan- yuayes-is an importanit aspect of human capital. Building up language capital begins early, with the development of oral fluency in one's native language. The development oflangIuage capital in the native language continues in school and elsewhere. The native language of most poor minority groups. however, is not the majority or dominant language spoken in the country. Not knowving the dominant language may limit a person's training opportunities, job mobility, and earnings and reduce that person's chances of escaping poverty. There is, 80 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD therefore, a labor-market incentive for acquiring skills in the dominant lan- guage (Chiswick 1991; Chiswick and Miller 1995). Whereas the primary education curriculum11 is relatively standar(d across coun1tries, secondary education curricula vary in duration (from two to six years). in the use of residential programs, in differentiation by streams (science. teacher trainiig, vocational. and so on). and in the nuLimber of courses offered (from I 0 to 200). Science education and vocational education raise particularly complex issues because of their perceived importanice and their cost. Science education is importanit for economic development and is increas- ingly incorporated in the curriculum. Advanced science education requires expensive laboratories and equipnlent, and teacher training in these subjecis is costly. Many countr-ies view all science education at the lower- and upper- secondary levels as "advanced" and restrict access to science education. In the Philippinies, for example, science is offered only' at special scilools. Yet, as OEC'D countries are recognizing, laboratory-based instruction is not essential for science education at primiiary or lower-secondary levels. Once laboratory use is reduced or eliminated, the costs of science education are no longer significantlv higher than those of other subjects, as evidence from Denmark shows. This means that instruction in a core subject need not be restricted because of expense. Students at lower levels still need to work with simple concrete objects and to see charts because they' need more help with conceptu- alizing than do older studenits. Vocational and technical skills are best impar-ted in the workplace, follow- ing general educationi. Thie private sector should be directly involved in the provision, finanlciig, and governanice of vocational schoolinig. At the second- ary level, vocational education and general education are converging, as such subjects as science. technology, mathematics, and English are added to the vocational curriculumii and as general seconclar-y education increasingly in- cludes basic techiological education. This convergence has yet to be evaluated in termis of labor market outcomes. It is. however, in harmony' with the trend toward rapidly changinlg labor markets that place an increased emphasis on trainability. It is also consistenit with earlier comparisons which showed clearly that the social rates of' retulIl to investimienit in very specialized vocational secondary education were lower than those to general secondary education, largely because of the much higher costs of the forimier (Psacharopoulos 1989). Curriculum reform policies typically tocus on changing the intended cur- riculum: the types of courses to be offered. the level at which they are to be introduced, and their duration. For example, Malawi introduces health educa- tion as early' as grade 2 and Kenya's curTicular and structural refonm of the mid-I 980s introduced several new subjects, raising to thirteen the number to be examined at the endl of the new primiiary cycle. However. curriculum reforms IMPROVING QUALITY 81 that focus on revisions of courses and timetables without concomitant revisions in standards and guidelines, instructional materials, teaching practice, and the incentives offered by tests and examinations are likely to have little impact. Many countries have adopted a two-pronged approach to curricular reform. First, performance standards for learning are established and outcomes are measured through examinations or national assessments. Second, within the general curriculum, local variation in the use of materials, in teaching methods. and in the allocation of time is encouraged. Kenya, for instance, has a national curriculum, but schools determine the language of instruction in the first four primary grades. India's Department of Education has developed a competency- based curriculum around the concept of minimal levels of learning, but states and districts are responsible for adapting materials and teacher training to local conditions. Wlhat Are the Necessaiy Inputs? A wide variety of policies and practices, chosen to fit local circumstances, can bring about effective schooling. In low- and middle-income countries school and classroom characteristics account for only about 40 percent of differences in learning achievement; the remainder, as noted earlier, is attributable to indi- vidual and family-background characteristics not typically amenable to school- level interventions. Recent reviews of the literature on correlates of learnmig in low- and middle- income countries show that the most consistently positive effects are found for teacher subject knowledge, instructional time, textbooks, and instructional materials (Fuller and Clarke 1994; Lockheed, Verspoor, and others 1991; Harbison and Hanushek 1992; Velez, Schiefelbein, and Valenzuela 1993). Inputs in these categories would have the highest priority for expenditure. However, the exact composition of the basket of inputs and their relative im- portance for a given school will vary widely in accordance with local condi- tions. For example, a recent study of reading achievement in twenty-five couII- tries (Postlethwaite and Ross 1992) found that of fifty-six inputs examined, only eleven contributed to learning in at least three of the four developing countries included (Hungary, Indonesia. Trinidad and Tobago, and Venezuela). In many countries, education systems have made a practice of investing in inputs that expand access (for example, hiring more teachers to reduce class size) instead of those that have a demonstrable effect on enhancing learning (Hanushek 1994). Such inputs as smaller classes and higher teacher salaries set on the basis of seniority and formal qualifications are cited less often in the research literature, however, and therefore probably deserve lower priority (figure 4.1). In addition, expensive inputs, such as laboratories, are not effective. 82 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD FIGURE 4.1 DETERMINANTS OF EFFECTIVE LEARNING AT THE PRIMARY LEVEL Percentage of studies showing positive effect 100 80- 60- 40- )on 0 Libraries Instruc- Home- Text- Teacher Teacher Labor- Teacher Class tional work books knowl- exper- atories salarv size time edge ience ".ote. Stmdies covered ilmore thani twemty-five couuitnes. Somr e: Fullerand Clarke 1994. TEACHERS KNOWLEDGE AND SKILLS. Teachers' subject knowledge. an in- tended outcome of preservice training, is strongly and consistently related to student performance. Teachers with a better knowledge of subject material and greater written and verbal language proficiency have better-performing stu- dents (Lockheed, Verspoor, and others 1991; Harbison and Hanushek 1992. for Brazil- Ross and Postlethwaite 1989, for Indonesia; Warwick and Reimers 1992, for Pakistan; Bashir 1994, for India). In both Brazil and Pakistan, teach- ers' own subject knowledge and formal education had more impact on student performance than did preservice training (Warwick and Reimers 1992). At the primary level, research suggests that the overall level of relevant knowledge is insufficient in many countries. For example, in India less than half of grade 4 teachers could correctly answer 80 percent of questions testing grade 4 mathematics knowledge (Bashir 1994). The curriculum for preservice educa- tion may need to be revised to stress knowledge of the subject matter. Related to teachers' subject knowledge is their pedagogical knowledge. Although no specific teaching practice is universally effective. teachers with a wide reper- IMPROVING QUALITY 83 toire of teaching skills appear to be more effective than those with a limited repertoire. The most effective strategy for ensuring that teachers have adequate subject knowledge is to recruit suitably educated teachers whose knowledge has been assessed. The assessment of learning outcomes for higher education-includ- ing teacher education-is as important as it is for the primary and secondary levels. In fact, subject matter knowledge is routinely tested for secondary and tertiary teachers but not for primary teachers. There are a few exceptions, such as Mexico. where teacher knowledge is related to pay at all levels of education. Harbison and Hanushek (1992) suggest a national teacher examination. based on their findings for Brazil and other countries. At a minimum, the recruitment of primary and secondary teachers could resemble more the recruitment of higher education teachers, which is almost entirely based on subject knowl- edge-as in France and Japan. where recruitment is highly selective. Well-designed, continuous in-service trainiing is a second strategy for im- proving teacher subject knowledge and related pedagogical practices. Recog- nized effective elements of in-service training include exposure to new theory or techniques, demonstrations of their application, practice by the teacher, feedback to the teacher, and coachinig over time (Joyce and Showers 1985, 1987, 1988; Joyce, Hersh, and McKibbin 1983: Joyce 1991). As these elements suggest, in-service training is most effective when it is Iinked directly to class- room practice by the teacher (Walberg 1991; Nitsaisook and Anderson 1989) and provided by the head teacher (Raudenbush, Bhumirat, and Kamali 1989). The effect of in-service training oni student achievement has been demonstrated for the Escuela Nueva program in Colombia (Colbert, Chiappe, and Arboleda 1993). for science in the Philippines (Lockheed, Fonacier, and Bianchi 1989), and for mathematics in Botswana (Fuller. Hua, and Snyder 1994). Distance education programs for in-service (and preservice) teacher training are typi- cally more cost-effective than residential programs. For example, in Sri Lanka distance education programs of up to four years in duration are more than five times as cost-effective as two-year programs in colleges of education or in teachers' colleges (Nielsen and Tatto 1991). In Botswana in-service training is a more efficient way to raise achievement than reducing class size or providing supplementary reading materials (Fuller. Hua, and Snyder 1994). TIME. The amount of actual time for learning is consistently related to achievement. More time spent on a wider coverage of the curriculum results in increased learning and less variation among achievement levels (Stevenson and Baker 1991; McKnight 1971). Internationally the school year averages 880 hours per year of instruction at the primary level. The official primary school year is shorter in low- and middle-income countries than in industrial countries, however. In addition. students in low- and middle-income countries spend 84 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD much less time being instructed than those in OECD countries as a consequence of unscheduled school closings, teacher and studenit absences, and miscella- neous disruptions (Lockheed. Verspoor. and others 1991). The first strategy for increasing the amount of instructional time is to in- crease the length of the official school year. if it falls significantly below the norm. There is no guarantee. however, that schools will implement the official school year. particularly if it makes no accommodationi for local conditionis that may affect the participation of teachers or stidents. In many countries. schools or regions are permitted flexible scheduling of the instructional day. week, or year to accommodate variations in demand associated with weather, agricul- tural seasons, religious holidays. and children's domestic chores. This strategy has been effective in both nonformal programs of basic education and formal programs supported by the World Bank in Bangladesh, Colombia, Costa Rica, and Ecuador. A second strategy for increasing learning time is to assign home- work-an approach that has been effective in OECD COUntries. TOOLS AND TECHNOLOGY. Instructional materials include the entire range of teaching tools, from chalk to computers. After blackboards and chalk, text- books are the most common and most significant instructional material in most countries. The availability of teaching tools at all levels in low- and middle- income countries is limited, particularly at the primary level. In some countries textbooks are supplemented by libraries, other print and graphics materials, audiocassettes, films, radios, television, and computers (Lockheed, Middleton, and Nettleton 1991). Particularly critical in improving reading achievement is the provision of supplementary reading material, since voluntary reading and library usage, together with teacher education, are correlated with reading achievement (Lundberg and Linnakyla 1992: Postlethwaite and Ross 1992). Hence the importance of libraries in figure 4. 1. Almost all studies of textbooks in low- and middle-income countries show that the books have a positive impact on student achievement (Heyneman, Farrell, and Sepulveda-Stuardo 1978; Fuller and Clarke 1994). In addition, interactive radio instruction has been found to have a positive and cost-effec- tive impact on mathematics, science, and English achievement at the primary level in Bolivia, Honduras, Lesotho, and Papua New Guinea (Tilson 1991). New technologies stand to improve the efficiency of education through software tools that improve student performance and through new means of providing instruction and educationial resources to underserved populations. Computers improve student achievement and attitudes at all levels (Thompson, Simonson, and Hargrave 1992), and small-scale experiments with computer- based instruction have been carried out in several low- and middle-income countries, including Chile. Mexico, and the Philippines. In industrial countries technologies are being innovatively combined at the primary and secondary IMPROVING QUALITY 85 levels to increase instructional effectiveness. Intelligent tutoring systems. CD-ROM, multimedia. and other applications have improved student achieve- ment in all disciplines (Sivin-Kachala and Bialo 1994). from early childhood programs through college preparatory classes. Broadcast and network tech- nologies allow teachers with specialized skills (such as Japanese or Russian language teachers) and educational resources (such as on-line libraries) to reach beyond the traditional limits of classrooms and schools. Teachers can reach students via interactive television, teleconferencing, computer conferencing. audiographics, voice and data transmission systems, shared electronic black- boards and light pens, fax, voice-mail. computer bulletin boards, and electronic mail. Transmission systems include satellites. microwave, fiber-optics, interac- tive cable systems. and microcomputers connected to local and international networks. Al the tertiary level technology can substitute, at least partially. for teach- ers. Correspondence courses and open universities, for example, can increase cost-effectiveness. Satellite and computer technology allow high-quality inter- active courses to be broadcast directly to places of work, in cooperation with firms that wish to improve their employees' skills. Students can save signifi- cant travel time to and from school ancl do not forgo income by inteffupting employment to earn a more advanced degree. For example. in the United States the National Technological University (NTU) offers 1.000 master's degree courses annually via satellite to over 100.000 regular and continuing education students. NTU's unique learning consortium includes 43 accredited universities and over 200 firms (National Techilological University 1994). This model is being copied to provide distance instruction in both industrial and developing countries. Even in industrial countries, many programs using educational technology are still in the pilot stage and depend on grant funding. Initial costs are gener- ally high. and the costs of adding additional users are low, but all costs are closely linked to the idiosyncrasies of individual technologies and to the quality and availability of local telecommunications infrastructure. As a rule, these programs depend on reliable, high-quality telecommunications connections (Derfler 1992), which are seldom found in developing countries. Even where programs are technically feasible, low- and middle-income countries may lack the experienced pedagogical and technical support necessary for successful implementation. The need to use scarce resources for other instructional mate- rials and for the improvement of existing educational programs. combined with lack of institutional support and of information about appropriate, durable, and cost-effective technologies. has limited the widespread use of advanced educa- tional technology in many low- and middle-income countries. Failure fully to use this technology carries the risk of further increasing the gap between these countries and industrial ones. 86 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD The World Bank has supported governments in their efforts to design, print and distribute textbooks "in-house." World Bank projects now routinely include funding for textbooks, which accounted for 6 percent of all education lending in fiscal 1990-94. compared with 3 percent a decade earlier. Support for textbooks is now broken down into development, production. distribution. and use. Textbook development needs to be closely linked with culTiculum development. Production and distribution are best handled by the private sec- tor. although governments need assistance with procuremenit policies and with measures for ensuring quality. The effective use of textbooks must involve training teachers in the use of new books and providing teacher guides. Some governments have also provided packages of instructional materials, such as wall charts, games. anatomic models, and science kits. In Mexico, for ex- ample. such packages have been provided under two consecutive World Bank- assisted primary education projects; the packages prepared under the second project were modified on the basis of experience with the first project. Flexibilitv in Provi(ding InpLits Setting standards and supporting effective inputs are important for raising learning achievement. Even more critical is the tlexibility to decide locally how to combine and manage inputs in schools and other educational institu- tions. Governments can support this flexibility through their method of pro- viding inputs and by encouraging the conditions that have been shown to promote student learning. Governments have principally employed two strat- egies for providing inputs to schools and institutions of higher education, and the World Bank has supported both. The fi-st strategy' has been direct provi- sion of packages of inputs. Experience shows that centrally provided packages are not used unless the inputs fit local conditions, teachers know how to use the inputs, and policies related to the use of inputs do not act as disincentives. The second strategy has been to provide budget transfers so that schools and other institutions can purchase what is most relevant to prevailing local conditions. No single package of inputs can be considered "most effective" or most cost-effective for all schools or all preexisting conditions, and it is difficult to specify in advance what will work in a particular situation. School personnel have the closest knowledge of preexisting conditions and are in the best posi- tion to select the most suitable package of inputs. Even when school personnel have not been delegated authority for budgets, their knowledge can be em- ployed to tailor packages of inputs to local conditions. In Jamaica the World Bank is supporting a comprehensive reform of secondary education that in- cludes curriculum revision for grades 7-9. teacher training in curriculum ob- jectives, provision of tools and technology for curriculum implementation, and assessment of student learning in the core curriculum areas. Before receiv- IMPROVING QUALITY 87 ing a package of instructional materials for the core curriculum subject, school personnel complete an inventory of materials already available in the school, and only the missing materials are provided. The results are significant cost savings over the alternative of providing all schools with all materials. In some countries schools have authority in cLirriculum choices and text- book selection but not over budgets and personnel. In most high-income coun- tries teachers and schools select books from an approved list. This practice is gradually being introduced elsewhere. notably in the transition economies of Europe and Central Asia. In only a few cases do school-level personiniel have full autonomy, with authority in all the critical aspects of school management: budget, curriculum, and personnel. Even this practice does not necessarily lead to higher student achievement, however. Preliminary evaluations of school- based managenment in the United States and Canada do not demonstrate any effect-positive or negative-of this reform on achievement (Summers and Johnson 1994: GAO 1994). A more promising approach. tested so far only in industrial countries, is school-based leadership that ensures an effective climate for learning (box 4.2). Even easier to achieve is the appointment of the best BOX 4.2. A PROMISING AVENUE: SCHOOL the behavioral and learning styles of LEADERSHIP students and when classes are disci- plined so that learning time is high, School characteristics that are impor- student achievement is usually higher. tant for achievement center on each Governments can foster school lead- school's leadership. Effective learning ership and classroom conditions which institutions in industrial countries have encourage learning by ensuring that leadership able to ensure that re- these factors are prominent in the se- sources are available, to communicate lection and training of teachers, school a vision for the school that includes principals, and administrators and that high expectations for students and an they are central objectives of school orderly environment, and to provide supervision, inspection, and support pedagogic guidance and support to activities (Brookover and Lezotte 1979; teachers. This support may not require Brubaker and Partine 1986; Carter and full autonomy over budgets, curricu- Klotz 1990; Chubb and Moe 1990; lum, and personnel. Dalin 1992; Frederick 1987; Gibbs Pedagogic leadership at the institu- 1989; Hallinger 1989; Joyce, Hersh, tional level needs to support the class- and McKibbin 1983; Levine 1990; room conditions that are known to fos- Levine and Lezotte 1990; Lezotte and ter learning. Students whose teachers others 1980; Lezotte and Bancroft have high expectations for their stu- 1985; Purkey and Smith 1983; dents and offer rewards and incentives Scheerens and Creemers 1989; Smith foracademicachievementlearnmore. and Andrews 1989; Steller 1988; When teaching methods respond to Wynne 1980). 88 THE RECORD OF EXPERIENCE AND THE TASKS AHEAD principals to the neediest schools, such as rural multigrade schools and those in urban slums. Many education systems in low- and middle-income countries are rigid, run in a centralized manner with, for instance, central selection and purchase of textbooks and central direction concerning classroomii instruction. Despite this, many schools have considerable de facto autonomy, at least about the method of instruction, if not about the deployment of teaching staff. However, manage- ment and supervisory links are often weak, and teachers work in isolation. especially in small schools. The consequence of this isolation is that the cur- riculum is not implemented. instructional time is reduced, and teaching tools are not used. Three factors are necessary for overcoming these shortcomings: shared local consensus about desired outcomes, professionalism among teach- ers, and school autonomy. These three factors combine to hold the school and its teachers accountable to parents and communlities for outcomes in the con- text of national or regional indicators of perfonnance, such as examinations and learning assessments. P A R T 11 Six Key Reforms T HE EDUCATION challenges described in chapter 2 can be met if reforms are introduced along the lines of the changes in the financing and management of education discussed in chapters 3 and 4. Six reforms, taken together. will go a long way toward enabling low- and middle-income coun- tries to meet the challenges in access. equity, quality, and pace of reform that they face today. These reforms are a hig-her priority for education; attention to outcomes; concentration of efficient public investment on basic education, coupled with more reliance on household financing for higher education: atten- tion to equity: household involvement in the education system: and autono- mous institutions that will pennit the flexible combination of instructional inputs. Since education challenges are present to varying degrees in individual countries, the six refonms will not all have the same priority everywhere. And. while much of the discussion necessarily involves the setting of subsectoral priorities, it must never be forgotten that the education system is indeed a system and that chaniges and investments in one subsector will have implica- tions for other subsectors and for the system as a whole. The six key reforms called for in this report will help to improve education and, in the poorer countries, to reduce illiteracy in the future. They will not. however, contribute significantly to solving the problem of adult illiteracy today. in a world with more than 900 million illiterates. Programs of adult 89 90 SIX KEY REFORMS education are necessary, but such programs have a poor track record. One study showed an effectiveness rate of just 13 percent for adult literacy campaigns conducted over the past thirty years (Abadzi 1994), and there has been little research into the benefits and costs of literacy programs. Several new ap- proaches to adult literacy appear promising, however. in large part because they address motivation-the key factor in all successful programs. Adult lit- eracy efforts have a better chance of success if they (a) have an initial objective other thani literacy, such as reading sacred literature like the Koran, acquiring health infonmation, or assisting with children's education: (b) distinguish among teenagers and older adults, as adults lean in different ways from adolescents; (c) include women as well as men (most unsuccessful campaigns have focused on male themes); and (d) use a participatory pedagogy sensitive to the local environmiient. In the REFLECT program. being developed with the help of the NGO ActionAid in Bangladesh, El Salvador, and Uganda, poor communities are encouraged to construct maps, calendars, matrices. and diagrams based on local circumstances and are helped to analyze and systematize their knowledge. The alphabet and literacy then become a more elaborate way of representing this local knowledge, and literacy is Iinked much more tightly to other aspects of development in the local area. These new approaches will be reviewed in detail in a future World Bank paper, prompted by challenges to the view that large-scale literacy programs are generally unsuccessful. The paper will also analyze the costs and benefits of literacy programs and the factors that were important in the successful implementation of programs that have been ex- panded from small experiments to the national level. This issue is not further discussed in this report, which focuses on mainstream fonrnal education. C H A P T E R F I V E A Higher Priority for Education GOVERNMENTS and peoples in almost all countries neecl to pay more attention to education. Education is usually the province of the ministry of education-sometimes also of a ministry of higher education-and other parts of government tend to leave it to this body. This method is short-sighted. for three reasons. * Continuous change in economies and labor markets is now normal. as the result of permanent economic reform and technological change that necessitate renewed and enduring attention to investment in both physical and human capital. * The rate of return to investments in education is high compared with other investments. * There are important synergies among investments in eclucation and other aspects of human capital formation, especially nutrition, health, and fertility. During the 1 980s and early 1 990s many countries began restructulring their economies, driven by macroeconomic imbalances. excessive external debt, and an increasingly competitive world economy. Economic reform programs have now brought positive results in the two regions where they were most neces- sary; economic growtil has resumed in both Africa and Latin America. African countries have made much progress in many key areas of macroeconomic 91 92 SIX KEY REFORMS reform, but these efforts must be sustained and expanded, especially to include the achievement of fiscal balance. Increasing international trade and the in- creasing mobility of capital and technology have made most economies more open and have created a more competitive environment for attracting global investment. China and India are the two largest examples of countries with increasingly competitive and open economies. (The record of East Asia was discussed in chapter 1.) Developments in the transition economies of Europe and Asia are even more dramatic, as wholesale shifts toward market economic structures occur. Now that economic reform is becoming a pemianent process. it is important for governments to focus on the factors, in addition to appropriate macroeconomic policies, that are necessary for sustaining growth and reducing poverty. Investment in production andl seivices is increasingly attracted to countries with the necessary physical infrastructure and with flexible labor forces. All governments must pay renewed attention to investment in infra- structure and to investment in people if they are to stimulate private sector investment and hence growth. The appropriate physical infrastructure and hu- man capital investments will vary from country to country according to the level of economic and educational developmenit. Investment in people is par- ticularly critical because of the lags between investmenits in education ancl the entry of new workers into the labor force. Delays in reforming education sys- tems therefore carry the risk of reducing future economic growthl. Investments in all levels of education yield high rates of return-above the opportunity cost of capital, which is usually thought to be around 8-10 percent. and comparable to (or, for lower levels of education. higher than) the rates of return to investments in agriculture. industry, and infrastructure (table 5.1). Moreover, these social rates of return to education are Linderestimates because they exclude benefits such as improved health and reduced fertility, as well as external economies such as threshold effects and technology acquisition and development. Investmnents in physical and human capital investment are comple- mentary: without investments in education, investments in physical capital will yield lower returns, and vice versa. Increased understanding of the relationships among education, nutritioll, health, and fertility walTants greater attention to education. Parents, especially mothers, with more education provide better nutritioni to their children, have healthiier children, are less fertile, and are more concerned that their children be educated. Education-in particular, female education-is key to reducing pov- erty and must be considered as much a part of a COUntry's healthi strategy as. say, programils of immunization and access to health clinics. Education is thus more important for economic development and poverty reduction thani it used to be or was understood to be. It deserves a higher priority A HIGHER PRIORITY FOR EDUCATION 93 TABLE 5.1 RATES OF RETURN TO INVESTMENTS IN DIFFERENT SECTORS OF THE ECONOMY Itcm /974-82 1983-92 1974-92 Educarion itnwestmzenrs Primary 20 Secondary 14 Higher I World t Batik pr-ojet s Agriculture 14 11 Industry 15 12 Infrastructure 18 16 All projects 17 15 (cs,ues. P]`acharopoulo 1'994: World Banik 1994ii; Woi Id B.tnk Operafions Evaluatioii Depanment databa.e. from goverinmenits as a whole--not only from ministries of education but also from ministries of finance and planning. The need for such an emphasis has long been realized by countries in East Asia and is increasingly coming to be understood in other areas, especially in Latin America and India. It is important that education receive more attention elsewhere, too, especially in Africa, South Asia, the Middle East. and the former socialist countries of Europe and Asia. most of which have until recently been appropriately preoccupied with shorter- term issues of economic reform. The retuin to a longer-term focus on develop- ment and poverty reduction implies a higher priority for education, with spe- cific policies and priorities within education varying according to country cir- cumstances. At the same time, too much must not be claimed for education. Its contribution to the reduction of poverty depends critically on complementary macroeconomic policies and investments in physical assets. C H A P T E R S I X Attention to Outcomes AN ORIENTATION toward outcomes means that priorities in edu- cation are determined through economic analysis. standard setting, and mea- surement of the attainment of standards. A sectoral approach is key for setting priorities. While govenmnents detennmie priorities for many reasons, economic analysis of education-in particular. rate of return analysis-is a diagnostic tool with which to start the process of setting priorities and considering alterna- tive ways of achieving objectives within a sectoral approach. The reasons for determining priorities vary country by country and even government by gov- ernment. and this report's call for more attention to the impact of eclucationial outcomes does not suggest that these other reasons are inappropriate. Rather. it argues that insufficient attention is generally given to outcomes. whether de- fined in labor market or in learning terms. Using Outcomes to Set and Monitor Public Priorities Most governments typically define how much education should be available for everyone implicitly. through legislation regarding the school starting age. compulsory school attendance laws, regulations on the minimum work age. constitutionial stipulations. and nationially ratified international convelitions. Countries fall short of their goals (table 6.1 ) largely because they do not allo- cate sufficient resources to achieve themii. Even where resources are available. 94 ATTENTION TO OUTCOMES 95 TABLE 6.1 COMPULSORY EDUCATION, ENROLLMENT RATIOS AND MINIMUM AGE FOR EMPLOYMENT, SELECTED COUNTRIES, 1990S Co)mphuulslr slainj mum cehicution Primary l grss enrollmnenlt a,,e for7 (duration ralio, 1990 (percent) em1pos,vluent Counl s in years) Total Male Female (1992) Bangladesh 5 77 83 71 12 C6te d'lvoire 6 69 81 58 14 El Salvador 9 79 78 79 14 Guatemala 6 79 84 74 14 Guinea-Bissau 6 60 77 42 14 Malawi 8 66 73 60 14 Morocco 9 65 77 53 1 2- Senegal 6 58 68 49 14 lSources: ILO 1992: LNESCO 1993b. however, the emphasis solely on years in school is misplaced. More appropri- ate would be an emphasis on knowledge and skills. School attendance is a means. not an end: it contributes to the acquisition of skills. knowledge. and attitudes. It is what students learn that is important. Basic education is the top priority in all countries because it provides the basic skills and knowledge necessary for civic order and full participation in society, as well as for all forms of work. The skills and knowledge acquired at the upper-secondary and tertiary levels, by contrast, are applied more explicitly in the labor market, and economic analysis can help to guide public sector investments at these levels. Economic analysis applied to education focuses on the assessment of ben- efits and costs, for individuals and for society as a whole. The costs of alterna- tive interventions to achieve a given educational objective are compared. and the relationship between benefits and costs is measured-usually, by calculat- ing the rate of return, taking as the benefit enhanced labor productivity as measured by differential wages (see box 1.1). Both the social rate of return and the clifference between the social and private rates of return can help in setting public sector priorities. The priorities for public investment determined by this type of economic analysis are those in which the social rate of return is highest and the level of public subsidization is lowest. The contrast between the social and private rates of return to investment in education, using wage differentials as the benefit measure, highlights the extent 96 SIX KEY REFORMS of public subsidization of education (see table 1.1). An index of public subsidi- zation (the percentage by which the private rate exceeds the social rate) can be calculated. In most countries higher education is the most heavily subsidized level of education. In Paraguay. for example, the private and social rates of return to primary education are 23.7 and 20.3 percent. respectively, while the private and social returns to higher education are 13.7 and 10.8 percent. The index of public subsidization in this case is 27 percent for higher education and only 17 percent for primary education (Psacharopoulos, Velez. and Patrinos 1994). These rates of return must be calculated in specific country circumstances and cannot be assumed. Methodological considerations and the practical prob- lems associated with the valuation of external benefits mean that it is prudent to exercise caution and use good judgment when applying cost-benefit analysis. For instance, rates of return are normally based on current average earnilgs differentials, whiclh are known to remain stable over long periods of time. Where possible. however. earnings differentials at the margin should be Lised for workers in sectors with free entry. Estimates of rates of return are also slow to reflect new developments in the labor market, such as growing imbalances between employers' demand and the output of the education system. Aside from these difficulties, the calculation of private rates of return to education is relatively straightforward. That of the social rates is more prob- lematic. No consensus exists on how to quantify and evaluate the social exter- nalities of education. The common practice is therefore simply to calculate social returns by adjusting private returns downward to allow for the public sector's net expenditures on education and to ignore potential offsetting exter- nal benefits. Again, judgment must be used; the cost-benefit framework of economic analysis provides a key diagnostic tool that points policymakers in certain directions, rather than a precision indicator for setting priorities. Once priorities have been set and financing arrangements put in place, it is necessary to pay close attention to the costs of educational investments and to attempt to reduce unit costs by improving efficiency. Cost-effectiveness analy- sis that compares alternative ways of achieving the same result is needed. The most cost-effective technique is the one that produces the desired result at minimum cost or yields the largest gain in educational achievement for a given cost. For example, cost-effectiveness analysis has shown that laboratories are not necessary for acquiring basic scientific competency. It has also been used to assess the cost of instructing students at a consolidated school or at multiple schools. Basic education provides the essential knowledge, skills, and attitudes for functioning effectively in society. Basic competencies in such general areas as verbal, computational, communication, and problem-solving skills can be applied ATTENTION TO OUTCOMES 97 in a wide range of work settings and can enable people to acquire job-specific skills and knowledge in the workplace (Becker 1964). This basic level typically requires about eight years of schooling. Basic skills are increasingly important in all societies. In the United States between 1978 and 1986 the wage premium associated with mastery of elemen- tary mathematics rose from $0.46 to $1.15 per hour for men and from $1.15 to $1.42 for women, for people with the same number of years of schooling (Murnane, Willet, and Levy 1993). East Africans with good basic skills are more likely to enter the modern wage labor market and to earn higher wages than are less literate and less numerate workers with the same number of years of schooling. Secondary-school leavers who scored in the top third on the school-leaving examination earn 50 percent more than those in the bottom third in Kenya and about 35 percent more in Tanzania (Boissi&re, Knight, and Sabot 1985). The high rates of return estinated for basic education in most developing countries strongly suggest that investments to expand enrollments and improve retention in basic education should generally be the highest-priority education investments in countries which have not yet achieved universal basic educa- tion. In many cases, expanded basic education coverage will require invest- ments to expand school capacity. to train qualified teachers, and to provide suitable educational materials. But in other cases, where insufficient capacity is not the binding constraint, the demand for education will need to be increased through actions designed to improve educational quality, to improve the school environment, or to defray the direct and indirect costs of school attendance. This is particularly so in poor settings where children contribute more to the household than they consume (Lindert 1976). Other education interventions may also merit a high priority. Investments to improve educational quality or efficiency often have high rates of return. In some cases these returns may be even higher than for investments to expand education coverage. The benefits of investments to improve efficiency in edu- cation-for example, through improved student retention or more intensive use of staff and facilities-can typically be expressed in ternis of reduced unit costs per student or per graduate. Although the benefits of improved educational quality-for example, as indicated by prospective earnings-are more difficult to measure, they should nonetheless be weighed explicitly in considering the relative priority of such investments. In addition to these investments, some improvemenits in educational effi- ciency or quality will often be possible through policy changes requiring no specific investments. For example, more efficient assignment of existing teach- ers could reduce the need for recruitment of new teachers in some education systems. 98 SIX KEY REFORMS Decisions about priorities for public spending on education beyond the basic level have to be taken within a broad sectoral approach. A distinction needs to be made between countries that have achieved. or almost achieved. universal basic education and those that have not. Countries that have achieved universal basic education are likely to consider upper-secondary and higher education the priorities for new public spending. Decisions about public priorities among these postcompulsory levels can be informed by the prudent use of economic analysis focusing on labor market outcomes and other social benefits. Economic analysis has shown. for example, that the average social rates of return to general secondary education are much higher than those to highly specialized vocational secondary education (Psacharopoulos 1987, 1994). This result is consistent with the constant and rapid changes in technology and labor markets that call for flexible, "trainable" workers able to acquire new skills as technology changes. The preferable way to achieve this goal is to emphasize learning skills and attitudes rather than specific job-related skills that are best taught in specific job settings. For ex- ample, in Indonesia, a rapidly industrializing country, the returns to academic and vocational secondary education fell between 1979 and 1986, but returns to academic secondary education nevertheless remained higher ( I percent. down from 32 percent) than those to vocational secondary educationi (9 percent, down fromn 18 percenlt). The low returns to secondary vocational education signal that additional investment under culTent conditions would be inefficient. Reducing the provi- sion of places is, however, not necessarily the correct policy response. For vocational education, the low returns could be attributable to higih costs rather than to a lack of demand for technically skilled labor. Therefore, one alternative would be to reduce costs in order to raise the returns. Possible reforms would include shortening the duration of the course and reducing unit operating costs. If the returns do not increase significantly, alternative arrangements for training skilled workers outside the formal school system might be appropriate (Milgat and Tan 1985). Complementary tools for assessing investment priorities include tracer stud- ies and aninual routines (Mingat and Tan 1985; Sapstord and Tzannatos 1993). These can be used to collect data relating to school programs and to monitor labor market trends affecting recent school leavers. This type of iiformation should be collected aniually, as it provides useful feedback on the effects of recenit policies so that adjustments can be made to the system. The annual routine can improve the responsiveness of investment decisions to new labor market conditions. It increases the external efficiency of educational invest- ments within the broad strategies suggested by the rate of return analysis. The advantage of economic analysis is that assumptions must be stated at the outset. In the above example. it is assumed that there is a demand for the ATTENTION TO OUTCOMES 99 output of the secondary vocational stream. But empirical analysis can be used to check this assumption. One direct method is through employer surveys, as in Indonesia. There, the results show an inconsistency between the public and private perceptions of the skills development problem: private companies do not view skill shortages as a problem, and most companies provide their own training (Dhanani 1993). Estimates of rates of return to investments in education are not very rel- evant when the labor market is not competitive-or does not exist, as was the case in the socialist economies of Eastern Europe. The dramatic economic changes in these countries are expected to result in a premium on entrepreneur- ial ability as workers for the first time face a competitive labor market where success is rewarded. As Nobel laureate T. W. Schultz has argued. entrepreneur- ial ability complements education and work experience. Therefore. relative returns to education should rise in the newly emerging market economies in comparison with retirns before the transition, when entrepreneurial skills were not needed. Empirical analysis of changes in the wage structure in Slovenia between 1987 and 1991 reveals that the returns to human capital (as measured, in this case, by years of schooling) rise dramatically during transitions (Orazem and Vodopivec 1994). In comparisons of educated and least-educated groups, those with four years of university education gained the most in relative earnings, followed by those with two years of university: those with vocational qualifica- tions gained least. Major changes in the economy-transition to a market economy, restructuring, trade liberalization, and new economic alliances such as the North American Free Trade Agreement-should therefore result in greater gains to those with more education and a more general education. Vocational education in high-income countries and in some upper-middle- income countries is accordingly becoming less specialized: it now typically includes a very extensive general education component. At the same time. general secondary education now typically also includes technology education. designed to facilitate entry into the world of work. Also increasingly common are school-to-work programs that combine general education at school with one or two days per week of work on the job and so accustom students to the attitudes required at work. This convergence of vocational and general second- ary education and the linkage of school and work in OECD countries has yet to be evaluated using economic analysis. but it seems likely to yield higher returns thani narrow vocational education. Many low- and middle-income countries with universal primary and lower secondary education now have large. highly specialized upper-secondary vo- cational programs. In the formier command economies of Eastern and Cential Europe. the bulk of Lipper-seconclar-y schooling consists of extremely special- ized vocational and technical education. In Poland. for instance, only 20 per- 100 SIX KEY REFORMS cent of secondary students are in general education, even though the emergiig market economy requires workers with a general education (World Bank 1992). Labor markets themselves are only now emerging in Eastemn Europe and Cen- tral Asia, and the relationship between them and the education system has still to develop. Countries cannot quickly reduce the size of such large vocational second- ary programs, but the programs should increasingly be made more general and linked to the development of attitudes and general skills, rather than specific skills, necessaiy for work. Evidence on labor market outcomes indicates that. over time, reducing the vocational share of secondary education is advisable. There are countries in which the returns to some types of specialized voca- tional education have been found at some times to be higher than those to general secondar-y education. This finding retlects shortages of certain skills in the labor market. In Chile, for example, the retumns to agricultural training. industrial skills, and commercial skills are all currently higher than those to general secondary education. Hence the central government subsidizes munici- pally run vocational centers, with the amount of the subsidy varying according to labor market needs. In 1993 agricultural schools received 200 percent of the subsidy to general secondary schools, industrial schools 150 percent, and coin- mercial schools 125 percent (Cox-Edwards 1994). Countries that have yet to achieve near-universal basic edLication face sini- pier decisions in deternlining priorities. Given the importance of achieving full coverage of basic education. economic analysis can help guide public choice about investments above these levels to select those that will clearly have a greater inmpact on labor productivity and other social benefits. In most countries the priority for new investmnent will clearly be basic education, but science at the upper-secondary level and science and engineering at the higher level may be exceptions. If the social rate of return to these courses, or to specialized vocational courses, is higher than that to primary aiid lower-secondary educa- tion. there will be a case for increased public investment. So far, few economic analyses have included this degree of specificity. but governments can use such analyses in specific country circumstances to guide investment decisions (McMahon and Jung 1989). Setting Stan(lardls ancl Moniitoring Peerfolo-riance Once the public sector has made decisions concemning allocation of public resources, an important step is to define the skills and competencies to be acquired at each level of publicly financed education and to monitor their acquisition. There is much scope for broader use of mechanisms for setting standards and monitoring learning outcomes (see chapter 4) and. ideally, for ATTENTION TO OUTCOMES 101 more use of internationally agreed definitions. The OECD, for instance. is pro- posing the continuoLus monitoring of three categories of standard outcome indicators for its member countries: student outcomes, system outcomes, and labor market outcomes. Student outcomes include performance in reading. mathematics, and science and gender differences in reading achievement: sys- tem outcomes include upper-secondary graduation, university graduation, sci- ence and engineering degrees. and science and engineering personnel. Labor market outcomes include unemployment, education, and earnings (Tuijnman and Postlethwaite 1994). After standards for performanice have been establishied, performance needs to be observed and linked to incentives. A variety of performanice indicators can be used, including, but not limited to, tests and examinations. What is tested tends to be what is taught. and public examinations, in particular. have consid- erable potential for improving the quality of student learning (Kellaghan and Greaniey 1992). But serious problems can arise if (a) examinations are linked to instruction in such a wav that the curriculum is narrowed, (b) an emphasis is placed on examinationi techniques and rote-memorized knowledge; and (c) past examinations start to dictate not only what is taught but how it is taught. Public examinations to raise qjuality cannot be the same as those for selection. as the latter do not take account of the needs of the majority of students who are not proceeding to the next level. Performance measures have both policy and pedagogical applications. They can be used to monitor progress toward national educational goals, evaluate the effectiveness and efficiency of specific policies and programs, hold schools accountable for performance of students, select and certify students, and pro- vide feedback to teachers about individual students' leamning needs (Larach and Lockheed 1992). They can also be linked with incentives to drive a system towarcl higher achievement. Internationally, there is greater experience with the use of performance indicators for individual accountability than for institutional accountability. Examinations for selection and certification, often but not always liiked to the culTicUlum, are found worldwide (Eckstein and Noah 1993). As a pedagogical practice, teachers typically use tests and quizzes to monitor student learning. The use of performance measures for institutional accountability is still fairly recent. Experience so far shows that such practice has potential. but also that this potential is limited. Holding schools fully accountable for their stu- dents' performance can be difficult because (a) statistically valid distinctions among schools cannot always be made- (b) comparisons of schools may not con-ect for differences in student intakes. in terms of socioeconomic status, or the social and physical conditions under which the schools operate: (c) school rankinigs can vary according to the particular outcome measure that is used: and 102 SIX KEY REFORMS (d) as an outcome of the publication of results, schools that are perceived to be doing well may attract students of high ability, to the detriment of schools that are perceived to be doing badly but may in fact be doing well for their intake and conditions (Greaney and Kellaghan 1995). More recently, performance measures have been used to target resources and have been linked to incentives for improvement. In Chile results from a national assessment system in four subjects have been combined with other social indicators to assist the Ministry of Education in targeting additional support to the poorest schools (Himmel 1995). Each potential recipient school is reviewed according to the average student performance on the SIMCE (the national assessment test), the school's socioeconomic level, rural-urban loca- tion, and number of primary grades offered; the test score couints as 50 percent of the school score. On the basis of the score. the schools are rated as "high." "'medium," or "low" risk: 46 percent of the available resources for school improvement are targeted to high-risk schools and another 46 percent to medium- risk schools. Within risk categories, schools compete for funds by proposing school improvement activities to be supported. In 1995 school-level scores on SIMCE will provide evidence regarding the effectiveness of four different edu- cational interventions to be tested on a pilot basis before widespread introduc- tion: interventions that do not boost learning will not be eligible for larger-scale implementation. Particularly important is the inclusion of socioeconomic sta- tus. Similarly, in New Zealand schools are funded in part in reverse proportion to the socioeconomic status of their students' families. Such mechanisms en- sure both incentives for higher achievement and the provision of resources to needier schools. In many countries national learning assessment systems are enabling minis- tries of edLication to monitor their own progress. evaluate the potential impact and cost-effectiveness of experimental programs, and improve the quality of their educational planning. Information from national assessments can inform teaching and learning processes when this informationi is disseminated broadly. The World Bank and other donors are assisting many countries to strengthell the institutionis responsible for national public examination and assessment systems (Larach and Lockheed 1992). This strengthened capacity will enable donors to monitor the effect of their support on an important development goal-the learning achievement of children. Networking among countries that are introducing learning assessments is also an increasingly important factor in improving their- imiplemnenltation. The World Bank is encouraging the use of performance and effectiveness indicators in the educationi projects that it helps finance. Especially importanit here are labor market and learning outcomes and the relationship of outcomes to inputs. C H A P T E R S E V E N Public Investment Focused on Basic Education G OVERNMENTS invest in education for many reasonis. This chap- ter looks at public investment in education fromil the strictly economic view- point of maximizing both efficiency and equity. In practice. other objectives are always present in decisions about public investment, but more attention to efficiency and equity in the allocation of new public investment on education would do much to meet the challenges that education systems face today. Such attention would lead to new public investmenit being focused on basic educa- tion in most countries. (This emphasis will clearly be least applicable to those countries that already have achieved near-uniiversal enrollment ratios in basic education.) To achieve efficiency, public resources should be concentrated in a cost- effective manner where the returns to investimient are highest. To achieve eq- uity. the government needs to ensure that no qualified student is denied access to education because of inability to pay. At the same time. and because the gap between the private and social returnis is larger for higher education thanl for basic education, advantage should be taken of any willingness to pay for higher education by sharing its costs with students and their parents. Governments canl also intervene. By bearing some of the risk, they can help correct the capital market failures that preclude financial institutions from lending 'for higher education. 103 104 SIX KEY REFORMS Combining these principles would result in a policy package of fees and efficient expenditure in the public sector. The elements of this package-which would have to be adapted to particular circumstances-would usually be: * Free public basic education. combined with targeted stipends for households that cannot afford to enroll their children and with cost-sharing with communi- ties. * Selective charging of fees for upper-secondary education, again combined with targeted scholarships. * Fees for all public higher education, combined with loan, tax and other schemes so that students who cannot afford to pay the fees out of their own or their parents' current income may defer payment until they have income them- selves. This fee system would be accompanied by a targeted scholarship scheme to overcome the reluctance of the poor to accumulate debt against future earn- ings of which they are not yet confident. * A goal of quality primary education for all children as the priority for public spending on education in all countries. * Improvement of access to quality general secondary education (initially lower- secondary and later all levels of secondary) over time. * Efficient spending at the school and institution levels in the public sector. Korea is an example of a country that follows most of these policies. Fees (including Parent-Teacher Association dues) account for only 2 percent of recurrenit costs at the primary level but for4l percent at the middle school level. 73 percent at the high school level, and 77 percent at the tertiary level. Overall, the private costs of education account for about 50 percent of recurrenit costs for the entire education system (Adams and Gottlieb 1993). Public expenditure is heavily concentrated on basic education: 44 percent for primary education. 34 percent for middle and high school education, and 7 percent for higher education. Pricing, Policy for Ptiblic Education Basic Education A complete basic education is normally provided free of fees, since it is essen- tial for the acquisition of the knowledge. skills, and attitudes needed by society. The definition of basic education is country-specific, but it typically encom- passes at least primary education and often lower-secondamy education, as well PUBLIC INVESTMENT FOCUSED ON EDUCATION 105 (although not always, as the example of Korea shows). The importance of primary education, in particular, is confirmed when externalities are taken into accoLint. To obtain the maximum gain for society as a whole, the top public priority is the acquisition of basic competencies by all students. Achievement of this goal requires increasing demand and ensuring access for every child through free basic education. Indonesia, Kenya. and Tanzania increased their enrollments significantly after abolishing primary school fees (Colletta and Sutton 1989; Lockheed, Verspoor, and others 1991). C6te d'lvoire made the mistake of introducing fees when the demand for primary schooling was al- ready declining. Unofficial fees and charges can be a barrier to enrollments in primary education, as in Ghana in 1992, when first-year intakes fell by more than 4 percent. Public schools should not, of course, be prohibited from mobi- lizing resources, in cash or in kind, from local communities when public fi- nancing is inadequate and such extra resources constitute the only means of achieving quality. Even when public basic education is free, there will be poor households that cannot afford to send their children to school or keep them in school because of direct and indirect costs, such as buying books or losing production around the home. Targeted stipends can help these households offset the income lost when children attend school. Such a scheme, for girls. operates in thirteen Guateina- ]an communities. Thailand provides bicycles to students in rural areas to enable them to reach distant schools (Lockheed, Verspoor, and others 1991). Many World Bank projects have financed free textbooks and uniforms for poor fami- lies, and a few have included experiments with direct stipends at the secondary level. Cost-sharing with communities is normally the only exception to free basic education. Even very poor communities are often willing to contribute toward the cost of education, especially at the primary level. In Nepal, for examnple, almost all primary and many secondary schools are built and inaintained by local communities. Matching-grant schemes may increase local involvement in schools, create a sense of ownership, and encourage greater private contribu- tions to education. Such schemes are increasingly common in Africa and in Latin America. Brazil. Ghana, India. and Tanzania are experimenting with matching-grant schemes for school construction. with World Bank support. In Brazil public funding is conditional on local communities achieving agreed targets for contributions of materials and labor for school construction. Many countries, including Bolivia. Cameroon, Ethiopia, Honduras, Senegal, Uganda, and Zambia. have set up social investment funds. These programs induce community participation, investment in employment generation, and provision of basic social services. World Bank projects support such funds in twenty-two countries. In Ghana the govemnment contributes up to two-thirds of 106 SIX KEY REFORMS the total project cost, and the communities supply labor and local materials. This approach has cut the construction time for new schools to about three months (World Bank 199 Ia). Through tripartite arrangements the World Bank- supported Social Rehabilitation Fund in Ethiopia finances building materials and school equipment, the community contributes labor, and the Ministry of Education provides a teacher. All these contributions are included in formal contracts. Upper-Secondary Education Since upper-secondary-school graduates will have higher earnings than those who leave school earlier. selectively charging tees for public secondary school can help to increase enrollments. Cost-sharing with communities can be en- couraged at the secondary, as well as the primary, level. Fees can be usually charged without affectilg overall enrollments, but enrollments of the poor and of girls do fall unless offsetting measures are taken. There is considerable evidence that houselhold demand for education is relatively price-inelastic- that is, unresponsive to increases in private costs (Jimenez 1987: World Bank 1986), except among the very poor. Indeed, inelasticity of demand could be a useful criterion for making decisions about charging fees. In Peru household survey data show that a moderate increase in fees does not affect the overall demand for secondary education, although it does discourage enrollments by the lowest income groups (Gertler and Glewwe 1989). Poorer families have difficulty meetilg the direct and ildirect costs of children's school attendance. To offset this hardship, secondary school fees can be combined with targeted scholarships and stipends to ensure equity in enrollments. If scholarships equiva- lent to fees and the direct costs of schooling were provided to the poorest 20 percent of Indonesian households, for instance. dropout rates at lower-second- ary schools could be cut in half (World Bank 1993c). The Bank has supported such targeted scholarship schemes for girls in Bangladesh and Pakistan and for poor families in Colombia. The charging of fees at one level can affect enrollmenits by other family members at other levels. A poor family that has to pay fees for an upper- secondary student may not be able to enroll other children in primary sclhool because the younger children's work is needed to generate the income from which the fees are to be paid (Chernichovskv 1985). This dilemma is precisely why fee charginig must be accompanied by targeted stipends to enable the enrollment of students from poor families. Fees used without compensatory measures will have a negative impact on the enrollment of such children. PUBLIC INVESTMENT FOCUSED ON EDUCATION 107 Higher Education In general, fees are justified at public institutions for higher education. Also acceptable is the elimination of subsidies for such noiiinstructional costs as student housing and meals, except where income taxation systems are very progressive or include a graduate tax (Colclough 1990); either measure can permit recovery of the costs of higher education from lifetime earnings. Most developing countries have neither an effective nor a progressive income tax. however. and a graduate tax or a system of fees and loans would be a more equitable way of recovering costs. Few countries have tried to use graduate taxes. As with upper-secondary students. the demand for higher education is rela- tively price-inelastic. A 10 percent increase in fees in Thailand would result in only a 2 percent drop in enrollment (ChLItikul 1986). An optimal policy would be full cost recovei-y by public higher education institutions, with students paying fees out of parental income and out of their own future incomes, through a loan scheme or a graduate tax. Such a policy is very distant in all countries, however, because fee levels are so low arid experience with loan schemes has been relatively disappointing. A good start in the right direction would be to charge fees covering 100 percent of the recurrent cost of student welfare services, such as food and housinig, and 30 percent of instructional costs. Several countries, including Chile, Jamaica. and Korea, are well on their way toward this target (table 7.1) Others have very far to go. In Senegal welfare payments to tertiary students accounted for nearly half of total public spendinig on higher education il the mid- 1 980s. In the Sahel countries of West Africa, welfare services account for 57 percenit of public spending on higher education, compared with 7 percent in Asia (table 7.2). Malaysia. by contrast, has contracted out student food seivices to private suppliers who recover all costs, and it is discussing privatizatiorn of student housilng. Student loan schemes are an essential complement to cost recovery and the charging of fees. Many students are unable to afford the cost of higher educa- tion out of their families' current income, and loan schemes permit them to pay out of their future earnings. About fifty countries. industrial and developing. have such schemes. More than half are in Latin America: other developing countries with loan schemes include China. Ghana, Egypt, India, Jordan. Kenya. Korea, Malawi, Malaysia, Morocco, Pakistan, the Philippines, and Sri Lanka. In most countries loans are repaid according to a fixed schedule: in a few. including Australia, Ghana, and Sweden, they are repaid as a proportion of a 108 SIX KEY REFORMS TABLE 7.1 FEES FOR PUBLIC HIGHER EDUCATION AS A SHARE OF UNIT OPERATING EXPENDITURE, SELECTED COUNTRIES Coun,rY (neald of'dlatai Percent Coiuintr Eindl veal- of diata Percent Algeria. 1990 0 Bolivia, 1991 3 Argentina, 1990 0 Hondulas. 1991 3 Bangladesh. mid-1980s 0 Sri Lanka, inid-1980s 3 Benin. 1991 0 Egypt, 1990 4 Brazil. 1991 0 Pakistan. 1990 4 Ghana, 1990 0 India, m,id-1980s 5 Guinea, 1990 0 Thailand, mid-1980s 5 Madagascar, 1990 0 Malaysia, mid-1980s 6 Malawi, 1990 0 China. 1991 9 Mexico, 1991 0 Japan, 1991 9 Morocco. 1990 0 Colombia, 1991 1() Niger, 1991 0 Nepal, mid-1980s 10 Nigeria, 1989 0 Kenya. 1991 12 Papua New Guinea, mnid-1980s 0 Barbados, 1991 15 Peru. 1991 0 Philippines. mid-1980s 15 Senegal, 1990 0 ULnited States, 1985 15 Sudan. 1987 0 Costa Rica. 1991 16 Uganda. 1991 0 Israel, 1991 20 United Kintdomil. 1990 0 Spain, 1988 20 Venezuela, 1991 0 Korea, Rep. of. mid-1980s 23 France, 1990 1 Viet Nam. 1991 23 Guatemala, 1991 2 Indonesia, 1989 25 Hungary, 1990 2 Jamaica, 1991 25 Chile, 1991 26 Note: Expeiidiliires on student housing are excluded. Siurte: Zider[iian and Albrechit 1995. graduate's income each year. Experience to date has been relatively disappoint- ing. Heavily subsidized interest rates. high default rates, and high administra- tive costs have led to very low recovery ratios (Albrecht and Zidermall 1991). Losses range from 30 percent in Sweden to 103 percent in Kenya, where the interest rate is heavily subsidized, administrative costs are high. and very few graduates repay their loans. In large part, the reason for this anemic perfor- mance is that loan schemes have been administered by govemment ministries and agencies rather than by financial institutions such as banks. PUBLIC INVESTMENT FOCUSED ON EDUCATION 109 TABLE 7.2 SHARE OF SECONDARY AND HIGHER EDUCATION BUDGET DEVOTED TO STUDENT WELFARE, SUB-SAHARAN AFRICA AND ASIA, ABOUT 1985 Couniftrv Se'ondar! Terliarv Sahelian countries 18.1 56.5 Burkina Faso 35.0 79.0 Chad 0.0 32.9 Mali 19.() 66.8 Mauritania 18.6 56.9 Niger 22.7 59.5 Senegal 13.2 43.9 Non-Sahelian SLub-Saharan Africa French speaking 24.9 57.0 English speaking 7.5 23.2 Asia 5.1 7.2 Son,,re: Jarou,se and Mingaa 1993. Loan schemes can be made financially sustainable, as the experiences of Quebec (Canada) and Colombia demonstrate. They require that the pLiblic sector bear some of the risk, since banks and other financial institutions are generally unwilling to accept students' likely future earnings as collateral. Public sector intervention is appropriate to offset this failure of the capital mnarket. The World Bank is supporting loan-scheme reform in Jamaica, Kenya. Malawi, the Philippines, Tunisia, and Venezuela. Sustainable loan schemes require an effective collection agency with incentives to minimize evasion and default. Interest rates must be positive in real terms. Income-contingent and graduated annual payment schemes are needed to encourage repayment com- mensurate with the student's future earnings, which will rise over time. Loan schemes alone will not suffice to enroll low-income students in higher education. Even though their future earnings will be high, students from poor backgrounds are understandably reluctant to take on debt against future earn- tigs that may not seem certain to them. In addition, while pursuing higher education, they forgo earnings that may be important for their families' in- come. Targeted scholarships and work-study programs are needed to overcome this problem. Work-study, in particular, can help low-income students finance their living costs, if not their tuition. In most countries sttLdenits come from 110 SIX KEY REFORMS relatively well-off backgrounds and have higih earnings prospects, and so the bulk of financial assistance should be provided through loans rather than schol- arships. Two alternatives to loan schemes are graduate taxes and nationial service. Graduate taxes are supplementary income tax payments made by university graduates. It is difficult to assess the potential of such tax schemes because no country has yet adopted one. In national service schemes students receive subsidies to attend higher education institutions and then work for society at below-market salaries. Nepal and Nigeria have had successful programs in which graduates went on to provide social services in rural areas. Several other countries. including the United States, are considering introducing such pro- grams. National service schemes can run the danger of turing into schemes of guaranteed public employment for higher education graduates. Priorities for Ptiblic Spending Basic education is the priority for public policy and hence also for public expenditure in all countries. The objective is usually to have all children enroll in and complete primary and, ultimately, iower-secondary school and to learn effectively in school so that they acquire basic skills. This goal is consistent with the objective adopted in 1990 by the World Conference on Education ftor All and supported by the World Bank. It is both efficient and equitable, having the highest returns and increasing educational and earnings opportunities for the whole population (World Bank 1990a). To achieve high enrollments and sustained performance, expenditure at the primary level may also need to be complemented by targeted expenditure on early childhood development for poor families. As primary er-ollment ratios go up, increasing public resources will be devoted to secondary education. Demographic change. in part fueled by pri- mary education for girls. can help this process. For example. Korea has been able to increase public secondary enrollments and increase spending per stL- dent at both primary and secondary levels without increasing the share of national income devoted to public spending on education. By the late 1980s lowered fertility rates allowed East Asian countries to spend a significantly lower share of GNP on primary and secondary education, compared with coun- tries with higher fertility rates (table 7.3). Countries that have largely achieved univer-sal primary and secondary edLu- cation face different challeniges in determining priorities for public spending on education. Higher education will take on a greater relative priority for public expenditure, as it does in OECD count-ies. It is important, however, that atten- tion to higher education not deflect spending from the basic levels, where PUBLIC INVESTMENT FOCUSED ON EDUCATION 111 TABLE 7.3 SAVINGS IN GNP ALLOCATED TO EDUCATION AS A RESULT OF LOWER FERTILITY RATES IN EAST ASIA Eypenditure oil l)itiliu v andl sevondarv Peo, entage o< GNP sav ed/ due fo ednhcation gro'wthI rates of school-age populatimit as ( po eiYeitalte thait were lower thati those, in: Et ononly vof GNP Ken 1ya Mexi o Pakistan Hong Kong 1975 2.0 1.2 1.0 1.0 1980-81 1.7 1.5 1.7 1.2 Jlapan 1975 4.2 4.0 3.8 3.8 1988-89 2.8 4.8 2.8 3.9 Korea, Rep. of 1975 1.9 0.6 0.4 0.4 1988-89 2.8 2.8 1.4 2.0 Malevsia 1980-81 4.4 1.3 0.4 0.4 1988-89 4.0 1.6 0.4 0.8 Sing7apore 1975 2.1 1.1 (.8 2.0 1980-81 2.2 2.0 1.3 1.3 Thailand 1975 27.8 0.6 (1.0 0.0 1988-89 2.6 1.3 0.3 (.8 Sout e. World Bank I993a. taible 5.2. quality remains relatively low in low- and middle-income countries and where economic restructuring threatens the maintenance of adequate funding. In Rus- sia, for instance. it is important to ensure that compulsory education remains adequately funded as financing responsibility shifts from the federal level to the state governments. It is also important to ensure that essential preschool programs, previously provided by public enterprises. be funded, at least for poor children, as the enterprises withdraw from this responsibility. 112 SIX KEY REFORMS Sustai nability A particularly important issue in the public financing of education is fiscal sustainability. To achieve this, policies are needed to increase the efficiency of public spending, to increase public spending when appropriate, and to comple- ment public with private financing (chapter 3). To ensure sustainability, it is always advisable to project the fiscal impact of expenditure measures several years into the future and to have sound financing plans. C H A P T E R E I G H T Attention to Equity ACHIEVING equity-an important goal for many governments- often requires more attention than it has received in the past. This is particularly true at the first level of schooling, especially when systems include private schools and private financing. The government has two fundamental concemns regarding equity. The first is to ensure that everyone has a basic education-the basic competencies nec- essary to function effectively in society. The second is to ensure that qualified potential students are not denied access to institutions because they are poor or female, are from ethnic minorities, live in geographically remote regions, or have special education needs. No qualified student should be unable to enroll because of inability to pay. To determine who is qualified at the postcompulsory level, a fair and valid means for assessing potential students' qualifications for entry is needed. Increased attention to equity will also increase efficiency. Considerable evidence now exists that improving the educational status of the poor, of women, and of indigenous people increases economic growth and reduces poverty. Investing in the education of girls from poor backgrounds sets off a process of intergenerational poverty reduction: educated women are more likely to send their own children to school. The well-known efficiency arguments for im- proving girls' schooling (Summers 1994) apply also to indigenous people. Had Guatemala invested in education to raise secondary enrollment ratios from 7 to 113 114 SIX KEY REFORMS 50 percent in 1960. for example, the cOuntry's per capita growth rate during 1960-85 would probably have been higher by 1.3 percentage points per year (Gould and Ruffin 1993; Barro 1991). Achieving equity at the first level of schooling is a matter of both increasing the demand for eclucation and meeting that demand through fhianlcing and special measures. Financing is important at all levels for those who cannlot afford to go to school-either because they and thei- parents cannlot pay the associated costs or because the household cannot afford to lose their labor services. Special measures tend to be concentrated on1 the lower levels of educa- tion. They include recruiting more female teachers to provide role models for girls, making special education available, providing bilingual education in coun- tries with linguistic diversitv. and conducting health and nutrition programs. Taken together, these measures amount to providing universal access to learn- ing (not just universal school attendance) at the primary level, which opens the way to equity at all levels of the education system. Fiiiaiicial Meastlres In most countries public primary education is free; lower-secondary education is often free as well. Even when there is no tuition charge. however, the direct and indirect costs for poor families can be too high to ensure enrollment and learninag. Direct costs can include transport, textbooks, exercise books, pencils, uniforms, and the like. If poor children lack these items, they may not attend school or may not learn. In Kenya during the 1980s the government made parents responsible for providing books. By 1990 in the poor arid and semiarid areas there was frequently only one book per class. The policy was reversed in 1992 for the poor areas of the country, with the support of a World Bank adjustment operation. In Morocco parents' reluctance to send daughters to school without proper clothing increases the direct cost of sending girls to school compared with boys. Funding assistance is also needed for the poor when the indirect costs of attending school are high in comparison with the child's contribution to the household economy. It is often more difficult for parents to send a girl to school than a boy because girls contribute more hours of work at home than do boys. In Burkina Faso girls age 7 ancd over spend 3.5 hours a day on household tasks, compared with 1 .5 hours for boys (Chowdhurv 1993). The use of child labor reduces the demand for schooling, again because of poor families' need for income. Children work for a variety of reasons, the most important being poverty and the pressure to escape from it. Although children are not well paid, they still make major contributions to family income in developing countries. For example, minors in Paraguay contribute almost a ATTENTION TO EQUITY 115 BOX 8.1 REDUCING THE HOUSEHOLD out rates in upper grades. Students COSTS OF GIRLS'EDUCATION must maintain a certain grade aver- IN BANGLADESH age to continue to earn stipends. The project also supports a number In Bangladesh the World Bank-sup- of other measures to encourage fe- ported Female Secondary School As- male school enrollment. These include sistance Project provides stipends to increasing the proportion of women girls. The stipend rates are structured teachers and conducting a community to reflect rising educational costs from awareness program to promote public lower to upper grades and to provide support for girls' education. extra incentives for reducing high drop- quarter of total family inicome (Patrinos and Psacharopoulos 1995). In rural Java, Indonesia, a typical I 3-year-old boy from a poor household earns about $11 a month as an agricultural laborer, contributinig. on average, about 40 percent to the household incoine. This income is more than twice the average per student direct expendliture on lower-secondary education among poor house- holds (Mason 1994). These figures are underestimates, since they do not in- clude the value of children's contribuLtions to home production. One reason why parenits' demand for educating daughters is low is the associated direct and opportunity costs in many countries, including Bangladesh. Egypt, Guatemala, Mali. Morocco, Peru. Tunisia. and Yemen. Some projects have cut these costs by waiving or reducilng fees, supplying free textbooks. providinig scholarships or stipends for ghils, offering flexible school hours, and establishing childcare centers. Such approaches, in addition to reducing costs to parents. improve school quality, reduce dropout rates, improve the effi- ciency of the school system, and significantly increase girls' effective partici- pation. Bangladesh and Guatemala have girls' scholarship programs in which tuition is free and stipends are paid to parents to compensate them for other direct costs, such as books, and for the loss of their daughters' time. Bank- supported projects provide stipends for girls at secondary school in Bangladesh (box 8. 1). scholarships for rLiral girls in Morocco and Mozambique, and incen- tives for girls in the Gambia to study science. More analysis is needed of the effect of these schemes on school quality and of their fiscal sustainability. Targeted scholarship schemes can be used to increase the demand for educa- tion among all disadvantaged groups, not just womnen. Several middle- income countries are experimenting with targeted scholarship schemes for stu- dents who cannot afford fees. These schemes cover the cost of tuition but do not provide any compensation to the family for the loss of the child's time. The 116 SIX KEY REFORMS targeted vouchers used in Colombia and in the U.S. commonwealth of Puerto Rico combine targeted subsidies with student choice of institutions. Fees at higher education institutions must be combined with student loan and scholar- ship programs to ensure that all who wish to borrow for their education are able to do so and to guarantee necessary financial support to academically qualified poor students. When the University of the Philippines raised tuition fees in the late 1980s. for example, it also provided a special fund to support qualified students from low-income families. Special Meastires Special measures are needed to increase the enrollments of girls. of the poor, of linguistic minorities, and of special populations. As poor parents do not always appreciate the value of educating their children. and many parents do not see the value of educating their daughters. investing in parents' education can be an important mechanism for increasing child schooling. Social marketing or aware- ness campaigns can help overcome lack of knowledge: examples are the com- munity female education awareness program in Bangladesh (see box 8.1), the Pacto Pela Infancia in Brazil for disadvantaged children, and the program in Guatemala to promote daughters' education among fathers. Changes in the location, schedule. staffing, content, or direct costs of education can make schooling more relevant to social and material conditions (Colletta and Perkins 1995). Such measures include recruiting more women teachers and more teachers from the local community. Gi-ls Parents in many countries would like their girls to be taught by women, and a shortage of female teachers can inhibit school attendance. In Kerala state, which has the highest female literacy and enrollment rates in India, more than 60 percent of teacher-s are women, compared with less than 20 percent in Bihar and Uttar Pradesh, the two states with the lowest enrollment rates for girls. Not having a school withinl easy reach of home also deters girls' enrollment because of parents' concerns about girls' safety. In Morocco the presence of a paved road increases by 40 percent the chance of a girl ever attending school and reduces the probability of her dropping out by 5 percent. Too often, girls do not go to school because of the lack of separate lavatories and common rooms. In some cultures girls' participation in school depends on whether single-sex schools are available. The basic policy instrument for expanding girls' enrollments is to increase school places for them. This can be done by reserving places and by expanding ATTENTION TO EQUITY 117 enrollments. If there are too few places, those available often go to boys. In Malawi one-third of all secondary school places are reserved for girls, and a Bank-assisted project to build secondary schools resulted in higher female enrollments than expected. Tanzania and Zambia have similar policies. Bangladesh, Chad, India, Pakistan, Senegal. and Yemen have made special efforts to expand classrooms or build new schools for girls. In Bangladesh and India these improvements include women's colleges and polytechnics at the tertiary level, as well as at the primary and secondary levels. Evidence shows that in many cultures girls' enrollment and performance improve if they attend single-sex schools rather than coeducational schools (Lee and Lockheed 1990). Care should be taken, however, that there are no differences in curricula in such cases. Some projects in Bangladesh and Pakistan are also providing separate sanitary facilities and constructing boundary walls around girls' schools. Lo- cating schools within easy access of children's homes can lessen parents' con- cerns about girls' personal safety and reduce the direct costs of transport and boarding. Morocco is providing small local schools for middle-level education. Other ways to increase girls' enrollment are to provide female teachers and childcare centers and to adjust school hours to fit girls' schedules. Cross-coun- try data suggest a strong positive correlation between the parity of enrolilment for bovs and girls and the proportion of female teachers (Psacharopoulos and Tzannatos 1992). Initiatives in World Bank projects to overcome the shortage of female teachers, especially in rural areas, include implementing a quota system to recruit more female teachers, removing age restrictions, recruiting and posting teachers locally, and building teacher-training institutions in rural areas. Experience fiom Bangladesh. Pakistan, and Nepal suggests that it is not hard to find good female teachers if required training is providecd and teachers are posted near their homes. The combination of locally recruited and moti- vated women teachers and active in-service training and supervision can re- duce the shortage of women teachers in rural areas. Ongoing Bank-financed projects in Bangladesh. China, India, Nepal. Pakistan, and Yemen are testing such strategies for increasing the proportion of female teachers. Childcare centers at or near schools and flexible hours can free girls to attend school. Childcare provision relieves girls from sibling care during the day and, when accompanied by nutrition programs, can help improve the health and school readiness of younger siblings. In Colombia. where single mothers head one-fifth of the poorest households and where 44 percent of poor children between ages 7 and II do not attend school. the community day-care program has freed many girls and women to attend school or join the work force. Adjusting school hours so that girls can more easily combine schooling with chores has also worked well in many countries. particularly in Nepal (World Bank 1994b). 118 SIX KEY REFORMS Special Populations The principal policy instrument for reducing the high incidenice of phvsical and learning impairments in developing countries is improvement of child nutrition and health. Special programs to improve the nutrition and health of schoolchil- dren can help increase access to and equity in schools. For example, school feeding programs can be designed to have a differential impact oni the enroll- ment and participation of girls, as in Ghana. Other programs, such as the treat- ment of parasites and micronutrient fortification or supplementation-both of which are relatively inexpenisive and easy to implement-can significantly improve the disadvantaged child's ability to take advantage of educational opportunities. Educating children with minor impainnents cloes not usually require costly facilities or programs. In India, for example, an Integrated Edu- cation for the Disabled project led to the identification and education of more than 13,000 children with special educational needs at a unit cost comparable to that of regular education (World Bank 1994j). Disabilities affect about 140 million children, about 15 percent of whom could have their sight, movement, or hearing enhanced at a unit cost of $25-$40 (Mittler 1992). Unit costs tor special education can be reduced by using community-based approaches, which also create better opportunities for children. Community-based rehabilitation programs exist in man\y countries, including India, Indonesia, Jamaica. Kenya, Malaysia. Nepal. the Philippines, and Zimbabwe. Costs can be sharecd with non0governmiental organizations. as in Indonesia, where the public education system provides 45 percent of the resources for special education and private voluntary agencies provide the other 55 percenit (World Bank 1994j). Language Diversiy In multilingual nations. reading comprehension is often greater for students taught in bilingual schools, who first learn to read in their native language and then transfer their reading skills to the second language. More than 40 percent of Guatemalans entering school do not know Spanish. In 1979 Guatemala established a national bilingual education program. with tISAID and World Bank support, to improve the quality of education for the indigenous popula- tion. The national cunriculum was adapted and translated into four Mayan languages for the preprimary level through grade 4. The program has led to an increase in student comprehension and has reduced student failure, repetition, and dropout rates compared with a control group of Mayan children being taught only in Spanish. Bilingual program students score higher in all subjects, including Spanish, and have a promotion rate 9 percent higher (World Bank ATTENTION TO EQUITY 119 1994d). Bilingual education also has the support of the children's parents and so increases the demandl for education (Richards and Richards 1990). Effective schools in multilingual societies may be those that are permitted flexibility in language of instruction (Eisemon 1989: Eisemon. Ratzlaff, and Patel 1992). The implementation of language policies should not be prescribed by national authorities, at least at the primary level. There, the focus should be on languiage learning outcomes and perhaps on the establishment of general objectives for the use of the mother tongue and other languages at particular grade levels and for particular subjects. Implementation should be a local. preferably school-level, responsibility. Institutional autonomily makes local imple- mentation easier, since local schools and communities know their own circum- stances best. Other Disadvantaaged Grou ps Special attention is also needed to ensure equity of access to other disadvan- taged groups, such as nomads, those who live in geographically remote re- gions, street children, and refugees. Strategies must vary from country to coun- try. and nonformal methods will often be more appropriate than formal school- ing. Particularly troublesome is the growing nulimber of refugee children in Africa: many of themii have no government to take the responsibility for provid- ing them with schooling. C H A P T E R N I N E Household Involvement EDUCATIONAL institutions may be more accountable for their perfonrance when households are more closely involved in the institutions that family members attend. Parents involved with a school are more likely to be satisfied and, even more important, to help make it more effective. Most house- holds already contribute, directly or indirectly, to the costs of education, but they could participate in school management and oversight. along with their wider communities, and they could be given the possibility of choosing among schools. School Governance Around the world, parents and communities are becoming more involved in the governance of their children's schools, just as students are in their higher education institutions. Sri Lanka fumnishes one example (box 9. 1). The elected boards of trustees that manage schools in New Zealand are drawn from parents of children at the school. In Mauritius parent-teacher associations have been so successful that government funds are now being used to stimulate the partner- ship further. Many countries have found that communities which participate in school management are more willing to assist in the financing of schooling. Jamaica has set up a major program to stimulate this tendency; Bangladesh's 120 HOUSEHOLD INVOLVEMENT 121 BOX 9.1 SCHOOL DEVELOPMENT * Moral development (promotion of BOARDS IN SRI LANKA cultural, religious, and moral activities) Legislation enacted in 1993 in Sri * Physical resources (infrastructure Lanka established school development development) boards (SDBS) with the purpose of pro- * Library and educational equipment moting community participation in (facilities improvement) school management (Commonwealth Secretariat 1994). Each SDB consists * Schoolbooks, midday meals, and of representatives from the school staff, uniforms (deciding on school require- parents, past students, and well-wish- ments) ers and is chaired by the school princi- C Welfare and community relations pal. Acting through ten subcommittees, (strengthening welfare activities) the SDB decides on and implements programs for the school's develop- m Communications (interaction with ment. The ten subcommittees are: media and the community) * Educational development (improve- * Finance (utilization and disburse- ment of academic curricula and modes ment of school funds) of teaching) * Student personality development * Co-curricular activities (promotion of (development of the personality of the extracurricular activities) pupils and the school). Social Mobilization Campaign. which involves the community in education, has been accompanied by a reactivation of school management committees throughout the country. El Salvador has started to involve communities in rural school management. with significant results in improved teacher attendance. Student achievement levels are comparable with those at traditional schools, even though the students tend to come from poor backgrounds. Effective involvement in school govemance does not come easily, how- ever. New Zealand realized after it had embarked on its reform that intensive training was necessary for the newly elected parent trustees. Jamaica is training parents to help manage schools. Botswana found it very difficult to attract sufficiently qualified people to lower-secondary school boards of govemors, especially in rural areas. Training can be effective both in literate communities, such as New Zealand, and in relatively illiterate ones, as in parts of Uganda. ActionAid in Uganda is providing community training in two districts for parent-teacher associations and school management committees. 122 SIX KEY REFORMS School Choice Several countries have a long tradition of parental choice. The Netherlanids has had it throughout the whole twentieth century. In poor African countries such as Uganda there has always been complete freedom of choice for parents. Increased experimentation with parental choice is anothier hallmark of re- cent educational reforms. particularly in Australia. Chile. England, the Nether- lands. New Zealand. Sweden, and the United States. This trend retlects both a more market-oriented perspective on education, in which consumers (parents ancl students) choose among suppliers (schools and institutions), and the "choosiness" of a growing number of parents and students. who will no longer accept being assigned to a particular public school but want to make their own decisions (OECD 1994b). For choice to be an effective concept. the following factors are important. * The student must have within reach either more than one possible school or institution or multiple programs within a single institution. * The institutions should have some distinguishing characteristics. * Schools and institutionls need to enjoy considerable autonomy in how thev teach. Diversity among institutions or programs can take the shape of differing emphases in the curriculum, styles of teachinig. higher-level course offerings. and ownership (public or private). The existence of a variety of types of pro- grams and institutiolIs makes it possible for parents and students to exercise some choice, which in turn shouild give an incentive for providing quality education cost-effectively. This strategy is most relevant at the upper-second- ary and higher levels of education. where choice among institutions can help to meet the growing unmet enrollmenit demand. In higher education, for instance. choice among institutions of clifferent types will typically imply the development of nonuniversity tertiary institutions and the encouragement of private and public institutions. Higher education institutions are much more varied in high-incomile countries thaii in low- and middle-incomie ones. Differentiation among developing countries is most ex- tensive and effective in Asia. In East Asia. for example. the average anilual growth of university enroliliments between 1980 and 1988 was 6 percent, but for nonuniversity institutions it was 10 percent. The principal advanitages of nonluniversity institutionis are lower program costs (retlecting shorter courses). lower dropout rates, and lower per student annual costs. In Bulgaria average costs at universities are 15 percent more than at the higher institLItes and 95 percent more thanl at the technical institutes. Nonnilliversity institutionis also HOUSEHOLDINVOLVEMENT 123 offer training opportunities that responld tlexibly to labor market demand rather than to supply-side factors. World Bank-assisted projects support the differen- tiation of higher education. In Tunisia, for instance, a project is helping to establish a network of two-year technology institutions. Greater separation of teaching and research will promote differences among and reduce costs at public universities. Much scientific research requires ex- pensive scientific equipment. and there are undoubted benefits from concen- trating efforts at a few institutions. The prevailing assumption that every public university should conduct research therefore needs to be reexamined. Effective choice also means having private as well as public schools avail- able. Most countries pennit private schools. but some, including Algeria. Latvia. Syria. and. until recently, Pakistan. do not. Other countries excessively regulate the establishment and operation of private schools and universities. In Nigeria, for instance, it takes well over a year to go through the cumbersome bureau- cratic process for establishing a private school. The proprietor must meet many ostensibly reasonable requirements (for example. showing that the school will be nonprofit and that there are sufficient resources to operate it for a specified period, posting- a bond, and so on). but the real purpose is to put difficulties in the way of establishing private schools. The ideal is a positive regulatory framework. In higher education such structures-in place in Colombia. Kenya. and Romania, for example-provide an appropriate legal basis and accreditation system for both private and public universities. In some countries, sucih as Chile. students may enroll at the pri- mary and secondary level in public or private institutions of the parents' choice, the state provides the funds. Another factor in ensuring effective household decisionmaking is educa- tional institutions' autonomy in how they teach. This freedom is related to the second ftactor, since autonomy permits different combinations of inputs that in turn produce institUtions with distinct characteristics. Private schools are auL- tonomous; public ones can be but often are not. Indeed, management differ- ences at the school level are probably one reason why private secondary' schools in a study of five developing countries appear to provide more learning for the same cost as public ones (table 9.1). even after controlling for the socioeco- nomic backgrounds of the students. Private schools per se are not necessarily more effective than public ones, however. In the Indian state of Tamil Nadu yovernment-aided schools are niore cost-effective in raising student achieve- ment in mathematics and reading. but fully private unaided schools are less cost-effective than public schools. School management practices in govern- ment-aided schools-notably. instructionial management by the prinicipal, the quality of textbooks, training for teachers in how to use textbooks. and the availability of instructional materials-go far to explain their better perfor- 124 SIX KEY REFORMS TABLE 9.1 RELATIVE AVERAGE COST AND EFFICIENCY OF PUBLIC AND PRIVATE SCHOOLS, EARLY 1980S Ratio f private cost Ratio of re/alwthv Ratio of Ie/ative co'itoo-v to phlic (OS! puc' sil eflCs to cost e(ost to ej eficive'nress Colombia (0.69 1.64 0.61 Dominican Republic 0.65 2.02 0.50 Philippines 0.83 1.20 0.83 Tanzania 0.69 1.68 0.59 Thailand 0.39 6.74 0.17 Source: Lockheed and Jimiieniez 1994. mance (Bashir 1994). School-level management, not private or public status, is what can affect outcomes and hence provide real choice to households. Risks Although there is a long tradition of school choice in many countries, little systematic research into its effects has been done to date. As yet, no evidence exists that the competition among schools and programs implicit in the concept of school choice improves or worsens school performance. However, "the dynamic of competing for pupils typically enhances some school characteris- tics associated with effectiveness. such as strong leadership and sense of mis- sionl" (OECD 1994b). This finding points in the direction of further cautious experimentation in focusing increased houselhold involvement on school choice. (No such ambiguity exists with regard to increased involvement in school governianice.) Increased household involvement carries several risks. It can make carry- ing out systemwide education policies and enforcing broader national objec- tives more difficult. Social segregation can also increase if the education sys- tem becomes polarized between prestigious schools for the academically able children of educated parents and schools with less impressive credentials for the children of the poor and uneducated. Equity can be reduced if schools and institutions start accepting students on the basis of their ability to pa) rather thani on their academic entrance qualifications. These risks can be mitigated relatively easily through policies for the provi- sion of public funding and, as in the Netherlands, through limits on fees at schools that receive public funds. Public finance for public schools or to assist private schools can be restricted to schools that follow certain overall policies, HOUSEHOLD INVOLVEMENT 125 such as adhering to a national curriculum and practicing nondiscriminatory enrollment policies. in addition to meeting basic health and safety require- ments. Public finance for the education of children from poorer backgrounds can be provided at a higher unit level than for children from better-off socio- economic groups. as is the practice in New Zealand, for instance. Public finalic- ing or provision of transport to school can be adjusted so that it is not difficult for a child to attend a school other than that nearest to the family home. Another risk is that parents may not have enough information to make effective judgments about quality. Many studies show that parents' decisions about schools are not made principally on the basis of well-informed compari- sons of educational quality (OECD 1994b). Although this risk can never be completely overcome. it can be mitigated through the provision of open and independent information about school quality. It is appropriate for the govenn- ment to supply such information, since there is an infonnation asymmetry between the education system and households: school administrators have ac- cess to performance and financial data that is unavailable to students and par- ents. The British government provides inspection reports and examination re- sults in a forn that is designed to be meaningful to parents. In the United States the Boston school system, having broken the link between place of residence and assignment of public school places, has set up infonnation centers to assist parents in making decisions among public schools. Since school choice was introduced in Sweden, schools have prepared information on curricula and finances. Governments in low- and middle-income countries may not be able to emulate all these practices, but they could make available infonnation from national examinations and from the national assessment systems that are in- creasingly being introduced. Kenya's Ministry of Education, for instance. pub- lishes a table ranking the performance of secondaiy schools on national examni- nations. C H A P T E R T E N Autonomous Institutions EDUCATIONAL quality can increase when schools are able to use instructional inputs according to local school and community conditions and when they are accountable to parents and communities. Increased household involvement to increase accountability was discussed in chapter 9. If effective use is to be made of instructional inputs. institutions must be autonomous. Such a strategy is relevant in all contexts, even remote rural areas. Fully autonomous educational institutions have authority to allocate their resources (not necessar- ily to raise them). and they are able to create an educational environment adapted to local conclitions inside and outside the school. It is important to note that school autonomy is not the same as either local financing or adminiistrative decentralization. although the three are often con- fused. Local financinig of education means that resources are raised locally. which can create problenms of equity among r icher and poorer localities. Decen- tralization is simply assignment of responsibility for education to an institution or level of government other than the central govenml1ent. InstitUtional au- tonomy can be encoulaged by both administrative and financial means. Administrative Meastires To obtain the necessary flexibility, school managenmenits (principals and gov- ernmig bodies) must have authority to allocate resources. This includes the authority to deploy personnel anid to deteniine such things as the timnig of the 126 AUTONOMOUS INSTITUTIONS T27 school day, the duration of the school year. and the language of instruction, to fit local conditions. Such authority will increase the efficiency of learnmig. Teachers must have authority to determine classroom practices-within limits set by a broad national curriculum, encouraged by examinations, and moni- tored and supported by standards. learning assessments, and school inspectors. Finally, school staff must be accountable to the local community. So long as schools are solely accountable to central bureaucracies. they will be organized in a management structure that limits school autonomy (Hannaway 1991 ). This framework reduces incentives to respond to parental and commu- nity concerns about school performance and costs, and it curbs schools' capac- ity to respond because clearances must be sought before schools can proceed with chaniges in their operations. This is trule for countries at all income levels. However, schools in low- and middle-income countries are much less autono- milous thani those in highi-income countries (tables 10.1 and 10.2). Relatively fewer decisions are made by principals and teachers, and too often. decisions are made without appropriate input from local authorities (Lockheed. Verspoor, and other-s 1991: OECD 1993). Local finanicinig and decentralization can contribute to autonomy and ac- countability, but this does not happeni automatically. as recent experience in Nicaragua, India. Chile and Russia demonstrates. As part of decentralization. the Nicaraguan Ministry of Education transfers funds to municipalities. which tihen hire, fire, and pay teachers. The potential benefit of this move has been nullified, however, by a law that stipulates a national teacher pay scale and by the insufficielIcy of the funds transfer-red from the ministry. Neither schools nor municipalities gain autonomy through such an arrangement. They do gain autollomy under a more promising refo.i-l: Nicaragua is transferring public secondary schools to private associations. To date. 20 of the country's 350 secondary schools have been transferred. Amendments to the Indian constitution in 1992 shifted authority, including responisibility for education, to locally elected bodies (revived Panchayati Raj institutions) at the villaoe, intermediate. and district levels withil states. In response, teachers in Andhra Praclesh state successfully petitioned to become state employees. to limit the authority that the panchayats will be able to exer- cise over them. After local financing was introduced. public spendinlg on educatioll de- cldned 17 percent a vear in Chile in 1985-9(0 and 9 percent a year in Mexico in 1982-90 (Prawda 1993). Local financing is also lowering spending in Russia as the federal goverinimient transfers responsibility downward. As these examples show. it is critical to the improvement of learning that resources not be reduced when local management and financing of schools are increased. Measures are needed to ensure that adequate resources are available to each school if local financiig is adopted. The purpose of increasing school 128 SIX KEY REFORMS TABLE 10.1 LOCUS OF DECISIONMAKING AUTHORITY IN PRIMARY EDUCATION SYSTEMS IN SELECTED DEVELOPING COUNTRIES (percent) Tvpe o?f decision) anid, deciding body Rep. of Korea Philippines Nigeria Authorizing major expenditures Central or regional body 9 66 42 School board 78 7 33 School principal 11 5 1 Teachers 0 0 0 Selecting principals Central or regional body 40 83 38 School board 39 3 45 School principal 0 0 0 Teachers 0 0 I Selecting teacher-s Central or regional body 6 63 37 School board 71 7 49 School principal 3 14 1 Teachers 0 1 0 Deternmining the range or type of science courses Central or regional body 61 82 82 School board 3 0 5 School principal 5 5 5 Teachers 28 5 5 Choosing science texts Central or regional body 89 76 59 School board 5 1 12 School principal 1 2 9 Teachers I 0 6 Note: Percentages are based on the responses of teachers and school administrators to questions about decisionmaking authority. The four categories of decisionmaker do not include all the options. so percentages may not total 1)00. Soure:e Lockheed, Verspoor. and others 1991. table 5-l. autonomy is to permit flexibility in the combination of inputs and hence im- prove quality-not to save resources. For this reason, institutional autonomy need not involve local generation of resources but only local control over their allocation. AUTONOMOUS INSTITUTIONS 129 TABLE 10.2 DECISIONS MADE AT THE SCHOOL LEVEL AS A SHARE OF ALL DECISIONS BY PUBLIC SCHOOLS IN OECD COUNTRIES, BY LEVEL OF EDUCATION, 1991 (percent) Level of education Prirnars Lower- secondary' Upper se ondar v Austria 44 44 47 Belgium 29 26 26 Denmark 39 39 42 Finland 41 38 59 France 17 35 35 Germany 32 32 32 Ireland 50 74 74 New Zealand 73 72 79 Norway 31 31 26 Portugal 33 42 42 Spain 28 28 28 Sweden 47 47 47 Switzerland' 9 9 23 United States 26 26 26 a. Almost all educaition decisions in Switzerland are made at the lowest level of goveriment. the canton. S0141 fe OECD 1993. Autonomy and accountability leading to flexibility also require that schools be allowed to manage themselves in ways that create conditions conducive to learning. For example. flexible use of multigrade teaching, at the option of the individual school, has been largely responsible for the success of the Escuelas Nuevas in Colombia, in which flexible techniques are applied to the formal education systein. Similarly. with the program of the Bangladesh Rural Ad- vancement Committee (box 10.1. flexibility has been effective on a large scale outside the formal system. There is an important place for increased teacher involvement in decisionmaking in schools. Teacher participation will improve learning qual- ity, but only as long as it is explicitly focused on instruction (Smylie 1994). Such a focus requires some outside influence and direction (David and Peterson 1984: Shavelson 1981). Without such direction, teachers' energies are likely to be diverted to areas at best only tangentially related to instruction (Haniaway 1993). The best external direction for focusing teachers on instruction is a national or regional curriculum. Clusters of schools. somnetimes called nucleos or school learning cells. facilitate professional interaction among teachers and decisionmaking about instruction. Indeed, professional interaction inay be more important than 130 SIX KEY REFORMS BOX10.1 INVOLVEMENTOFNGOS In 1992 more than 8,000 schools IN EDUCATION: THE BRAC STORY were operating, and plans are being implemented to expand the NFPE pro- In diversifying the supply of education gram to 50,000 schools nationwide by in Bangladesh, the government has 1995. Throughout, BRAc has been able recognized the contribution nongov- to balance its expansion program with ernmental organizations (NGOS) can its quality goals. make to efforts to expand access and Internationally, BRAC is a model for improvethequalityofeducationforthe the potential of the nongovernmental country's children. NGOS already play sector in educational expansion. It also a national role in Bangladesh's health illustrates how a combination of tar- and population programs. geting, school design, flexibility, and The Bangladesh Rural Advance- follow-through can dramatically in- ment Committee (BRAC), the largest crease girls' primary-school participa- NGO in Bangladesh, is well known for tion rates. While national education its rural development, credit, and health systems everywhere have the formal programs. In 1985, in response to re- mandate to provide quality education quests from participants in its rural de- to their populaces, such NGOS as BRAC, velopment programs, BRAC started the which have greater flexibility than gov- nonformal primary education (NFPE) ernment bureaucracies, may some- program for 8-to-10-year-olds in times be able to reach target groups twenty-two villages. Girls were given more effectively. Furthermore, BRAC'S special emphasis. By late 1991,6,003 expansion program illustrates that schools, serving 11 -to-1 6-year-olds as NGOS need not necessarily be limited well as the NFPE age group, had been to small pilot projects but can also carry established. The program is free to stu- out larger-scale delivery programs dents, except for community contribu- (Ahmed and others 1993). tions for school construction. decisionmiiaking authority for the motivation, learning, and social control of teachers' work (Hannaway 1993). Periodic conferences or workshops can pro- vide opportunities for teachers representing different clusters to share what they are doing with their- colleagues. School cluster-s have been used success- fully in Costa Rica to develop new curricular imlaterials in local languages and in India and Sri Lanka to share demonstration lessons amon,g teachel-s (Bray 1987.) Financial Nleasuires Public finance can be used to encourage autonomy and accountability. The standarcl pattern of public education financinig is to raise revenues throughl AUTONOMOUS INSTITUTIONS 131 general taxes and to allocate expenditure centrally. through direct payments for inputs such as teachers' salaries and textbooks. The more that schools them- selves can control the allocation of resources, however, the greater the possi- bilities ol' effective schooling. And the more that households are involved, the greater will be the incentives for the schools to improve quality. Public fiianice mechanisms for achieving these purposes include the use of local rathier thani centr-al government taxation: cost-sharing with local commu- nities; the use of block grants: the charging of fees at higher levels of education: the encouragemenit of revenue diversification; the use of "portable" capitation grants, vouchers. and stLldeit loans: and funding basecl on output and quality. A variety of mechanisms are available for specific circumstances and levels of education. Local Taxation Fundinig of education through local taxes can increase the accountability of schools and institutions to parents and students. Local taxes are often used to fund school systems: in the United States. for instance, local property taxes are usually the principal source of revenues for school districts. There are two principal drawbacks to the use of local revenues for education. First. local governments may have less capacity thani national ones to administer tax sys- tems. Second, localities vary in their resource mobilization capacities. and this can lead to different per student funding levels and to inequalities in access, quality, retention. and learning outcomes. The advantages of cost-sharing with local communities have already been discussed. This subject is partictilarly relevant at present in Eastern Europe. where responsibility for education at the primary and secondary levels is being transferTed from central to state and local governments at the same time as fiscal federalism is being introcluced. Revenue-sharing formulas could be adopted, as in Australia, to offset differences in state governments' fiscal ca- pacities. Local fuLidinig is not essential to increased accountability. however: local control of centrally financed expenditure can have the same effect. Bloc-k Gr ants In Australia primary education is principally a state responsibility, but central government funds are allocated redistributively to states and districts on a per student basis in direct relation to the relative poverty of the district. In New Zealand the central government provides grants for operatting costs directly to the school: the funds are adminiistered by a locally elected board of trustees. Suchi mechanlisms permit local control of resouLlces for education without plac- 132 SIX KEY REFORMS ing all the burden of resource mobilization on the local community or govern- ment. They can also offset differences in the socioeconomic status of students. In New Zealand, for instance, 80 percent of school funding is related to the nunmber of students and 20 percent to the students' socioeconomic status. Poorer students thus attend schools that receive more funding per student. Fees Charging fees can create accountability between parents and students and school managements at the higher levels of education. Scholarships can be used to encourage enrollment of students from low-income families. Even at the pri- mary level, the charging of fees need not be incompatible with the principle of free primary education. so long as those fees are regulated and are met by parents out of vouchers financed by the state. as is now done in Chile. Revenue Div'ersification Encoul-aging public educational institutions to diversify thei- sources of rev- enue and allowing them to keep such revenues can encourage autonomy'. The scope for this is greatest in higher education. The practice of attracting re- sources from alumni and private industry is standard among private schools and universities and is beginning to spread to public ones. The University of the West Indies, for instance, has in recent years obtained alumni funding to estab- lish a scholarship fund. In Chile, Indonesia. Thailand, and Venezuela private industry provides scholarships or subsidized loans for talented university stu- dents. Tax regimes can encourage such donations. In India 100 percent of individual and corporate contributionis to universities are tax-deductible; in Chile, 50 percent. Public schools and universities can also use their facilities to provide il- come. Universities in Uganlda and Senegal generate 4 to 5 percent of annual expenditure by renting out facilities (Zidennan and Albrecht 1995). Chinia, Mongolia. and Viet Nam encourage schools to rent out premises. run short courses, and provide services to industry. Such income amounts to 5 percent of the Monlgoliani education budget, 12 percent of the Chinese higher education budget and 14 percent of the Vietnamese budget (Wu 1993: Tsang 1993; Zidernan and Albrecht 1995). "Portable" Stiudent Fuinding Capitation grants, vouchers, and loans have the potential to encourage au- tonomy and competition, but experienice with these mechanisms is limited. All AUTONOMOUS INSTITUTIONS 133 operate under the same principle: the state makes available to the student a voucher or loan, which the student may then use to pay for his or her education at anv institution, public or private. These mechanisms thus finance the demand side of education. encouraging a marketplace in which suppliers must meet the demand. They establish a situation in which public subsidies increase the edu- cational purchasinig power of poor students and put them in the same situation relative to the providers of education as those who pay for their higher educa- tion from their own or family funds. Primary and secondary education in Chile is now financed publicly through the use of capitation grants; parents may enroll their children at any private or public school, and the school then receives funding from the government ac- cording to the number who enroll. Since the system was introduced in the early 1980s, it has led to a major increase in the number of private schools and to increased private school enrollment. In the Netherlands parents have the right to a free primary and secondary education for their children at any public or religious private school of their choosing, and the schools are funded by capita- tion grants on the basis of enrollmenits. Preschool education in New Zealand is funded by the state on the basis of a fixed sum per student, which can be received by any accredited institution or person. In the United States. Minne- sota allows public high school students in their final year to enroll in postsecondary institutions (with tax funds being paid to the institutions) and peniits enrollments outside the school district in which the student lives. A few experiments with voucher schemes have been tried in World Bank-financecl projects, as in Colombia and Pakistan (for girls), but evaluation is still at an early stage. In Hungary, Indonesia, Mongolia, Nigeria, and Viet Nam public higher education institutions are funded on the basis of enrollment figures, with allow- ances for different ulnit costs for different courses. Unless accompanied by limitations on admissions or on the number of students who can receive such funding, these schemes can, in theory, lead to open-ended budgetary commit- ments. They also fail to provide sufficient incentives for efficiency. While relatively small in that it covers only 6 percent of the student body, the Colom- bian student loan scheme is available to low-income students, and the loans are portable: they may be used to finance education at public and private institu- tions, not just in Colombia but also abroad. Output-Based and Qualih1-Based Funding Output-based schemes fund institutions by the number of graduates they pro- duce, rather than the number of students they enroll. This reduces wastage and improves overall efficiency. Such schemes are relatively rare and are limited to 134 SIX KEY REFORMS higher education in Australia, Denmark, Finland, Israel, and the Netherlands. No developing country yet uses such a plan, although Brazil is moving in this direction. In the Netherlands universities receive 4.5 years of annual unit cost funding per graduate. regardless of how long it takes students to complete their studies. After the scheme was introduced, the graduation rate improved from 48 percent in 1980 to 80 percent in 1987. Quality-based funding has been tried in only one country, Chile. Highier education institutions receive a financial award from the government for each entering student who scored among the top 27,500 in the university aptitude test. The objective of this scheme is to stimulate competition among institutions to improve their quality and thus to attract the best students, although it runs the risk of favoring those institutions that attract students from the highest socio- economic groups. In Hungary the World Bank is supporting reform of higher education through a "fund for new ilitiatives," access to which is competitive on the basis of institutions' proposals to strengthen the quality, efficiency, and relevance of thei- programs. Research has been funded in this way in World Bank-assisted projects in Brazil. China. Egypt. and Korea. Risks In higher education the benefits of autonomy are clear. At the school level. some caution is needed. Among the countries that have experimented most with increasing school autonomy in recent years are Chile. New Zealand. and the United Kingdom. In all three, there is still little evidence available on how the increased school-level flexibility that has resulted from autonomy has af- fected overall quality. The risks of school-level autonomy arise particularly with regard to in- equalities in educational opportunities and adherence to nationial standards and the curriculum. They can be largely mitigated by clearly separating school- level management and control over resource allocation from exclusive reliance on local financing and by ensuring that some functions are maintained external to the school. at the national or regional level. Among the particular-ly impor- tant tasks in this category are standaid setting: development of curricula and performance assessment mechanisms, such as public examinationls and learn- ing assessments; and use of national mechanisms to offset regional inequalities in resources if schools are fuLniced locally. Thle cuLTiculum and the financing of education have not been left to the school or even to the local level in anv of these three pioneering countries. P A R T III Implementing Change IN MOST COUNTRIES vested interests are roadblocks to financial anid manal;gerial change in ecducation. For both governments and the World Bank. a sectoral approach, rooted in country circumstanices. is essential. For countries. a sectoral approacih implies working for maximulm1 efficiency in the allocation and use of resources so as to improve the quality and increase the qLiantity of education. For the World Bank, a sectoral approach implies. il addition to those improvemlients in quality and quantity that Bank lendinig sup- ports directly, attentioni to the policy environment andl to institution bLlildiig so that Bank fiianicing and assistance help develop the sector as a whole. In all countries Bank lending will be concerned with the policy envi-on- ment and with the creation and strenigtheniing of institutions. The subsectoral allocation of lendinii will usually follow counlities' own resource allocation priorities. Primar-y and lower-secondary education will therefor-e contilue to receive the highest priority in the Bank's educationi lending to countries that have not yet achieved universal literacy and adequate access, equity. and qual- ity at these levels. In some cases it may be necessary to adjust the rate of increase of enrollments in order to ensure that the quality of schooling k-eeps pace with them. In the countries that have yet to achieve ulliversal literacy. the Bank's involvemenit in higher education will continue to be confined mailly' to makinig the financinig of higher education more equitable and cost-effective so that primilary and secondar-y education can receive increased resources. 135 136 IMPLEMENTING CHANGE As the basic education system develops in coverage and effectiveness. more attention can be devoted to the upper-secondary and higher levels. Bank lending for higher education will support countries' efforts to adopt policy reforms that will allow the subsector to operate more efficiently and at lower public cost. Countries prepared to adopt a higher education policy framework that stresses a differentiated institutional structure and a diversified resource base, with greater emphasis on private providers and private funding. will continue to receive priority. The transition economies of Eastern and Central Europe form a special category. Their primary and secondary enrollment ratios are high, but they need to adjust the entire education system toward the needs of a market economy. Particularly important are efforts to maintain funding levels for compulsory (primary and secondary) education: to shift away from overspecialization in vocational, technical, and higher education institutions; and to reform the gov- ernance and financing of higher education. In many ways, reform of higher education is the key starting point for much of the broader sectoral reform that is involved for these economies. C H A P T E R E L E V E N The Political and Social Context of Change REFORMS of education, whatever their technical merit, will not take hold unless they are politically and socially acceptable and unless the pace of reform is appropriate. Education is intensely political because it affects the majority of citizens, involves all levels of government, is almost always the single largest component of public spending, and carries public subsidies that are biased in favor of the elite. Prevailing systems of education expenditure and management often protect the interests of teachers' unions, university students, the elite, and the central government at the expense of parents. communities, and the poor. The pace of reform must therefore take account of those vested interests and also of the need for adequate resources to sustain the reform. (The appendix to this chapter discusses reform priorities in the special case of the transition economies.) Teachers are usually the single largest group of civilian public employees in developing countries. Because educational finance and management are typically the responsibility of central government. teachers' unions are impor- tant actors on the national political stage. In Latin America, Eastern Europe, and some Asian countries, for instance, they have established their own politi- cal parties or have formed alliances with parties representing trade union move- ments. When governments fail to reach agreement with strong central unions over conditions of employment for teachers, collective action can disrupt edu- cation and sometimes lead to political paralysis. as has happened in Bolivia, Peru, and other countries in recent years. 137 138 IMPLEMENTING CHANGE The relationship between higher education students and the govemnment can be oppositional, as well. The contflict arises because of the centralized nature of university financing and governance and because higher education students, who come disproportionately from tipper socioeconomilic households. are a vocal and articulate political constituency. When students have griev- ances. usually only national governimients can address them. In Romnania in 1 993, for instance, university students besieged the national Ministry of Educa- tion and the parliament to protest overcrowdinig in student hostels. In Africa heads of state are usually univer sity chancellors. institutionalizing the potential for political opposition when studcents have grievances. In Kenya and Uganda students have several times brought before ministers and the head of state complaints about the food served in cafeterias and the introductioll of tuition charges. Such action has often precipitated political crisis and Liniversity closures. Chaniging the centralized pattern of university fiianlce and admilistratioll can ease the oppositional relationship between students and the government. Whereas teachers and higher education students are politically influential as a result of centralized patterns of education finance and management, par- ents and commlillunities are relatively weak. In most developing countries since the end of the colonial period, national govenmenits have assumed responsibil- ity for education systems, at least at the primary level. They have taken up this role with the best of intentions: to overcome the inequities that characterized the preindependence pattern of private. voluntary. and local education: to widen social access to education: and to build "nationalism" through a single govern- ment-controlled curriculum. Emphasis on these goals has come at the cost of shifting responsibility away from communities and parents. Where there is a tradition of local responsibility, the education system can be very responlsive to parents andI communities. In the Indian state of Kerala, for instance, almost all schools have active parenit-teacher associations that involve parents in school manlagement, noon feeding programs. fund raising, and facilities development. The proelite bias of public spending on education, particularly tertiary education, makes reform difficult. The relatively affluent are naturally loath to give up their privileges, as has been shown in many countries when govern- ments have introduced or increased fees at public universities and other higher education institutions. Successful reform of education finance and management requires a signifi- cant expansioni of educational opportunities, extensive consultation with cur- rent and potential stakeholders, mechanisms to increase the say of parents and communities in the system, and thorough design of the reforms that includes public finance. Educationi financing and management changes are best intro- duced as educational opportunities are expanded (box 11. I ). Sometimes the change itself does this, as when prohibitions on the private sector are lifted. The expansion of cost-sharing in public higher education, for instance, is politically THE POLITICAL AND SOCIAL CONTEXT OF CHANGE 139 BOX 11.1 TRADEOFFS BETWEEN ACCESS lack of trained teachers and facilities, AND QUALITY: LESSONS FROM KENYA and the need to prepare new curricula AND THAILAND and textbooks. Nevertheless, the gov- ernment hastily implemented the re- Have expanding enrollments caused form, although few schools outside the quality of schooling to suffer? The Kenya's principal cities and towns evidence cited to support that asser- could marshal the teaching and other tion includes the poor academic per- instructional resources necessary for formance of low- and middle-income the new program of studies when it country students, the high incidence was introduced. More than 18,000 un- of wastage and repetition, increasing trained teachers were eventually em- reliance on untrained teachers, use of ployed to enable schools to extend in- double shifts, and lower investments struction by an additional year, setting in critical learning inputs such as text- back the significant progress that had books. In many countries, enrollments been made since 1963 in providing have expanded more rapidly than primary schools with trained teachers the financial and other resources that (Eisemon 1988). are needed to support greater school Educational access and attainment coverage while maintaining school can be increased, with attention to quality. school quality, when plans for expan- This is particularly so in Sub-Sa- sion take account of implementation haran African countries, which tradi- issues. In Thailand a similar reform, tionally had low gross enrollment ra- announced in 1988, would have tios and which, since becoming inde- lengthened compulsory schooling to pendent, have given high priority to nine years and integrated lower-sec- increasing access to schooling. In ondary and primary schooling. After some of these countries primary-level more careful consideration of what was enrollments have grown by 5 to 10 required to successfully implement percent a year; in Kenya, for instance, such a reform, the government decided primary school enrollments increased to introduce the nine-year schools dramaticallyaftertuitionandotherfees gradually. In the pilot phase 718 of were gradually abolished beginning in these schools were established in 38 1974. In 1984-85 the duration of pri- economically disadvantaged prov- mary schooling was lengthened, and inces. In 1990 the experiment was ex- enrollment expanded by about 583,000 tended to 122 schools in 73 provinces, in one year. and it is expected that by 1996 the Each such measure to increase par- number of schools will have increased ticipation and raise educational attain- to 4,187, with an enrollment of 750,000. ment has had deleterious implications The schools are experimenting with the for school financing and instruction. new curricula and textbooks and with Ministerial task forces established to new models of school management, study the implications of the new eight- including community involvement in year primary cycle urged caution and designing the school program gradualism, citing inadequate finance, (Holsinger 1994). 140 IMPLEMENTING CHANGE most feasible when it is explicitly linked to the expansion of opportunities for higher education. In Chile and Hungary the reform of higher education was successful because overall enrollments increased. In Chile fees were intro- duced, and enrollments were increased through an expansioni and differelitia- tion of the higher education system and the introduction of student choice. A World Bank loan to Hungary is supporting a sweeping reform of the higher education svstem in which public institutions will start to raise fees and oppor- tunities will expand. The reform involves the encouragement of private institu- tions and the direct funding of students at private and public institutions through a state student-grant fund and through loans. In countries as diverse as Bolivia, the Dominican Republic, Ghana. Guinea. India. Jordan, Mauritius, Mozambique, Romania, and Thailand, education re- form has made a good start because stakeholders have been involved in devel- oping and implementing the reforms. In Bolivia and the Dominican Republic the UNDP has financed consultation mechanisms to encourage the development of national consensus around education reform. In both cases a reform policy document and an accompanying public investment program were endorsed by the teachers' union, parents' representatives, and the major political parties and were successfully presented at meetings of donors. including the World Bank. In Bolivia the refonm program prepared by the previous government has re- mained intact since the opposition came to power in the 1993 election. In Ghana a process of national consultation extended from the head of state down to every community through a series of "town meetings." In Mauritius a far- reachinig edLIcation master plan is being implemented, following an extensive process of public consultationi (box I 1.2). Less extensive processes have also provecd useful in Jordan and Thailand, where reforn committees included rep- resentatives of teachers' unions, education ministries, and school and univer- sitv administrators, although parent and student involvement was more limited (Haddad 1994). Effective collaboration between government and teachers' unions to achieve refonn has been demonstrated in several countries, including Ghana, Korea, Singapore, and Zimbabwe. Reform is also successful when involvement by communities, parents, and students is increased. Community and parent control, when accompanied by measures to ensure equity in the provision of resources, can offset much of the power of vested interests, such as teachers' unions and the elite. Parent and community involvement can be increased by decentralization, school level autonomy, and accountability of the school to local people. At least in urban areas, it can be enhanced by the use of market mechanisms that increase ac- countability and choice. In both urban and rural areas it can also be enhanced by the use of participatory methods in designing and implementing refonms (Colleta and Perkins 1995). THE POLITICAL AND SOCIAL CONTEXT OF CHANGE 141 BOX 11.2 STAKEHOLDER PARTNERSHIP ment, reduce wastage, and reduce IN MAURITIUS variations in quality by lengthening the cycle of compulsory education to nine Mauritius is a multiracial, multilingual years, revising curricula and assess- society whose education system had ment practices, and strengthening not changed significantly since the school-level management. It will affect colonial period. In 1990 its minister of the powers and responsibilities of education began consultations on re- teachers and private schools, as well form of the system. A steering com- as many interests within the education mittee, a working group, and sectoral bureaucracy, including the examina- subcommittees were created to collect tions syndicate, teacher training insti- opinions on the subject and hold pub- tutions, and various technical depart- lic hearings to study options. Techni- ments of the Ministry of Education. cal studies were commissioned, and a "The intense internal struggle and strategy paper was circulated for pub- strong opposition from some groups lic debate. Special efforts were made has not been reconciled," a recent to solicit the views of teachers on par- evaluation of the reform process ticularly contentious issues, such as acknowledges. "The important shift their role in curriculum development introduced by this reform process, how- and continuous assessment, in-service ever, was to move educational policy- training, and performance evaluation. making from the close province of the A televised national seminar was held professional educators (and govern- on the reform plans to ensure country- ment) to a more open forum involving wide participation and consultation. parents, vested interests, lobbyists, The reform plan that emerged was unions, and the community at large." It adopted by the government with sup- is still too early to assess the impact of port from the World Bank and other the reform itself, now being put into donors. The plan will increase attain- effect (Bhowon and Chinapah 1993). It is important that reform efforts clearly define the new roles and responsi- bilities of the actors in the education system. Critically important here is the complementarity of policy reform and public finance. Half measures-too rapid policy refonrs without financial mechanisms-do not work, as illus- trated by decentralization in Ghana and the regulation of private institutions in Korea. In the colonial period Ghana had strong local governments with inde- pendent tax bases that provided high-quality primary education. Following years of centralized control after independence, recent reforms have returned the responsibility for financing basic education to local governments but with- out giving them the authority to raise revenue (Associates in Rural Develop- ment 1993). In Korea during the 1970s the goveniment tried to influence the 142 IMPLEMENTING CHANGE level and distribution of enrollments in private educational institutions through the accreditation process. It attempted to Imnit total private enrollments and to shift their balance toward science and engineering. Since public finance was not available to the private institutions, the govenmment had no effective instru- ment for enforcing its policy, and the regulations were ultimately abandoned. Appendix. Priorities for Educationial Reform in Easter-n and Central Eulrope In Eastern and Central Europe. strategies for educational refolrm must antici- pate the establishment of competitive market economies and pluralistic, legiti- mate. and stable political systems. These conditions hardly exist in many coun- tries in the region. Large state-ownecd enterpr-ises, for example, continue to provide most employment, and frequent changes in governments have inhib- ited economic and educational reform. Radical refomis-including new curTicula and innovative methods of in- struction-are required in the management. financing, and structure of educa- tion to restore these countries to political and economic health. Comprehensive educational reforms will accelerate recovery, provide a long-term foundation for growth. and support development of democratic political and social institu- tions. The benefits of an open economy and a participatory political system cannot be captured by individuals without market ancd citizenship skills. Na- tional income would be very much higher if education were restructured now. without delay. High priority should be placed on refonn of compulsory education. The challenge is enormous. Real per pupil expenditures on compulsory schooling have been declining precipitously in most countries in the region. while enroll- ment has been stable or increasing. For example, in Russia per student expendi- tures on compulsory schooling decreased 29 percent between 1991 and 1992. although enrollmlienit sliglitly increased. Investnent in physical facilities dropped by 23 percent during this period and supplies of textbooks by 16 percent. Teacher salaries fell to about two-thi-ds of average industr-ial wages, prompting widespread teacher strikes that in 1992 accounted for the majority of all days lost to strikes. In Russia and many other countries in the region, reforms are needed to protect compulsory education fromii fiscal instability and ensure equity at least in nonsalary expenditures as more responsibility for financing is devolved onto local government authorities. To promote democratization, local authorities and schools should have greater control of their budgets-for establishing teacher salaries and conditions of employment, for allocating expenditure be- THE POLITICAL AND SOCIAL CONTEXT OF CHANGE 143 tween capital and recurrent costs, and for obtaining textbooks and other learn- ing resources. To accomplish this. the functions of national and local educational authori- ties will need to be re-examined. At the national level, transfer payment schemes will have to be designed to take into account the different resource-generating capacities of local governmenits and to promote fiscal effort and local initiative. Mechanisms need to be developed at the national level to manage compLulsory education effectively. allowing variation in implementationi at the local level. The critical functions of national govenmments include setting curricular objec- tives for core subjects, formulatinig minimal standards for instructional facili- ties and the distribution of texts and learning resources, elaborating a regulatory framework to facilitate the development of private education, monlitorilng stu- dent performance. and protecting the rights of linguistic and ethnic minorities. Reforming the overspecialized programs of vocational, technical. and highel education institutions will require even bolder initiatives. At the begilining of the transition, a very high proportion of stucdents in secondary and higher education was enrolled in programs designed to supply to state-owned enter- prises and public services graduates with specialized skills that the nascent labor market cannot absorb. In Poland in 1990-91. 76 percent of secondary school students were enrolled in vocational and technical programs. In Roma- nia in 1989-90. about two-thirds of all higher educationi students were enrolled in narrowly defined engineering programs. Vocational and technical schools. which accounted for more than 80 percent of upper secondary enrollment. offered 354 cun-icular specializations. In 1991, 50 percent of students enrolled in secondary tecihinical schools and 25 percent of those in vocational programs were training in machine trades and metal work. although job vacancies in these specialties represented only 5 percent of vacancy listings. The number of students in specialized enginieerinig, vocational, and techni- cal programs has ciropped sharply during the transition, and the distribution of enrollment has shifted in response to student demand. For example. the propor- tion of students in engineering in Romanian universities declined from 65 to 38 percent between 1989-90 and 1992-93. Romania's higher eclucation system has experienced rapid growth in recent years. The majority of students are nlow enrolled in foreign language programs, law, economics, managemenit, and other social sciences. But where little progress has been made in reforming the struc- ture of secondary and higher education programs. total enrollments have usu- ally declined, as has happenecd in Russia. Governments in Eastern and Central Europe should give priority to expand- ing enrollment in general secondary education: to increasing the amount of academic instruction students receive, especially in foreign languages: and to 144 IMPLEMENTING CHANGE introducing new subjects such as computer training that a competitive market economy requires. Above all, refonns should increase student choice, through promoting curricular flexibility in general programs and in vocational and technical programs, so that the system will react more rapidly to changes in employment opportunities. At the higher level, government policies should encourage private provi- sion and increased private financing of public higher education to stimulate competition. innovation, and responsiveness to the labor market. The political freedoms public universities obtained after the collapse of socialism should be accompanied by more autonomy in using public funds and in mobilizing addi- tional resources. At the same time, governments should establish open. trans- parent mechanisms for allocating public funding, with incentives for improv- ing efficiency and restructuring academic programs. New policy structures will also have to be created to ensure quality control of both public and private institutions and to guide the development of the higher education system as a whole (Eisemon and others 1995; Heyneman 1994: Laporte and Schweitzer 1994; Sadlak 1993; Spagat 1994; World Bank 1992. 1994k, 19941). C H A P T E R T W E L V E The World Bank and Education T HE WORLD BANK is today the largest single source of external finanice for education in developing countries, accoulitillg for about a quar ter of all external support (table 12. 1). Since its first educationi pro ject in 1963. tile Bank has continuously expandedl its financing of education projects, in both absolute and relative terimis. as part of its mission to reduce poverty. Total lendinlog for educationi over the past thirty year-s through fiscal 1994 amoulits to $19.2 billioni, throuoh more than 500 projects in more thani 100 countries. Lendinig commitments are currently around $2 billion each vear. Developments since 1980 Six matjor shifts characterize the decade and a half since the last World Bank e(lucation sector policy paper wvas issued. in 1980. The total volume of educa- tion lending has tripled, and its share in overall Bank lending has doubled. Primiiary and secondlary educationi are increasingly important and in fiscal 1993 and 1994 represenited half of all education lending. Lending. once concenitrated in Africa, East Asia. and the Midcdle East. is nlow significant in all regiolns. Girls' educatioll is receiving- explicit emphasis. Bank fuLids are today used less focr buildings and more tfor other- educational inputs. A narrow project approach is increasingly givino way to a broad sectoral one. 145 146 IMPLEMENTING CHANGE TABLE 12.1 EXTERNAL SUPPORT FOR EDUCATION, 1975-90 Iten,i 1975 1980 /985 1986 1987 19,8 1989 199( Amiount (nilflions ,ofdl4lars) Total 2.018 4,496 4.255 4.644 4.584 5.528 5.838 6.035 Bilateral 1.490 3,595 2.679 3.169 3,512 3.950 3.790 3,640 Multilateral 528 901 1,576 1.475 1.072 1.578 2.048 2.395 World Bank '224 440 928 829 440 864 964 1,487 Worldi Batik sloici-e (percent) Of total support I 1() 22 18 1( 16 17 25 Of multilateral sLuppOIt 42 49 59 56 41 55 47 62 .Sourlc iLNESCO 199ib. Lending Volume Lending for education has iicr-eased significantly since 1980, both in absolute terms and as a share of total World Bank lending (figure 12.1). In the early 1980s lending commitments for education averaged about $0.6 billiont a year and represented 4 percent of total Bank lending. They have now tripled in volume, to about $2.0 billion a year, with annual fluctuations, and the share has doubled to more thanl 8 percent. These figures exclude project-related training included in Bank projects in other sectors. The Bank is strongly committed to continLied support for education. A particularly interestilng development in fiscal 1994 was the first education loan by the Intemnational Finance Corpora- tion (IFC), to the private Rainbow Academy secondary school in Uganda. Pr-ior-ities within? the Ed ucationi Sector Lendinig for primary education has grown rapidly since 1980. and particularly since the late 1980s (figure 12.2), reflecting the growing realization of the itupor-tance of this level of education for economic growth and the reduction of poverty, the influence of the 1990 policy paper on primary education, and the Bank's commitimient to the objectives of the 1990 Education for All Confer- ence. The Bank was one of the sponsors of the conference and continuLles to participate in Education for All follow-up activities. During fiscal 1990-94 a third of all Bank lending for education was for primary educatioll, more thanl double the share a decade before, with an emphasis on both access and quality. Future lending plans indicate a likely continuation of this trenld and an increase THE WORLD BANK AND EDUCATION 147 FIGURE 12.1 WORLD BANK EDUCATION LENDING, FISCAL 1980-94 Share of World Bank len(ling (percent) Billions of dollars 10 8 6 4 2 0 0 l).5 Hl( 1.5 2.0 2.5 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1992 1993 1994 So,,,, Il ,u,, i,wl Rc1i a. .irious year. in the share of secondary education. The emphasis on primary and secondary education has led to a slightly reduced share for higher education in overall lendinig. In addition, the Bankli as in the 1990s started to lend for early chilcl- hood development. inclucinlg early childhood education. although these projects are still too new to be evaluated. Despite the major shifts in subsectoral empha- sis, the overall increase in lenidinig volume has led to an absolute increase in lending for all subsectors. except vocational education. which has cleclined in both relative antd absolute ternls. Primary education projects have focused on access. equity. intemnal effi- ciency, and quality. Access is being improved through projects that target poor regions. as in Chinla and Mexico (box 12. 1), and girls and minority grouLps, as in 148 IMPLEMENTING CHANGE FIGURE 12.2 WORLD BANK EDUCATION LENDING BY SUBSECTOR, FISCAL 1964-94 Per cent 60 50- 40- 30 -~ .(01 1964-69 1970-74 1975-79 1980-84 1985-89 1990-94 * Primiiary education U General secondary U Vocational secondary eCLucath on edlucatioln D Vocational post- C Teacher training, t Higher educationi secondarv education S,Swt(c: WVm (id Bank dalzr. Bangladeshl India, Mexico, and Pakistan. Internial efficiency gains are the ob- jective of projects that support increased student-teacher ratios (Barbados), militigrade schools (Colombia). double-shift teaching (Triniidad and Tobago), decentralizationl (Brazil). and communllity participation (Ghanla). Qualitv im- provements are supported by the development of national learning assessmlient systems. which are now included in 27 percent of primar-y school projects. compared with only 3 percent twenty years ago, and by an increased emphasis on iilportant iptLIts other thani school buildinigs. such as textbooks, laboratory equipmenit, and teacher traininig. Lending for secotidarv edLication decliined between 1980 and 1990. wheni it represented 10 percenit of education lending. Since 1990. however. it has in- creased modestly, to 12 percent of all education lending. with 30 percent of education projects now containing a secondary education coiuponient (Demsky 1994). Seconclary education lending is still low compar-ed with the 1960s and THE WORLD BANK AND EDUCATION 149 BOX 12.1 MASSIVE EDUCATION REFORMS supervisors in providing pedagogical IN MEXICO'S POOR SOUTHERN STATES assistance to teachers in classroom teaching. Second, an educational ma- Mexico has made strides in national terial resources component will pro- education over the past few decades, vide educational materials for both and the country recently adopted ag- teachers and students; supply class- gressive reform policies to boost room library and reference books and school achievement. Yet the poor promotetheiruse;rehabilitate,replace, states of the South, which lack the re- or build urgently needed educational sources and administrative capability facilities to replace inadequate schools to exploit national advances, remain and meet new enrollment demand; and seriously behind the national average. design (in consultation with represen- Now the Mexican government is tativesofeachindigenouscommunity) pushing its reform agenda farther and deliver bilingual textbooks and south. With help from a $412 million material for indigenous schools. Third, World Bank loan-one of the largest an institutional strengthening compo- Bank social sector loans ever-and nent will strengthen education man- with more than $200 million of its own, agement capacity at both central and Mexico is launching an ambitious pri- state levels by providing management mary education project in ten of the training for planning and policy analy- poorest southern states. Because of sis, installing project monitoring, evalu- the project's targeted nature within the ation, and information systems, con- government's social compensatory ducting education studies to prepare program, project benefits will include future investments at the preschool and a more equal distribution of economic secondary education levels, and sup- opportunities among Mexican children, porting project promotion and diffusion especially for underserved indigenous activities. This component will also populations. improve the supervision system and This Second Primary Education strengthen the textbook and materials Project consists of three major com- distribution system. Finally, the project ponents. First, a human resources will provide incentives for teachers to development component will provide work in remote indigenous or hardship in-service training and assistance to areas; absenteeism will be monitored upgrade the skills of primary school directly by the local community and teachers, principals. and supervisors, school council (World Bank 1994i). emphasizing the role of principals and 1 970s, when it accouLntedi for more than half of all Bank education lending, but it is Lrowing fast. Lendin2g in recenit vears has reflectecl selective SUppOrt for seconidary ecducation. particulally to improve equity for the poor and girls. to improve quality, and to hicrease external efficiency. A secondary educatioll project in Colombia. for instance. iicludes a voucher program to enable poor 150 IMPLEMENTING CHANGE studenits to attend private schools. The first free-standing female education project is for secondary education in Bangladesh. Bank support for postsecondary education shows a mixed pattern. Lending for postsecondary vocational education and training has declined, consistenit with Bank policy that this type of education is generally best provided on the job by employers (World Bank 199 1c). Lending for teacher traininig remained roughly constant, at about 9 percent of total sector lending until fiscal 1990-94, when it fell to 7 percent and shifted away from preservice toward in-service training. During fiscal 1970-74. for instance. 49 percent of Bank projects contained preservice training components and 35 percent in-service compo- nents. For fiscal 1990-94 the comparable figures were 39 and 65 percent. Teacher traininlg projects now support the development of traininig curricula, teacher training, and the development of instructional materials. whereas they were formerly mainly Iimited to the provision of teacher training facilities. Lendinig for university and polytechnic-level higher education peaked at 36 percenit of total lending in the mid- 1 980s: it has since fallen to about 26 percent. Originally directed mainily toward institutions that train professionals and tech- nicians for the economy (such as agricultural universities). highier education projects increasingly support universities and institutions responsible for ad- vanced scientific training and research. Several recent projects have also sup- ported the improvement of Iinks between industrial development and teaching and research in science and technology (Korea and China), the expansion of cost-sharing at public universities and the development of studenit loan and scholarship schemes (Kenya, Philippines, Tunisia. and Venezuela). and limita- tion of access to public universities (C6te d'lvoire and Morocco). The bulk of postsecondary education projects in all categories have been in Africa and East Asia and the fewest in Latin America. Postsecondary education is expected to be an important componienit of future Bank education lending for the transition economies of Europe and Central Asia. Regional Composition Bank lending in the early 1980s was heavily concentrated in Africa. East Asia. anid the Middle East. These regions' relative shares have declined as lending has expanded for primary education in South Asia and Latin America (figure 12.3). Absolute lending has iicr-eased in all regions, however. Lending is still relatively small in the Europe and Cential Asia region. In the early years of Bank membership of the transition economies in this region. huliman resource projects tended to focus on social safety nets, labor markets, and health system rehabilitation. THE WORLD BANK AND EDUCATION 151 FIGURE 12.3 WORLD BANK EDUCATION LENDING BY REGION, FISCAL 1964-94 Percent 50 40 30 20 10- SUb-Saliarant East Europe Latin Middle SoLith Africa Asia anid Aitterica East and Asia and tlhe Cerntral and the North Pacitic Asia Caribbean Africa _ 1964f-69 1970574 l95-79 = I84-84 _ 19X5-89 [= 199(l-(4 .5ow, e Worldc Bank data. Gil/s' Editcation F:emale education is receiving increased attention il Bank projects. Less thanl 15 percent of projects in the 1 980s contained components to educate girls. but since 1990 this share has increased to 22 percent. and the trend is expected to coitillue. The regional distribution of female education components reflects regional differences in gender disparities. Lending has been concentrated on the Middle East (44 percent of all education projects address female education), South Asia (39 percent), and Africa (16 percent). Gender gaps in education are much less significant in East Asia. Europe and Central Asia, and Latin America, and hence there is little specific lending for girls in these regions. Use of Bank Finanming In the early days ol Bank lending for edUcation and thr-oughout the 1 970s, more than half the funds lent by the Bank were used for civil works-to construct schools. colleges. and adimiinistration buildings. Since the imnid-1980s this pro- 152 IMPLEMENTING CHANGE FIGURE 12.4 WORLD BANK EDUCATION LENDING BY EXPENDITURE CATEGORY, FISCAL 1964-94 Percent 8( 71() \-Civil vork's 5(0 - Fumiture, vehicles. andlc equLipnment 40 - Textbooks 8nd(1 20 - eduCittiOrnal nateria Is (l ..., ... * * * * | * - - -...... --Oher 1964-69 1970-74 1975-79 198(0-84 1985-89 199(9-94 w-,ue World B.1i3k :ini. portioni has ftallen to about a quarter (figure 12.4). Batik loans increasineylv finanice inputs dlesigned to itprove the quality and the administratioin of educa- tion: textbooks, teacher trainlinig, laboratory equipImenit. learninLg assessmenits. examinatioti systems. educationial administration, technical assistancle, and re- search. Despite this shift. Bank projects were always conce-ilred with qutality. even when imlost finanice went to buildings. A few loans also finalice the de- ttiald for educationi throitgh voLIchers for the poor. as in Colombia: scholar- ships for girls. as in Bangladesh: and student loans. as in Venezuela. Overall. however, almost all Bank lending is for the supply of educatioll services. as opposed to the demand. Sectoral Approach The use of Batik fitiancing for itiputs that impr-ove quality' as well as iticrease enr-ollilieilts -eflects a wider tren( toward a sectoral approaci to educatioll. This ttend is seeti even toore clearly in the inteLration of education withiin econotilic THE WORLD BANK AND EDUCATION 153 reform programs. in the emergence of education sectoral adjustment an(l sectoral investmenit loans. in the increasiliw incILision of SUppOrt for sectorwidle refol-ms in specific investment loais, and in the Banik's continiued attention to inStitution buildinge. Economnic reforiml programs supported by the World Bank no0w routinely incIude measures to ensure that primary education is protected during fiscal adjustment. Recent structural adjustment loan.s to Guatemala, India, and Zim- babwe. for instance, support increases in public spending on primary education and other basic social services. There have been six education sector adjust- menit loans, the first in Morocco in fiscal 1986 ancl the others in Sub-Saharaln Africa. These loans have tackled such1 sectoral issues as the allocation of public spendling. In Guinea, for instanice. a sector adjustment loan supports the rede- ployment of primary sch0ool teachers to uliderserved rural areas: in Kenya a loan supports the introduction of fees at public universities and the reversal of a trenid towarcl ever-lower student-teacher- ratios at prinary and secondal-y schools. Thii-ty-ninie sector- investment loans have been made since 1979. mainly in Asia and Latin America, where institutions are relatively strong. These loalns tocus on institutional and policy reforms and use Banik finanice to fund a share of an overall education public investment programi. In the Philippines a sector loan suppo rts a policv shift from the expansioni of primary education toward a tocus on its quality, efficiency, and equity. Procedures for efficient and equi- table regionial budget allocationis in Colonibia are supported by a sector invest- ment loan. Specific education investmenit projects also include attention to sectorwide policies. The Fourth Education Project in Burkinia Faso includes agreements to redluce spendinig on secondar-y and higher educatioll subsidies by I() percent annually. Ftittire Bank Support tor Educatioti The volume of external assistance to education is impressive. but it is minol in comparison with spendilng on educationi by developing countries-by their governments and their people. A conservative estimate of public spending at 4 percenlt ofGNP and private spendinig at 2 perceit yields a total annual education expenditure of $270) billion for- all low- and middle-inicomile countries. All exterinl.l fhianice togcthel- amounts to only 2.2 percenit of this toial. and World Bank financing to only 0.6 perceit. This low share of total spending meanis that the Bank shouild concentrate on providing advice designed to bI-p goveri- menits develolp educationl policies suitable foir the cilcuinstances of their ownl coUlltri es. Futur-e operaztionis will theret'ore adlopt an even more explicit sectorwide policy focus to support chlaLIges in educational finance and management. This 154 IMPLEMENTING CHANGE strategy may increase both the resources and the tine needed to prepare projects. as key stakeholders must be involved in preparation. In increasingly decentral- ized contexts, these stakeholders will include not only central governments but also other levels of government, as well as commuLnities. parents. teachers. and employers. The design of new operations will focus on the conditions needed to ensure successful implementation, which is the principal indicator of the Bank's development effectiveness. Effective implementation is especially important given the explosion in lending for education and the increased awareness of the importance of human resources. Banik lending is already moving in the direction of the following six key areas of reforn, described in part 11 of this report. * The Bank will continue to encourage its low- and middle-income coulitry clients to give a higher priority to education and education reform (chapter 5) as an important complement to economic reform programs. * Education projects are taking more account of outcomes (chapter 6) and their relationship to educational inputs and processes at the institutional level. Accordingly. (a) more explicit use is being made of participatory methods in sector work and project design to ensure that all relevant clients are involved and that there is agreement on desired project outcomes and institutional-level coonditionis to produce the outcomes: (b) learning assessments are being em- ployed to measure projects' impact on learning and on institutionial environ- merits; (c) niore attention is being given to the collection of policy-related data: (d) monitorinig and evaluationi during and after project implementation are being iniproved; and (e) effectiveness indicators are being used more system- atically. In addition, cost-benefit and cost-effectiveness analysis will be more systematically used both in sector work and in the identification. design. and appraisal of education projects. * The trend toward allocation of an increased share of lending to primiiary and lower-secondary education is expected to contilue (chapter 7), with particular emphasis on the poorest countries that receive IDA funds and especially on Africa and South Asia. This emphasis will fit within the context of a sectoral approach that recognizes the importance of the various parts of the education system, the interdependencies aniong these parts. anid the need to ensure that the focus, as well as the nature, of Bank assistance is based on a determinationi of where the Bank can be most useful in the particular ci-cumstanices of each country. Within basic education, quality is being more systeniatically encour- aged, along with access and equity. In all regions increased private finalicilng and private provision of higher education are being encouraged in the conitext of increased attention to fiscal sustaiiability. THE WORLD BANK AND EDUCATION 155 * Equity is also receiving more systematic attention (chapter 8), to ensure a direct impact on poverty reduction. Increasing attention to girls' education is now accompanied by more attention to children from poor families and to disadvantaged groups, including ethnic minorities. Targeted early childhood development programs are becoming more important as a means of improving equity and later school performance. * Projects support more household involvemnent in education (chapter 9), in- cluding participation in school management and (so far. only rarely) experi- mentation with school choice. Demand-side interventions, such as targeted scholarships for the poor, stipends for girls. and higher education loan schemes. are becoming more common. Since choice can run both equity and quality risks. experimentation requires more attention than in the past to the regulatory framework for education, including, in particular, quality-enhancing mecha- nisms, such as standard-setting, outcome monitoring, school inspection, and. at the highier level, accreditation. * To enhance quality, projects are encouraging experimentation with the flex- ible and autonomous management of instructional resources (chapter 10) at the institutional level, complemented by attention to incentives for performance, such as examination systems. and to quality-enhancing mechanisms. These six key areas of education system reform will be supported by future World Bank projects in the context of the guiding principles that the Bank adopted in 1994 to guide its operations: selectivity, partnership, client orienta- tion. resLIlts orientationi, cost effectiveness. and financial integrity (World Bank 1994h). For example: Partniers-hilp will mean cooperation with other donors and agencies-par- ticularly important as both multilateral and bilateral agencies increasingly fo- cus their aid on the humani resource sectors. Already the Latin America and the Caribbean region of the Bank is working in close partlership with the Inter- American Development Bank on social sector projects. as are the two Asia regions with the Asian Development Bank. The Human Development Depart- mnent is working with UNESCO to improve the quality of international statistics on education. A possible difficulty in donor cooperation is the increasing adop- tion of a sectoral approach by many donors, not just the Bank. This makes cooperation on the policy framework imperative, placing new demands on many donors' capacity to provide analytical policy advice on the education system as a whole. Client orienztationi is reflected in this report's focus on stakeholder involve- iment in educational reform. Several regions, particularly Africa. Latin America and the Caribbean. and South Asia. are enlarging their education presence in 156 IMPLEMENTING CHANGE field offices to achieve such involvement. The Europe and Cential Asia region has established a human resources ' hub" in Hungary. In all iegions there is increasing emphasis on the participation of potential beneficiaries in project design and implemelitation. Results orienitation in the educationi sector will meani even more attention to project outcomes. This will involve more intense efforts to help COUlntr-ies im- prove their education data: inclusioni of more leariniig assessmenit components; more links betweeni sector workl and lending: more Lise of benefit-cost analysis in project identification. design. and appraisal: increasedl attention to monitor- ing and evaluation durinig amid after project implemiielitation: anld renewecd ef- forts to strengthien borrowers' educational adminiistiations. References Ahadzi, Helen. 1994. Wlhat Wc' Koabo)uiI)oIt Adlult Lit,r-tacy:IN Tlhc're Ho,lK:' World Banlk Discussioll Paper 245. Washinigtoni. D.C. Adams. Don, and Esther- F. Gottlieb. 1993. Education a/ndl Social C'hc,,u,e in Korea. New York: Garland. Ahimled, Manzoor. Colette Chabbott. Aruni Joshi. and Rohini Pan1de. 1993. Prinarv Edcwationfi f All. Lea,-ningfi Jr-omn tl BRAC Experience. A Case Studv. Project ABEL (Acivanicinig Basic Education and Literacy). Washington. D.C.: Academy for Educational Developmenlt. Ainsworthi. Martha. Mead Over. and A. A. Rwegarulira. 1992. 'Economnic Impact of AIDS on Oirphanecd Childreni: What Does the Evidenice Show'?" Prepared for the Expert Meeting on Family and Developmiienlt, Washington. D.C.. July 16-17. Na- tionazl Academy of Sciences, Washingtoni, D.C. Albreciht. Douglas, and Adrian Ziderman. 1991 . Del u7rred Cost Recov crv fwr Hig-her Edhcation:- Stludent Loun Progratns in Dec'el1,pinng Counutrics. World Bank Discus- sion Paper 137. Washingtoni, D.C. Arrow, Keninethi J. 1993. 'Excellence and Equity in Higher Education.' Education Economcic, 1(11:5-12. Associates in RuL-al Development. 1993. "Predicanients of Decenitralization."For-no,2r- Advanc ing, Basic Education and LiteracN ' 3 1):12-13 . Azariadis. Costas. and Allen Drazeni. 1991). "Thieshold Externalities in Economilic De- velopienit. Quarterll.1ournal of Ec onotmic.s 10(2)2 ':5(01-26. Barinett, W. S. 1992. "Benefits of Compenisatory Preschool Educationl.' Journal of Huimianzoz Resources 27(2 ): 279-3 2. BaLn. Nicholas. 1993. The Economics oft tlhe Uelfcn- State. Stanford, Calif.: Stanford Univer-sity Press. BaiTo. R. J. 1991. "Economiiic Growth in a Cross-Sectioni of CoLiuntries." QuarterlY .Jorntal of Econmicis I (6(2):4(7-44. Bartel, Anin P.. and Frank R. Liclitenherg. 1987. "'The Comparative Advantage of Educated Woriker-s in Implementing New Technology." Revicin of Economicsv aInd Stti.sticS 69(lJ: I -11. 157 158 REFERENCES Bashir, Sajitha. 1994. 'Public versus Private in Primary Educationl: Comparisons of School Effectiveness and Costs in Tamil Nadu." Ph.D. dissertation. University of Londoni. Becker. Gary S. 1964. Humau7n Capital: A Theoretical and Empfirical Ana7lysis, with Spe(ial Reflrence to Education. General Series 30. New York: Columbia Univer- sity Press. Benavot. Aaroni. and David Kamnens. 1989. "The Curricular Content of Primary Educa- tion in Developing Counitries." Policy Research Working Paper 237. World Bank. Education and Social Policy Department, Washington, D.C. Berg, Alan. 1987. Malnutrition: What Can Be Done? Lessons from World Bank EApe- rience. Baltimore, Md.: Johns Hopkins University Press. Bhowon, Rajayswar, and V. Chinapah. 1993. "Reform of Basic Education in Mauritius: The Process of Information Gathering, Consultation and Decision-Making." In David W. Chapnman and Lars 0. Malck. eds.. Iiofbrniationi Systenms in Educational Planning. From Data to Action. Paris: UNESCO, International Institute for Educa- tional Planning. Blackburn. M. L. 1990. "What Can Explain the Increase in Earnings Inequality among Males'?" Induistu-ial Relations 29(3):441-56. Blackburn, M. L.. D. E. Bloom, and R. B. Freeman. 1990. "The Declining Economic Position of Less Skilled American Males." In Gary T. Burtless, ed., A Future of LousyvJobs?' The Changing Srr-uc(tl//re of U.S. Wages. Washington, D.C.: Brookings Institution. Blaug. Mark. 1976. "The Empirical Status of Human Capital Theory: A Slightly Jaun- diced Survey." Journal of Economic Liler-atur-e 14:82'7-55. Boissiere, Maurice, J. B. Knight, and R. H. Sabot. 1985. "Earnings, Schooling, Ability. and Cognitive Skills." Amier-ican Economic Revie' 75 (December): 1016-30. Bound, John, and George Johnson. 1992. 'Changes in the Structure of Wages during the 1980s: An Evaluation of Alternative Explanations." Amnerican Economic Review 82(3):371-92. Bowman, Mary Jean. and C. Arnold Anderson. 1963. "Concerning the Role of Educa- tion in Development." In Clifford Geertz, ed.. Old Societies anzd Neun States: The Questfir Modernity in Asia and Africa. New York: Free Press of Glencoe. Bray, Mark. 1987. School Clusters in the Third World: Making Them Work. Paris: LUNESCO-UNICEF Cooperative Program. . 1990. "The Economics of Multiple-Shift Schooling: Research Evidence and Research Gaps." Inter-nationall ouirnal of Educational Developmient 10(1/3): 181 - 87. Bray. Mark, with Kevin Lilli. 1988. Commu,nity Financing Of Education: Issues and Policy Implications in Less Developed Counitr-ies. New York: Pergamon Press. Bridge, Gary R., Charles M. Judd. and Peter R. Moock. 1979. The Deteroi7iiatits of Ed cational Out omes: The Impa ct of Families, PeeL-s. Tea hers and Sc hools. Cambridge, Mass.: Ballinger. Brookover, Wilbur B., and L. W. Lezotte. 1979. Changes in School Chalracterisics Coincident with Changes in Studen7t A chievenient. Institute for Research on Teach- ing Occasional Paper 17. East Lansing, Mich.: Michigan State University Press. REFERENCES 159 Brubaker. 1H., and R. Partine. 1986. 'Implementing Effective Schools Research: The Audit Process. High School Observations." Paper presented at conference of Amelri- can Educational Research Association, San Francisco, April 16-20. Bryant. D. M., and C. T. Ramey. 1987. "An Analysis of the Effectiveness of Early Inteivention Programs for Environmentally At-Risk Children.'' In Michael J. Guralnick and Forrest C. Bennett, e ds.. Th7e Effictiei'es.s of Early Inte'rventiolftr At-Risk and Handicupped Chilir-en. Orlando. Hla.: Academic Press. Bundv. D. A. P. and others. 1990. "Control of Geohelminths by Delivery of Targeted Chemotherapy through Schools." Traiisactions of the Royal Societv of Tropical Med:'cine ,i nd Hygiene 84:115 -20. Butt. Abdul Rauf, and Mohammed Amjad Sheikh. 1988. 'An Analysis of the Gap between Demand for and Supply of Higher Education in Pakistan.' Pakistan Eco- 11017ioc and Social Review 26(1 ):41-56. Caldwell. J. C. 1979. "Education as a Factor in Mortality Decline: An Examination of Nigerian Data." Poplarition Studies 33(3):395-413. Carter. C.. and J. Klotz. 1990. 'What Principals Must Know Before Assuming the Role of Instructional Leader." NASSP Blulletin 74(525):36-41. Chaturvedi, S.. B. C. Srivastava. J. V. Singh, and M. Prasad. 1987. "Impact of Six Years Exposure to ICDS Scheme on1 Psycho-Social Development." Indinil Pediatrics '- 4(2): 153-60. Chemichovsky. Dov. 1985. "Socioeconomic and Demographic Aspects of School En- rollnienl and Attendance in RuLal Botswana." Economic Develop)mewent and Cill- turul Change 32:319-32. Chiswick, Barry R. 1991. "Speaking. Reading and Earnings among Low-Skilled Immi- grants. " Jomiinanl'?t Labor Economics 9(2): 149-70. Chiswick. Barrv R., and Paul W. Miller. 1995. "The Endogeneitv between Language and Earnings: International Analyses." .lIouriial o Labor Econ0o7iCs 13(2):'-46-88. Chowdhur\'. Kowar. 1993. "Women's Education: Barriers anid Solutions." World Banlk. Education and Social Policy Department. Washington. D.C. Chubb. John. and Terry M. Moe. 1990. Politics, Markets Ind Am4ierica'.s Schools. Wash- ington, D.C.: Brookings Institution. Chutikul, Sirilaksana. 1986. Thie Eff'cl of Tluition Fee Increases on 1he Denmndjfbrl Higher Edlueation: A Caise Studx' oJ a Higher Education ilnstitiutioni in Tlihilad. Sussex, U.K.: University of Sussex. Cleland, John, and Chr-istopher Wilson. 1987. "Demand Theories of the Fertility Traii- sition: An Iconoclastic View." Population Stiuldies 41:5-30. Colbert, Vicky. C. Chiappe. and J. Arboleda. 1993. "The New School Program: More and Better Primar-y Education for Childiren in Colombia." In Henry NM. Levin and Marlaine E. Lockheed, eds.. Ejfective Schools In Develofinsg Countries. London: Falmer Press. Colclough. Christopher. 1990. "Raising Additional Resources for Education in Devel- oping Countries: Are Graduate Payroll Taxes Superior to Student Loans'?" Interant- tion,al Journal of Edhicaional Developnme't 10(2/3):169-80. Colletta, Nat J.. and Gillian Perkins. 1995. '-Participation in Education." Participation Series Paper. World Bank, Environment Department, Washington, D.C. 160 REFERENCES Colletta. Nat J.. and Marelaret SuLtton. 1989. 'Achiievinio aind Sustainine Universal Pri- marv Education: International Experience Relevant to India." Policy Researclh Workinig Paper 166. World Batik. Populationi and Humani Resoul-ces Departmen1lt. Washington. D.C. Common wealthI Secretariat. 1994. ''The Changing Role of the State in Education: Poli- tics and Partnerships. An Oveerview of CounItry Papers." Agenda Item 2. Twelfthl Coltferenice ot'Commoniwealtlh Educatioll Ministers. Islamabad. Pakistan. Londoni. Cox-Edwards. Alejandra. 1994. 'CChile: Strategy for Rural Areas Enhancing Agricul- ritral Competitiveness and Alleviating Rural Poverty." In "Rulal Labor Markets." Report 12776-Chile. World Bank;. Washintgton, D.C. Dalin. Per, in cooperation with Tekle Ayano. Anbeso Biazeti, MumtIaz Jahan. Matthew Miles andl Carlos Rojas. 1992. H-ow Sc/cools Impro'voe. Oslo: Interinationial Move- menits toward Educational Chan.ge (tMTEC). David, J. L., and S. M. Petersoni. 1984. Can Sc/hools Imirove Themsel/es?.' A Studv of Sc /thool-Bc,scd Iomprovemenrt Prog cois. San1 Francisco: Bay' Area Research Group. Davis. S. J. 1992. Cross-Countrv Patterns of Change in Relative Wages.' NBER Maccroc'c'otocs Aniitiiccl 1992. Cambridec. Mass.: MIT Press. Demisky, Teny. 1994. "World Bank Lendinie for Secondary Educationi: A General Operational Review.." World Batik, Education and Social Policy Departmiient. Wash- ington, D.C. Denisoni. Edward F. 1967. Wh v Growt/t Roe's Difftr: Post-Woe Erperieccce in Ni/n' Western Cocntries. Washinigtoni, D.C.: Brookin2s Institutioin. Dertler, Frank. 1992. PC Goiide to C('mcctris'i. Emleryville. Calif.: Ziff-Davis. DhananLi. S. 1993. "Indonlesian Manufacturing Employmilent and Training." Vol. 1. "Major Findinigs of the 1992 West Java Enterprise Scitvey." Technical Report 2. BAPPENAS, Bureau of Manpower. Regional Manpower Planninig andl( Traininig Pro'ject (World Bank Professional HuLm1an Resoitrces Developmncit Loan 31 34-tD). Jakarta. Indonesia. Donors to African Educaltioll. 1994. .A Stactisticl Profile of E(hwchccio ico Pt Sio-S.h ctc.' Aitfica in tih 1980,s. Paris. DIotcher, Nadine. in collaborationi witlh G. R. Tucker. 1994. 'The Use of First and Secondl Languages in Educaltioll: A Review of Educationlal Experience.' Woricl Bank. East Asia aind the Pacific Region. Courttry Department III. Washilgtoll. D.C. Easterlini, Richaid A. 1981 . 'Why Isi't the W\hole World Developed?" .lomctc:l of Ec otocnc' Hristocv 41( 1- H15. Easteriv, William, and Stanley Fischer. 1994. "What We Can Learnii fromii the Soviet Collapse.' Filziatce ac1d Decl eoloepm 31(4): 2-5. Eckstein., Max A.. and Harold J. Noah. 1993. Sec onldrs'v Sch ool Em,niticions: Iictci - lfl- tioIIcil Pers/pceti es on Polic ies cocci Prcl ices. New Haven. Conn.: Yale University Press. Eisernon. Thomas Oweni. 1988. Bectefiticg otjo Blcsic Ecat/ oic/O.hoe ciol Qiiilitci . cocci1 Fcc/cc Ncaclonl Lither v' in Ken va. Comroparative and Interinationial Educaltioll Series 2. Oxford: Perel-mon Pres.s. REFERENCES 161 ._ I')X'). "Whlat LaniMuawe Shotilcd be [ sed tor TeachinW? Languatg Policv and School Reforml in Buru ndi .' .oonail of VMltilin,g,i,l and Mulic lmioal Dcv tIop- meot 10(6):473-97. Eisernon. Thomlio s Oweni. and( John Schwil le. 1991. 'Priniarv SchoolinLc in BuLLindi and Kenya: Preparation for Seconidary EcLuca1tionl or for Self-Employmilenit?" Ele,,,,- tarsx .S'ool .1Sh 101Zo / l 92l) :23-39. Eisenion. Thomas Owen. J. Ratzlaiff'. and V. L. Patel. 1992. "Reading Instructions fir Usieig C'oin niercia l Medic iies.' Annals /ftilt' Amr or ian At advi'n ol P/lilt ica and Sot ieal Scicnc' 520: 76-90. Eisemoni. Thomia.zs Owen. Johni Schwille. aLid Robent Prout11. 1989. "'Empirical Results and ConveitionaL;l Wisdoni: Strategies for Inccreasin- Primarv School Eftectiveness in BuI liidi.'' Dratt report. BRID(iES Proiect. Harvard IiIstitite for International Development. Camiibridge. Mass. . 1992. "Cani Schooling Make a Better Farmer? Outcomiles of AgricultUral Education in Burulidi. In G. G. Madles. ed.. Primairz Stchool Agriculilure ill Sidb-Schair-a Alriutt. Eschboihor: Dent scihe Geselischaft fil r Techlniscihe Zusamnmenarbeit. Eisenionl. Thonias Owen. I. MihailesCL. L. Vlasceaniu C. Davis .J. Sheeha;ln. and( C. Zamfir-. Forthcoinieii '"Higher Educationi Reforim in Romania." Hit1,,hr Euiceitioin. Frcledeick. J. NI. 1987. Mc'Ia.mir in , Stl ool Effetisvene.zs.: Guidelic'nes fin Educ ational Pra410itioller v. ERIC Clearinie HOUSe on1 Test Measlrement andl Evaluation. Ecducat- tional TestinLg Services. Princetoni. N.J. Fuller. Bruce. aLid P. Clarke. 1994. Raisine School Effects While tgnoring Culture? Local Condlitionis and( the Influelice of Classroom Tools. Rules and Pedagogy.- Re view of Eduhocatlonal Reseo hi 64(1:1 I19-57. Fuller. Bruce. H. Hui. and C. W. Snyder. 1994. Wheni Girls Learn More thalil Boys: The Influence of'Timiie in School anid Pedagomv in Botswana.' Com/parative Edlo,- Oti(toll Review 38(3):347-77. CiAO (U.S. General AccouLntin1) Office). 1994. fldncttio, Re/orm,,. School-Bs,ed Maclo- a.Iementl RsIlt's ill Cht,'Utes in Instruction an,,dl Budteliint,. Washington. D.C'. Gertler. PauLl. and PaMl GlesWWC. 1989. The 14 illij/Ij,,.ss to Pav It / Educatlion in Detel- op/il04 Cotoorrie's.c .'Ldile eo /rtn Rurtial Perl(. Livin.- Standards Measluremnent StLIcil Working Paper 54. Washington. D.C'.: WoriIcl Bank. Gibhs. Greg K. I 989.Effittit e Sc hl Reteart 1,: 1 he Pt/c/a i.s 1i illc tiol Lecoler. Washington. D.C.: U.S. Depar tmncit ot' Edlocaticii. Gill. I[idlerimit. andc Michiellc Riboud. 1993. 'Produlctivitv Cirowthi and Wage Structures ill MeXCO alid the U.S.: SuLrllisilig thei Emiployimienlt Effects of' the N AH A.. World Banik. Latini Americ; aand ilic Ca ibbean Regaion. Counitry Depart menlit 11. Wwal ing- ton. D.C. Gonizhlez-Su:irez. NIMirla. 1987. Barriers to Femiallc Achiieveiiieilt: Sex Stereotvlpes in Texthooks.'' Paper presenited to t le Corn paraltiVe aid tliternationalI EdctICllioll Soci- ety. Washinlgton. D.C. Gouild. David NI.. and( Roy J. Rutt'in. 1993. "Whatt Determines Ecoiolinlic Growth?' Et om/ibc Rev iet. Fcder cil Rest xi / e of l)all. (21 2l quarler). 162 REFERENCES Greaney, Vincent. and Thomas Kellaghan. 1995. Equity I.ssotes in Puhblij ExaminationsN in Developinig Cointr-ies. World Bank Technical Paper 272. Washington, D.C. Griffin, Peter, and Alejandra Cox-Edwards. 1993. "Rates of Return to Education in Brazil: Do Labor Market Conditions Matter?" Economics of Education Review I t(3):245-57. Haddad, Wadi D. 1994. The Dynamnics of Education Polic makin!g: Case Stuldies o' Buirk-ina Faso. Jordan. Per u and Thazilanld. EDI Development Policy Case Studies Series, Analytical Case Studies 10. WVashington. D.C.: World Bank. Haddad, Wadi D., Martin Canioy. Rosemary Rinaldi, and Omporn Regel. 1990. Educa- tioa and Dev elopment: Evidence fin'New Priorities. World Bank Discussion Paper 95. Washington. D.C. Hallinger, Phillip. 1989. "Developing Instiuctionial LeadershiipTeamils in High Schools.' In Bert Creemers, Toniy Peters. and David Reynolds. eds-.Schoo/ E.#ectiveness and School Improvement: Proceedings N e the SeCOnd Intet'rlatinItal Con(greVSS. Amsterdam: Swets and Zeitlinger. Halpeni. Robert. 1986. "Effects of Early Childhood Inter-vention on Primary School Progress in Latin America." Comtparati'e Education Review 30(2): 193-215. Hannaway, Jane. 1991. "The Organization and Management of Public and Catholic Schools: Looking Inside the 'Black Box."' Internati mal ootonal oJ Educational Research 16:463-81. . 1993. "Decentralization in Two Districts: Challenging the Standard Paradigm." In Jane Hannaway and M. Camoy, eds., De entraiziationt and/ Scltool IS ro c0- miew/: Can We Fu/ill the Promtise:' San Frniicisco: Jossey-Bass. Hanushek, Eric A. 1994. "Outcomes. Costs and Incentives in Schools." Paper prepared for the conference on Improving the Performanice of America's Schools: Economic Choices. U.S. National Research Council, Washington, D.C.. Harbison. Ralph W.. and Eric A. Hanushek. 1992. Educalional Peri)rtnance at the Poor: Lessonsfromn Rural Northeast Bra:il. New York: Oxford University Press. Herz, Barbara, K. Subbarao, Masooma Habib, and Laura Raney. 1991. Lelting Girls Learn: Promising Approaches in Primnarv ind Secondary Education. World Bank Discussion Paper 133. Washington, D.C. Heyneman. Stephen P. 1994. 'Education in the Europe and Central Asia Region: Poli- cies of Adjustment and Excellence." Internal Discussion Paper 145. World Bank. Europe and Cential Asia Region. Washington. D.C. Heyneman, Stephen P.. Joseph P. Fariell, and Manuel A. Sepulveda-Stuardo. 1978. Textbooks (alnd A(chtiev'em't: Wlhat W Knonw. World Bank Staff Working Paper 1.98. Washington. D.C. Himmel, Erika. 1995. Case Studv of National Assessment in Chile." EDI Developmenlt Policy Case Series. World Bank. Washington, D.C. Hobcraft. John. 1993. "Women's Education. Child Welfare, and Child Survival: A Review of the Evidence." Health Transition Rev iew, 3(2): 159-75. Holsinger. Donald. 1994. 'Thailand: Proposed Secondary Education and TeacherTraininr Project." World Bank, Education and Social Policy Department, Washington, D.C. Holsinger. Donaid, and David Baker. 1993. 'The Size and Structure of Secondary Education in Developing Countr-ies." ESP Discussion Paper 7. World Bank. Educa- tion and Social Policy Department. Washington. D.C. REFERENCES 163 Hyde. Karin. 1989. Improving Women's Education in Sub-Saharan Africa: A Review of the Literature.' PHREE Background Paper Series 15. World Bank. Education and Social Policy Department, Washington, D.C. IEA (International Association for the Evaluation of Educational Achievement). 1994. Indicators oJ Between-School Differences in Reading Achievement. Prepared by Andres Schleicher and Jean Yip. The Hague. ILO (Inteniational Labour Office). 1992. World Lalbor- Report. Geneva. Jain, Balbir. 1991. "Returns to Education: Further Analysis of Cross-Counitry Data." Economics of Education Revievw 10(3):253-58. Jarousse. J. P., and Alain Mingat. 1993. Optionsfir Accelerated Develop)Ment oJPri- mnary Educ'ationi in the Sahel. 1REDU-CNRS, Universite de Bourgogne. France. Jimenez. Emmanuel. 1987. Pricinig Po?licy in the Social Sectors: Cost Recovery for Education aid Health in Developing Countries. Baltimore, Md.: Johns Hopkins University Press. Joyce, Bruce R. 1991. "The Doors to School Improvement." Educcationlc Leadership 48(8):59-62. Joyce. Bruce R., and Beverly Showers. 1985. Power in Saffi Developnment through Research on Trainling. Alexandria, Va.: Association for Supervision and Curricu- lum Development. 1987. "Low-Cost Arrangements for Peer-Coaching." Educational Leadeership 45(2):22-24. 1988. Stude'lnt Achievecent Through Stff Decelop7cent. New York: Longman. Joyce. Bruce R., Richard H. Hersh, and Michael McKibbin. 1983. The Structure of School Improvement. New York: Longman. Kagitcibasi, Cigdem, D. Sunar, and S. Bekman. 1987. "Comprehensive Preschool Edu- cation Project: Final Report." International Development Research Centre, Ottawa. Kellaghan, Thomas. and Vincent Greaney. 1992. Using Examinations to Improve Edlu- cacition: A Studv of Fourteen African Countries. World Bank Technical Paper 165. Washinigton, D.C. Kelly, Michael J., Eileen B. Nkwanga, L. Henry' Kaluba. Paul P. W. Achola, and K. Nilsson. 1986. "The Provision of Education for All: Final Report of the Education Refomii Implementationi Project.' University of Zambia, School of Education. Lusaka. Knight, John B., and Richard H. Sabot. 1990. Ecihcc'ctioni. Productivity. and Inecquality: The East Afiican Natur-al Experiment. New York: Oxford University Press. Korea. Ministry of Education. 1992. Education in Korea: 1991-1992. National Institute of Educationial Research and Training, Seoul. Laporte, Bruno, and Julian Schweitzer. 1994. "Education and Training." In Nicholas Barr, ed., Labor Mart-kefs and Social Polic X in Centr-al anid Easter n Europce: The Tranlsitioni and Beyond. New York: Oxford lUniversity Press. Larach. Linda. and Marlaine E. Lockheed. 1992. 'World Bank Lending for Educational Testinig." PHREE Background Paper 91/62R. World Bank, Education and Social Policy Department. Washinigton. D.C. Lau, Lawrence J., Dean T. Jamison, and Frederic F. Louat. 1991. "Education and Productivity in Developing Counitries: AII Aggregate Production Function Ap- proach." Policy Research Working Paper 612. World Bank, Development Eco- nomics and Population and Human Resouices Department. Washington, D.C. 164 REFERENCES Lau. Lawrence J., Dean T. Jarnisoin, S. C. Liu. ancl S. Rivkin. 1993. "Education and Econoimiic Growth: Some Cross-Sectional Evictence fromii Brazil.'' .foumol o., De- )e/lo/liimewt E onom,n,i 41:45-70. Lee. Valer ie E.. andl Marlainle E. Lockheed. 1990. "'The Et'fects ot Single-Sex Schoolineg on Studelt Acihieveine ilt in Attittides in Nigeriat.' Comipualive Echicalion Review 34( __):2(09-3. Lco-R Ilviie. E. 198I. Report *o, 7/li S/,fi .Sv.temt i,, JIamaicoo Sclhno/l.s. University of the West In dies. Scthool of EdCucatiOll, Mona. Jam1aica. ILevii, HetirvN M.. and MNarlaine E. Lockheed. eds.. Effi'ctime Sc hools In Develop/niny Countries. Londoni: Fit liner Press. Levine. Daniel li. 1990. "Update on Effective Schools: Findings and Implications thom Research and Practice.'' .Jonwonl of ;eg,ro Edmin/ioll 59(4):577-84. Levine, DaLniel lI_ and Lawrence W. Lezotte. 1990(. "Unusually Effective Schools: A Review and Analvsis of Research and Practice.' National Center for Et'fective Schools Research and Developument. Madison, Wis. LeViie, Robert. S. E. Levine. A. Richlimianl. F. M. Tapia Uribe. C'. S. Correa, anzi P. NM. Miller. 1)991. ''Womien's Sclhoo[ing- and Chiil Care in the Demographic Tranisiti oil: A Mexican Case StIIdV.'' Popl/ation (II tl Dc vc/opmnenlt Rel w1v 17(3):459-96. Levinger. Beryl. 1992. "Nutritioni. Health and Learning: Curenit Trends and Issues." School Nutlitilo and Health NeLwork Monograph Series 1. Education Develop- nienit Center, Newtonl. Mass. Lezotte, L. W.. and B. A. Bancroft. 1985. 'Growing Use of the Effective School Model." E4/uic o>ioml Leac/ci s/i//i 42 (6):23-27. Lezotte. L. W.. and others. 198(0.Sel/ool Leo Rili,' Climate c andl At ademic Ac/iievcment: A ,Soc ial Systcm.s Ap/roach/i to im lrease StOcimc/t Learo/uing. Site-Specific Technical Assistance Center. Florida State Univeisity Foundation, Tallahassee. Fla. Ltindenbatim, Shirlev, Manisha Chakraborty. and Mohammed Elias. 1989. "The Influ- enice ot Maternal Education on Intfant amid Cliilci Mortality in Bangladesh ." In Joliii C. Calciwell anid Giti Santow. edls .,Sele(ited/ Reac/iii-s in t/li, Cnd/ro,/ Social and/ Bcl/oe ioral Deternimants ! ., Heolt/. HeaIlth Transition Series 1. Canberra: High- land Press. Lindert. Paul H. 1976. "C'hild Costs and Economilic Developmenit." In R. A. Easterlin. edl.. Popl/tioo cad Ei Econoonmic Can,,,ew i, Dev cl/pinit,i Commtrie.s. Cliicago: Univei- sity of Chicago Press. Lockheed, Marlaine E., anid Eric Hanushek. 1988. "Imnproving Educational Efficient's in Developing Countries: Whalt Do We Know'?" Comipre 18(1 ):21-38. Lockheed, Marlaine E., and Emmanmuel J imlenez. 1994. "'Public and Private Secondary Schools in Developing Countries." Human Resouices Developmenit anld Opera- tions Policy Working Paper 43. World Bank, Washington. D.C. Lockheed. Niarlaine E.. Josefina Fonacier, amid Leonard J. Bianchi. 1989. "Ei'fective Primary Level Science Teacliig ini the Philippines." Policy Research Working Paper 208. World Bank, Education anid Social Policy Departmnemnt. Wash iligton, D.C. REFERENCES 165 Lockheed. Marlaine E.. Dean T. Jamison. anld Lawrence LaLl. I 980. 'Farimer Educatioll anid Fatrimi Efficiencv: A SLuIvey." E onomnic De ve/opmenht mid Cdnliro ('Chaiq'e 29 (Octoher3:37-76. Lockheedl. Marlaine E.. Johni Middleton, and Greta S. Nettleton. eds. 199 1. Educationi Technology: Sustainable and El'fective Use.'' PIIREE Backgroul1d Plaper 91/32. World Bank. Eduicationi and Social PolicV Depatmnienit. Washington. D.C. Lockheed. Marlaine E., Adriaan M. Verspoor. and others. 1991. Iproi in jl Princarv Edincatjioo in DL)'elin C(ountrie.s. New York: Oxford University Press. Loh. Eng SenLg. I 992. *'Technological Changes. Trainingg. anid the Inter-Industry Wage StriCtUcN." Qalu-tcrv Rev'iew of Ec onohifs. C1li0Fjoanc e 32(4):-6-44. Lomnbard. As i ma. I 994.Stt ts.c Be inS. (io Home.: Thle Post, Present and1(1 Iiitil'c of th/c HIome lllntrm tiill Pro'rb'aIll /l'r PreVsc/l0) Y01111.tSS. D)Ushkill. Con n.: GUilford. Loulr. G. I 977. 'A Dyniamic Theorv of Racial Incomile Dit'ferenices." In P. A. Wallace and A. M. LaMonid. cd .. Voiten,i 1inoIritie.s. nid Emlvinvcinit lDiv( limillntioli. Ncss York: Macmillan. I 987. "Why Should We Care about Group IneqLlaUliY I So( lcld Phil0oophy Vald Poidj'! 5:249-71. Lucas. R. E. 1988. 'On1 the Mechanics of EconIom11ic Development.Joilrnlclil of MAll- (ctorv Econ)1m0) 2 :3-21 Lundberg. Ingvar. andl Pirjo Linnakyla. 1992. Teat hitig Reading AroInd the WUorldI. The Hague: Interinationial Association for the Evaluationl nf Educational Achievemilenit. MadaLus. G. F.. aid V. Greaney. 1985. '-The Irish Experience in Comilpetenicv Education: I ipliczationls foIr Americani EduIcatioll." Ameri(c a Jourail of'Educatiolo 93:268- 94. Martin. C. L., and(l Ci. Levy. 1994. 'Gender Roles andL Preschool Education." Iberna- tio/nal Emcv/ lopedia of Edotatiin. vol. 4. 2d ed. New York: Pergamoio Press. Mason, Andrew David. 1994. "Schoolimg Decisions. Basic Educatioll a;d tile Poor in Ruit ral Java.'' Ph.D. dissertatiotn. Stan ford University, Foo(d Researicl InstitLIte. Stanford, Calif. McKniiglt, Allan. 19771. Sti(ntists Abroad: A Stu' o?f' tothe International tllcmcntient oft Per.No01s ill Sc'ic/itn e and Ttchnoh '1gv. Paris.: UNESC(. McMahoni. Walter W., and Boediono. 1992. "Market Signals and Labor Market Analy- sis: A New View of Manpower SuL)plies and Demanlds.' In Waltel W. McMallon alid Boedinoio. eds.. Etdcationt ad) the Ecrnomvy. Jalkarta: U.SAID. McMIalonl. Walter W.. adl(i JUtig. H. J. 1989. 'Vocational and Technical Educationl inl Indonesia: TheoreticaLl Analysis and Evidence oti Rates of Rettir[i." Facult\ Work- iig Paper 89- 1582. University of llliniois. College of Comifilerce and Business Adm il istrat ion, L rban a-Champaignc, Ill.. McNamara. Robert S. 1992. 'The Post-Cold War World: Implications for IMilitary Expenit itLIl'C ill the Developing CoUntitrie,.'' Int PlntO'C''Tdllg'.\ of t/h' AmAlit//l Confl'- cIc'l 'i)1 Di'\vlopwn ent E(conoumic' /1991). Washingtong , D.C.: Worldl B ank. Midd letonl. Jotil, Adri an Ziclettilian. ard Airvil Vatn Adatiis. 1993.ASki/ls fir Prodl/nc ii - if% .- I oc ati11on1a Ectmoato cIcl111(1 Trainjing} ifOn Dev'llillo , Couicntries. New Y ork: Ok- fordl U liversity Press. 166 REFERENCES Mincer. Jacob. 1989. "Human Capital and the Labor Market: A Review of Current Research." Educational Researcher (May):27-34. Mingat. Alain. and Jee-Peng Tan. 1985. On Equity in Education Again: An Intena- tional Comparison." lJournal of Human Resources 20:298-308. . 1994. "International Perspectives on Education: Some Ideas for Data Analy- sis." World Bank. Education and Social Policy Department, Washington, D.C. Mittler, Peter. 1992. "International Visions of Excellence for Children with Disabili- ties." Intel-narional Jouarial of Disability, Dev elopmient anid Education 39(2):115- 26. Mittler. Peter. R. Brouillette. and D. Han-is. eds. 1993. World Year Book of Education. London: Kegan Paul. Moock, Peter. 1994. "Education and Agricultural Productivity." Inter-nationial Encvelo- pedia of Education 1:244-54. Oxford: Pergamon Press. Moock, Peter. Philip Musgrove, and Morton Stelcner. 1990. Education anid Earnings in Peru's lqta niral NVonlirtn Famnily Enterplises. Living Standards Measurement Study Working Paper 64. Washington, D.C.: World Bank. Murnane. Richard J.. John B. Willett. and Frank Levy. 1993. The Growing Importance qf Cognitive Skills in Wage Determination. Harvard University, Graduate School of Education and Massachusetts Institute of Technology, Department of Urban Studies and Planning. Cambridge, Mass. Myers, Robert. 1992. The Twelve Who Survive: Strengtlhening Progranmmes of Early Childh7ood Delvelopmnent in the 7hir Weolld. London: Routledge. Myers. Robert, and others. 1985. "Pre-School Education as a Catalyst for Community Development: An Evaluation." USAID, Lima. NCERT (National Council for Education Research and Training). 1994. Minimum Learn- ing Continuiuim. New Delhi. National Technological University. 1994. "The National Technological University: Background." Fort Collins. Colo. Nielsen, H. Dean, and NI. T. Tatto. 1991. The Cost EJf1ctiveness ojfDistance Education Jor Teacher Training. BRIDGES Research Report 9. Hiarvard Institute for Intenia- tional Development, Cambridge, Mass. Nitsaisook. Malee, and Lorin W. Anderson. 1989. "An Experimental Investigation of the Effectiveness of In-Service Teacher Education in Thailand." Teacher an7d Train- ing Edu-cationi 5(4):287-302. Noss, Andrew. 1991. "Education and Adjustment: A Review of the Literature." Policy Research Working Paper 701. World Bank, Education and Social Policy Depart- ment, Washingtoni, D.C. OECD (Organisation for Economic Co-operation and Development). 1993. Education at a Glance: OECD Indicators. Paris: OECD. Centre for Educational Research and Innovation. . 1994a. The Markets for Learning and Educational Services. Paris: OECD, Cen- tre for Educational Research and Innovation. 1994b. School. A Mlatter of Clhoice. Paris: OECD, Centre for Educational Re- search and Innovation. REFERENCES 167 Orazem. Peter F., and Milan Vodopivec. 1994. "Winners and Losers in Transition: Returns to Education, Experience, and Gender in Slovenia." IPolicy Research Work- ing Paper 1342. World Bank, Policy Research Departimient, Washington, D.C. Patrinos. Harry A. 1994. "Education and Earinigs Differentials During the 1980s." World Bank, Education and Social Policy Departmenit. Washington. D.C. Patrinos, Harry A.. and George Psacharopoulos. 1995. "Educational Performalice and Child Labor in Paraguay." Itnrerinational Journal of Educational Development 15t I):47-60J. Patrinos. Harry A.. Eduardo Velez. and George Psacharopoulos. 1994. "Language. Educationi and Earninigs in Asuncion. Paraguay." Journal of Developing Areas 29(1):57-68. Pollitt, Ernesto. 1990. MAflnutrition and Inte ' cion in tile Clarssroom. Paris: UNNESCO. Postlethwaite. T. N.. and Kenieth Ross. 1992. E7ff'ctive Schools in, Reading: Imiplica- tionis fi- Educilationial Placniers. The Hague: International Association for the Evalu- ation of Educational Achievement. Prawda, Juan. 1993. "Educationial Decentralization in Latin America: Lessons Learned." Inlternational lournal of Educatinial Dev'elopment 13:253-64. Psachar-opoulos. George. 1987. "To Vocationalize or Not to Vocationalize?' That Is the Curriculum Question. International .lomnal of Edm ation De elopmelnt 33(2): 1 87- 2 I 1. .19X9. "Time Trends of the Returns to Education: Cross-National Evidence." Economics of'EduIaotion7 Resview 8(3):225-.3 . 1994. 'Returns to Investment in Education: A Global Update." Wor-ld Develop- ment 22(9):1325-43. Psacharopoulos, George. and Harry A. Patrinos. 1994. Indigenous People ande Pov'erty in Latin Amnierc: A4n Empirical Anal.Ysis. A World Bank Regional and Sectoral Study. Washinigton. D.C. Psacharopoulos, George. and Zafiris Tzannatos. 1992-. Case Stllies on7 Women's Em- plovnient and Pay in Latin Anmerica. A World Bank Regional and Sectoral Study. Washinigton. D.C. Psachalopoulos, George, and Eduardo Velez. 1993. "Educatiolnal Quality and Labor Market Outcomes: Evidence from Bogota, Colombia." Sociology, of Education 66:130-45. Psacharopoulos. George. and Maureen Woodliall. 1985. Education finr Development: An Analysis of Investment Clhoic es. New York: Oxtord University Press. Psacharopoulos, George, Carlos Rojas, and Eduardo Velez. 1993. "Achievement Evalu- ation of Colombia's Escuela Nueva: Is Multigrade the Answer?" Compar'tive Education ReVieC 37(3):263-76. Psacharopoulos, George. Eduardo Velez, and HalTy Anihonly Patrinos. 1994. "EdIica- tion and Earnings in Paraguay."' Economics oJ Edacation Rev'ieu 13(4):3'2 1-27. Purkey, Stewart C., and Marshall S. Smith. 1983. "Etfective Schools: A Review." Elemsewtarv SchoolJ.Iournal 8314):427-52. Purvear, Jel'frey. 1995. "International Education Statistics and Research: Status and Problemils." ntesnational .oual of Ecicational Developient 15(1 ):79-9 I 168 REFERENCES Rauidenbush, Stephen W.. C. Bhumirat. and M. Kamali. 1989. Predictors a.d I Coose- quen ees of. 7ca he,l. Koolii'edlge i, Thahilhad. BRIDGES Project. Harvard Inistittite for International Development. Cambridge. Mass. Richards. M., and J. Richar-dls. 1990. Langouaugcs aitlb Commimitiics Emompusti7 sed by! Guet'emeh,st NVatijonla Bilion'oal Educati jon P t,IS. ISAID in conjUnOlion Witlh the M inisteirio de Educaci6i de (Giuatemalat. RodrlfLuez GernMill. and John Cleland. 1980. 'Socio-Economic Determinants of Mari- tal Fertility in Twenty Countries: A Multivariate Analx'sis.' WVorld Ferlility S,,rcy Cof6erence 198'U: Recoid of Proceediol's.Vol. 2. London: World Fertilitv Suivev. Romer. P. M. 1986. ''increazsing Retu-rs ancl Long-Run Growth.'' Journal of PoliticUl Economv 94:1002-37. Rosenzweig. Mark R. 1995. "'Why Are There Returl-Is to Schoolin-?' AmuricHa Eto- 11omi0 Rei iv 855('):I5 -5S8 Ross, Kenneth N.. mid Lars M;ihIck. 1990. Plaimuing Me/i Qudyitv o ElEdation: The Coliectioni (111(1 Use of Data for Il ormted Dec,isio-Makin o. Parlis: tINESCO. Ross, Kennethi. and T. N. Postlethwaite. 1989. 11doo1e.sid. Qufalitt of' Baoshi Education. Jakarta: Ministry of Education anid CulIt tire. Sadlak. Jan. 199:. ''Legacy and Change: Higher Education anid the Restoration of Academic Work in Romania." Techuiologv it So( iety 15( 1):75-10)(). Salhi. Jamil. 1991. "Perspectives on the Financing of Higher Education.' PHREE Back- erouLICn Paper 91/45. World Bank. Edlucation and Social Policy Departmienit. Wash- ington. D.C. Sapsford, Davidc and Zafiris Tzannatos. 1993. The Economics of'tie Latibouir Mairket. Basingstoke, U.K.: Macmiillain. Sathar. Zeba A.. aid Cynthia B. Llovd. 1993. Wlho Gets Pi iml St hlloolio im Pakistan :' Inequalities 011mo00 awll't VI itloli Fulodlie.s. Working Paper 52. New York: Popula- tiOnl CouIllcil. Schaeffcr. E. S. 1987. "Parelital Modernity and Child Academiiic Competence: Toward a Theorv of Individual and Societal Developuieitl.'' EorlK Chiildhoou Dc [el' h gine ,e alwil Catr' 27:373-89. Scheerens. Jaap. and( Bert Creeiner-s. 1989. "Towardcs a More Comprehensive C'onceptualiza1tionl of School Effectiveness.' In Bert Creemers. T. Peters. and D. Reynoldls. eds.. St hoo*lEf/Nltt toesZ.s. titil .S boo (vImptot'e oteobt. Amisterdlami: Sw ets and Zeitlinger. Schultz. T. Paul. 1993. 'I ivestinents in Schoolig anid Health ot' Womeil and Men: Quantities and Return."Jo /onaIl of l-liHima ResttIt'i rct 281t40:694-734. . 1994. "'Inteeraltecl Approaches to Human Resoulce Developuelit.' World Bank. Huitiiaii Resottirces Develolipitent and Operationis Policy Vice-Presidency. Wash- iitgton. D.C. SChuLItz. Theodore W. 1961. 'Education and Ecoi01jC CirlOwth1. lit N. B. Henry. ed.. bSotiLd Foists in/t,7l ie;t i,' aAlll('iu an Edin ttit/Oll. Chlicao: Liii wsc sits of Cl icacgo Press . 1975. The Value ot tle Ability to Deal with Di.Sequii ibria.'' .10rnall of /l m,oii Literainre I 3(3):'27-46. REFERENCES 169 - 1)82. Iiic(7/ilL iii Peo/ple: The Econ(omi(i ol'P1p dl/lhion Qull/he. Berkeley: LJniversity of California Press. Schweinhart. L. J.. and J. J. Koshel. 1986. Polit, OnY n.r fbi Pl,es hoo Pl giaorr. Ypsilanti. Mich.: High/Scope Education Research Foundation. Selovsk% . Mareelo. 1979. WhIlo Benrfht li om Gove, n, lo Ea/pc,ditmi o,ns:' A Cas. SteSn(nl ofColomobia. New York: Oxford Uniiiversity Press. 1980. "Preschool Age Investment in HuLimani Capital." In John Simimionis. ed.. The Edw inmn i)ilemmatn. Oxl'ord: Pergarrion Press. 18. I 8. " Nutritioni. Health and Education: The Econiomiiic Si,nificance of Conipieinentarities at Early Ages.' In Pauil Sttreeten and Harrv Maiei. eds.. Con- (epns. Meal ,oeinctf, andi Lon,,o-Rim Pc-.spectiv e. Vol 2. of' H,looa Resorne 5(L. Emptph tent, iodl Dv eelopmn lent.: PI-,)( eedillN. st /ofte Si-t/l World Collg'l-e.is f the Internationail E onomlic A.s7o iation Helelh in 14es M .i Citv. 195/). New York: St. Martin's Press. Slhatffe, S.Sheldoni. 1993. The Iompat otf H/i .A4DS ol Echicaltio .SY.tefnl.N. Ediluctional Horimiros 71(4 :171-74. Shavelson, Richard J. 198 1. ('Conclusions nod Polie vI nlqupia(ctiona.. Vol. 7 of A S/oud of A l1ternalto.s in A.ome, itiotm Edi,catiott. Santa Monicat. Calif.: Rand. Siighi. Janlrdndll P. 1990. "Analysis of Project Costs in SLib-Szliahar Africa in SelecLted Sectors.' In EL onoic (t 1dl Sctomal Polit e /sLyes. World Bank Lonu-Terit Per- spectiVe Stuidv ynf Sul1-SaLa ran Africa Background Paper 2. Washlinlgtoni. D.C. Sivin-KachalIa. Jay. and Ellen R. Bialo. 1994.Reporton the Efttctivtwes.% oflTe(Imo)loLg in St hools, 19)0-(94. Interactive Educationial SYstems Design. New York. S ititii. Vil ma F.. and Richard L. Antdre% s. 1 989. IioslOL tiionn/il Leadei-.vhip. Hem PH in - Lipul.s Make ci Diffirellme. Alexandria. Va.: Association for SuApel-visioll aId CuLr- HicUtliUM [)eveloprn1enlt. Smvlie. M. A. 1994. "Redesi,nine, Teachers' Work: Coninectionis to the Classr-oonit." In L. Darliie H ammiliooid. ed.. Re view of Re s1eas II iii Etillc isioIl 21):1 9-78. Spagat. N'Michael. 1994. "HUitmanl Capital atid Long-Run Grotwtli in Russia." Brown lI llivelsitly Departtment of Economnics. Providenice. R.I. Steller. A rt hutr W. 1988. Et/i . SL hoc c/i ReL.i-LLL./h: Pi-ec /e, LiiL! P,l,inii Sc. Fas tbhack 276. Bloornimn gtnon. lid.: Kappa Educationial Foundlation0. Stevenson. David Lee. and Datvid 1P. Baker. 199'1. State Control of the CurriCulumli1 and Cl assrooMlt Instructioti.'' Sot iotl0vx of Edctwioit 64:1 -1i(. St-0111ItlUist, N. P. I1994. Getider- Mid Ed ucation.- Ictt/-ntLitiintl EllIeples)pLdit of Et/lLca- tioo. vol. 4. 2d ed. Nev York: Perpaninon Press. Sunimiiers. Anita A.. andt AmyV W. Johlison. 1994. "A Review of the Evidence oni the Effects of'School-Based Matnagement Plan.'- Paper prepared for the coniterenice oti Improving the Perforlimnice otf America's Schools: Economilic Choices. U.S. Na- tionial Research Counlcil. Washin-ton. D.C. Summl1ers. Lawrence H. 1992. "Investine in All the People.' Policy Researcil Working Paper Serics 905. World Bank. Office of the Vice Presidenit antd Chief Econwomiist. Wa'hlirgton, D.C. 170 REFERENCES . 1994. Investing in All The People: Educating Womeni in Developing Caountrlies. ED] Seminar Paper 45. Washington. D.C.: World Bank. Tan. Jee-Peng. 1991. "Thailand's Education Sector at a Crossroads: Selected Issues." In "Decision and Change in Thailand: Three Studies in Support of the Seventh Plan." World Bank, Asia Country Department 11. Washington. D.C. Tan, Jee-Peng, and Alain Mingat. 1992. Education in Asia: A Conmparative Studl ojf Cost and Financigw. A World Bank Regional and Sectoral Study. Washington, D.C. Thomas, Christopher. and Christopher Shaw. 1992. Issvues in the Developmentf oJlMul- ti,raode Schools. World Bank Technical Paper 172. Washington. D.C. Thompson. Ann, M. R. Sinionson1. and C. Hargrave. 1992. Edutatioaszl Technolol-k: A Re wiew of the Research l 1(6):71- 1. Tilak, Jandhyala B. G. 1 989. Educ ationi ond Its Relationi to Econonmic Grow th, Poverty. ant Inco m11oe Distribution: Past Evidence (ildl Further Analvsis. World Bank Dis- cussion Paper 46. Washington. D.C. 1993. "Financing Higher Educiation in India." Higher Educalion 244(t):43-67. . 1994. "On Pricing Higher Education." National Institute of Educational Plan- ning and Administration, New Delhi. Tilson. Thomas. 1991. "Sustainability in Four Interactive Radio Prolects: Bolivia. Hon- duras. Lesotho and Papua New Guiniea." In Miarlaine E. Lockheed. Johni Middletoll, and Greta S. Nettleton, eds.. "Education Technology: Sustainiable and Effective lJse." PHREE Background Paper 91/32. World Bank, Education and Social Policy Department. Washingtoni, D.C. Tsang. Mun. 1993. "Case Studies in Financing Quality Basic Education." Paper pre- pared for the second meetinig of the Intemzational Consultative Forum on Education for All. New Delhi. World Bank. Education and Social Policy Department, Wash- ingtoni. D.C. Tuijnnian. A. C., and N. Bottani. eds. 1994. Makingi Education Counft: Developing o,d UIsi,u- International Indliators. Paris: OECD. Centre for Educational Research and Innovation. Tuijniman. A. C. and T. N. Postlethwaite. eds. 1994. Monitoring the Standards af Edhncatio,i. Oxford: Pergamon Press. LINDP (United Nations Developmenit Programme)/tMPAcT. 1991. Disahilitv Prevenition: A Priority f)or the 9()s. United Nations Interiniational Centre, Vienna. UNESCO (Ulinited Nations Educational. Scientific and Cultural Organiizatiol). 1990. Con- pendiuon of 'Statistic.s on Illiteracv. Paris. 1 993a. Trends and Projections of Enrollment by Level of Educatiolln hy Age and bh SeA. 1960-202 7 Paris. 1 993b. World Education Report 1993. Paris. UNICEF (United Nationis Children's Fund). 1993. Th7e Progr,ess of Naoions. New York. UJnited Nations. 1986. "Education and Fertility: Selected Findines trom the World Fertility Survey Data." United Nations Population Division Working Paper ESA/P/ wp/96. New York. 1987. 'Education and Fertility." In Fertilily Belhaviour in the ConteXt foDevel- opmnent: Evidence from thie Wo,-Id1 Fertility Srove.y. New York. REFERENCES 171 Velez, Eduardo. Ernesto Schiefelbein, and Jorge Valenzuela. 1993. "Factors Affecting Achievement in Primary Education: A Review of the Literature for Latin America and the Caribbean." Human Resources Developmenit and Operations Policy Work- ing Paper 2. World Bank, Washington, D.C. Vlasceanu, Lazar. 1993. "Trends. Developments and Needs of Higher Education Sys- temns of the Central and Eastern European Countries." European Centre for Higher Education, Buclharest. Walberg. Herbert J. 1991. "Synthesis of Research on Teaching." In Merlin C. Wittrock. ed.. Tlhird Hnldbook of Researh c oni Teaching. New York: Macmi llain. Warwick, Donald. and Fernando Reimers. 1992. "Teacher Training in Pakistan: Value Added or Money Wasted'?" Paper read at the BRIDGES/1EES Confer-ence on School- ing Effectiveness: Cross National Findings. Harvar-d Institute for Intemiational Developmenit. Cambridge. Mass. Weale, Martini. 1993. "A Critical Evaluation of Rate of Return Analysis.'' Economnic ./{JIoorn(Ic 103 (418):729-37. Westoff, Charles F. 1992. Acf afit Mairriuaeue. Age cat First Bir-thi. and Fe,r'tilif'v in Afi ica. World Bank Techinical Paper 169. Washington, D.C. World Bank. 1980. Edhucation. Sector Policy Paper. Washinigton, D.C. . I _1986. Financincg Eduication in DevelopingJ Couo/1t1 ies: An1 Exp/loration of Polic ! Optionis. Washington. D.C. . 1988. Ecatticion in Suh1-Sahlaaelo Aftica: Policies fin Adjusten'tI. Revitalic-a- tio00 andc7 E.vpaosioe. Washiniigtoni. D.C. -____ 1990a. Primarv Educatioo. A World Banik Policy Paper. Washington. D.C. 1 1990b. Worlel Developmec'nc' RepOrt: Povertlt. New York: Oxford lIniversity Press. 1991 a. "Republic of Ghana: Community Secondary Schools Construction Project." Staff Appraisatl Report 9556-GH. Africa Region. CouLItry Department IV. Washington, D.C. . 1991b. "Romania: Accelerating the Transition: Hunian Resources Strateoy ftor the 1990s." Report 9577-Ro. Europe and Centr-al Asia Region. Country Depart- menit 1. Washington, D.C. . 1991c. V'ocatioacil cocc Technical Echicc,tion andr/ Trainhin,. A World Bank Policy Paper. Washington. D.C. ________. 1991 d. WoricIDec'clojmcecctReport: Tue C/calleccgc' of D'lccpocecct New York: Oxford Uni versitv Press. . 1992. "Poland: Reorienting Investmenits in Human Capital: A Critical Review of Secondarv Education and Training Systems." Report 10697-POL. Europe andl Cential Asia Region. Country Department 11, Washington, D.C. 1 I 993a. TIhe Ecast A.siatn Miracle: Economnic Grout/c cnmd Public Pclicv. A Worldc Bank Policy Research Report. New York: Oxford University Press. . 1 993b. "Female Secondary School Assistance Project: Bangladesh." Staff Ap- praisal Report II 386-BD. South Asia Region. Country Department 1. Washington. D.C. ______. 1 993c. "IIndoniesia: Public Expenditures, Prices and the Poor." lIndonesia Resi- dent Mission Report 11293-ID. East Asia and Pacific Region. Country Departmelit 111. Washington. D.C. 172 REFERENCES 1 1 993d. "Seconclary Education Project for Colombia." Report I 1834-co. Latin America andic the Caribbean Region. Countrv Department III. Washington. D.C. 1 1993e. "Venezuela 2000: Education tor Growvth and Social EqIuity." Report I I 130-VE. latin America aid the Caribbean Region. Couitiv Department I. Wash- ineton. D.C. 1 1993f. World Dc'ielolmic' Rel)rt: Inestlint' ill Health. New York: Oxtord University Press. 1 1994a. 'Colonibia Poverty Assessmenit Report." Draft. Latin America aLid the Caribbean Region. Country- Departnient 111 Washington. D.C. 1 1 994b. Enihalc in/01 Womien 's Paurti( i/ia!ion in E(oinooic Dcv e/cl//icita. Wash - iliCtoll. D.C'. ._19)94c. Enicshinmg Livs s: O c/iomi0n10 Iitaoioi and"l Milnel Ma/ilni titioil in DBev'c1o/itii,' Cou,itrie'.%. Washi ngton. D.C. 1 1994d. "Guateniala- Basic Education Strategy: Equity anid Efficiencv in Edu- cation.' Report 13304-GLJ. Latin America and the Caribbean Region. Country Department 11. Washington, D.C. Available fiom the World Bank Piiblic Inforimia- tion Center as pwt573. 1 1 994e. HIi/ichr Edimatian: The Lessoms of Eaerieince. Washinitoln. D.C. ._ i994f. "Kenya Povertv Assessment." Africa Region. Country Department il Washingtoon D.C. . 1994g. "Kenya Public Expenditule Review." Draft. Africa Region, Country Departimielit It. Washington. D.C. . 1994h. The World Bank. Leariiin, fi-omi the Past. Embi/-acing the lFtn/r. Washinigtoin. D.C. 1 i994i. "'Mexico: Seconlcd Primary Educatlion Project.' Report 17-529-ME. Latill America and the Caribbean Region. Countiv Departimient 11. Washington. D.C. Available f'romil the World Bank Public Information Center as PIc 153. * 1994i. Pro .siofi for Children wit/i Spec ial Ed/iiatiilill N eed/. /// til' Asia Re-ionI: So//lI' DCv c'o/i//iltm s itI CLI//uC/t Pr/icticc. World Bank Techlinical Paper 261. Asia Technical Departmiient. Washington. D.C. 1 i994k. 'Romania: Reforim (lf HiTher Education and Research." Eurlope and Central Asia Region. Country Depiaitinieit 1. Washingtonl. D.C. . 19941. 'Russia: Education in the Transition." Eulope aid Cential Asia Region. CouLi1try' DepaLtilenit 111. Washington. D.C. 19 I 94nm. "LUanlda Social Sector Strategv.' Report 10765-ucc. Africa Region. Co iiitry Department 11. Washington. D.C. .______ I 994n . U'on/l1lDc ic'c///Ol/ll tR kc'f)(l/t /ira/tllln tl/'fl/ L)c'/)( c/cl/li/11(1t. Newv Yorki: Oxford Universitv Press. 1 1995. W'orld DI'l I/op/loe/it Rejport 1995: Workerl i/i aii i/ItVlclilt" / olll. New York: Oxford ULniversity Press. V Various years. Worl /d Ba/iLk Aninial Rcp/7ct. Washington. D.C. WuL. Kin Bing. 1993. Moinlolia: Fina/ici/ng Ed/ucation dI/oiing Ec onoIi/I T/e/isition/. World Baiik DiSCuSSioll Paper 226. Washington. D.C. Wyvine, E. A. 198(). Loaklin, at .School.s. Good. Bci and1( l//ilhie,r/it. Lexington. Mass.: Heath. REFERENCES 173 Younig, Mary Eming. 1994. "integrated Early Child Development: Challeng-es ancl Oppontutnities." Humain Resources Developmilenit and Operations Policv Worhkings Paper 40. World Bank. Washineton. D.C. Zidernilazi. Adrian. and Douglas Albrecht. 1995. Fioa,u in", 1l"niversitie.s ill Dcreloping Conitries. London: Falhier Press. Other Recent Development in Practice Books Toward Gender Equalitq: The Role of Publhic Policy Better Ulb17anii Services: Finding the Rigiht Incentives (also available in French and Spanish) Strengthening the Effctciveness of Aidl: LessonsfJor Donlors Enriching Lives: Overcominig V itamini and Minerail Milnutrition in Developilng Countries (also available in Frenich and Spanish) A Nevw A-enda for II'omen's Health (iaid Nuitritioni (also available in Frenci) Population eiiid DeI)elo0n71et: ImpIliCations for the W0old Bank East Asia's Trade and Investment: Regional and Globlal Gailis ftoio Libercili:ation Governfance: The WorIld Bank's Experience Hioiher Education: The Lessons of Exerience (also available in French and Spanish) Better Health in Africi: Experience and Lessons Learnied (also available in Frenichi) A igefina sPrivadaion Protioram: Eperieuce. Issues, and Les^.sons Susltaining Raipid Development in Eaist Asiai aind the Pacific ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I ID -,. l1 1| -*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ rfi ; _l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I g1~~~~~~~~~~~~~~~~~~~ 7882 5= 3 3 3 1 1 2r;=