1 EDUCATION PUBLIC EXPENDITURE REVIEW GUIDELINES Education Global Practice 2 Education Public Expenditure Review Guidelines © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e- mail: pubrights@worldbank.org. i TABLE OF CONTENTS ii ACRONYMS iv PREFACE v USER’S GUIDE 1 PART I: CHECKLIST FOR EDUCATION PER STEPS 2 Introduction 3 Public Expenditure Review Steps 16 PART II: CHECKLIST FOR AN EDUCATION PER ANALYSIS 18 Section 1: Background and Overview of the Education System 20 Section 2: Overview of Education Financing and Spending 25 Section 3: Financial Accountability 30 Section 4: Adequacy and Sustainability 34 Section 5: Efficiency and Effectiveness 38 Section 6: Equity 41 TECHNICAL NOTES 85 PART III: EXAMPLES 206 NOTES 210 REFERENCES ii Education Public Expenditure Review Guidelines ACRONYMS ACS Average Class Size LGA Local Government Authority AY Academic Year LGU Local Government Units BIA Benefit Incidence Analysis LLECE Laboratorio Latinoamericano CBA Cost Benefit Analysis de Evaluación de Calidad de la Educación, or Latin American CCT Conditional Cash Transfer Laboratory for Evaluating the CEA Cost-Effectiveness Analysis Quality of Education CGE Computable General LSMS Living Standards Equilibrium Measurement Study COFOG Classification of Functions of MAMS Maquette for MDG Government Simulations CSOs Civil Society Organizations MFM Macroeconomics and Fiscal DD Difference in Differences Management DEA Data Envelopment Analysis MICS Multiple Indicator Cluster Survey DHS Demographic and Health Survey MOF Ministry of Finance ECD Early Childhood Development MTEF Medium-term Expenditure Framework EFA Education for All NEA National Education Accounts EMIS Education Management Information Systems NER Net Enrollment Rate EPSSim Education Policy and Strategy NGOs Non-governmental Simulation Model Organizations FTE Full-time Equivalent NSOs National Statistical Offices GDP Gross Domestic Product OECD Organisation for Economic Co- operation and Development GEM-Education The General Equilibrium Model for Education PASEC Programme d’Analyse des Systèmes Educatifs de la GER Gross Enrollment Rate CONFEMEN (the CONFEMEN GNI Gross National Income Programme for the Analysis of GNP Gross National Product Educational Systems) GP Global Practice PCF Per Capita Financing HNPGP Health, Nutrition, and PEFA Public Expenditure and Population Global Practice Financial Accountability IGR Internally Generated Revenue PER Public Expenditure Review INESM Inter-agency Network on PETS Public Expenditure Tracking Education Simulation Models Survey ISCED International Standard PFM Public Financial Management Classification of Education PIAAC Programme for the IT Information Technology International Assessment of Adult Competencies iii PIRLS Progress in International TVET Technical and Vocational Reading Literacy Study Education and Training PISA Program for International UIS UNESCO Institute of Statistics Student Assessment UNESCO United Nations Educational, PPP Purchasing Power Parity Scientific and Cultural PTA Parent-Teacher Association Organization PTCA Parent-Teacher Community WB-MAMS World Bank’s Maquette for Association MDG Simulations PTR Pupil-Teacher Ratio WDI World Development Indicators QSDS Quantitative Service Delivery Surveys VFM Value for Money RDD Regression Discontinuity Design SA Social Assistance SABER Systems Approach for Better Education Results SACMEQ Southern Africa Consortium for Monitoring of Education Quality SBM School-based Management SDGs Sustainable Development Goals SDI Service Delivery Indicators SERCE Second Regional Comparative and Explanatory Study SES Socioeconomic Status SNA System of National Accounts STEP The STEP Skills Measurement Program STR Student-Teacher Ratio SWAP Sector-wide Approach Program TERCE Third Regional Comparative and Explanatory Study TIMSS Trends in International Mathematics and Science Study iv Education Public Expenditure Review Guidelines PREFACE Public expenditure reviews are one of the World Bank’s core diagnostic tools for informing various stakeholders about the state of education financing in a country. Such reviews assess the efficiency, effectiveness, and equity of expenditures on education and their adequacy and sustainability relative to the country’s educational goals. They review not only public spending, but also private and donor spending. Guidelines for public expenditure reviews in education establish content and quality standards for such reviews in the sector, using technical notes and examples to deepen the user’s understanding of how to meet these standards. The World Bank prepared the last public expenditure review guidance for education in 2004 and revised it in 2009 as part of the Human Development sector-wide initiative. The document included a cross-sectoral Core Guidance and a sector-specific Guidance for each sector. This Guidance covers both cross-sectoral and education sector-specific issues in one volume. The new Guidelines update the contents of the earlier guidelines to reflect recent developments in education financing and respond to demands for more hands-on advice, as follows: (i) they adopt the World Bank’s new initiative—Systems Approach for Better Education Results (SABER) School Finance Framework—as the basis for the analytical dimensions; (ii) they use a decision-tree approach, providing step-by-step guidance for conducting public expenditure reviews under different context, scopes, and types of analytical tools; (iii) they analyze policy recommendations in completed public expenditure reviews, assessing why each is good or weak; (iv) they introduce a new analytical tool, BOOST, which converts government budget data into a useful format, makes analysis relatively easy, and contributes to the development of the comprehensive education-finance data accounts called National Education Accounts; (v) they enhance discussions of governance and public financial management issues that partly determine effective spending; (vi) they strengthen technical notes by including more details on data sources and definitions of concepts and variables; and (vii) they update good analytical examples from completed public expenditure reviews, categorized by topic. Sachiko Kataoka led a team under the guidance of Luis Benveniste (Director, Education Global Practice) to develop the new guidelines. The team included Sue Berryman, Yulia Makarova, and Aliya Bigarinova, and also relied on a number of colleagues who provided insightful comments throughout the process. Specifically, the team is grateful to: Dina Abu-Ghaida, Samer Al-Samarrai, Mohammed Audah, William Dorotinsky, Kebede Feda, Katia Marina Herrera Sosa, Margo Hoftijzer, Jennifer Klein, Shinsaku Nomura, Anna Olefir, Shawn Powers, Holy Tiana Rame, Furqan Saleem, Lars Sondergaard, and Ryoko Tomita Wilcox. The team also would like to thank Husein Abdul-Hamid, Melissa Adelman, Jung- Hwan Choi, Jennifer Klein, Laura Gregory, and Kirsten Majgaard for helping to compile recent public expenditure reviews to develop a database. v USER’S GUIDE This guidance note applies to all countries, regardless of their circumstances. It seeks to remind the analyst of the main features that are normally included in an education public expenditure review (PER). However, analysts should not use it as a checklist to which they must rigidly adhere. Every review must be selective in what it covers, based on such factors as what is needed for the country dialogue; what is already known and available elsewhere; and what is able to be accomplished given constraints of time, data, and funding. Analysts may omit topics, but with a justification in mind. In addition to agreeing in the concept note on the planned topic coverage, it is useful to convey to the reader of the full report the reasons for omitting any major themes. Similarly, the depth of treatment of any included topic will need to be considered, agreed upon, and explained. Conditions such as fragility, conflict, or violence in the country under review could affect whether a PER can be conducted at all. Such factors can also affect the quality of the data needed to conduct a review, as well as the ability of the country to implement any recommendations that the review suggests. CHECKLISTS Part I of the Guidelines provides a checklist for steps, from preparation to dissemination, of an education PER, including cross-sectoral questions, such as the overall budget and potential tradeoffs between social and other sectors. The PER team will first go through a decision tree to determine the scope, objectives, analytical tools, and policy questions for a particular review. The next steps will probably vary for different circumstances. Part II provides a checklist for an analysis to be completed as part of an education PER, organized by six key questions:1 1. Who finances education and how are funds channeled? 2. How much does the government spend and on what? 3. Is the public financial management system set up to enhance financial accountability? 4. Relative to the government’s policies and standards, how much is needed now (adequacy), and what can be afforded in the medium and long term (sustainability)? 5. Are public resources being used efficiently and effectively? 6. Does public spending promote equity? vi Education Public Expenditure Review Guidelines To keep the checklists as skeletal as possible, three different types of links are used to let review teams explore a topic in more depth. Two types of links are internal to this document. They take the analyst to in-depth technical notes and examples of good treatments of a PER topic, with red for Technical Notes and blue for Examples. The third type of link, in purple, signals a hyperlink to an external resource. Users of these Guidelines may also consult Solutions Notes, which provide concrete solutions for specific education financing issues, to help them frame good policy recommendations. These Notes will be made available at http://www.worldbank.org/en/topic/ education. Examples and Solutions Notes will be regularly updated. Technical Note: A Technical Note elaborates on a concept or provides detailed guidance on analytic methods, indicators, and sources relevant to the topic. Example: An example shows an especially good treatment of a topic. It often includes the description of analytic methods used and other guidance. Examples are taken from recent PERs and will be updated as more good case studies become available. If the analyst is working from an electronic version of these Guidelines, clicking on a color-coded oval that appears next to a topic in the checklist will take the reader to the material connected to that link. For hard-copy versions of the Guidelines, all Technical Notes germane to Part I and Part II appear at the end of Part II in numerical sequence. Part III consists of the Examples. PART I: CHECKLIST FOR EDUCATION PER STEPS 2 Education Public Expenditure Review Guidelines Introduction Why is education finance important? All education systems rely on financing to function. Education finance systems pay for the inputs required to implement education policies, such as teachers, school buildings, and learning materials. Availability of financial resources does not guarantee a quality education, but a quality education is impossible to achieve without adequate resources. Some uses of education expenditures can make a marked difference in learning, particularly in the cases of inputs that directly benefit students or resources that compensate for challenges arising from low-income settings. The same money can be wasted if it is allocated to input factors that only marginally affect learning or if policymakers fail to consider the conditions that must be met for factors to translate into learning gains. Governments are under increasing pressure to use education resources efficiently, but often lack guidance on the optimal ways to invest and manage their school finance systems. Research findings have shown that learning outcomes are not strongly related to spending levels (except in cases where education budget is very small), suggesting that the way money is spent—and not simply how much is spent—matters in education finance. Meeting the World Bank’s twin goals of poverty reduction and shared prosperity in the education sector implies the need to use country and donor resources effectively, efficiently, and equitably. A sound PER assesses how resources are used relative to these goals. In Part II, Sections 5 (Effectiveness and Efficiency) and 6 (Equity) are especially germane to these analytic standards. 3 Public Expenditure Review Steps This section provides guidance on steps involved in conducting a PER, from the preparation to dissemination of results, as summarized in Figure 1. It highlights issues common to such reviews, regardless of the sector, and addresses shared considerations and such cross-sectoral questions as the overall budget and potential tradeoffs between social and other sectors.2 Figure 1: Public expenditure review steps 1 STEP 1: Understanding the context and motivation Understand the country context and motivation of the PER. See Figure 2 for seven possible scenarios which may direct the PER team to a particular path. 2 STEP 2: Defining the scope and objectives Review key policy documents such as the education strategy, sector analysis, and latest PER recommendations to define the possible scope, objectives, and policy questions for the review. Agree on them with the client. 3 STEP 3: Securing access to data and information Identify available financing and non-financing datasets as well as policy and legal documents. Depending on data availability, the scope may need to be modified. 4 STEP 4: Analyzing data and information Assess the country’s education finance system, financial accountability, adequacy and sustainability, efficiency and effectiveness, and equity of education financing. PART II will discuss the analysis in depth. 5 STEP 5: Validating key findings and policy recommendations Discuss and validate key findings and policy recommendations with the client, possibly holding separate sessions for technical Ministry of Finance (MOF) and Ministry of Education (MOE) staff and for MOF and MOE policymakers. Disseminate the final report to various stakeholders. 4 Education Public Expenditure Review Guidelines STEP 1: Understanding the context and motivation The very first step is to understand the country context and motivation of the public expenditure review. Figure 2 suggests seven possible scenarios under which the review could be conducted. The scope, objectives, and key questions may vary widely, depending on the scenario. Figure 2: Seven scenarios for country context and motivation What are the country context and motivation? 1. (Re)new 7. Capacity engagement building 2. Emergency 6. Governance support challenges 3. Project 5. Fiscal preparation sustainability 4. Efficiency gains Scenario (1): (Re)new engagement. We have not been engaged in the education sector of the country for a long time or have never been engaged, and need to (re)gain our knowledge about its education financing to renew or initiate our engagement. • When was the last education public expenditure review conducted for the country? • If a review has not been conducted in the last five years, does other sector analytical work exist that can inform you about key issues to be addressed? • Can those available documents give you a good understanding of key challenges to be addressed? ӺӺ These questions affect the decisions on scope and objectives. To explore key issues, focus on all six sections of the guidelines. Scenario (2): Emergency support. The country has recently experienced a crisis (natural, political, or economic), and the Bank needs to assess quickly what kind of financial support the government needs. • What are the major funding items that need urgent support? Possibilities might include repair of damaged schools to restore access, payment of teacher salaries to avoid teacher absenteeism, and financing of cash transfers or school meals to avoid dropouts. • Can the Bank, in collaboration with other development partners, establish appropriate funding mechanisms? ӺӺ Under a crisis situation, the Reference Guide on External Education Financing, published by Inter-Agency Network for Education in Emergencies (INEE), may be more suitable than these Guidelines. 5 Scenario (3): Project preparation. We want to conduct a public expenditure review to inform the preparation of a new education project. • Has the subsector for the new project already been decided? ӺӺ If yes, the review should focus on the chosen subsector, with a thorough examination of six key questions for the subsector. ӺӺ If no, the review should analyze existing key documents and overall education spending, examine subsectoral allocations, and identify education finance policy issues to be addressed, by going through all six sections. Scenario (4): Efficiency gains. The government is facing budgetary constraints and has asked the Bank to analyze public expenditures on education to find fiscal space.3 • Does the government (Ministry of Finance) have a budget-reduction target for the education sector? How imminent is such a cut? ӺӺ In the short term, there is probably little that the Ministry of Education can do besides cutting capital and non-salary, recurrent spending because any reductions in personnel costs usually must be phased in. ӺӺ In the medium term, the public expenditure review should explore potential savings across the entire sector, including rationalizing human resources. At the same time, the review may play a role in defending the education budget if an analysis finds little room for efficiency gains without harming access and quality. Focus on Section 5: Efficiency and Effectiveness. Scenario (5): Fiscal sustainability. The government plans costly reforms, such as an extension of general education, and needs to know the capital and recurrent cost implications of the plans, as well as their fiscal sustainability and equity implications. • A simulation analysis would be essential to inform the government about the feasibility of such a plan ӺӺ Focus on Section 4: Adequacy and Sustainability and Section 6: Equity. Scenario (6): Governance challenges. The government is concerned that public funding is not reaching schools and wants to identify possible issues with public-finance management. • What are major spending problems? ӺӺ It is likely that certain public financial management (PFM) weaknesses are underlying causes of the spending problems. Focus on Section 3: Fiscal Accountability. ӺӺ Also consider other analytical instruments such as Public Expenditure Tracking Surveys (PETS) and Quantitative Service Delivery Surveys (QSDS), provided that there is sufficient budget to conduct them. Go to Technical Note 12: Public Expenditure Tracking Surveys (PETS), Quantitative Service Delivery Surveys (QSDS), and Service Delivery Indicators (SDI). Scenario (7): Capacity building. The government would like to strengthen its analytical capacity and wants the Bank to help it conduct a PER. • In this case, the main objective is to support the government. The client’s engagement and ownership are as important as the final output. This will require more time and budget for close communication with the client. 6 Education Public Expenditure Review Guidelines STEP 2: Defining the scope and objectives There are many factors to consider when defining the scope and objectives of an education PER. Such a review is primarily concerned with public revenues and expenditures as expressions of public policy and public involvement in the economy. A public expenditure review also examines private spending. Rarely can public resources meet all financial needs, and private spending may play a major role in providing educational services. Many kinds of analyses can be conducted in a public expenditure review, but obviously, such analyses must be aligned with the project’s budget, time frame, and data availability, as well as its overall focus and goals. The team needs to: • Review key relevant documents such as the education strategy, sector analyses, and recommendations of the last public expenditure review, to identify education finance policy issues to be addressed. • Decide the subsector focus. The review can have broad coverage (all subsectors, possibly even including adult education) or narrow coverage (a single subsector). State your focus clearly, why you decided on it, what has been excluded, why certain issues or subsectors have been excluded, and whether or how these exclusions might affect the due diligence purpose of the review. Broad, unfocused public expenditure reviews will be less effective than sharply focused ones. • Place education spending within a macro context. Confer with the Bank’s country economist— and, as needed, with the Ministry of Finance (MOF)—about the country’s macroeconomic situation and its implications for education budgets downstream. • Make sure that your education public expenditure review—if part of a multisectoral (comprehensive) PER—fits the purposes of the comprehensive PER, but still respects your main objectives. In the multisectoral case, where several sector-specific reviews feed into a comprehensive PER, the concept note for the overall review should clarify the objectives of the overall PER and the expectations for each of the sector-specific reviews. Although the expectations for the sector-specific reviews are often poorly defined, be sure to define the focus of the education PER and then clear it with the team leader for the comprehensive exercise. • Work to ensure that a comprehensive public expenditure review is more than the sum of its parts. Typically, each sector does its analysis in isolation, and the team leader then combines all sector- specific reviews into one document, including three to five recommendations per sector. By itself, the education review team cannot create incentives for members of comprehensive PER teams to realize the potential synergies within the context of a comprehensive public expenditure review. However, the leader of the education review can at least ask the leaders of the comprehensive review and of the other sector-specific reviews how the overall team might benefit from this specific, comprehensive public expenditure review. 7 • Check that the time and budget4 allocated to you for the education review are in line with your intended scope and its data requirements. Ultimately, the budget and time frame available to prepare and undertake a PER will be decisive in specifying its scope and in setting priorities. Be aware that resources may be needed for the following: ▷ Missions to assemble data and information (travel costs and staff days). ▷ Extensive data recoding and cleaning or new data collection. The early childhood development (ECD), technical and vocational education and training (TVET) and higher education subsectors often lack data required to conduct a review. If the scope of the review includes any of these subsectors, the team may need to collect new data. ▷ Days required to analyze data, write the draft public expenditure review, and revise it after peer reviews. ▷ Missions to discuss results and disseminate them more broadly, such as running a workshop for stakeholders (travel costs and limited staff days). ▷ Follow-up policy discussions. • Once you predefine the scope, discuss it with the key counterparts (typically, the Ministries of Education and Finance) and agree on the scope, objectives, and policy questions for the review during the first mission. Based on the discussion with counterparts, you may need to modify the scope or objectives. This process will help enhance the client’s ownership of the report and bolster the chances that findings and policy recommendations will be accepted and implemented. The following meetings may prove helpful: ▷ Connect with high-level policymakers (Ministers of Education and of Finance, their Deputy Ministers, or equivalent) to explain the objectives of this work and potential benefits to them, and to obtain their approval for this work. Note that since the Ministry of Finance is the primary client for a PER, the Ministry of Education may know little or nothing about this work. Thus, it is essential to obtain education policymakers’ approval before asking their staff to compile and provide data for the project. ▷ After you get the decision maker’s go-ahead, set up meetings with (i) the Education Management Information System (EMIS) team, finance or accounting team, assessments team, and any other unit that has data that you need; and (ii) heads of departments over seeing the levels or topics of education that you plan to analyze to ask them what they see as the main challenges in their respective areas (e.g., early childhood development, secondary education, or teacher policy). ▷ Set up meetings and reach out to key players outside the Ministry to get a broader perspective. Try to visit at least a few schools (ideally, ones different enough to give you a broad picture of the system, e.g., urban and rural, primary and secondary, public and private, wealthy and poor). These school visits can be combined with conversations with local officials who oversee schools at the subnational level, as well as with principals, teachers, and even parents. If you think teacher unions or civil society organizations (CSOs) might offer a useful perspective, ask to meet with them as well. 8 Education Public Expenditure Review Guidelines STEP 3: Securing access to data and information The nature and number of subsectors to be analyzed within the scope of the review helps to determine the datasets and documents that the PER team will need; data availability, in turn, affects decisions on the final scope of the project. Once there is an overall agreement on the scope, the team needs to secure access to reliable data. Good-quality data are accurate, timely, and relevant for decision-making.5 Be clear ahead of time about data sources, their availability, and their consistency, because they can constrain the scope of the report and its quality. When the education review is part of a comprehensive public expenditure review, discuss with the overall team leader how to maximize efficiencies, such as hiring a shared consultant or research assistant for data collection and analysis. Also determine if the team conducting the macro analysis for a comprehensive review will create a government-budget database that the education review team can use (e.g., BOOST). A public expenditure and financial accountability (PEFA) country report might be available, which could prove useful for an assessment of the Public Financial Management (PFM) system relevant to education. In addition, examine the websites of the country’s Ministries of Finance and Education and national statistics agency. These websites often post significant amounts of data, or, at the least, alert you to the existence of databases to which you can request access. Public expenditure reviews use various types of data. Financial data usually come from the Ministry of Finance and sometimes from the Ministry of Education’s budget department. If financing is decentralized for any subsectors within the scope of the review, financing data may be available by subnational unit at the national-level Ministry of Finance. In other cases, you may need to sample subnational units to assemble the needed data. If your scope includes autonomous institutions, such as universities or technical and vocational schools, you will probably have to obtain financing and expenditure data from the institutions themselves. Private and household contributions may be recorded separately, or estimated based on household survey data. Non-financial data, such as the numbers of students, teachers, schools, or classes, usually come from the education management information systems (EMIS) or equivalent, managed by the Ministry of Education. Detailed student assessment data are useful for measuring spending against performance. Whether for financial or non-financial data, you should get both aggregate data at the national level and data as disaggregated as possible at the school or provider level, or at least at the locality or municipality level, for the level(s) of education on which you plan to focus your analysis. In sum, public expenditure reviews use the following six types of data (see Figure 3). The Technical Notes for these six types of data describe common sources for each type and assess the usual quality of each: 9 Figure 3: Six types of data that can be used for PER analysis Government’s official Cross-national data documents (policies, laws, (international learning and regulations) assessment, cross-national statistical databases) Country system-level data Research reports PER (budget and sectoral data) (donor reports, academic analysis research papers) National surveys (census, New data collection household surveys) ӹӹ Government’s official documents, such as policies, laws, and regulations governing inputs to, and financial arrangements for, the learning process (see Technical Note 1). ӹ Country system-level data (e.g., BOOST, EMIS, national assessments; see Technical Note 2). ӹ National surveys such as census and household surveys (see Technical Note 3). ӹ Cross-national data such as international learning assessment and cross-national statistical databases (see Technical Note 4). ӹӹ Research reports (see Technical Note 5). ӹӹ New data collection (see Technical Note 6). The team also needs data for comparisons. Many indicators, especially various measures of spending on education, do not carry an intrinsic absolute value. We do not know, for example, what “adequate” levels of spending on education would be without situating the PER country within the range of values associated with other countries. Hence, it is essential that a public expenditure review also analyzes the following: • Statistical comparisons with other countries in the same region or at similar income levels, and with other countries that serve as aspiration goals for the country, such as those in the Organisation or Economic Co-operation and Development (OECD). Comparisons can be affected by demography (cohort sizes for school-age individuals), participation rates by subsector, unit costs by subsector, and cross-country differences in the structures of education systems.6 Obviously, financing requirements for small school-age cohorts or cohorts with low enrollment rates are less than those for large cohorts or cohorts with high enrollment rates (especially in the more costly subsectors). These comparisons are descriptive by nature and provide neither a diagnosis nor an explanation. Nevertheless, comparisons can show that a country performs significantly worse (or better) than “expected.” Such comparisons can help set some “reasonable” quantitative targets for policymakers and even provide an incentive to reach them. Technical Note 4: Cross-national data • Trend data that allow comparisons over time. These are essential for gauging how quickly the country may be reaching its target. If improvements are slow in coming, trend data can provoke a deeper exploration into the reasons for the lagging performance. 10 Education Public Expenditure Review Guidelines STEP 4: Analyzing data and information These Guidelines provide guidance on how to conduct a public expenditure review to answer the following six key questions: Key Questions 1. Who finances education and how are funds channeled? 2. How much does the government spend and on what? 3. Is the public financial management system set up to enhance financial accountability? 4. Relative to the government’s policies and standards, how much is needed now (adequacy), and what can be afforded in the medium and long term (sustainability)? 5. Are public resources being used efficiently and effectively? 6. Does public spending promote equity? The key concepts of each dimension of education finance are briefly described below and discussed in more detail in Part II. ▶▶ Questions 1–2: The financing system is the foundation for a country’s public investment in education. These questions ask how the education budget is financed, who spends it, and how it is spent. ▶▶ Question 3: Financial accountability in terms of education finance sheds lights on public financial management (PFM) systems, and assesses whether such systems are set up in such a way that government policies get implemented as intended and achieve their objectives. ▶▶ Question 4: Adequacy of education finance responds to the question of “How much is enough now?”, given the government’s policies and standards. Sustainability of education finance assesses whether planned spending levels on education are affordable in the longer term, given the country’s macroeconomic prospects, sector policies, and demography. ▶▶ Question 5: Efficiency and effectiveness in education finance require that systems invest in those inputs with the largest marginal returns, as measured through outputs and outcomes relative to costs, given a country’s particular stage of development. ▶▶ Question 6: Equity in educational opportunity is a key goal of public education systems. It measures the availability of a quality education to all students, regardless of background. 11 STEP 5: Validating key findings and policy recommendations You want the public expenditure review to impact the client’s policies. Impact usually increases if the review team respects certain principles. • Make evidence-based policy recommendations with an explicit link between the analysis and findings and recommendations. Avoid: ▷ Findings and recommendations that do not follow from the analysis ▷ Analyses that do not lead to any findings or recommendations • Prioritize your recommendations. Policymakers will not act on a laundry list of recommendations. Identify a limited number of key issues—three to five per sector—where getting some traction is most critical. “Nesting” your recommendations in hierarchies can be helpful. The main recommendation (e.g., improve equity of access) can be aimed at policymakers, with the technical specifics aimed at technocrats (e.g., change the balance between programs A and B; improve program C by taking actions 1, 2, and 3). Phasing—short-, medium- and long-term—is another way to “chunk” recommendations into digestible form. • Make the conclusions and recommendations appropriately specific, not bland and general. Place recommendations in a feasible social, political, and administrative context. The World Bank’s public expenditure reviews have a tendency to preach the good and the moral without an appreciation of the realistic and feasible. Can each recommendation be made operational and actionable for the government? If possible, estimate the fiscal costs of any major reforms proposed. What political costs and implementation barriers are involved in each? • Keep report recommendations short and accessible. It may be difficult to do this when laying out the analytical grounds for recommendations in a report that covers multiple sectors. One potential way to tackle this issue is to include a main report and a second volume with the technical background papers. • Discuss the draft of the main findings and policy recommendations with the client to validate the findings and assess the feasibility of policy recommendations. While the review should present critical observations supported by analytical evidence, it also needs to be acceptable to the government to boost the likelihood of policymakers implementing the recommendations. • Disseminate the final report to various stakeholders in the education community of the country, ideally through an in-country workshop. It is not easy to meet these criteria for good policy recommendations. Good examples of policy recommendations that meet most of the criteria (being tied to findings, prioritized, concrete, realistic, feasible, and sometimes costed) can be found in Example 1: Policy recommendations. Tables 1 and 2 show brief examples and assessments of recommendations from the universe of completed PERs, with Table 1 focused on relatively useful ones and Table 2 on less useful ones. 12 Education Public Expenditure Review Guidelines Table 1: Are these recommendations useful? Probably. Recommendation Why is this useful? With respect to developing an equalization The recommendation clearly states the problem mechanism, creative solutions are needed to and describes three concrete options for solving ensure adequate financing for schools with low it. These options vary in their political economy capacity to raise private funds. Although public and budget implications, giving the Government funds are distributed in a generally progressive some flexibility in trading off between fiscal and manner, private fees raised by schools flow political costs. overwhelmingly toward the more well-resourced institutions. Given the overwhelming share of private financing for primary and secondary education, an equalization mechanism is clearly needed to counteract the regressive nature of current funding flows. A wealth of international evidence exists on such mechanisms. One option could be to examine the current BSP program and evaluate whether it can be repurposed to allow a small share of school-raised funds to flow from more affluent schools to less affluent ones (within a district or across neighboring districts). A second option could be to revive the defunct Equalization Grant that was previously used to provide additional resources to underfunded schools. A third option could be to increase the relative funding coefficients for P3 and S3 schools under the Per Capita Grant. One approach currently being considered by the MoE is to introduce a new Building Levy, which would collect a small amount from each SDC budget into a central pool of funds, which would then be used to equalize resources among schools without increasing the fees charged to households. Regardless of the sources of funds—whether public or private—the MoE needs to ensure that all of the country’s schools have at least a basic minimum level of resources available to them to provide a quality education. 13 Recommendation Why is this useful? Continue effort to right-size the school network This review was done in response to a and teaching force. A simple estimation suggests request by the country’s Ministry of Finance to that if the sector could attain student-teacher the Bank for sequenced and targeted advice, ratios comparable to the OECD countries over during a period of austerity, on fiscal reform the medium term, it could save approximately options across priority budget areas, including 11 percent of the education budget, or around education. The recommendation shows the cost 0.6 percent of GDP per annum, in the wage bill to the country of its low student-teacher ratios. It alone. Additional savings could be achieved by suggests a practical way to start the process of consolidating the school network (including increasing student-teacher ratios. It shows the within-schools consolidation of classes). As an savings from such a step, and flags the politically initial step towards school size optimization, the attractive possibility that some of these savings sector could increase the student-teacher ratio in could be used to raise the salaries of the urban schools to the average OECD level, which, remaining teachers. It does not discuss options for according to Ministry of Education estimates, could shrinking the teaching force (e.g., hiring freeze vs. result in savings in amount of 0.03 percent of GDP. early retirement options) or lay out how long it These savings could be used to upgrade facilities, would take each alternative to achieve the raise teacher salaries, invest in teacher qualification cost savings sought. and procure other learning equipment, such as computers. Increase the allocation to the education sector This recommendation goes to one of the root budget. Many of the key issues facing the education causes behind the multiple problems besieging sector stem directly and indirectly from underfunding the sector. Companion recommendations of the education budget. The education budget as a discuss concrete means by which the sector share of GDP stands at about 1.8 percent executed could efficiently ramp up inputs should the sector (or 2.3 percent allocated), which is below the be better funded. This recommendation does not recommended GPE levels as well as the SSA average discuss how the sector budget could be (4.7 percent). Our estimations show that an increase increased because the country has the required to 4.7 percent (in line with SSA average) would be resources. The problem in this case is political will. sufficient to help the sector address three key issues: (i) it would cover the estimated cost of absorbing the out-of-school children into the education sector, (ii) it would allow the full onboarding of all teachers who are currently not in the system, and (iii) it would allow an additional reduction in fees and other costs passed on to households that now operate as barriers to schooling for children from poor households. It is imperative that the government effectively prioritizes the education sector in its budget allocation process as outlined in the MTEF. In order to do so, the spending on the education sector as a share of the total spending should also be revised upwards, closer to the recommended 20 percent, almost doubling the current allocation share. Sources: Cited from unspecified, but actual, public expenditure reviews. 14 Education Public Expenditure Review Guidelines Table 2: Are these recommendations useful? Probably not. Recommendation Why is this not useful? Expand pre-school education, while carefully This recommendation is not connected to the considering the efficiency and effectiveness of problem that it addresses. It gives no sense of the different models. rate at which the subsector could feasibly expand. It fails to provide any specifics on what constitutes an attractive preschool model, or any data to support its recommendation to expand preschool education, such as the relative cost-effectiveness of such programs. Unlike comparators, the sector allocates a relatively This recommendation is vague. What does higher share of its total spending on education to “analysis of impact of spending” mean? Given post-secondary education. After a more detailed the political costs, what are realistic options analysis of the impact of spending at post- for expanding cost-sharing arrangements? Is secondary level, consider introducing cost-sharing there any reason to think that the sector can arrangements at this level so that public spending retain savings at the post-secondary level in can be reallocated to lower levels. order to reallocate them to lower levels? Reform the curriculum to strengthen foundational This recommendation is not tied to the problems skills at primary level and increase actual hours that it addresses. It remains unclear whether of instruction. the instructional time issue is due to low hours as specified by policy or a gap between policy and provision. The recommendation gives no guidance on fruitful curricular-reform options. If policy changes are needed to address the instructional-time problem, it gives no guidance on the potential costs of increasing instructional time, or advice on how to allocate the additional instructional time among subjects. If the problem is due to a gap between policy and provision, the recommendation fails to propose any incentives or penalties that might be used to close the gap. 15 Recommendation Why is this not useful? Better targeting of vulnerable schools and pupils This recommendation, although tied to a problem, would contribute to reducing inequity in education is also vague. Are some “additional resources” more outcomes. In a context of high inequality in learning important than others? Does “better targeting” achievements, as evidenced by the PISA results, indicate that there are now programs in place a revised targeting approach seems necessary focused on vulnerable schools and pupils, but that to ensure that vulnerable pupils and schools these resources have been misdirected in some are provided with additional resources that can way? If so, where are the leakages, and how should contribute to improving their learning outcomes. they be reduced? If there are no such programs, Such targeting would need to be well coordinated what options might be pursued? with other measures to support vulnerable households, which are often managed outside of the Ministry. Address early dropouts and gender disparity This recommendation flags that early dropouts and through the development of supply and demand- gender gaps are problems whose solution requires side interventions. a two-pronged supply-and-demand approach.  It gives no guidance on the types of approaches or tradeoffs among them.   PART II: CHECKLIST FOR AN EDUCATION PER ANALYSIS 17 Part II takes users through what a comprehensive public expenditure review of the education sector should examine. There is no standard outline for such reviews. A review may focus on selected topics or specific subsections, depending on the client’s priorities, the World Bank’s due diligence concerns, the deadline for completing the PER, and budget. In other words, these Guidelines address the six key questions discussed in Part I in a comprehensive manner, but most public expenditure reviews may cover only some of these topics and subsectors. Part II is structured to respond to the six questions in order. Section 1 starts with background and an overview of the education system. Section 2 discusses an overview of education financing and spending (Questions 1–2). Sections 1 and 2 are descriptive and provide a broad picture of the education finance system. Section 3 reviews financial accountability mechanisms qualitatively (Question 3). Sections 4–6 are heavily analytical and examine three principles of education finance: adequacy and sustainability, efficiency and effectiveness, and equity (Questions 4–6). Key Questions 1. Who finances education and how are funds channeled? 2. How much does the government spend and on what? 3. Is the public financial management system set up to enhance financial accountability? 4. Relative to the government’s policies and standards, how much is needed now (adequacy), and what can be afforded in the medium and long term (sustainability)? 5. Are public resources being used efficiently and effectively? 6. Does public spending promote equity? 18 Education Public Expenditure Review Guidelines Section Background and Overview of the 1 Education System Background A background section should provide the sectoral background information as well as the objectives of the public expenditure review. • Sectoral background. Provide a brief summary of sector characteristics relevant to the review. This section should be concise and focused on information relevant to the review. • Major progress and remaining challenges since the last public expenditure review. Acknowledge the client’s reform efforts and achievements, identify why some recommendations have not been implemented, and flag persistent challenges. • Objectives and scope of this review. Define the review’s objectives and scope, including the key policy questions that it addresses. Give the reasons for excluding any important subsectors or topics, and assess how these exclusions might affect a comprehensive analysis of the sector’s financing. • Data sources and analytic methods. Describe data sources (see Technical Notes 1–6) and analytic methods. This section should also address any challenges related to data availability and quality. Example 2: Data sources Overview of the education system An overview of the education system should be presented at the outset to provide basic information about the system and to set the context. • A diagram of the structure of the country’s education system helps the review team and readers of the completed review understand how it is organized. ▷ Structure. Levels of education by grade and official ages for each grade and level. What are the government’s policies on multiple shifts and multi-grade? ▷ Flows (pathways) among levels and types of education. Allowable pathways usually have implications for equity, efficiency, and learning outcomes. For example, students may not be allowed to enter university from an upper-secondary vocational education program. ▷ Type of schools. What types of education providers manage schools? Schools include public, private (financially dependent on the government), private (privately managed without government financing), and religious. Private providers can include nonprofit entities such as non-governmental organizations, or for-profit groups. 19 • Does the education system require private school students to take the same examinations as public school students? How bureaucratically easy or difficult is it for private providers to go into business? Does the state have policies to let private providers access capital more easily? • The overview should provide basic educational statistics (see Technical Note 7: Definitions and notes on indicators). Internationally comparable indicators are available in World Bank EdStats (see Technical Note 4: Cross-national data). Relevant statistics include: ▷ Number of schools by level, location, and type ▷ Number of students in school by level, gender, school location, and school type ▷ Net enrollment rate (NER) by level; if this rate is not available, gross enrollment rate (GER) by level ▷ Number of out-of-school children ▷ Dropout rate by grade ▷ Repetition rate by grade 20 Education Public Expenditure Review Guidelines Section Overview of Education Financing 2 and Spending Key Question 1: Who finances education and how are funds channeled? Education funds can come from government or non-governmental and external sources, and can be channeled through a variety of agents to a broad range of education providers. It is essential to review overall education spending by financing sources, which may include not only public funding, but also private and donor funding. Public expenditure reviews traditionally have not examined the revenue side of the education budget, but it is important to explore how countries raise revenues to finance education. Public funding may be raised at central and local government levels, as well as at the school level. Consider creating a flow chart of financing sources and channels. Policies on fiscal decentralization and school autonomy, together with intergovernmental financing arrangements, define how education funding is channeled down to schools or generated at the school level. These policies can also result in horizontal and vertical imbalances,7 and define how much of education spending is discretionary at each level of government and at the school level. The central government may allocate the education budget to local governments conditionally or unconditionally, based on a transparent methodology or based on negotiation. Schools may or may not have autonomy over budget. Private funding can account for a significant portion of education funding, typically at preschool and tertiary education levels, often at the secondary level, and sometimes for technical and vocational programs. Household surveys can be used to estimate families’ education expenditures. Donor funding may play an important role in financing education in some countries. It could be integrated into the government budget or be off-budget. • How much is spent on education in total, and who finances it, i.e., public, private, or international sources? Technical Note 8: Definition of source of funds Example 3: Analysis of source of funds • How does the government raise revenues to finance education? ▷ At the national level, which sources generate revenues for the national education budget by law (e.g., general revenue or taxes, profits from natural resources, profits from nationalized industries, revenue from lottery, dedicated sources, stabilization funds)? ▷ Do local governments raise revenues, or are all taxes collected by the central government? Example 4: Analysis of revenue sources • Which levels of government finance education? ▷ If financing is split between the central and subnational levels, which levels of government pay for what? Local governments may fund certain levels of education, such as preschool and basic education, but not upper secondary education. They may fund certain inputs, such as school maintenance, but not capital expenditures. Example 5: Analysis of decentralized financing 21 ▷ Are there fiscal transfers from the central government to local governments for financing education? ■ Are these grants earmarked for education or unconditional—i.e., local government can spend the money as it wishes and, in theory, could decide to use none of it on education? ■ How are the grants calculated? ■ Is there evidence of vertical imbalances? ■ Is there evidence of horizontal fiscal imbalances? If so, do they emerge from variations in Example 6: Analysis of education financing by level of government and intergovernmental fiscal transfers • Do schools receive grants? If yes, how are they allocated? How are allocation formulae defined? • Do schools charge parents fees for school? What types of school fees are charged at the primary and secondary levels? Technical Note 9: School fees and other informal payments • Are schools allowed to keep school-generated revenues, such as fees or revenues from entrepreneurial activities, the sale of products in vocational schools, or the rental of school space? For what purposes can public schools use private or international revenues? Example 7: Analysis of school budget by financing source • How much do households with children in school spend on their schooling and for what goods and services? What share of their expenditures consists of mandatory payments (e.g., tuition fees) versus voluntary spending? Example 8: Analysis of household surveys on private spending • Do households receive cash transfers, including vouchers to attend private schools? • Does the private sector, or do other actors, contribute to education spending, e.g., co-financing technical and vocational education and training, or cash transfers to households? • How much do international actors finance education? What is the nature of this spending? How dependent is the government on donors? Example 9: Analysis of donor funding 22 Education Public Expenditure Review Guidelines Key Question 2: How much does the government spend and on what? Rationale for public intervention in the education sector The public expenditure review must scrutinize and justify public intervention in the education sector. Assessments of the following issues should be a subtext of your report: • Proponents of government intervention in education cite a need to ensure equity and efficiency in the sector. The public sector has an equity role because markets generally fail to ensure equal opportunities for all citizens—and, in fact, often create inequities. The efficiency concerns emerge from market failures and imperfections that are commonly associated with information asymmetries,8 externalities,9 and economies of scale. • Government involvement can include regulation, the provision of information, financing, and service provision. Although the government always has regulatory and information-dissemination roles, and is almost always involved in financing education, it does not have to provide such services themselves to ensure equitable access and quality programs. In fact, if the private or nonprofit sector provides education to the country’s standards at lower cost, the government may better serve consumers of education services by subsidizing the consumers, the non-public providers, or both. • Opponents of public intervention in education cite governments’ failure to meet the goals of such intervention or to ensure sustainable financing. Worse, critics say, the governmental involvement can crowd out potentially efficient and equitable private investment and activity. • Where government should intervene, the key challenge is to find ways to prevent failure due to bureaucratic inefficiency, rent-seeking,10 elite capture,11 and other abuses. Overview of public spending on education The public expenditure review first analyzes public spending on education at the aggregate level. What is the total public expenditure on education? As a percentage of gross domestic product (GDP)? As a share of total public expenditure? • Work with the consolidated budget (central government and local government budgets). • Check for off-budget expenditures. • Work only with executed budgets for the past fiscal years because planned is not executed. If you cannot access recent data for executed budgets and decide to include the preliminary figures, clearly label them as preliminary. • Use real expenditures as opposed to nominal ones if you want to track annual percentage changes in the public funds flowing into the sector. Technical Note 10: Calculating the share of a nation’s resources going to education Example 10: Analysis of total public education expenditure The World Bank Group 23 23 Public spending on education can be broken down by functional, economic, administrative (organization), and program classifications.12 Functional classification. A functional classification of expenses organizes government activities according to the socioeconomic objectives that policymakers want to achieve through various kinds of expenditure. Many, but not all, countries adopt the internationally comparable Classification of Functions of Government (COFOG). This classification follows the level categories of the 1997 International Standard Classification of Education (ISCED-97) of the United Nations Educational, Scientific and Cultural Organization (UNESCO). The UNESCO document breaks down education levels into pre-primary and primary, secondary, postsecondary non-tertiary, tertiary, education not definable by level, subsidiary services to education, R&D education, and education services not-elsewhere- classified (or n.e.c.), such as administration, operation, or support of activities. Note that in countries that have not adopted the Classification of Functions of Government, subsectors may be defined differently. Even in countries that have adopted the classification, the length of a given education program, such as primary education, may vary. • Look for anomalies, such as a disproportionate financing share for tertiary education relative to basic education. Example 11: Analysis of education spending by functional classification • Calculate the public expenditure per student as a percentage of per capita gross domestic product (GDP), by level of education? Functional classifications are also useful in analyzing the allocation of resources among sectors. The review should explore if marginal investments in other sectors can help to further education goals. For example, investing in rural roads or improved access to clean water can help to address barriers preventing girls from attending school, and may have more effect on school enrollment at less cost than direct investments in the education sector. If funding for human capital development—e.g., for technical and vocational education and training programs—is dispersed across sectors, other ministries besides the Ministry of Education may provide better value for money. Economic classification. The economic classification of expense identifies the types of expense incurred according to the economic process involved. It includes compensation of employees (wages and salaries and employers’ social contributions), use of goods and services, consumption of fixed capital, interest, subsidies, grants, social benefits, and other expenses. • Consider how education expenditure is allocated by economic factors, such as capital versus recurrent expenditures; staff compensation versus non-staff recurrent expenditures; expenditures spent on teachers versus non-teaching staff? Example 12: Analysis of education spending by economic classification • Be alert to important complementarities among inputs. The educational effectiveness of certain inputs depends on the simultaneous provision of other inputs. Thus, classrooms need teachers; teachers need textbooks. Not infrequently, wages and benefits for labor in the sector crowd out recurrent expenditures for items such as textbooks and other learning materials or infrastructure 24 Education Public Expenditure Review Guidelines maintenance. Capital investments can be paltry relative to the additional school seats required due to increased enrollment in an education subsector. Capital investments can be misallocated among education subsectors. For example, heavy capital investments to build new university campuses can starve the capital budget needed for pre-tertiary education. For vocational education and training, it is important to check whether the budget covers the costs of regular upgrades of workshops and equipment. ▷ Are salaries crowding out the non-salary recurrent budget? ▷ How much is available for routine maintenance per school? How adequate is this amount, compared with estimates of the costs of routine maintenance for the average school? ▷ Considering the government’s textbook policy, is an adequate amount spent per child on textbooks?13 ▷ Are teachers given incentives tied to performance and services? Administrative (organizational) classification. The administrative classification identifies the entity that is responsible for managing the public funds concerned. Public education budgets may be spent by various ministries, different levels of governments, as well as on- and off-budget. • Figure out which levels of government are responsible for spending resources on education. This question is different from who finances education. For instance, the central government may transfer (finance) education block grants to local governments, but the latter is responsible for spending the block grants. • Check the budgets of all ministries that might have education expenditures for all types and levels of education addressed by the public expenditure review. For example, sometimes the budget for early childhood education falls under a ministry responsible for the welfare of women and families. All, or some, of the budget for vocational education and training may come under the Ministry for Labor. Program classification. The program classification gives detailed costs of every activity or program that is to be carried out together with objectives and expected results of a proposed program. Reports on program-based budget execution provide extremely useful data on the extent to which education policies and plans have been successful, and they can stimulate policy changes and in-year course corrections. However, this type of budgeting requires strong expenditure tracking, monitoring, and accountability arrangements that are often not in place in our client countries. Transitioning to a program-based budgeting system not only requires improvements in the government accounting system, but also a shift in internal controls and a shift in accountability from inputs to program outputs. • Are the expected outputs and outcomes specified for each program? • Is the government accounting system capable of providing data on budget execution by programs and sub-programs? • Are the data on actual outputs and outcomes achieved credible? • Do budget-holders and implementers have the flexibility to reallocate among budgetary items in order to achieve expected outputs and outcomes? 25 Section Financial Accountability 3 Key Question 3: Is the public financial management system set up to enhance financial accountability? Effective systems of public financial management (PFM) can contribute to providing better inputs and improving accountability. The World Development Report 200414 stressed that better public financial management can ease several key challenges facing the education sector, including inequitable access, dysfunctional schools, low quality of instruction, low client responsiveness, and stagnant productivity. More recently, the World Bank Group’s Education Strategy 2020 (2015) recognized the important link between better public financial management and improved service delivery outcomes. For instance, equitable allocation of resources can help improve access to education for the poorest families; performance-based incentives can motivate teachers; and efficient spending can lead to savings that can, in turn, help pay for improvements in the learning environment. While theoretical links between improved public financial management processes and service delivery outcomes are well established, empirical evidence has been more limited in terms of when, and how, such reforms might contribute to better services (Welham et al. 2013; World Bank 2012). A number of country examples show strong links between education outcomes and public financial management improvements, but the evidence remains circumstantial.15 A range of public financial management analyses is available to help users assess whether such a system established for the education sector is successful in ensuring that government-earmarked funds are spent on policies as intended. These analyses include the overall Public Expenditure and Financial Accountability (PEFA) Framework,16 sector-level assessments such as SABER School Finance, Public Expenditure Tracking Surveys (PETS), Quantitative Service Delivery Surveys (QSDS), Service Delivery Indicators (SDI) surveys, project fiduciary assessments, and audit reports. The PEFA Framework is helpful in identifying common public financial management issues across sectors. Sector-level and project fiduciary assessments would provide more details on how issues involving public financial management systems specifically affect the implementation of the education program. Internal and external audit reports of the education sector and a walk-through of public financial management processes—all critical business processes related to management and use of funds and assets—could help identify potential weaknesses in these processes. Technical Note 11: Public expenditure and financial accountability (PEFA) country reports Technical Note 12: Public Expenditure Tracking Surveys (PETS), Quantitative Service Delivery Surveys (QSDS), and Service Delivery Indicators (SDI) 26 Education Public Expenditure Review Guidelines The public expenditure review team should use a three-dimensional approach to analyze potential public financial management weaknesses underlying spending problems, as appropriate . Public financial management laws, policies, and procedures determine and regulate the behavior of public officials and organizations implementing them (Allen et al. 2013). Where laws and procedures are sufficiently appropriate but practices lag, it is also necessary to consider the capacity of actors who implement these laws, and the process through which the actors bargain over the design and implementation of policies within a specific institutional setting (World Bank 2017). A three- dimensional approach to a public financial management analysis considers the following factors: (i) Public financial management legal framework—the adequacy of laws, policies, and procedures. What are the shortcomings and how do these affect education sector spending? Are these laws, policies, and procedures complied with? What are the key reasons for any deviations in both public financial management processes and outputs? Are certain political economy factors at play? Why do oversight agencies fail to pick up these deviations? (ii) Public financial management capacity—capacities at central and decentralized levels. What public financial management training and reform mechanisms are in place at the country level and specifically in the education sector? How well does the Financial Management Information System enable the implementation of the public financial management legal framework? (iii) Institutional setup—who does what and the mapping of public financial management roles of central ministries, decentralized agencies, and other factors. Are financial and transaction authority adequately delegated? What role and impact do civil society, unions, media, and the international community have as pressure groups to influence policy formulation and program implementation? Map the impact of institutional and individual actors’ interests on public financial management. Every spending problem is likely to have a unique set of underlying weaknesses in the public financial management process. Linking such weaknesses to key spending problems can be extremely useful in stimulating necessary reforms. How do public financial management processes, such as budget release, funds flow, and internal controls, apply to each of the main spending problems? What do the links between PFM arrangements and sectoral spending problems imply about the reforms needed? For example, textbook purchases generally require a one-off bulk procurement followed by a logistics process of distributing books from central level to schools. Schools then reuse these books over several years. In contrast, school grants do not require a bulk procurement or logistics involving physical assets. However, the issuance of such grants often entails a long (and often complex) funds-flow from central- or provincial-level governments to schools—with the process then relying more on local accountability mechanisms to ensure the funds’ effective use. Table 3 summarizes major spending problems and potential underlying weaknesses in the public financial management system. Example 13: Issues related to budget formation and execution 27 Table 3: Major spending problems and potential underlying PFM weaknesses Spending problem Potential public financial management weaknesses Disconnect • Financial and human resource data are not used for policy- and decision-making. between the • Sector plans do not define institutional responsibilities for different levels education policy of the government for each programs and subprogram. intent and • The government budgeting and accounting system does not track implementation program and subprogram expenditures due to several system and capacity issues, including challenges in apportioning joint costs over multiple (sub)programs. • The central government (Ministries of Finance and Education) has limited authority over, or capacity to handle, the implementation of sector plans to influence decisions made at subnational and local levels. • There is limited or no incentives or accountability linked to results at subnational and local levels. • Budget-holders who are responsible for the best use of the available budget and implementers have limited flexibility to reallocate the budget among input categories for mid-year course corrections. • Delayed budget releases and mid-year cash rationing with cumbersome funds flow (including of school grants) to lower levels are holding up non- salary and capital expenditures. • Weak procurement and internal controls result in delayed or cancelled implementation of activities. Inadequate • Links are weak among target outputs, outcomes, and the budgets allocated. budget allocation • The multi-year sector plan does not adequately cost out stated education policies or medium-term strategic plans. • Education resources are diverted to other functions due to weaknesses in the chart of accounts or lack of transparency in financial reporting. • The sector has budget arrears. • The sector plan has not fully integrated donor-financed projects. Inequitable • Block grants to subnational and local governments are inadequate for budget allocation equalizing fiscal imbalances among the localities (i.e. horizontal equalization). • Education spending results in disproportionate allocations to poor and vulnerable students. Poor human • Delays in receiving salaries lower the motivation of teachers. resource • Lack of hardship incentives limits teacher deployment to rural or remote areas. management • Weak, or lack of, management system for leave record leads to inadequate docking of salaries for absenteeism. (Table continued on next page) 28 Education Public Expenditure Review Guidelines Spending Problem Potential public financial management weaknesses • Sanctions and other action in response to audit findings are limited, even in cases of repetitive or permanent absenteeism. • Lack of independent oversight of teacher attendance increases the possibility of collusion with school directors and district education staff. • Internal and external auditors, monitors, inspectors, and supervisors apply duplicate and document-based controls in an uncoordinated manner, with no, or minimal, unannounced physical checks. • Personnel and payroll records for public servants in the education sector are not reconciled regularly to account for a possible difference between the two. • School directors or teachers cannot perform their duties in school because they are assigned to administrative functions at the district level or are given political party duties. • Teacher unions are politicized and represent a significant pressure group. Lack of school • School directors’ inability to make independent decisions about hiring autonomy temporary teachers results in an inadequate or excess supply of teachers. • School councils are not sufficiently empowered to participate in decisions on the use of school grants, or to monitor teacher attendance. • School councils have no, or limited, involvement in monitoring school performance. Insufficient • Procurement of textbooks, which is often centralized, is delayed. and/or delayed • Storage and distribution logistics are poorly managed, resulting in availability of pilferage of textbooks. teaching and • Controls and incentives that encourage the reuse of textbooks over learning materials several years are limited. • Teaching materials are often procured locally, but lengthy and cumbersome procurement procedures apply. • Delays take place in the receipt of school grants. Poor school • Budget allocation is inadequate for ongoing or planned construction projects. infrastructure • School construction is not completed as planned, with no return on capital investment, and with exposure to rapid deterioration and cost escalations. • Politicized allocation of funds to schools leads to suboptimal regional coverage. • Inadequate technical supervision by the public works staff leads to poor construction quality. • Funding for school maintenance is inadequate and responsibilities are unclear. • To simplify contracting, authorities use expensive, standard school-building designs countrywide, instead of applying climate-suitable adaptations. (Table continued on next page) 29 Spending Problem Potential public financial management weaknesses Over-reliance on • Policy and capacity to regulate and support private schools (e.g. grants, public provision access to finance) are inadequate, not transparent, and not predictable in a way that would incentivize them to provide good services. • Private schools lack accounting capacity. • Regulatory authorities for private schools engage in limited coordination with the Ministry of Education. 30 Education Public Expenditure Review Guidelines Section Adequacy and Sustainability 4 Key Question 4: Relative to the government’s policies and standards, how much is needed now (adequacy), and what can be afforded in the medium and long term (sustainability)? Although international and regional benchmarks are useful in terms of advocacy and cross-country analyses, the adequacy of the budget for a specific country must be carefully assessed. The Third International Conference on Financing for Development (in Addis Ababa, July 2015)17 set the following international benchmarks for education spending: at least 4 percent to 6 percent of gross domestic product (GDP), and at least 15 percent to 20 percent of total public expenditure to the education sector. Worldwide, the former target has been met, but not the latter: In 2012, countries, on average, allocated 5.0 percent of GDP and 13.7 percent of public expenditure to education (UNESCO 2015c). However, benchmarking should be used cautiously because many factors affect total expenditure in the sector, such as the government’s financial capacity, demography, enrollment rates by subsector, quality and prices of basic inputs, geographical challenges, and policies on public versus private financing. A comparison with neighboring countries may be helpful, but regional neighbors may share the same difficulties as the country in question, and hence, may not be appropriate benchmarks. Two other concepts are useful for measuring the adequacy of education spending. Countries have different education goals and standards, and their cost of achieving them will vary, independent of price differences. In the short term, adequacy can be measured relative to the costs of the inputs required, if the country meets its own provision standards for monetary and non-monetary inputs. These standards include those governing unit cost per student, teacher compensation, and the ratios of students to teachers, classrooms, or textbooks. Detailed descriptions of inputs are discussed below. The second concept for measuring adequacy is obvious shortfalls, as evidenced by arrears in the sector or inadequate expenditures on inputs complementary to teachers, such as teaching and learning materials, school maintenance, and needed capital expenditures. Teacher compensation is typically the largest expense in the education sector, and personnel costs are often met by underfinancing non-salary expenses. Key input indicators Besides macro indicators such as education spending as a percent of GDP and of total public spending, the adequacy of monetary inputs can be measured in two ways. One way is to calculate per student spending as a percentage of per capita gross domestic product (GDP), by level of education for public and private schools. The second way is to measure variations in per student spending among subnational units or schools and between public and private schools. Use comparative data to get some sense of whether, as indicated by costs, the country’s provision of inputs—on average and as they are distributed across schools—is “out of range,” either in terms 31 of apparent over-provision or under-provision. Since country- and sector-specific conditions affect allocations among differently priced inputs, comparisons can only flag if there might be a question to answer. Several different ways are available to compute per student spending, or unit costs. Technical Note 13: Definitions and notes on monetary-input indicators Technical Note 14: Calculating unit costs from aggregate and itemized spending Technical Note 15: Calculating unit costs when the country’s fiscal year and school year do not coincide Example 14: Analysis of per student spending Example 15: Minimum norms and standards for resource allocation Example 16: Analysis of cost of teachers The adequacy of non-monetary inputs (human resources, classrooms, and infrastructure) can be measured by scrutinizing government standards and by assessing their actual provision. Governments, including the Ministry of Education and sometimes parliamentarians, set standards governing inputs to education. These include standards for where to build new schools relative to population settlements, construction designs and standards, student-classroom ratios, student- teacher ratios, student-textbook ratios, annual hours of teacher training, annual instructional time overall and per subject, annual duration of schooling, and so forth. However, actual provision of inputs may or may not meet those standards. Therefore, it is important to examine both standards and actual provision, and identify whether either, or possibly both, may be inadequate to deliver quality educational services. Both are problems that PERs can address, but their solutions differ. The former requires changes in policies that define standards; the latter, changes in how inputs are distributed. Obtain information on input standards by interviewing the Ministry’s policymakers or the senior technical staff, backed by relevant policy documents. Data on actual provision of inputs should be disaggregated by level of education or grade (whichever appropriate) and by region, district, and school (using the smallest unit for which data are available), within a country. They should also be compared over time within the country, and with data for comparator countries. Technical Note 16: Definitions and notes on non-monetary input indicators Technical Note 17: Definitions and notes on research indicators Check these aspects of the three main types of non-monetary inputs: • Human resources ▷ Student-teacher ratios ▷ Percentage of qualified teachers ▷ Ratio of teachers to non-teaching staff ▷ Non-teaching staff per type of school ▷ Organization of teachers’ working time Example 17: Analysis of teacher distribution 32 Education Public Expenditure Review Guidelines • Classroom inputs ▷ Average class size ▷ Student-textbook ratios for core subjects ▷ Percentage of schools meeting minimum standards requirements for educational inputs ▷ Annual instructional time Technical Note 18: Efficiency of the curriculum • Infrastructure ▷ Average school size (number of students per school) ▷ Unit cost of building a classroom ▷ Standards and schedule for infrastructure maintenance ▷ Percentage of schools that run double and triple shifts ▷ Percentage of schools that use multigrade classrooms ▷ Percentage of schools meeting minimum standards requirements for learning conditions Comparing the projected costs of education with the anticipated size of the resource envelope can shed light on whether planned spending levels are realistic. Beyond considering how much is needed now, it is essential to examine what happens to costs in the medium and longer term, given the government’s sectoral goals, demographic projections for the school-age population, assumptions about enrollment rates by level, and macroeconomic and budget forecasts. Projected costs of education provision are often spelled out in the Medium-term Expenditure Framework (MTEF). If the country does not have such a framework in place, a costing exercise will be required. Example 18: Cost projections Example 19: Fiscal sustainability analysis • Sectoral goals. Medium- to long-term education goals and targets may include extending the length of compulsory education, increasing education access for the poor, hiring more qualified and, thus, expensive teachers, converting community teachers to regular teachers, introducing computer-assisted instruction, retrofitting facilities to protect against earthquakes, expanding vocational education and training, and introducing cost recovery for tertiary education. It is essential to examine whether these goals are fiscally feasible. • Country’s demographic structure and trends, including urbanization, and projected enrollment rates by subsectors. Rapid urbanization can increase future enrollment rates. What do the trends in the school-age population and assumptions about trends in enrollment rates by subsector imply about the inputs required now and in future? Technical Note 19: Demographic trends and enrollment projections Example 20: Demographic trends and enrollment projections • Country’s macroeconomic projections and government’s strategic priorities. The former affect the total resource envelope available; the latter, the government’s decision on how to allocate resources among different sectors. 33 If at all possible, do some simple modeling to assess what the joint implications are for projected costs and revenues and, thus, how realistic the government’s plans are for the sector. For instance, UNESCO’s Education Policy and Strategy Simulation Model (EPSSim) is a sector- wide and goals-based generic simulation model that is driven by demographic trends. Enrollment targets are taken as a priori and the simulation calculates the corresponding financial resource implications. The World Bank’s Maquette for MDG Simulations (WB-MAMS) is a Computable General Equilibrium (CGE) model and can help assess the broad, economy-wide effects of alternative education scenarios. Be sure to work closely with the country economist or the macroeconomist on the public expenditure review team. Technical Note 20: UNESCO’s Education Policy and Strategy Simulation Model (EPSSim) Technical Note 21: Simulating the economy-wide effects of alternative education scenarios 34 Education Public Expenditure Review Guidelines Section Efficiency and Effectiveness 5 Key Question 5: Are public resources being used efficiently and effectively? Resources are scarce, and an important purpose of any PER is determining value for money from education investments, regardless of their financing source. The purpose here is similar to the “value for money” (VfM) approach that the United Kingdom’s Department for International Development (DfiD) has adopted for making aid decisions;18 that the World Bank has pursued for decades in its economic analyses of investments; and that studies such as cost-benefit and cost- effectiveness analyses and impact evaluations support. The basic value for money concept is that of obtaining the maximum benefit over time with the resources available. Value for money is high when an optimum balance exists among three elements: when costs of inputs are relatively low, productivity is high (or efficient), and successful outcomes have been achieved (or effective).19 It should be noted that efficiency and effectiveness are not the same. A service or good may be efficiently produced, but not effective. Similarly, it may be effective, but not efficiently produced. For example, a teacher-training program may produce a large number of graduates at a small cost, yielding an efficient cost-output ratio. However, if these teachers’ classroom performances shows no discernible improvement, the training is efficient but ineffective. A student-to-qualified teacher ratio of 15/1 may result in learning gains. However, if increasing the ratio to 25/1 achieves the same learning gains, using qualified teachers is effective, but at the 15/1 ratio, inefficient. Output and outcome data To assess efficiency and effectiveness, we need input, output, and outcome data. Section 4 above defined monetary and non-monetary inputs. The primary output and outcome data needed are the following, including trend data for the PER country and comparative data: Technical Note 22: Definitions and notes on output and outcome indicators • Participation rates ▷ Gross and net enrollment ratios by level of education ▷ Dropout rate by grade ▷ Repetition rate by grade ▷ Completion rate for each educational program • Learning outcomes for core subjects Technical Note 4: Cross-national data • Employment and wage rates by level of education ▷ Labor force participation rate ▷ Employmentandunemploymentrates Several different types of economic efficiency analyses exist, but those most useful for education public expenditure reviews are analyses of allocative efficiency, technical efficiency, internal efficiency, and external efficiency. Technical Note 23: Concepts of effectiveness and efficiency 35 Allocative efficiency Allocative efficiency asks whether the sector is allocating its resources among subsectors “so as to maximize the welfare of the community.” Section 2 asked for an analysis of how public expenditure is allocated among education levels or subsectors. Subsector allocations for the country undergoing a public expenditure review can differ for good or poor reasons from those of countries with which it is being compared. For example, the definitions of education levels can vary across countries. Eliminating school fees usually creates enrollment bulges that require high “catch-up” allocations to the newly “free” subsector. Significant success in obtaining high primary-completion rates usually translates into sharp spikes in demand for secondary school seats that then require major increases in capital budgets for the subsector. However, high and increasing expenditures for tertiary education may crowd out allocations for pre-tertiary-level programs—a trend that ultimately favors the children of wealthier families. Assess the allocative efficiency of the standards themselves, independent of how well the public expenditure review country meets them, to separate inefficient objectives from inefficient implementation. As discussed in Section 4, the sector usually sets standards for inputs, such as student- teacher ratios, teacher compensation schedule, student-textbook ratios, and student-classroom ratios. Comparative data for these variables will reflect, not comparators’ standards, but how comparators implement their own standards. However, these data will give some idea of reasonable ranges for standards in the country undergoing review. Figure 4: Distribution of student-teacher Start with simple histograms to display the distribution ratios for primary schools in a province of an input across regions, schools, or classrooms. Highly variable deployments of inputs always signal allocative inefficiencies. For example, if the sector standard for the Share of schools (%) student-teacher ratio in a primary class is 35 students per teacher, the hypothetical example in Figure 4 shows wide (but not unusual) variances in these ratios between schools— inherently and relative to the 35:1 standard. Completely aside from whether the standard itself is or is not reasonably efficient, the very uneven deployment of a key resource is inefficient Student-teacher ratio (and inequitable). When significantly variable deployments are observed, the review must try to determine the sources of Note: A hypothetical example the variability. Technical efficiency The sector achieves value for money when it gets the best outcomes at least cost. This result can be achieved in either of two ways. The inputs used by a given intervention can be reduced to the minimum required to achieve the outcome sought. Or a different intervention can be used— one with a different combination of resources that achieves, at less cost, the outcome sought, as well as, or better than, the alternatives. Example 21: Technical efficiency of inputs (efficiency indicators) 36 Education Public Expenditure Review Guidelines Analyses of technical efficiency look at costs, input mixes, and results. Cost-effectiveness analysis (CEA) relates monetary inputs and non-monetary outputs and outcomes. Cost-benefit analysis (CBA) is used when both the costs and the outcomes can be monetized. A public expenditure review is not usually expected to conduct either type of analysis. However, the review may use findings of international and in-country cost-effectiveness analysis and cost-benefit analysis to help identify instances of probable low value for money. For instance, in examining six recent systematic reviews or meta-analyses of interventions that improve learning outcomes in low- and middle-income countries, Evans and Popova (2015) observed a wide variation in conclusions across the studies, with much of the variation driven by variation within categories of interventions. Thus, the specific details of the intervention determined its effectiveness. Based on a careful examination of the details of interventions in the six studies, they identified three categories of interventions that were broadly supported across the studies: “(i) pedagogical interventions that match teaching to student’s learning, including through the use of computers or technology; (ii) individualized, long-term teacher training; and (iii) accountability-boosting interventions, such as teacher performance incentives and contract teachers” (Evans and Popova 2015, 3). These findings may help the PER team identify a potentially inefficient area that the sector is strongly advised to investigate. Technical Note 24: Cost-effectiveness analysis Example 22: Analysis of unit costs and outcomes Technical Note 25: Cost-benefit analysis Example 23: Cost-benefit analysis Use cost data to sense whether the sector may be overpaying for an input relative to the value obtained. Especially relevant data are costs of teachers relative to in-country comparators, costs of textbooks relative to comparators, and in-country variations in unit costs for classroom construction. For example, private school teachers may be considerably cheaper than public school teachers. If learning outcomes in private schools are similar to, or better than, those in public schools, controlling on family socioeconomic status (SES), the sector may want to consider shifting more provision of education to the private sector via vouchers or subsidies. Significant in-country variations in unit costs for classroom construction also raise questions about value for money. Theunynck (2009) finds that the most efficient designs are ones that can be mass produced at low cost on the local market, using small- and medium-size enterprises in the formal and informal sectors. He also alerts us to the fact that alternative procurement and contract management arrangements differ in their cost-effectiveness. A relatively complex analytic technique called data envelopment analysis (DEA) can be used to measure the efficiency of multiple service delivery units by comparing the mix and volume of resources used (inputs) and services provided (outcomes) by each unit. DEA is a linear programming methodology. The techniques can be used to define and estimate efficiency as the distance from the observed input-output combinations to an efficient frontier, which is the maximum attainable output for a given input level. It is often used to compare countries relative to an efficiency frontier. For example, for a large panel of countries, total government education expenditures as a percent of GDP (or GNI) can be plotted against different education outcomes, such as primary completion rates, enrollment rates, or learning outcomes.20 Example 24: Data envelopment analysis 37 Internal efficiency21 Internal efficiency measures the percentage of children who complete an educational cycle (e.g., primary education or lower secondary education) as a share of those who start the cycle or as a percentage of those who finish the cycle in the minimum number of years. The first definition allows the calculation of the dropout rate—i.e., the number of those who start minus the number who ultimately complete, as a share of those who start. The second definition measures the dropout rate plus the repetition rate. If the data show either relatively high dropout rates, high repetition rates, or both, the sector has an expensive internal-efficiency problem. Dropping out imposes costs on individuals and countries in the form of unrealized human capital. Repetition imposes costs on the sector in the form of its having to pay double (or triple) the unit cost of a year of school per repeater. In cases of significant internal inefficiency, the difference between unit costs per completer of an education program and unit costs per student who completes the program without interruption are better measures of costs than unit costs. Technical Note 26: Calculating the budget costs of one completer and of one uninterrupted completer Example 25: Internal efficiency indicators External efficiency22 External efficiency measures the returns to individuals, employers, and the country of public and private investments in education. It depends on a match between the type and quality of skills and knowledge that school leavers acquire in school relative to the skills and knowledge needed and paid for by employers. Does education improve the employability and wages of school leavers? Does public investment in education and training contribute to the country’s growth and economic development? Measuring and linking employment and wage returns to education is particularly important for vocational education and training, and tertiary education. Technical Note 27: Estimating private rates of return to education Example 26: Rate of returns to education 38 Education Public Expenditure Review Guidelines Section Equity 6 Key Question 6: Does public spending promote equity? A fundamental responsibility of the state is ensuring equity and managing redistribution. Public policy, including educational finance policy, can help minimize subgroup differences in educational access and achievement. This section explores: (i) how to identify inequity, if any; (ii) whether, and how, the government spends its budget to promote equity in education; and (iii) how households are responding to public policies and filling the funding gap between their needs and public spending. Equity in education financing can be assessed in terms of two main principles: horizontal and vertical equity. Horizontal equity is defined as the equal treatment of equals and is used to justify similar levels of funding across comparable schools or subnational divisions. Accordingly, a mechanism may operate to equalize education spending across subnational divisions to preserve fiscal neutrality, so that the amount of available resources for education is not positively correlated with the wealth of where a student lives. Vertical equity supports the unequal treatment of “unequals” (Underwood 1995). For example, progressive spending may be necessary to provide equivalent education to students whose native language is different than the language of instruction or to students with special education needs. Other examples include targeted support programs (such as conditional cash transfers and scholarships), or student weights to differentiate spending for certain types of students. Vertical and horizontal equity have a special meaning in tertiary education. According to Salmi and Bassett (2012): The vertical dimension looks at who enters tertiary education and who graduates from tertiary education. The horizontal dimension looks at what kind of institution they attend, and what labor market opportunities are offered to graduates with various types of qualifications and levels of degrees. The first analytical step is to diagnose the country’s trends related to equity in education. Disparities in access to education, completion, and learning achievements across different student groups indicate existing problems of inequity in education. 39 • How do school enrollment rates, completion rates, and learning outcomes vary by gender, household income, geographical location, and ethnic or religious group? Example 27: Analysis of inequity • Do the data assembled for Section 4 on the distribution of inputs by school, district, or region indicate substantial variation among geographical areas in terms of children’s opportunities to learn? • What is public spending per student by level of education, subnational division, school type, and student group (e.g., geographical location, income quintile, gender, ethnicity, language, religion, and special needs)? Technical Note 28: Per capita financing Example 28: Analysis of per capita financing The second step is to examine what role the state plays in mitigating or exacerbating inequity. • What demand-side and supply-side financial interventions does the government adopt to promote equity? The most common programs that address equity include: ▷ Demand-oriented interventions, such as conditional cash transfers, school feeding, vouchers, scholarships or student loans, universal and targeted child benefits, and full or partial subsidies for school supplies, transport, and boarding Technical Note 29: Targeting mechanisms, coverage, and depth of programs Example 29: Analysis of cash transfer programs ▷ Supply-oriented interventions, such as an expansion of the education system to reach the poor, and the provision of additional funding for disadvantaged students23 • How progressive or regressive is the state’s financing of education? ▷ What are benefit incidences across different groups of households? Technical Note 30: Benefit incidence analysis Example 30: Benefit incidence analysis • What actions other than financing does the government take to increase parental demand for education? If parents fear for their daughters’ safety during travel to school, does the government implement measures aimed at protecting girls? For families speaking a minority language, does the government publish textbooks and provide teachers who offer instruction in that language?24 • Does the state finance private education through partial or full subsidies to providers or consumers? What are the rules governing these subsidies? For example, what percentage of the estimated per student cost for public schools does a subsidized private school receive? Can a private school receiving public subsidies also charge fees? Depending on how they are designed, public subsidies may implicitly subsidize the wealthy’s preference for private education. Technical Note 31: Subsidies 40 Education Public Expenditure Review Guidelines • Are there corrupt practices that affect access, grades, or graduation—e.g., bribes to university faculty to secure entry into a particular department, or parental “gifts,” such as new computers, given to gain admission to a prestigious secondary school?25 If so, how widespread are these practices? Has the state taken any actions to stop them? Does the state regulate the practice of the student’s teachers, or teachers within the student’s school, providing private lessons to the student? Bribes, gifts, and private lessons penalize the poor. • What public policies govern students’ progression through the educational system? If tertiary enrollments are rationed, examinations during the pre-tertiary years are often used to “weed” students out of the system. Pathways into tertiary education may be highly restricted. Students may have had to complete the academic program at the upper secondary level, and access to this program may be highly restricted, as well. Such policies favor wealthier families. The third step is to analyze private spending. Use household survey data on households’ education expenditures to assess whether educational disadvantages are related to private costs for education (financial barriers). • What do families in different consumption quintiles pay for education by level of education as a percentage of their average consumption? Consider both formal and informal payments. Technical Note 9: School fees and other informal payments Example 31: Analysis of private spending by income quintile TECHNICAL NOTES 42 Education Public Expenditure Review Guidelines Technical Note 1: Government’s official documents Government’s official documents—such as policies, laws, and regulations that specify standards for inputs to, and financial arrangements for, the learning process—are necessary to differentiate policy intent and implementation. For discussion of inputs, see Section 4: Adequacy and Sustainability. The SABER-school finance data collection instrument26 can provide a more detailed list, as needed. Technical Note 2: Country system-level data Data generated by country systems include: (i) government budget documents (central government consolidated accounts; Ministry of Education budgets; state or provincial budgets, if separate from consolidated government accounts; institution-level financing data; medium-term expenditure framework documents); (ii) education management information systems (EMIS); (iii) national learning assessments; and (iv) sector-specific databases, such as school mapping and teacher databases. It should be noted that data generated by country systems tend to be the most problematic and need to be treated with caution. (i) It is not easy to obtain reliable financing data from the government. Wherever possible, try to obtain data from the Ministry of Finance, instead of the Ministry of Education. BOOST, a World Bank initiative launched in 2010, draws detailed government expenditure data from government financial management information systems and creates easy-to-use databases. The program strengthens public-expenditure policy outcomes and accountability by improving the quality of expenditure data, facilitating rigorous expenditure analysis, and improving fiscal transparency. Experience indicates that BOOST is most useful if the raw data from the Ministry of Finance is sufficiently disaggregated. The Governance Global Practice’s BOOST team could support the public expenditure review team by creating an education module in an Excel format that combines expenditure data from the BOOST database with education statistics and other information on public institutions, service delivery, and households.27 The ease of access to, and preparation of, analytical reports supports decision- making for the purposes of planning, budgeting, monitoring, and evaluation. BOOST often does not have expenditures by programs and sub-programs because governments typically do not have relevant disaggregated data. In collaboration with the World Bank, more than 70 countries have developed a BOOST government budget database to date. If the counterpart does not yet have a BOOST, consider working with the Governance Global Practice and Macroeconomics and Fiscal Management (MFM) Global Practice colleagues to encourage the counterpart to develop a BOOST database. If the education public expenditure review is part of a comprehensive expenditure review, the Macroeconomics and Fiscal Management team is most likely to create a BOOST that the education team can then use. However, teams need to be aware that building a BOOST may take a long time. Also, BOOST tends to lack disaggregated data at the local level. If education functions and financing are decentralized, the education team may need to collect disaggregated data separately from local governments. 43 (ii) Especially in low-capacity countries, education management and information system (EMIS) data can be problematic in terms of validity and reliability, especially when it comes to recent data. The team needs to be alert to incentives in the system that encourage misreporting of the numbers. For example, per capita financing creates incentives to inflate enrollment numbers. In such cases, the public expenditure review team will have to triangulate among data sources to estimate values of school-specific variables, enrollment rates, or teacher absenteeism rates. (iii) Designing and administering a good learning assessment is quite technical. Although it is generally preferable on quality grounds to use well-established regional or international learning assessments, national learning assessments can meet technical design standards. Relative to cross-national learning assessments, they can also be better aligned with the country’s curriculum. However, confer with the Bank’s country education team on the technical quality of these assessments. In using the data, flag concerns raised by the Bank’s team. (iv) The quality of other sector-specific databases, such as school mapping and personnel rosters, varies across countries. However, if they are of reasonable quality, they are a good source for determining the distribution of inputs relative to government standards—for example, infrastructure relative to population settlements, or teachers relative to student enrollments. Technical Note 3: National surveys National surveys measure variables important for PERs and can be a goldmine. For example, household consumption surveys are essential to creating poverty maps and to measuring household payments for education. Census data are essential for estimating changes in the size of school-age cohorts that the education system will have to accommodate. Labor force surveys can be used to estimate employment and wage returns to different levels of education or the educational attainment of the working-age population or active labor force. Although the team needs to determine quirks and data-quality problems with any survey, the quality of national surveys tends to be adequate and may be excellent. The statistics unit of a country tends to conduct such surveys, although they may use internationally established frameworks and processes. Data are collected under the same protocol. The staffs of units conducting these surveys benefit from international experience with the design, administration, and analysis of such surveys. Donors also often fund technical assistance to help such units professionalize the conduct of the surveys under their jurisdiction. Census data and reports. Country-specific estimates and projections can be checked with population experts in the Health, Nutrition, and Population Global Practice (HNPGP). Household and other surveys. Household surveys usually collect data on enrollments and completion by level of education, age, gender, and residential location, and private spending on education. Country-specific survey data include the following common and highly developed sources: • Living Standards Measurement Study (LSMS): The LSMS is a household survey program housed within the Survey Unit of the World Bank’s Development Data Group that provides technical assistance to national statistical offices in the design and implementation of multi-topic household 44 Education Public Expenditure Review Guidelines surveys. Since its inception in the early 1980s, the LSMS program has worked with dozens of statistics offices around the world: generating high-quality data, incorporating innovative technologies and improved survey methodologies, and building technical capacity. • The Skills Towards Employability and Productivity Program (STEP) program: The World Bank’s STEP measures skills in low- and middle-income countries. It provides policy-relevant data to enable a better understanding of skill requirements in the labor market; backward linkages among skills acquisition and educational achievement, personality, and social background; and forward linkages among skills acquisition and living standards, reductions in inequality and poverty, social inclusion, and economic growth. The STEP program includes a household-based survey and an employer-based survey. All relevant survey documentation is provided along with the datasets. The “STEP Methodology Note” presents key concepts and describes the STEP survey instruments. It also provides guidance on how to use the data. • Demographic and Health Surveys (DHS): These nationally representative household surveys provide data on a wide range of monitoring and impact evaluation indicators in the areas of population, health, and nutrition. Education is a key background indicator in demographic and health surveys, which help contextualize a country’s health and development situation. • Multiple Indicator Cluster Surveys (MICS): UNICEF supports governments in carrying out these surveys through a global program of methodological research and technical assistance. Cluster survey findings have been used extensively as a basis for policy decisions and program interventions, and for the purpose of influencing public opinion on the situation of children and women around the world. All available results and datasets from these surveys can be accessed on mics.unicef.org. The results from the most recent MICS-5 surveys, carried out in 2012–15, are becoming progressively available. (MICS-6 was launched in October 2016. It will cover five of the household survey-based indicators for Sustainable Development Goal (SDG) 4, Education 2030.)28 • EdStat’s Education Equality: Household surveys can provide detailed information on gender, income, and geographic inequalities in education access, progression, attainment, and expenditures. EdStats gives users access to household survey data through the following tools and resources: ▷ The Education Equality Query holds household survey data from DHS and MICS. Indicator names beginning with the labels “DHS” and “MICS” were generated by EdStats using the ADePT Education tool. Indicator names beginning with “UIS” were generated by the UNESCO Institute for Statistics (UIS) using its stated methodology. Variances may exist in data from differing sources based on methodological differences. ▷ Education Equality Country Profiles are detailed Excel file reports for all available DHS, MICS, and LSMS. They include a series of graphs and a wider variety of indicators than are currently available in the Education Equality Query. ▷ The Education Equality Dashboard is a data visualization tool that allows users to visualize disparities in attendance rates, completion rates, educational attainment, and out-of-school children. 45 ▷ The ADePT Education tool allows users to derive education indicators from household survey data and create customized reports and graphs. The program contains built-in settings for DHS surveys, but also accepts other types of surveys, and determines automatically what tables can be built from the available data. • University of Oxford’s Young Lives survey: This survey includes a household questionnaire with special items focused on the children and a community questionnaire. It covers Ethiopia, India, Peru, and Vietnam. It is often complemented by a school survey and the collection of in-depth, qualitative longitudinal data on some of the children. The household survey collects data similar to the World Bank’s Living Standards Measurement Study. These include information on household composition, livelihood and assets, household expenditure, children’s health and access to basic services, and education. This is supplemented with additional questions about caregiver perceptions, attitudes, and aspirations for the children and the family. The Young Lives survey also collects detailed data on how all family members use their time, information about the children’s weight and height (and similar information for caregivers), and data on children’s learning outcomes (language comprehension and mathematics). It asks the children about their daily activities, their experiences, and attitudes towards work and school, their likes and dislikes, how they feel they are treated by other people, and their hopes and aspirations for the future. The community questionnaire provides background information about the social, economic, and environmental context of each community. It covers topics such as ethnicity, religion, economic activity and employment, infrastructure and services, political representation and community networks, crime, and environmental changes. Technical Note 4: Cross-national data The two main types of cross-national data systems are international learning assessments and statistical databases that can show where a country sits within the range of practice. In addition to national assessments of learning outcomes, check to see whether the country has participated in any of the international or regional learning assessments. EdStats’ Learning Outcome Dashboard/By Country has a table of the assessments in which each country has participated. These assessments usually measure the gender composition and characteristics of the home that can serve as a proxy for socioeconomic status and sometimes other characteristics that indicate subgroup membership. Cross-national learning assessments also provide comparisons and tend to meet higher design standards and better-tested administration and data-cleaning procedures than national learning assessments. 46 Education Public Expenditure Review Guidelines • Program for International Student Assessment (PISA) • Trends in International Mathematics and Science Study (TIMSS) and Progress in International Reading Literacy Study (PIRLS) • OECD’s Programme for the International Assessment of Adult Competencies (PIAAC) and other adult literacy surveys • Early Grade Reading Assessment (EGRA): Applications and Interventions to Improve Basic Literacy and Early Grade Mathematics Assessment (EGMA): A Conceptual Framework Based on Mathematics Skills Development in Children, developed by the Research Triangle Institute (RTI) • Southern Africa Consortium for Monitoring of Education Quality (SACMEQ) • Programme d’Analyse des Systèmes Educatifs des Pays de la CONFEMEN (PASEC), or the CONFEMEN Programme for the Analysis of Education Systems (i.e., conference of education ministers for countries sharing the French language) • Laboratorio Latinoamericano de Evaluación de Calidad de la Educación (LLECE), or Latin American Laboratory for Evaluating the Quality of Education The quality of cross-national statistical databases is mixed. • OECD's annual Education at a Glance: These OECD data are of high quality because of the exceptional processes in place that produce them. • UNESCO Institute of Statistics (UIS): Although the Institute works persistently with governments to help them improve the quality of data generated by their education management information systems, it ultimately has to depend on country-specific data with all of their problems. As a result, the quality of Institute data is uneven. • World Development Indicators (WDI): These are the World Bank’s primary collection of development indicators, compiled from officially recognized international sources. They assemble the most current and accurate global development data available and include national, regional, and global estimates. Six themes are used to organize indicators: world view, people, environment, economy, states and markets, and global links. For education, the World Development Indicators cover five types of variables: education inputs (e.g., government expenditure per student, government expenditure on education as a percentage of GDP and as a percentage of total public expenditure, trained teachers, and student-teacher ratios); participation in education; education efficiency (e.g., cohort survival rates, repetition rates, transition rates); education completion and learning outcomes; and education gaps by income, gender, and area (country and region). The data sources for the first four types of variables are almost entirely from the UNESCO Institute of Statistics. As noted above, these must be used cautiously because the data depend on country- specific education management information systems (EMIS) data. The data for the fifth type (gaps by income and gender) depend heavily on demographic and health surveys, and multiple indicator cluster surveys. 47 • World Bank EdStats (Education Statistics): EdStats All Indicator Query holds more than 4,000 internationally comparable indicators that describe education access, progression, completion, literacy, teachers, population, and expenditures. The indicators cover the education cycle from pre- primary to vocational and tertiary education. Most EdStats data come from the UNESCO Institute of Statistics. EdStats also includes learning outcome data from international and regional learning assessments, equity data from household surveys, and projection and attainment data to 2050. ▷ Education Expenditures: Education-expenditure data reside in two databases on the EdStats website: (i) the EdStats Query–Education Expenditures; and (ii) the World Bank Education Expenditure Database, which has been created using data extracted from World Bank public expenditure review documents. Table TN1 summarizes the differences between the databases. Table TN1: Education-expenditure data base EdStats Query World Bank Education Education Expenditures Expenditure Database Use of Download core expenditure Download detailed expenditure Database indicators and compare data on one country. Data cannot expenditure data across countries. be used to compare countries. Number of Indicators 93 More than 800 Data Source UNESCO Institute for Statistics World Bank PER Documentsa a. The PER documents are available in http://datatopics.worldbank.org/education/wDataQuery/ExpBackground.aspx. ▷ EdStats Dashboards: The Key Education Indicators Dashboard presents a broad portrait of all levels of a selected country’s education system from pre-primary to tertiary education. It includes gender, regional, and income-group comparisons. ▷ The State of Education–Expenditure Dashboard: The State of Education’s Expenditure Dashboard presents a global view of education spending through dynamic maps, charts, and accompanying analysis. It presents not only key expenditure indicators such as government spending on education as a percentage of gross domestic product, but also the correlation between government spending and outputs such as enrollment rate and learning outcomes. 48 Education Public Expenditure Review Guidelines • National Education Accounts (NEA): This tool takes into account multiple education-financing data from different sources, including public, private, and donor funding, in a given country, and seeks to enable international comparisons of education financing (UNESCO 2015a). Drawing a complete picture of education financing in a given country does not necessarily allow international comparisons due to differences in budget classifications. To enable international comparisons, the National Education Accounts methodology (UNESCO Institute for Statistics, UNESCO International Institute for Educational Planning, and IIEP Pôle de Dakar 2016a and 2016b) has been developed on the principles of existing international standards, such as the System of National Accounts (SNA 2008), and the International Standard Classification of Education (ISCED 2011). Development of a National Education Account requires a careful matching of country and international classifications, and only a limited number of countries have implemented this tool to date. A long-term goal for the international education community is to enhance the use of this tool to enable international comparisons of education financing across countries. • World Economic Forum annual Global Competitiveness Report: This report has survey data collected from employers in each country. Respondents identify barriers to doing business, including an inadequately educated labor force, and rate the quantity and quality of a number of human-capital measures. • World Bank's Doing Business Survey: This survey is less useful for education than the World Economic Forum survey or the World Bank’s Enterprise Survey. It sometimes assesses labor market regulation. • World Bank’s Enterprise Survey: An Enterprise Survey is a firm-level survey of a representative sample of an economy’s private sector. The surveys cover a broad range of business-environment topics, including access to finance, corruption, infrastructure, crime, competition, labor, and performance measures. The labor module assesses the characteristics of the firm’s employees (e.g., their educational attainment), the firm’s labor policies (e.g., employee-training programs) and the employer’s views of the extent to which inadequately trained workers constrain their business. Since 2005–06, more than 125,000 interviews in 139 countries have taken place under the Global Methodology. Enterprise Surveys implemented in Eastern Europe and Central Asian countries, conducted jointly with the European Bank for Reconstruction and Development, are also known as Business Environment and Enterprise Performance Surveys (BEEPS). • Transparency International: This organization publishes data on perceived corruption by country. In 2013, it published an analysis of the sources of corruption by level of education, including extensive treatment of corruption in higher education. 49 Technical Note 5: Research reports Before starting a public expenditure review, be sure to scan for relevant research reports published by the World Bank, donor partners, non-governmental organizations (NGOs), or the international research community. The Bank’s education public expenditure reviews tend to underuse the international evaluation literature, such as meta-analyses of rigorous evaluations of the effects of inputs on students’ participation and learning outcomes (Glewwe et al. 2011; Snilstveit et al. 2016). These studies provide important data on the likely effectiveness of different investments and the conditions under which effectiveness occurs. As such, they are useful for public expenditure reviews’ analyses of the efficiency and effectiveness of inputs. To locate studies, use these links: • World Bank research and publications: Look for Systematic Country Diagnostics, Country Economic Memoranda, poverty assessments, earlier public expenditure reviews, Public Expenditure Tracking Surveys (PETS), Quantitative Service Delivery Surveys (QSDS), Service Delivery Indicators (SDI) (see Technical Note 12 for details on PETS, QSDS and SDI), public expenditure and financial accountability (PEFA) country reports (see Technical Note 11 for details on PEFA), and other analytical reports. • Other sources: Check the websites of international donors and non-governmental organizations active in the country in question, websites of the OECD, UNICEF and UNDP, particularly the annual Human Development Report. • External publications, especially academic research and evaluation studies: Type in JOLIS, scroll down, and choose EconLit under “Popular Resources,” then search by topic, author, or title. This source of data varies in its quality. Agencies such as the World Bank and academic journals have standards and processes, such as peer review requirements, that at least create a floor on quality. However, small donor groups, such as non-governmental organizations, may lack such processes. Studies conducted by groups such as these can vary widely in quality, depending on the individual doing the study. The only way to judge data from these studies is to read the original studies. Technical Note 6: New data collection Strongly consider collecting your own data when the topic is important and the data required to assess it are unavailable or unusable. • When new data have to be collected, design decisions by the PER team will define its quality. For example, if sampling is required, the team’s choice of its sampling frame and sampling criteria for selecting the units to be surveyed will determine if the results can be properly generalized to the universe. • When new data collection is essential, the team should, as needed, renegotiate the Bank’s budget for the expenditure review to cover the costs of new data collection or seek substantial trust funding to cover its costs. 50 Education Public Expenditure Review Guidelines • If time and budget do not permit new data collection, consider assembling a panel of experts or a focus group to give you a sense of the shape and magnitude of the issue. Methods certainly exist for increasing the validity and reliability of qualitative data collected through expert panels or focus groups. However, the intent here is to determine whether the PER should recommend new data collection if future operations or analytic work must address the issue in question. • If you can do nothing or little about important data gaps, describe them in the review and flag the need for future work. Finesse the data gaps as best you can. Technical Note 7: Definitions and notes on indicators Core Indicators Definitions / Notes Number of schools • By level, location, type • “School” is a service point (or campus that is part of a larger educational institution) that provides instructional or education- related services to a group of pupils. A school may have a single administrative unit with several service points (or group of branch schools or satellite school or campuses). An administrative unit refers to any school, or group of schools, under a single director or a single administration. A service point refers to any location that provides a service for pupils or students, whether it is a single entity or part of a larger administrative unit (UIS). Number of students • By level, gender, school location, school type in school • Total number of students in the theoretical age group for a given level of education who are enrolled in that level, expressed as a percentage of the total population in that age group (UIS). Number of students in • Number of students enrolled in college and university programs tertiary education in a given academic year, per 100,000 inhabitants (UIS). Net enrollment rate • By level; if net enrollment rate (NER) is not available, gross enrollment rate (GER) • Total number of students in the theoretical age group for a given level of education who are enrolled in that level, expressed as a percentage of the total population in that age group (UIS). Number of out-of- • By level school children • Children in the official primary school age range who are not enrolled in either primary or secondary schools (UIS). 51 Core Indicators Definitions / Notes Dropout rate • Proportion of pupils from a cohort enrolled in a given grade at by grade a given school year who are no longer enrolled in the following school year (UIS) (except for those graduating). Repetition rate • Number of repeaters in a given grade in a given school year, by grade expressed as a percentage of enrollment in that grade the previous school year (UIS). • Administrative data usually calculate the number of dropouts as those individuals who neither transition to the next grade nor are repeaters. But the literature shows that repetition is systematically underestimated, producing overestimates of dropout rates. Significant internal or external migration also poses measurement problems. Students who move are counted as dropouts from their school of origin, but this does not mean that they do not re-enroll in a school at their new destination. Additional Indicators: Private educational • Private educational institutions that are controlled and institutions managed by a non-governmental organization (e.g., a church, a trade union, a business enterprise, or a foreign or international agency), or its governing board consists mostly of members who have not been selected by a public agency (UIS). Net intake rate to Grade • New entrants to Grade 1 of primary education who are of the 1 of primary education official primary school entrance age, expressed as a percentage of the population of the same age (UIS). New entrants • Students who, during the course of the reference school or academic year, enter a program at a given level of education for the first time, irrespective of whether the students enter the program at the beginning or at an advanced stage of the program (UIS). Note: The UIS Glossary includes terms related to education, science, technology and innovation, culture, and communication and information, and can be found at http://uis.unesco.org/glossary. 52 Education Public Expenditure Review Guidelines Technical Note 8: Definition of source of funds Sources of funds include public, private, and international sources (UIS). Private entities include households, firms and business enterprises, and nonprofit organizations (including religious organizations) which, although their principal activity is non-educational, might finance activities in the domain of education (UIS). A useful source for the relative proportions of public and private expenditure on educational institutions by level of education is OECD, Education at a Glance (2016), Table B3.1a. Technical Note 9: School fees and other informal payments The burden on households as a result of fee payments can be significant. For instance, the demand response in countries that have abolished fees at the primary level (e.g., Malawi, Tanzania, Kenya, Uganda, and Cameroon) is strong evidence that tuition fees curtail demand. As Kattan and Burnett (2004) point out, in addition to tuition fees, households frequently face a wide range of user fees for publicly provided primary education, including textbook fees, compulsory uniforms, parent-teacher association (PTA) dues, and various special fees, such as exam fees and community contributions to district education boards. In many countries, private tutoring adds to the household costs of primary education. Typically, the poorer the family, the greater the burden of education spending. In Thailand, for instance, poor households spend 47 percent of their consumption on education, while the average for all households is 16 percent. The World Bank conducted surveys on user fees in 2001 and 2005, with a focus on fees in primary education.29 The 2005 survey included questions about lower secondary education, as well as primary education fees. For the primary school level, the Bank collected data for 79 countries in 2001 and 93 countries in 2005 (out of a total of 144 World Bank client countries). In 2005, it collected data at the lower secondary level for 76 countries. The 2005 survey results showed that user fees were common at the primary level. Of the 93 countries surveyed, only 16 countries had no fees. Five countries had all five types of fees (tuition, textbook, uniforms, financial contributions, and other school-based activity fees): the Dominican Republic, Haiti, Honduras, Indonesia, and Vietnam. In 59 countries, or 62 percent of those surveyed, national policy did not address the elimination of fees. More fees were collected at the lower secondary level (94 percent of surveyed countries) than at the primary level (81 percent of countries). In addition, fee levels were generally significantly higher in lower secondary than at the primary level. Of the 76 countries for which responses were received regarding secondary school fees, 94 percent reported that at least one type of fee was collected. Most countries generally had several types of fees; just 10 countries required only PTA or community contributions. Fifty-one countries had more than one type of fee in lower secondary education, and seven countries had all five types of fees for lower secondary education: Bangladesh, Bolivia, Honduras, Indonesia, Lesotho, Mozambique, and Uganda (Kattan 2006). 53 Informal payments in education are charges for education services or supplies that are meant to be provided for free or are paid “under the table” directly to public officials or teachers to obtain specific favors. These are generally measured as the fraction of survey respondents reporting that they made payments to a public education entity for education services intended to be free of charge. Household surveys and perception surveys of citizens and public officials are the most common sources of information. More detailed surveys may also include data on the average value of payments made, the recipients of the payments, and the specific services for which the payments were made. Types of informal payments include, but are not restricted to, payments for admission, advancement, preferential access to resources, and specific grades. Data on informal payments in education are increasingly being collected, but household surveys vary in whether, and how well, they measure informal payments (Lewis and Pettersson 2009). Technical Note 10: Calculating the share of a nation’s resources going to education Three options are gross domestic product (GDP), gross national product (GNP), and gross national income (GNI). They differ, sometimes noticeably, depending on the country’s economic arrangements. All three measures reflect the national output and income of an economy. The main differences are that gross national product takes into account net income receipts from abroad (gross domestic product + net property income from abroad). In other words, GNP measures the value of all goods and services produced by nationals whether in the country or not. Net income from abroad includes dividends, interest, and profit. Gross national income is based on a principle similar to gross national product; the World Bank defines GNI as “the sum of value added by all resident producers plus any product taxes (minus subsidies) not included in the valuation of output, plus net receipts of primary income (compensation of employees and property income) from abroad” (see http://data.worldbank.org/indicator/NY.GNP.ATLS.CD). The World Bank now uses GNI to classify economies into income groupings. Gross national income is possibly a better metric for the overall economic condition of countries whose economies include substantial foreign investments. However, major comparative databases for education use gross domestic product in their financial measures, such as unit costs. These include OECD’s Education at a Glance and UNESCO’s Institute of Statistics. Although gross national income might be more accurate for certain types of countries (e.g., China), losing comparability to data in the major education databases is of real concern. 54 Education Public Expenditure Review Guidelines Technical Note 11: Public expenditure and financial accountability (PEFA) country reports PEFA is a methodology for assessing public financial management performance. It identifies 94 characteristics (dimensions) across 31 key components of public financial management (indicators) in seven broad areas of activity (pillars). The program provides a framework for assessing and reporting on the strengths and weaknesses of public financial management (PFM) using quantitative indicators to measure performance. The purpose of a good PFM system is to ensure that the policies of governments are implemented as intended and that they achieve their objectives. An open and orderly public expenditure and financial accountability system is one of the enabling elements needed for desirable fiscal and budgetary outcomes. Public expenditure and financial accountability country reports provide a snapshot of a country’s performance in this area at specific points in time using a methodology that can be replicated in successive assessments, giving a summary of changes over time. Technical Note 12: Public Expenditure Tracking Surveys (PETS), Quantitative Service Delivery Surveys (QSDS), and Service Delivery Indicators (SDI) Good public expenditure management requires attention to the level of aggregate spending, allocation of public funds, and actual service delivery. The Bank’s public expenditure reviews focus on the first two issues, but the third tends to receive less attention primarily because of a lack of relevant data. Public spending data, irrespective of category, tends to be a poor proxy for actual service delivery. Lack of information on actual delivery also creates an identification problem when the efficacy of public capital or services needs to be evaluated. The case where public capital or services actually created by public funds are highly productive, but the supply system is not, cannot be distinguished from the case where the supply system is effective, but the goods and services being produced yield few benefits. PETS, QSDS, and SDI are designed to provide the missing information from different tiers of government and frontline service facilities, using the sample survey approach. The former collect data at each tier of government to create a picture of how funds and other resources are flowing and where they may be leaking.30 The latter collect a wide variety of information from schools and other sources to answer a range of questions about service delivery. Methodology As their names imply, both of these tools rely primarily on surveys and require the same technical expertise as that required by any properly conducted survey. This includes skills in constructing sampling; designing and pretesting survey instruments; training and monitoring the performance of data collectors; establishing efficient routines for cleaning the data; and analyzing the data. The QSDS, especially, also uses public accounts sample data, preferably panel data, on government spending and information on outputs of service providers at ministerial, regional, local, and service-provider levels. 55 Public Expenditure Tracking Survey A Public Expenditure Tracking Survey is useful when a resource has to travel from one source to the beneficiary. An example is the money that a secondary student earns after school to help pay his school fees, although even here parents can divert the money to buy goods and services for the household, or even to gamble or drink. Even when the source and beneficiary are close to the same, as when PTAs raise funds for particular uses by their schools, the money can be diverted through corruption or used for unintended purposes. As the chain between the initial source and the ultimate beneficiary lengthens, the chances that the resource will fail to reach the beneficiaries increase. For example, the chain from the procurement of textbooks by the central level to students in schools can be long, with substantial loss through poor storage, lack of distribution from subnational storehouses and school storerooms, or diversion for sale on the private market.31 The first Uganda education Public Expenditure Tracking Survey developed the use of surveys to track resource flows (Reinikka and Svensson 2004). The survey collected data from central ministries, local governments (districts), and schools. It tracked the delivery of capitation grants to cover schools’ non- wage expenditures. Using panel data from a unique survey of primary schools, the study assessed the extent to which the grant actually reached the intended end-user (schools). The survey data revealed that during 1991–95, the schools, on average, received only 13 percent of the grants. Most schools received nothing. The bulk of the school grants were captured by local officials. The data also showed considerable variation in grants received across schools, suggesting that some schools used their bargaining power to secure greater shares of funding, rather than being passive recipients of flows from the government. Specifically, schools in better-off communities managed to claim a higher share of their entitlements. As a result, actual education spending, in contrast to budget allocations, proved regressive. The Public Expenditure Tracking Survey for Peru’s “glass of milk,” or Vaso de Leche, program presents an excellent example of how the flow of complex resources can be traced through multiple levels (World Bank 2002). This program distributed to low-income families milk in any form, milk substitutes, or other foods such as soybean, oatmeal, quinoa, and kiwicha. Its main goal was to improve the nutritional level of infants, small children (school age or younger), and pregnant or breastfeeding mothers, and to improve the quality of life of the poorest segments of the population. The project measured leakages through surveys at each level in the transfer process from the central authority down to the household. Specifically, it traced the flow and leakage of central funds from the top of the chain to the last link at the bottom, by using survey data at the level of the municipality, at the level of the local milk-distribution committees, and, finally, at the level of the beneficiary household. The methodology is very complex, not only because it involved multilevel comparisons, but because the input itself was transformed from cash to commodities as the funds moved from the top to the bottom, and as “the commodity” itself actually became commodities, since the program was not limited to milk or milk products alone. The product was then transformed at the household level, as the food products were mixed with other foods before being served. Yet, despite this complexity, the survey was able to determine the relative magnitude of leakages at each level. 56 Education Public Expenditure Review Guidelines The tracking survey found that targeted beneficiaries received, on average, the equivalent of 29 cents of each dollar initially transferred by the Central Government. This does not mean that 71 cents from each dollar were fully lost in corruption costs. Rather, the diverted resources leaked away through a combination of factors: off-budget administrative costs; expenditure on non-eligible products; in-kind deliveries to non-beneficiaries; fees for overpriced items; and, last but not least, sheer corruption. Surprisingly, the leakages were much higher at the bottom levels (the Vaso de Leche committees and households) than at the top (central government and municipalities). The findings turn on its head the conventional belief that every local body is necessarily more accountable than the national and public authorities. In addition, the Peru case highlights the importance of good program design and of transparent and accountable local organizations. The relationship between the direct beneficiaries and the local Vaso de Leche Committees32 has two important features. First, final beneficiaries had limited information about decisions made by the committees, the amount of resources to which households were entitled, and effective ways to secure the resources. Second, the committees operated without transparency or accountability to either the beneficiaries or the upper echelons of government. In fact, the committees so dominated the running of the program at the local level that they could divert resources from their original purpose, without being held accountable or sanctioned for doing so, since neither the higher authorities nor the intended beneficiaries knew about it. The committees then distributed the resources at their own discretion and sometimes ended up diluting, even unwittingly, the program’s expected effects. These findings highlight the design flaws of the program. The local committees were not accountable to the beneficiaries, and they were frequently dominated by rent-seeking “representatives” of the program beneficiaries. Upon reaching the households, the resources sometimes underwent additional dilution. On average, target beneficiaries only received 41 percent of the ration that arrived at the household (not taking into account all of the losses associated with earlier leakages). This dilution effect was possible because, in most cases, the beneficiaries did not receive their rations directly from the committee; rather, the children received the rations through their mothers (or, in some cases, their fathers), who picked up the total rations allocated to the household. Consistent with evidence from studies of other nutritional assistance programs worldwide, the official distribution criteria are very difficult, if not impossible, to enforce. In most cases, it is impossible to exclude non-targeted members of the household from utilizing the resources. Furthermore, in about 60 percent of the committees visited, the products were distributed in unprepared forms. These unprepared products are frequently mixed into dishes that feed the whole family. 57 Quantitative Service Delivery Surveys A Quantitative Service Delivery Survey examines the efficiency of public spending and incentives, and various dimensions of service delivery in provider organizations, especially at the level of the service facility. It measures the resources delivered, such as the actual hours of instruction per day that teachers provide, the student-textbook ratio at the classroom level, the receipt of fiscal transfers from national and subnational levels of government, the size and use of the school’s own revenues, and so forth. It also quantifies the factors affecting the quality of service, such as incentives, accountability mechanisms, and the relationship between agents and principals. Typically, the facility or frontline service provider is the main unit of analysis, much in the same way that the firm is the unit of observation in enterprise surveys and the household in household surveys. In each case, the surveyor collects quantitative data both through interviews and directly from the service provider’s records. Facility data can be “triangulated” by also surveying local governments, umbrella non-governmental organizations, and private provider associations. The compilation of facility-level quantitative data typically requires much more effort than, say, a perception survey of service users, which makes this type of survey more costly and time- consuming to implement than its qualitative alternatives. The Zambia Public Expenditure Tracking Survey and Quantitative Service Delivery Survey illustrates the types of data (Table TN2) and the range of survey modules (Table TN3) used to produce a Quantitative Service Delivery Survey. Table TN2. Data Source Data Description PETS-QSDS 2014 See sub-section PETS-QSDS 2014 NAS 2014 National assessment of Grade 5 students and teachers LCMS 2010 and 2015, GER, NER, and out-of-school children and  ZDHS 2013-14 ESB 2013 Enrollment and school numbers, repetition and dropout rates, teacher statistics Yellow and Bluebooks Government financial statement C and budget Interviews and Meetings GoZ officials and CPs Source: World Bank 2015, table 2.1, p.16. Note: PETS = Public Expenditure Tracking Survey; QSDS = Quantitative Service Delivery Survey; NAS = National Assessment Survey; LCMS = Living Conditions Measurement Survey; ZDHS = Zambia Demographics and Health Survey; GER = gross enrollment rate; NER = net enrollment rate; ESB = Education Statistical Bulletin; GOZ = Government of Zambia ; CPs = Cooperating Partners. 58 Education Public Expenditure Review Guidelines Table TN3: Survey modules, contents, and respondents Data Primary Respondent/Source Description Teacher List Supervisor based on registry Listing of all teachers (for grades 2, 5, 7, 9 (teacher register) and 11) and their information. Student and Supervisor A list sample students and teachers, and Teacher Selection mapping of students and teachers Module Teacher Head teacher based on Basic teacher information and teacher Attendance I registry (attendance book) attendance for all teachers listed (First Visit) General School, Head teacher based on registry General school information, location, Part A (student and teacher registers facilities, enrollment, and repetition and attendance books) General School, Head teacher/financial School financing, fund flow, expenditure, Part B administrator based on and decision-making registry (accounting books) Head Teacher, Head teacher Head teacher information Part A Head Teacher, Head teacher Head teacher personality and motivation Part B Teacher, Sample teacher Detailed teacher information and Part A (up to three teachers) characteristics Teacher, Sample teacher Teacher personality and motivation Part B (up to three teachers) Student, Sample students Detailed student information and Part A (up to 20 students) characteristics for selected students Student, Sample students Student personality and motivation Part B (up to 20 students) (Table continued on next page) 59 Data Primary Respondent/Source Description Household Parents of sample students Household demographics, education, and economics status Classroom Observer Observation of classroom of Grade 5 Observation sample teachers Teacher Observer Second unannounced school visit to check Attendance II attendance of 10 random, sample teacher (Second Visit) PEO Supervisor PEO office information and PETS DEBS Supervisor DEBS office information and PETS Grade 9 Student Sample students selected for Grade 9 student assessment Assessment interview module 7 Grade 9 Teacher Sample teachers selected for Grade 9 teacher assessment Assessment interview module 6 Source: World Bank 2015, table 2.2, p.16. Note: DEBS = District Education Board Secretaries; PEO= Provincial Education Office. Another good example of a Public Expenditure Tracking Survey and Quantitative Service Delivery Survey is Philippines: Assessing Basic Education Service Delivery in the Philippines: Public Education Expenditure Tracking and Quantitative Service Delivery Study. Service Delivery Indicators (SDI) These data are used to assess the quality and performance of education (and health) services over time. Decision makers can use these data to track progress, and citizens can use them to hold governments accountable for public spending. The indicators are broken down into three categories: (i) provider competence and knowledge; (ii) proxies for effort; and (iii) availability of key infrastructure and inputs. The indicators are quantitative and ordinal in nature (to allow cross-country and country specific comparisons).33 60 Education Public Expenditure Review Guidelines Technical Note 13: Definitions and notes on monetary-input indicators Core Indicators Definitions / Notes Public expenditure OECD, Education at a Glance (2016), Table B4.1 on education Public expenditure on education covers direct public expenditure on educational institutions, as well as public support to households (e.g., scholarships and loans to students for tuition fees and student living costs) and to other private entities for education (e.g., subsidies to companies or labor organizations that operate apprenticeship programs). It includes expenditure by all public entities, including ministries other than ministries of education; local and regional governments; and other public agencies (OECD 2016, 228–9) Public expenditure UIS (EdStats) on education as a OECD, Education at a Glance (2016), Table B4.2 percentage of total public expenditure and As defined above, “public expenditure on education” includes as a percentage of GDP, public subsidies to households for living costs, which are not spent by level of education in educational institutions. Therefore, the figures presented here exceed those on public spending on education institutions found in Table B2.3 below (OECD 2016, 231) Government (i) Average total general government expenditure (current, capital, expenditure per student and transfers) per student in the given level of education, expressed (i) as a percentage of per as a percentage of per capita GDP (UIS) capita GDP; (ii) in U.S. http://data.worldbank.org/indicator/SE.XPD.PRIM.PC.ZS (primary); dollars; and (iii) in local http://data.worldbank.org/indicator/SE.XPD.SECO.PC.ZS?end=201 currency, by level of 1&start=2011&view=chart (secondary); education http://data.worldbank.org/indicator/SE.XPD.TERT.PC.ZS? end=2011&start=2011&view=chart (tertiary) (ii) Average total general government expenditure (current, capital, and transfers) per student in the given level of education, expressed in nominal U.S. dollars at market exchange rates (UIS). Wils (2015) and UNESCO (2015d) project per student spending by country (iii) Average total general government expenditure (current, capital, and transfers) per student in the given level of education, expressed in local currency (UIS) (Table continued on next page) 61 Core Indicators Definitions / Notes Expenditure on UIS (EdStats) educational institutions OECD, Education at a Glance (2016), Table B2.3. as a percentage of GDP, by source of funding (public and private), and by level of education Annual expenditure per OECD, Education at a Glance (2016), Table B1.1. student by educational institutions for all In equivalent U.S. dollars, converted using purchasing power parity services, by level (PPP) for GDP, by level of education, based on full-time equivalents of education Note that this OECD indicator includes private spending on tuition fees that occur at educational institutions. In reality, it might be difficult to include private spending at educational institutions when calculating per student spending Per student spending can be broken down by subnational unit if the budget data and education statistics are available at subnational level. If school-level budget data and education statistics are available, and schools’ locations (typically, urban or rural) can be identified, per student spending by geographical locations can be computed. If schools may be categorized by language of instruction, or religion, breakdown by those categories can be done Annual expenditure per OECD, Education at a Glance (2016), Table B1.4. student by educational institutions for all services, In percentage of per capita GDP relative to per capita GDP, by level of education Relative expenditure per Per student spending can be expressed relatively between levels of student (unit cost) by education. The expenditure for primary is treated as the "base case". level of education The expenditure for each other level/type of education is expressed as a percent of the base case. E.g., primary 1.0; secondary 1.2 (20 percent higher than primary); tertiary 1.5 (50 percent higher than primary). This indicator can show the relative cost of different levels of education easily Public expenditure on OECD, Education at a Glance (2016), education by economic Table B6.1. Share of current and capital expenditure by education level classification Table B6.2. Share of current expenditure by resource category (compensation of teachers, compensation of other staff, and other current expenditure) (Table continued on next page) 62 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Capital expenditure: expenditure for education goods or assets that yield benefits for a period of more than one year. It includes expenditure for construction, renovation and major repairs of buildings, and the purchase of heavy equipment or vehicles. It represents the value of assets acquired or created–i.e., the amount of capital formation–during the year in which the expenditure occurs (UIS). Current expenditure: expenditure for educational goods and services consumed within the current year and which would have to be renewed if needed in the following year (UIS). Ideally, salary data should include all “staff compensation” as defined below: expenditure on teaching staff and non-teaching staff salaries; contributions by employers or public authorities for staff retirement and pension programs, and social insurance; and other allowances and benefits (UIS). Current expenditure other than for staff compensation: expenditure on school books and teaching materials, ancillary services, and administration and other activities (UIS). Teachers or teaching staff: persons employed full-time or part-time in an official capacity to guide and direct the learning experience of pupils and students, irrespective of their qualifications or the delivery mechanism (i.e., face-to-face or distance learning). This definition excludes educational personnel who have no active teaching duties (e.g., headmasters, headmistresses, or principals who do not teach), or who work occasionally or in a voluntary capacity in educational institutions (UIS). Non-teaching staff: persons employed by educational institutions who have no instructional responsibilities. Although the definition can vary from one country to another, non-teaching staff generally include head teachers, principals and other administrators of schools, supervisors, counselors, school psychologists, school health personnel, librarians or educational-media specialists, curriculum developers, inspectors, education administrators at the local, regional, and national level, clerical personnel, building operations and maintenance staff, security personnel, transportation workers, and catering staff (UIS). (Table continued on next page) 63 Core Indicators Definitions / Notes Salary cost of teachers OECD, Education at a Glance (2016), Table B7.1. per student by level of education Teachers’ compensation usually consists of the largest part of education spending, and thus, spending per student. The salary cost of teachers per student is a function of (i) the instruction time of students, (ii) the teaching time of teachers, (iii) teachers’ salaries, and (iv) the number of teachers needed to teach students, which depends on estimated class size (OECD 2015). Differences among countries in these four factors may explain, to a large extent, differences in spending per student. Conversely, a similar level of spending per student may be associated with different combinations of these factors. In other words, governments may be able to improve efficiency by changing combinations of these factors. For a detailed definition of this indicator, see “Box B7.1. Relationship between salary cost of teachers per student and instruction time of students, teaching time of teachers, teachers’ salaries and class size” (OECD 2016, 264). Teacher salary relative OECD, Education at a Glance (2016), Table B7.2. to public sector wages, per capita GDP, and This indicator helps judge the adequacy of teacher salary level those with similar relative to the country’s economic capacity and context. Data from qualifications a labor force or household survey can be used to compare wages among public servants and between similarly qualified individuals in the private and public sectors. It is also important to understand how teacher salaries are set and paid. Teachers’ salary cost per student as a percentage of per capita GDP, by level of education. Teachers’ statutory OECD, Education at a Glance (2016), Table D3.1. salaries, based on typical qualifications, Annual teachers’ salaries, in equivalent U.S. dollars, converted using PPPs at different points in for private consumption. teachers’ careers Teachers' salaries are expressed as statutory salaries, which are scheduled salaries according to official pay scales. The salaries reported are defined as gross salaries (total sum of money that is paid by the employer for the labor supplied) minus the employers' contribution to social security and pension (according to existing salary scales) (OECD). Salaries at starting; after 10 years of experience; after 15 years of experience; at top of scale, by level of education. (Table continued on next page) 64 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Teachers’ actual salaries OECD, Education at a Glance (2016), Table D3.2a. relative to wages of tertiary- educated workers Ratio of salary, using annual average salaries (including bonuses and allowances) of teachers in public institutions relative to the wages of workers with similar educational attainment (weighted average,) and relative to the wages of full-time, full-year workers with tertiary education. Comparison of teachers’ OECD, Education at a Glance (2016), Table D3.3a. statutory salaries, based on typical qualifications Ratio of salaries at different points in teachers’ careers, and salary per hour in U.S. dollars converted using PPPs for private consumption. Ratio of salary at top of scale to starting salary. Salary per hour of net contact (teaching) time after 15 years of experience (in U.S. dollars). Technical Note 14: Calculating unit costs from aggregate and itemized spending Mingat, Tan, and Sosale (2003) identify two basic ways to compute unit costs. Both should yield consistent results. In one approach, unit costs are calculated by dividing aggregate spending, such as that reported in budget documents, by the number of students. This is easy to implement, but the method can have problems. First, the aggregate data may be organized under rubrics that prevent clear-cut attribution of spending by level or type of education. For example, administrative expenditures may appear as one entry, with no distinction by level of education. A second problem is that the aggregate data may be organized according to sources of funds or to the structure of the government bureaucracy. As a result, expenditure for a given level of education may appear in several places in the budget, possibly without any detail regarding functional categories. For example, in some countries, the budget documents show spending supported by external donors separately, even though for analytical purposes, the expenditure may belong in the same category as the government’s own spending. In addition, the data may not distinguish between capital investments and recurrent spending, making it difficult to compute meaningful indicators of costs. Given the potential shortcomings and incompleteness of the foregoing approach, it is useful to check the estimates against those obtained through another approach, namely, by building up from the constituent parts of costs. In this approach, the cost components are identified, evaluated, and then aggregated to obtain the desired estimates. In primary education, for example, teachers and pedagogical materials are two of the main components of costs. Thus, the unit cost of these components is calculated separately and then added together to obtain an estimate of the overall unit cost of primary education. Furthermore, instead of dividing aggregate spending on each component by the number of students, other data can be used to make the estimates. For example, to obtain the unit cost of teacher inputs, we would use data on teacher salaries and pupil-teacher ratios. This approach yields more detailed analysis of education costs and provides a basis for simulating the cost implications of alternative choices in the delivery of education services. 65 Technical Note 15: Calculating unit costs when country's fiscal year and school year do not coincide Although countries can differ in their definition of their fiscal year, it usually runs January 1 to December 31. The school year is almost always split across fiscal years. The unit cost can be calculated for a school year or for a fiscal year. However, the school year is usually the desired estimate. Take this example: The fiscal year runs from January 1 to December 31. The school year is from September to June (where teachers are paid for 10 months), and September to August (where teachers are paid for 12 months). Assume that in Fiscal 2001, total expenditures in primary education are $100; in Fiscal 2002, $110. In School Year 2001/02, the total number of students enrolled in primary education is 15 students. Figure 1 shows how the school year splits across the fiscal years. Figure TN1: Unit costs for one school year and two fiscal years MONTH 1 2 3 4 5 6 7 8 9 10 11 12 FY 2001 expenditures for primary education: US$100 School year 2001/02: primary FY 2002 expenditures for primary education: US$110 Enrollment: 15 students The unit costs for School Year 2001/02 are: - 0.4 of US$100 + 0.6 of US$110)/15 students = US$7.07/per capita (where teachers are paid for 10 months a year) - 0.4 of US$100 + 0.8 of US$110)/15 students = US$8.53/per capita (where teachers are paid for 12 months a year) Technical Note 16: Definitions and notes on non-monetary input indicators Core Indicators Definitions / Notes Human resources inputs Pupil/student-teacher Average number of pupils per teacher at a given level of education, ratio (PTR or STR) by based on headcounts of both pupils and teachers (UIS) level, geographical location, and group Governments may have a policy defining a minimum or maximum of schools student-teacher ratio. If this ratio continues declining, the government may need to require or incentivize subnational governments or schools to reduce the number of teachers to maintain the target size (Table continued on next page) 66 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Percentage of qualified A qualified teacher is one who has the minimum academic teachers qualifications necessary to teach at a specific level of education in a given country. This is usually related to the subject(s) they teach (UIS). Ratio of teachers to non- Whereas PERs tend to focus on efficiency of distribution of teachers, teaching staff that of non-teaching staff can be also important in countries where there is a generous supply of non-teaching staff.  Non-teaching staff per Some countries define norms for non-teaching staff per school size type of school (number of students), school area, type of school, etc. Those norms may be very generous, indicating possible room for efficiency gains. Organization of teachers’ OECD, Education at a Glance (2016), Table D4.1 working time   The proportion of statutory working time spent teaching provides information on the amount of time available for non-teaching activities, such as lesson preparation, correction, in-service training, and staff meetings. A large proportion of statutory working time spent teaching may indicate that less time is devoted to tasks such as assessing students and preparing lessons. It also could indicate that teachers have to perform these tasks on their own time and, hence, to work more hours than required by statutory working time. Actual teaching time is the annual average number of hours that full-time teachers teach a group or class of students, including all extra hours, such as overtime. Statutory teaching time is defined as the scheduled number of 60-minute hours per year that a full-time teacher teaches a group or class of students, as set by policy, teachers’ contracts of employment, or other official documents. Teaching time can be defined on a weekly or annual basis. Annual teaching time is normally calculated as the number of teaching days per year, multiplied by the number of hours a teacher teaches per day (excluding preparation time and periods of time formally allowed for breaks between lessons or groups of lessons). At the primary level, short breaks between lessons are included if the classroom teacher is responsible for the class during these breaks. (Table continued on next page) 67 Core Indicators Definitions / Notes Classroom inputs Average class size The average class size refers to the number of enrolled students divided by the number of classes for the whole country. To ensure comparability among countries, special-needs programs are excluded. Data include only regular programs at primary and lower secondary levels of education and exclude teaching in subgroups outside the regular classroom setting (UIS). Governments may have a policy defining a minimum and maximum for average class size. If this ratio continues declining, the government may need to require or incentivize subnational governments or schools to consolidate schools and classes to maintain the target size. Student-textbook ratio This indicator may be important in countries where textbook supply is an issue. Percentage of schools Minimum standards for educational inputs may include standards on meeting minimum prescribed textbooks, and access to computers for learning purposes.  standards requirements for  educational inputs Annual instructional This consists of the required number of hours of instruction per time year. It is not the same as hours in school because those hours may include lunch and inter-class breaks. What is the required number of instructional hours per week for core subjects by grade? For vocational and educational training, what are theoretical and practical instruction hours?  Infrastructure inputs Average school size This indicator may suggest a possible need to rationalize the school network in some parts of the country. It is often observed in countries where the school-age population is decreasing, or where rapid urbanization is taking place so that rural schools are losing students. Factors that influence a decision on whether to build a school include standards for maximum distance that primary- and secondary-level students should walk to school, and for minimum population in the catchment area required to establish a school. (Table continued on next page) 68 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Unit cost of building The unit cost is primarily determined by construction materials, a classroom school design, and equipment. Vocational education and training programs require occupation-specific workshops and equipment. Standards and schedule for It is essential that the education budget include facility maintenance infrastructure maintenance and is fully disbursed for maintenance purposes. Percentage of schools Multiple-shift schools may be prevalent where there are not enough that run double shifts, school buildings to offer single-shift schools. It is an efficient way triple shifts of using existing infrastructure, but not ideal, particularly if a school runs more than two shifts. Percentage of schools Multi-grade schools may be used to manage small student that use multi-grade populations in rural areas. It is efficient to combine grades, but classrooms teacher training on multi-grade teaching is essential to provide good-quality teaching. Without proper training, such an approach may lead to poor-quality education and outcomes. Percentage of schools Minimum standards for learning conditions may include standards on meeting minimum potable water, functional hygienic facilities, electricity, and libraries. standards requirements for learning conditions Technical Note 17: Definitions and notes on research indicators Core Indicators Definitions / Notes Doctorate productivity The number of doctorate degrees, relative to the number of full-time equivalent (FTE) academic staff. Research publications The number of research publications attributed to the department (absolute numbers) and that are indexed in the Web of Science Core Collection database, where at least one author is affiliated with the source university or higher-education institution. Citation rate The average number of times the department’s research publications from the period 2010–13 are cited in other research (published in 2010–15), adjusted (normalized) at the global level for the field of science, and the year in which a publication appeared. (Table continued on next page) 69 Core Indicators Definitions / Notes Top-cited papers The proportion of the department’s research publications that, compared to other publications in the same field and in the same year, belong to the top 10 percent of most frequently cited papers. Interdisciplinary Percentage of department’s research publications that belong to the field’s publications top 10 percent of publications with the highest interdisciplinary scores. Research orientation The degree to which research in the field informs the education offered of teaching by the institution (based on a survey of students in the program). Postdoctoral positions The number of postdoctoral positions relative to the number of academic staff who work full-time or equivalent. External research income Revenue for research that is not part of a core (or base) grant received from the government. Includes research grants from national and international funding agencies, research councils, research foundations, charities, and other nonprofit organizations. Measured in €1,000s, using Purchasing Power Parities (PPP). Expressed per full-time equivalent academic staff. Research publications The number of research publications indexed in the Web of Science (size-normalized) database, where at least one author is affiliated with the university, relative to the number of students. Publication output Number of all research publications included in the institution’s publications databases, where at least one author is affiliated with the institution (per full-time equivalent academic staff) Art-related output The number of scholarly outputs in the creative and performing arts, relative to the full-time equivalent number of academic staff. Citation rate The average number of times the university’s research publications are cited in other research (during 2011–14); adjusted (normalized) at the global level to take into account differences in publication years and to allow for differences in citation customs across academic fields. (Table continued on next page) 70 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Interdisciplinary Extent to which reference lists of the university’s publications reflect publications cited publications in journals from different scientific disciplines. Research publications The number of university research publications (indexed in the Web (absolute numbers) of Science Core Collections database), where at least one author is affiliated with the source university or higher-education institution. Technical Note 18: Efficiency of the curriculum Focused on a few, versus many, subjects. A curriculum fragmented across multiple subjects increases the number of different subject-matter specialists required. In this situation, teachers are less apt to be deployed efficiently, either because they are assigned classes out of their specialty (competence mismatches), or because specialty teachers teach fewer than the usual hours. A fragmented curriculum also increases the textbook varieties required, and disperses students’ learning instead of focusing it on a few core subjects. Focused on a limited, versus large, number of topics in each subject. A curriculum can cover a large number of topics in each subject, but inevitably, superficially. This approach usually necessitates revisiting the topic over several grades. Or a curriculum can address only a few topics per subject, but in depth, and then leave the topic. Studies show that the latter approach is much more efficient in terms of the use of instructional time and student's learning. Technical Note 19: Demographic trends and enrollment projections A country’s demographic structure and trends significantly affect the sustainability of education spending. When examining the demographics data, it is important to understand what trends in the school-age population may imply about the inputs required now and in future. If the population is trending down, do trend data on the size of the teaching force and number of classrooms in use indicate that the government is downsizing inputs into the system—for example, reducing the number of teachers, or closing schools or classrooms? Population data by single-year or five-year age groups—preferably for male, female, and total between 0 and 29 years of age, for the last five years and as projected for the next decade—will be useful. Population data is usually available in the government’s population census statistics book (typically, General Statistics Office), or UN World Population Prospects. EdStats also has this kind of population data in the Core Indicator Query. 71 Technical Note 20: UNESCO’s Education Policy and Strategy Simulation Model (EPSSim) EPSSim is a sector-wide and goals-based generic model comprising the common features of all modern school systems, together with a set of optional components to be included or excluded as required to achieve a first approximation of a country’s education system development. It is driven by demographic trends; enrollment targets are taken as a priori and the simulation calculates the corresponding financial-resource implications. At the national level, EPSSim can accompany countries through all stages of the strategic planning and management cycle. The model was conceived with a view to providing technical and methodological support to national administrations and specialists in education ministries in their efforts to formulate credible education development plans and programs, including in the context of the Education for All (EFA) goals. The aim of the model is to provide a potentially self-contained package (including self-training features) that can be deployed by country planners without external support and can be adapted within a typical range of structural alternatives with minimal expertise. For instance, the model contains some built-in training modules which provide exercises on the key indicators used in the model and enable users to conduct a simplified simulation using hypothetical data. EPSSim starts by computing the projected intake, enrollment, and flow rates on the basis of population data, enrollment status, and policy objectives. The number of enrollments by level and type of education, combined with the current and future modalities of resource utilization (teaching staff, equipment, infrastructure, etc.), enable the estimation of future requirements for teachers, non- teaching staff, instructional materials, educational facilities, and so forth. These projected requirements, together with cost-related data and hypotheses, provide information on financial requirements and the possible financing gaps associated with certain education policy goals. The EPSSim software can be downloaded from UNESCO’s Inter-Agency Network on Education Simulation Models (INESM) website. For EPSSim file downloads, see: http://archive.is/U6qrO. For user guidance and more information on the model, see: http://unesdoc.unesco.org/images/0022/002201/220198E.pdf. Technical Note 21: Simulating the economy-wide effects of alternative education scenarios Appropriately designed Computable General Equilibrium (CGE) models may be the best tool for assessing the broad effects of alternative education scenarios. If such a model is used, it is essential that the Computable General Equilibrium analysis and the rest of the public expenditure review be closely integrated. One such model is MAMS (Maquette for MDG Simulations), developed at the World Bank to assess strategies for achieving the 2015 Millennium Development Goals (MDG) and currently being broadened for the Sustainable Development Goals (SDGs).34 In most applications, MAMS divides 72 Education Public Expenditure Review Guidelines education into three levels (primary, secondary, and tertiary). It considers the impact of private and government educational services on educational attainment and views financing of the government budget in the context of the projected growth in domestic tax revenues, foreign aid, and non- education demands on government resources (which sometimes may be reduced thanks to improved government efficiency). The level and distribution of labor market entrants across different educational levels influence production in different sectors, wages, trade, and gross domestic product. Such a detailed, economy-wide assessment of educational policies may have particularly high payoffs in low-income countries that are expected to see sharp increases in enrollment at different levels. In those countries, while government spending on education may increase to meet growing demand for teachers and other education workers, the actual supply of such workers may increase with a time lag, potentially resulting in supply-and-demand mismatches. Depending on the country context and data availability, it may alternatively be better to use a more macro-oriented model that needs only data that are available for virtually any country. Such a model, labeled GEM-Education (the General Equilibrium Model for Education), is currently being piloted. The model splits the economy into two sectors (private and public). Then, it assesses the impact of alternative scenarios for enrollment and government-education services—which are translated into changes in government consumption and investment spending—on a set of key economic indicators. These include: household consumption and other final demands, gross domestic product, poverty, and the government budget. The channels through which different education scenarios impact the economy are: (i) labor productivity (mainly by influencing the educational composition of the labor force); and (ii) government spending, which requires adjustments in financing from some combination of domestic and foreign sources (taxes, foreign grants, and foreign and domestic borrowing) (World Bank 2013). 73 Technical Note 22: Definitions and notes on output and outcome indicators Core Indicators Definitions / Notes Gross enrollment rate Number of students enrolled in a given level of education, regardless of age, expressed as a percentage of the official school- age population corresponding to the same level of education. For the tertiary level, the population used is the five-year age group starting from the official secondary school graduation age (UIS). Net enrollment rate Total number of students in the theoretical age group for a given level of education enrolled in that level, expressed as a percentage of the total population in that age group (UIS). Dropout rate by grade Proportion of pupils from a cohort enrolled in a given grade at a given school year who are no longer enrolled in the following school year, except for those graduating (UIS). Repetition rate by grade Number of repeaters in a given grade in a given school year, expressed as a percentage of enrollment in that grade the previous school year (UIS). Completion of an Participation in all components of an educational program educational program (including final exams, if any), irrespective of the result of any potential assessment of achievement of learning objectives (UIS). Education attainment The highest International Standard Classification of Education rate (ISCED) level of education an individual has successfully completed. This is usually measured with respect to the highest educational program successfully completed which is typically certified by a recognized qualification. Recognized intermediate qualifications are classified at a lower level than the program itself (UIS). Attendance rate In contrast to the enrollment rate, this indicator sheds light on the frequency with which students attend school or education programs. Attendance-rate data are collected in household surveys such as Demographic and Health Surveys (DHS), Multiple Indicator Cluster Surveys (MICS), and Living Standards Measurement Study (LSMS). (Table continued on next page) 74 Education Public Expenditure Review Guidelines Core Indicators Definitions / Notes Labor force participation Employment and wage data for recent graduates are often used to shed light on this question. But be aware that these data reflect Employment and supply-and-demand interactions, not necessarily the adequacy of unemployment rates the skills developed by the educational system. External research Research revenue that is not part of a core (or base) grant income received from the government. It includes research grants from national and international funding agencies, research councils, research foundations, charities, and other nonprofit organizations. Measured in €1,000s using Purchasing Power Parities (PPP). Expressed per full-time equivalent academic staff. Technical Note 23: Concepts of effectiveness and efficiency Concept Definition Effectiveness  When something is deemed effective, it has met an intended, desired, or expected outcome. Unlike the concept of efficiency, effectiveness alone is determined without reference to costs. In popular discourse, effectiveness means “doing the right thing,” and efficiency means “doing the thing right.” When effectiveness is combined with cost, “cost-effectiveness” relates monetary inputs and non-monetary outputs and outcomes. Technical and Technical and productive efficiency are two closely related concepts that are productive often lumped together under the term technical efficiency. In both cases, we efficiency are looking for the best outcomes at least cost, a result that may be achieved in either of two ways: (i) by reducing the inputs used by a given intervention to the minimum required to achieve the outcome sought (technical efficiency), or (ii) by selecting a different intervention—i.e., one with a different combination of resources—that achieves at less cost the outcome sought as well as, or better than, the alternatives (productive efficiency). Technical efficiency refers to the physical relation between a set of resources (capital and labor) and an education outcome. A technically efficient position is achieved when the maximum possible improvement in the outcome is obtained from the resource inputs. An intervention is technically inefficient if the same (or greater) outcome could be produced with less of one type of input. For example, if research shows that a student-textbook ratio of 2/1 results in the same learning gains as a ratio of 1/1, the higher ratio is technically more efficient. Technical efficiency cannot directly compare alternative interventions. Productive efficiency, closely related to the concept of technical efficiency, directly compares alternative interventions. It asks about the cost benefit or cost-effectiveness ratios of resource combination A versus alternative resource combinations. For example, if the objective is to improve teachers’ classroom (Table continued on next page) 75 Concept Definition performance, educators can try alternative training regimes. Teachers can attend a week of training at a teacher college. They can participate with their same-grade or same-subject colleagues in weekly, facilitated training sessions at their schools. They can take online courses at home beamed via the government television station. What does each option cost, and how well does it improve teachers’ classroom performances? Since different combinations of inputs are being used, the choice between interventions is based on the relative costs of these different inputs relative to the same outcome (better teacher classroom performance). Productive efficiency enables assessment of the relative value for money of interventions with directly comparable outcomes. It cannot address the impact of reallocating resources at a broader level—for example, from teacher training to infrastructure—because the education outcomes are incommensurate. Allocative Allocative efficiency takes account of the productive efficiency with which efficiency education resources are used to produce education outcomes and how these outcomes are distributed among the community. Such a societal perspective is rooted in welfare economics. The World Bank defines the concept of allocative efficiency as the capacity of government to distribute resources on the basis of how well its public programs meet its strategic objectives (Schick 1998; Shand 2000). This includes the ability to shift resources from old priorities to new ones, and from less- to more-effective programs. Allocative efficiency thus requires that governments establish and prioritize objectives, and that they assess the actual or expected contribution of public expenditures to those objectives. Shand’s (2000) summary of the basic principles of allocative efficiency states that expenditures should be affordable in the medium term and be based on government priorities and the effectiveness of public programs. Internal Internal efficiency measures the percentage of children who complete an efficiency educational cycle (e.g., primary education or lower secondary education) as a share of those who start the cycle or as a percentage of those who finish the cycle in the minimum number of years. The first definition allows the calculation of the dropout rate—i.e., the number of those who start, minus the number who ultimately complete as a share of those who start. The second definition measures the dropout rate plus the repetition rate. Dropping out imposes costs on individuals and countries in the form of unrealized human capital. Repetition imposes costs on the sector in the form of its having to pay double (or triple) the unit cost of a year of school per repeater. External External efficiency measures the returns to individuals, employers, and the efficiency country of public and private investments in education. It depends on a match between the type and quality of skills and knowledge that school leavers acquire in school relative to the skills and knowledge needed by the country and needed and paid for by employers. Does education improve the employability and wages of school leavers? Does public investment in education and training contribute to the country’s growth and economic development? Measuring and linking employment and wage returns to education is particularly important for vocational education and training, and for tertiary education. 76 Education Public Expenditure Review Guidelines Technical Note 24: Cost-effectiveness analysis Cost-effectiveness relates monetary inputs and non-monetary outputs and outcomes. An increasing number of impact evaluations of various interventions have assessed their cost-effectiveness relative to student learning and other outcomes. For instance, to provide comparable cost-effectiveness estimates across different policy options, the Abdul Latif Jameel Poverty Action Lab (J-PAL) has adopted a standard methodology for conducting cost-effectiveness analysis of randomized trials of 29 programs.35 The trials found that the effectiveness and costs of numerous strategies to improve student learning vary considerably. Some programs achieve learning gains with much greater cost- effectiveness than others. Be alert to the quality of cost-effectiveness analyses. The risks of correlations being interpreted as causal relationships rise substantially when the methodology of the study does not use a randomized controlled trial, a difference in differences (DD) regression design, a regression discontinuity design (RDD), or matching methods. For example, teachers with formal degrees may increase their students’ learning more than teachers without such qualifications. However, teachers who obtain formal degrees may differ in important ways from those who do not, and these differences may account more for their students’ learning gains than the teachers’ formal degrees. Most research studies in the development literature do not protect well against bias and cannot be properly used to draw causal inferences (X causes Y). At the same time, if these are the only studies available, flag the potential for bias, triangulate where possible, and be very careful to state conclusions in correlational, not causal, terms. For critical policy questions, suggest that a small, randomized trial be conducted to sort out correlation from causation. Careful meta-analyses36 of the effects of inputs on participation in school and learning find numerous instances where the “common wisdom” about the effects of inputs is either wrong or fails to take into account the conditions that have to be in place for those effects to occur. Take the example of textbooks. The 3ie study reported a relatively consistent pattern of textbooks having no effect on learning outcomes, as measured by math, language, and composite test scores (Snilstveit 2016). However, the study found that many of these programs experienced implementation challenges—e.g., the books did not reach the students because they were locked up for “safekeeping." For their methodologically less-rigorous sample of 79 studies, Glewwe et al. (2011) found that most studies showed positive effects, and that most of these effects were statistically significant. This evidence strongly suggests that textbooks and similar materials (workbooks, exercise books) increased student learning. However, when the analysis was restricted to the methodologically more-rigorous sample of 43 studies, the estimated effects of textbooks were far from unanimous. Slightly less than half of the estimates showed positive effects, and only three of these were significantly positive (and one was significantly negative). Thus, after dropping less-rigorous studies, the evidence that textbooks and similar materials (workbooks, exercise books) increased student learning was quite weak. 77 The one “gold standard” randomized, controlled trial in the Glewwe et al. (2011) sample found no impact of providing textbooks. However, the authors of that study conducted further analyses to determine why textbooks did not raise scores. The results of the study did not appear to be statistical artifacts. The treatment and comparison schools were similar in geographic location, enrollment, and pre-program test scores. Neither selection nor attrition bias appeared to drive the results. They found that the effectiveness of textbooks depended on prior conditions: textbooks improved the scores of students with higher pre-test scores. In other words, there was an interaction effect between pre-test scores and assignment to the textbook program that had a highly significant, positive correlation with post-test scores. Initially better students could benefit from the textbooks much more than initially weak students. The authors also found that the official Kenyan government textbooks were of limited use to many students. English is the medium of instruction in Kenyan schools, but it was the third language for many pupils, including those examined in this study. The study showed that many students could not read the textbooks. Technical Note 25: Cost-benefit analysis Cost-benefit analysis, or CBA, is used when both the costs and outcomes can be monetized.37 Rate-of-return analysis is the type of cost-benefit analysis most frequently applied to education. Rates of return to investments in schooling have been estimated since the late 1950s. George Psacharopoulos’s (1973) findings on the rates of return to education—that primary education ought to be the main focus of national school systems, since its rate of return was found to be the highest among all education levels—continues to play an important role in the formation of significant global educational policies. In the intervening 40-plus years, however, the economies of many developing countries have matured, changing the relative returns to different levels of education. More recent discussions of the rates of return include Psacharopoulos and Patrinos (2004) and Montenegro and Patrinos (2014). 78 Education Public Expenditure Review Guidelines Technical Note 26: Calculating the budget costs of one completer and of one uninterrupted completer This is a hypothetical example, although it uses real data for dropout rates by primary grade and for average repetition rates for primary education. It shows how to calculate what it costs to get one completer, who may, or may not have repeated one or more grades, and what it costs to get one uninterrupted completer who repeats no grades. The example does not calculate the substantial social and individual costs of dropping out of school, although, by making certain assumptions, this could be done. Table TN4: Progression through primary school of Cohort A that starts with 100 grade 1 entrants Grade 1 2 3 4 5 6 7 8 # Number of new entrants to grade 100 67 50 37 27 20 15 10 326 Percentage of cohort who drop 21 12 11 14 11 11 20 14 out during year Number of students who drop 21 8 6 5 3 2 3 3 51 out during year Total in class at end of year 79 59 44 32 24 18 12 7 Using 15% repetition rate, number that do not move to 12 9 7 5 4 3 2 1 43 next grade Total who progress to next grade 67 50 37 27 20 15 10 6 232 Assumptions The per capita cost of a year of primary school in constant dollars is $50. The per capita cost, in constant dollars, is the same across grade and across time. The percentage of students who drop out by grade is taken from real data for a real country. The average percentage of students who repeat a grade are taken from real data for the same country. A student who repeats a grade repeats that grade only once, although he or she may repeat A student who drops out in a year costs $50, regardless of when he or she drops out in the year, because recurrent costs (e.g., number of teachers) are sunk costs by the time the student drops out. (viii) Once a student drops out, the costs of all post-dropout years for that student are as budget savings to the system. 79 Table TN5: Costs of one completer and one uninterrupted completer Cost and Savings U.S. Dollars If no students drop out or repeat and unit costs are $50 per capita in constant dollars, $40,000 then the cost of a primary education (grades 1–8) for Cohort A is 100 × 8 × 50 Per capita cost of 8 years of education (8 × 50) $400 Total costs of those who start each of the eight grades, whether or not $29,100 they subsequently repeat or drop out (326 students) + repeaters (43). The figure for repeaters includes the additional cost of the repeated year + all remaining years of primary education until completion. Costs of repetition alone (43 years repeated × $50 ÷ year) $2,150 Savings from dropouts. Budgetary costs to a system with high dropout rates $13,050 are lower than to a system with no dropout rates if we assume that when a student drops out, the system saves his or her full, $50÷year cost in each post- dropout grade. Social and individual and household costs, not computed here, are obviously substantial. Cost for one completer: 29,100 ÷ 49. Total completers = 49 (6 uninterrupted $594 completers + 43 completers with one repeated grade). The cost is about 50% more than the cost without repetition, which is 8 × 50, or $400. Cost for one completer without interruption ($29,100 ÷ 6 completers who did $4,850 not repeat any grade). The cost of one uninterrupted completion is about eight times the cost of a completion that includes repeaters. Technical Note 27: Estimating private rates of return to education The private rate of return compares the costs and benefits of schooling as incurred and realized by the individual student who undertakes the investment. The models and methods used to calculate private rates of return to education depend on the policy questions of interest and the quality of the available data. For example, allowing for heterogeneous returns to education across individuals with the same level of education--e.g., as a result of variations in cognitive achievements or regional variations in wage structures--is more data-intensive than assuming homogeneous returns. In comparisons between countries of returns to education, Montenegro and Patrinos (2014) provide guidance for calculating private returns to education. Noting that the now-standard method for estimating private returns per year of schooling is the Mincerian earnings function method, they state that this means estimating log earnings equations of the form: Technical Note 1: Estimating private rates of return to education The Technical private rate Technical Note 1: Estimating of return compares Note 1: Estimating private private rates and benefits the costs of return to education rates of return to education of schooling as incurred and realized by the individual student who undertakes the investment. The models and methods used to calculate private rates of Technical Note 1: Estimating private rates return to education depend on the policy return to of ofquestions education interest and the quality of the available The private rate of return compares the costs and benefits of schoolingacross as incurred and realized by the 80 Education PublicThe data. individual For example, Expenditure private rate of return compares student who undertakes allowing Review for heterogeneous Guidelines the costs and benefits returns to of education individuals with the same schooling as incurred and realized by the level of education--e.g., individual student who undertakes as a result the investment. the investment. The models and methods of variations in cognitive achievements The models and methods used to variations or regional calculate private used to calculate private in wage rates The of private structures--is return to education depend on the rate of return compares more data-intensive than assuming homogeneous the policy costs questions and benefits ofof interest schooling returns. and the as quality of the incurred and realized by available the rates of return to education depend on the policy questions of interest and the quality of the available data. individualFor example, student allowing who undertakes for heterogeneous the investment. returns to education The models andacross methods used to with individuals the same calculate private data. For example, allowing for heterogeneous returns to education across individuals with the same level In rates of education--e.g., of return to education depend on the comparisons between as a result countries of of variations returns tocognitive achievements in policy questions Montenegro education, of interest or regional and the variations quality of the in wage available level of education--e.g., as a result of variations in cognitive achievements orand Patrinos regional (2014) variations inprovide wage structures--is data. For guidance for more data-intensive than example, calculating allowing private assuming homogeneous for heterogeneous returns to returns to education. Noting returns. education that across the nowindividuals standard with method the same to structures--is more data-intensive than assuming homogeneous returns. level of education--e.g., estimating private returns asper a resultyear of schooling in of variations cognitive achievements is the Mincerian earnings function method, or regional variations they instate wage In comparisons structures--is that this means between countries of returns more data-intensive than estimating log earnings equations to education, assuming homogeneous of the form: Montenegro returns. and Patrinos (2014) provide In comparisons between countries of returns to education, Montenegro and Patrinos (2014) provide guidance for calculating private returns to education. Noting that the now standard method to guidance for calculating private returns to education. Noting that the now standard method to estimating In comparisons private between returns per year of of countries schooling returns is tothe Mincerian earnings education, Montenegro function method, !and Patrinos (2014) they providestate estimating private returns that this guidance means calculating (1) for estimating log earnings per year private !of = to schooling equations returns + is the of !the education. ! ! + + !Mincerian earnings form: Noting that! the ! + now !standard function method, they state method to that this means estimating log earnings equations of the form: estimating private returns per year of schooling is the Mincerian earnings function method, they state that this means estimating log earnings (1) ! hourly= equations +annual, of !the+ form: ! + data) ! !! ! + ! for the i individual; where ! is the natural (1) ! = or log (of + ! + ! ! ! depending on ! ! + ! !earnings + ! th is years of schooling (as a continuous variable); ! is labor market potential where Ln(wi) is the natural log (of hourly or annual, depending on experience (estimated as ! ! (1) ! = + ! ! + ! ! + ! !!data) + !earnings for the ith individual; Si agei - ! - 6); ! is potential experience-squared; and ! is a random disturbance term reflecting is yearswhere ! is the natural of schooling a continuous (asTherefore, log (of hourly variable); or annual, X is labor market depending on data) earnings potential for the ith individual; (estimated as age experience where is the natural unobserved abilities. log (of ! can hourly or annual, be viewed as depending on the average rate of earnings data) return to for years the ofithschooling to individual; i ! is 2years of !schooling (as a continuous variable); ! is ilabor market potential experience (estimated as - Si - 6); wage age !X is years where ii - is potential employment. of schooling (as - 6); ! ! experience-squared; The list ! the natural is is of control variable); a continuous potential log (of variables hourly experience-squared; or and is! kept annual, isμ and labor depending on i is a random deliberately market is a random disturbance small potential data) to avoid overcorrecting experience (estimated earnings disturbance for term the term i th reflecting reflecting for as individual; unobserved age factors Therefore, ! i - !that - 6); are !!correlated is potential with years of schooling. experience-squared; and This !!is is alsoa random as the “Mincerian” known disturbance term reflecting method abilities. (Mincer 1974). β1can ! is years of schooling (as unobserved abilities. be a continuous Therefore, viewed as variable); ! can be viewed the average as rate ! is labor the average rate of of return market potential return to to years years ofofschooling schooling to as experience (estimated to wage unobserved abilities. !Therefore, ! can be viewed as the average rate of return to years of schooling to wage - ! - 6); !The ageiemployment. is potential list of control variables is kept and experience-squared; ! is a random deliberately small to disturbance term reflecting avoid overcorrecting for employment. wage employment. factorsunobserved abilities. that are The list of Therefore, The correlated control list of control with years variables ! can variables of is kept schooling. be viewed is kept deliberately deliberately small small to to avoid avoid overcorrecting overcorrecting for for factors that The factors earnings that are function correlated method with years can beof used schooling.to as This is also the average rate of estimate This is also returns known known as return at different the to “Mincerian” years of schooling to schooling as the “Mincerian” method levels method by are correlated (Mincer (Mincer wage converting 1974). with employment. 1974). years The of schooling. list of control This variables is also is kept known the continuous years of schooling variable (S) into a series of dummy variables, say Dp, Ds deliberately as the small“Mincerian” to avoid method overcorrecting (Mincer for 1974). factors and Dt (where that are p is correlated primary schooling, with years of schooling. s is secondary This is also schooling and known t is tertiary) as tothe “Mincerian” method denote the fact that a The (Mincer person earnings has 1974). functionthat achieved method level can be used The of schooling. to estimate omitted level returns is peopleat different with no schooling schooling levels and that by The earnings function method can be used to estimate returns at different schooling levels by The earnings converting dummy is converting function the not continuous the in the method continuous years of can equation to avoid matrix be used schooling years of schooling variable to estimate variable singularity. (S) into (S) intoThe a a returns series of dummyat different estimation equation in this series of dummy variables, variables, schooling say is case say Dp Dp , Ds of the , Ds levels by converting and The Dt (where earnings p is primary function schooling, method can s is secondary be used to schooling and estimate t returns is tertiary) at to different denote the fact schooling that levels a by the continuous form: and Dt (where person converting has achieved years p is primary the continuous ofthat schooling, variable schooling levelyears of schooling. s is secondary of schooling The ( S)omitted variable into schooling anda)level (S series into a t isoftertiary) people is series dummyofwith dummy variables, to denote the fact no schooling variables, and say that say Dp Dp, that a Ds and Dt (where , Ds person has achieved that level of schooling. The omitted level is people with no schooling and that p is primary dummy and Dt isschooling, not in the (where s is secondary schooling, s schooling p is equation to avoid matrix primary singularity. is secondary and The schooling andt is tertiary) to t is tertiary) estimation equation in this denote to denote the fact !the case factis ofthat the a that a person has person has(2) ! =level equation to avoid matrix + of ! ! + ! ! + dummy is not in the singularity. The ! is ! + ! !with estimation equation in this + ! ! + case is of the ! and form: achieved that schooling. The omitted level people no schooling that achievedform:that level of schooling. The omitted level is people with no schooling and that dummy is not in dummy is not in the equation to avoid matrix singularity. The estimation equation in this case is of the the equation form: to (2)avoid matrix = singularity. + The + estimation + equation+ in +this ! + is case ! of the form: (2) ! ! ! ! ! ! ! ! ! ! ! ! ! = + ! ! + ! ! + ! ! + ! ! + ! ! + ! After fitting this “extended earnings function” (using the above dummies instead of years!of schooling in (2) ! private the earnings function), the = + of rate ! ! + ! ! + levels return to different ! of ! + ! ! + schooling ! can be ! + ! the derived from following formulas: After fitting After fitting this this “extended earnings “extended earnings function” function” (using the above (using the dummies aboveinstead of dummies years of schooling instead in of years of schooling in After fitting this “extended earnings function” (using the above dummies instead of years of schooling in the earnings function), the private rate of return to different levels of schooling can be derived from the the earnings the earnings following formulas: After fitting function), function), the private this “extended earnings rate the private rate of (3) return to different function” of return ! = ( (using the ! ) / ( to different levels ! above dummies ) of schooling levels can be of schooling instead of derived fromcan bein years of schooling the derived from the following formulas: following earnings function), the private rate of return to different levels of schooling can be derived from the the formulas: following formulas: (4) (3) ! = (=! − !)) ( //(( !)− ! ) (3) ! ! ! = (! )/(! ) ! (3) (4) ! = (!−=(! )/(! ) ! )/(! − ! ) (4) ! = (! ! − ! )/(! − ! ) (5) (4) ! ! = (! ! − ! )/(!! − ! )/( − ! ) !) - where S , S and St stand p s Sp, Ss and where for the St stand for thetotal number total number of years of years of schooling of schooling for each for each successive successive level. level. Care has Care has to to be taken regarding the foregone earnings of primary school-aged children. In the empirical analysis be taken regarding the foregone earnings of primary school-aged children. In the empirical analysis that follows we have assigned only three years of foregone earnings to this group, following tradition that follows, we have (Psacharopoulos assigned only three 2004). = (!of (5) ! years −foregone ! )/(! − earnings ! ) to this group, following tradition (Psacharopoulos 2004). The costs incurred by the individual are her foregone- earnings while studying, plus any tuition fees or where Sp incidental , Ss and Stincurred expenses stand forduring schooling. the total numberSince of yearsschooling is mostly of schooling provided free by for each successive the Care level. state, at to has The costs incurred least be at the basic by taken regarding the education the individual level, then foregone are earnings his of or her foregone in practice the only primary earnings cost in a private rate of school-aged children. while In the studying, return calculation empirical analysisplus is any tuition fees that the foregone follows we earnings. have assignedThe private benefits only three amount years to whatearnings of foregone a more educated individual to this group, earns tradition (after following or incidental taxes), expenses above (Psacharopoulos incurred a comparable 2004). during group of schooling. individuals with less Since schooling.schooling This more isor mostly provided less refers to adjacentfree by the state, at least at the basic education level, then in practice the only cost in a private rate of return calculation is levels of schooling; for example, tertiary versus secondary school graduates. Although convenient because The costs it requires incurred lessbydata, this method is the individual slightly are her foregoneinferior to the full earnings discount while method studying, (Psacharopoulos plus any tuition fees or the foregone 2004); inearnings. incidental fact, it assumes expenses The private flat incurred benefits age-earnings during amount profiles schooling. Since to what for different schooling levels aof more is mostly educated schooling provided individual (Psacharopoulos free by and at the state, earns (after taxes), above a comparable group of individuals with less schooling. This more or less refers to adjacent levels Layard least 1979). at the basic education level, then in practice the only cost in a private rate of return calculation is the foregone earnings. The private benefits amount to what a more educated individual earns (after of schooling; From for equation taxes), example, above (1) the tertiary versus return to potential a comparable group of secondary experience individuals school is given by: with less schooling. This more Although graduates. to adjacent because it or less refersconvenient requires less data, this method is slightly inferior to the full discount method (Psacharopoulos levels of schooling; for example, tertiary versus secondary school graduates. Although convenient 2004); in fact, because it requires less data, this method is (6) ! + slightly 2! to the inferior full discount method (Psacharopoulos it assumes flat age-earnings profiles for different levels of! schooling (Psacharopoulos and Layard 1979). 2004); in fact, it assumes flat age-earnings profiles for different levels of schooling (Psacharopoulos and Layard 1979). From equation From which (1)to the equation needs return (1) the be toat evaluated potential return to potential experience experience a given value of Xi. is For is given by: given by: each sample we use the average years of potential experience as the evaluation point. It is important to stress that in the empirical part when we (6) refer to potential experience we are referring to the ! ! + 2!based on equation (1). estimates which needs to be evaluated at a given value of Xi. For each sample we use the average years of potential which needs to be evaluated at a given value of Xi. For each sample we use the average years of experience as the evaluation point. It is important to stress that in the empirical part when we refer to potential experience as the evaluation point. It is important to stress that in the empirical part when we experience potentialrefer we to potential are referring experience to the to the we are referring estimates based estimates on equation (1). based on equation (1). 81 Technical Note 28: Per capita financing Per capita formula funding is widely used to help achieve both horizontal and vertical equity by adjusting a standard unit cost (horizontal equity) with coefficients to provide more funding for the disadvantaged (vertical equity). Whether per capita financing is in the form of a funding formula or other funding mechanism, it is important to examine how the unit cost is determined, what factors are taken into account to address vertical equity (e.g., special needs), and how funding is indeed implemented.38 Technical Note 29: Targeting mechanisms, coverage, and depth of programs Are there methods in place to identify the needs of disadvantaged students for education transfers or subsidies, such as scholarships and free textbooks? How are eligible students identified—e.g., projection from historical levels, geographic targeting, analysis of household survey data within the last five years, analysis of individual student data? Is there targeting to certain income, gender, or ethnic and religious groups? When educational merit is the basis, the subsidy or transfer will tend to favor wealthier families because of often observed learning gaps at lower levels of education by income quintile. Per student spending on targeted programs to provide various types of subsidies for disadvantaged students, such as schoolfee discounts or waivers, free lunch, or scholarships, can be separately computed. If education spending is allocated based on a formula, we need to assess whether the formula takes into account vertical equity by setting a higher unit cost (or coefficients) for special needs students, rural schools, and other types of disadvantages. Technical Note 30: Benefit incidence analysis Benefit incidence analysis (BIA) is an analytical tool that helps examine whether the benefits of public expenditure are distributed across population groups by wealth or other socioeconomic or geographic characteristics. This type of analysis can be used to (Narayan 2014): (i) inform policymakers about the current incidence of social spending, i.e., the extent to which different segments of population (e.g., the poor or the rich) are benefiting from the current allocation of social spending, and changes in the incidence of spending over time; (ii) establish a benchmark for cross-country comparison of distribution of public spending on social services; (iii) analyze if specific policy reforms in the past may have accounted for the current observed incidence, or changes in the incidence of spending, over time; and (iv) demonstrate whether a pro-poor benefit incidence is actually translated into better social outcomes for the poor). 82 Education Public Expenditure Review Guidelines Minimum practice—known as average (or simple) benefit incidence analysis—is to report average public education financing per child or per household by household consumption (e.g., quintile, decile), or by other socioeconomic or geographic characteristics (urban vs. rural, male vs. female, etc.). In this type of analysis, benefits are calculated, on average, on the basis of unit costs (aggregate expenditure divided by the number of beneficiaries), multiplied by the number of users of the service in a specific group (quintiles or deciles, etc.). In progressive or “pro-poor” public spending, poorer quintiles or deciles get a disproportionately higher share of the total benefit compared to their share in the national income distribution; for example, the bottom 40 percent receive more than 40 percent of the total funds. The choice of quintile definition (household versus population quintiles) may affect the analysis results due to variation in the numbers of individuals in each quintile. When quintiles are defined over the population, the population size of each quintile is defined to be equal. However, the population size of each household quintile varies, depending on the household-size characteristics of the quintile. Because poor households typically have a larger household size, when the household, rather than the population, is used for the analysis, the distribution of spending appears more progressive than it actually is (Demery 2000). Government subsidies for services may vary significantly by region and across groups. Thus, aggregating unit subsidies may mask inequality. The best benefit incidence analysis practice involves more sophisticated analyses as opposed to the average benefit incidence analysis. These include analyses based on disaggregated unit subsidies (as opposed to average unit subsidies), where specific costs for specific interventions are accounted for separately. Demery (2000) gives examples of using disaggregated unit subsidies specified for different geographical areas and education levels. Figure TN2: Concentration curve: government spending on education and various benchmarks Cumulative share of government spending on education (percent) Pro-poor spending 45-degree line - Equality Progressive Lorenz Curve of income or consumption Regressive (poorly targeted) Cumulative share of population (percent) Source: Narayan (2014) 83 Concentration curve: A concentration curve shows the relationship between the cumulative share of government spending on education (y-axis) against the cumulative share of the population, ranked by consumption or income level from the poorest to the richest (x-axis). Lorenz curve: A Lorenz curve shows the relationship between the cumulative distribution of total household expenditures against the cumulative population, ranked by consumption or income level. Targeting: Public spending is pro-poor if the concentration curve for education benefits is above the 45-degree line; otherwise, pro-rich spending. Progressivity: Public spending is progressive if the concentration curve for these benefits is above the Lorenz curve for income or consumption, but below the 45-degree line. The results of a benefit incidence analysis can be presented in two ways: tabular or graphical (Lorenz or Benefit Concentration Curves). The expenditure Lorenz curve is derived from tracking the cumulative distribution of total household expenditures against the cumulative population, ranked by per capita expenditures. It provides a point of comparison against which to judge the distribution of education spending shown in the concentration curves (Demery 2000). Public spending on education tends to be more pro-poor at lower education levels (e.g., primary), but becomes pro-rich for subsequent levels of education. Two main factors account for this phenomenon: (i) Poorer household quintiles tend to have more children than richer quintiles and may receive a disproportionately higher share of public education resources, at least at the primary level, and possibly at lower secondary level; and (ii) any poor children do not progress to secondary or higher levels of education, where per capita spending tends to be higher and where a disproportionately lower share of resources goes. Other types of benefit incidence analysis include marginal benefit incidence analysis, which estimates the incidence of the last (or the next) unit of benefit, and behavioral benefit incidence analysis, which estimates behavioral responses to a policy change. Narayan (2014) and Demery (2003) provide a good basic explanation of some of the issues and concepts in incidence analysis; Mingat, Tan, and Sosale (2003), in Chapter 7, provide a general discussion on calculating the distribution of public subsidies for education; van de Walle (2003) offers a more advanced treatment of some of the methodological approaches. Other useful information on the topic, including manuals, guidelines and a STATA (data analysis and statistical software) program, which includes a module to perform benefit incidence analysis, can be found at the Open Budgets Portal. 84 Education Public Expenditure Review Guidelines Technical Note 31: Subsidies Why do we care about subsidies for private schools? There is no obvious efficiency rationale for fully funding non-public schools unless they are shown to achieve better outcomes than public schools. Paying the full unit cost saves the state nothing, but partial subsidies can create incentives for private provision that save the state money because it does not have to pay the full cost of educating the child. Depending on how they are designed, public subsidies may implicitly subsidize the wealthy’s preference for private education. PART III: EXAMPLES 86 Education Public Expenditure Review Guidelines Process and criteria for selecting examples The World Bank team examined about 80 education and cross-sectoral public expenditure review documents published between 2010 and 2016. A complete PER database, which also includes those published between 2002 and 2009, is available at http://datatopics.worldbank.org/education/ wDataQuery/ByExpenditures.aspx. The team adopted the following criteria to identify good examples of different topics of reviews: Analysis that is relevant, detailed, and comprehensive Analysis that provides a methodology Analysis that includes international and regional comparisons and benchmarks Analysis based on disaggregated data (if available), including subnational or school-level data, or data on specific student groups Analysis with clear and original presentation of results (graphs, tables, diagrams, etc.) Most recent analysis For good examples of policy recommendations, the team selected reviews that provide clear, prioritized, specific, and implementable, and costed recommendations. The team selected several country public expenditure review documents for more than one example because they show a relatively better treatment of a specific topic, compared to other country reviews. This section will be updated as more good examples become available. Table of Contents Examples No. Topic Example Page Albania (2014); Ethiopia 1 Policy recommendations 88 (2015) Myanmar (2015); Samoa 2 Data sources 96 PER Notes (2014) 3 Analysis of sources of funds Nigeria (2015) 99 4 Analysis of revenue sources Nigeria (2015) 100 5 Analysis of decentralized financing Sudan (2014) 103 Analysis of education financing by level of 6 Indonesia (2013) 104 government and intergovernmental fiscal transfers 7 Analysis of school budget by financing source Philippines (2013) 108 8 Analysis of household surveys on private spending Mali (2016) 112 9 Analysis of donor funding Mali (2016) 115 10 Analysis of total public education spending Albania (2014) 117 Honduras (2015); 11 Analysis of functional classification 118 Kenya (2014) (Table of Contents: Examples continued) 87 No. Topic Example Page Analysis of education spending by economic 12 Honduras (2015) 120 classification 13 Issues related to budget formation and execution Solomon Islands (2011) 121 14 Analysis of per student spending Armenia (2011) 124 15 Minimum norms and standards for resource allocation Belarus (2013); 126 Kosovo (2014); Jordan 16 Analysis of cost of teachers 130 (2016) 17 Analysis of teacher distribution Indonesia (2013) 140 Tajikistan (2013); Jordan 18 Cost projections (2016); Democratic 144 Republic of Congo (2015) Albania Volume II (2014); 19 Fiscal sustainability analysis 149 Madagascar (2015) Guinea (2015); Belarus 20 Demographic trends and enrollment projections 154 (2013); Georgia (2015) 21 Technical efficiency of inputs (efficiency indicators) Belarus (2013) 161 Albania (2014); Belarus 22 Analysis of unit costs and outcomes (2013); Indonesia (2013); 163 Madagascar (2015) Ethiopia Policy Research 23 Cost-benefit analysis 173 Working Paper (2008) Kenya (2013); 24 Data envelopment analysis Democratic Republic of 176 Congo (2015) 25 Internal efficiency indicators Bangladesh (2010) 182 26 Rate of returns to education Sri Lanka (2011) 183 27 Analysis of inequity Costa Rica (2015) 185 Tajikistan (2013); 28 Analysis of per capita financing 188 Mauritania (2016) Costa Rica (2015); 29 Analysis of cash transfer programs 193 Indonesia (2012) Democratic Republic of 30 Benefit incidence analysis 197 Congo (2015) 31 Analysis of private spending by income quintile Madagascar (2015) 199 88 Education Public Expenditure Review Guidelines Example 1: Policy Recommendations Overall comment: The Albania and Ethiopia public expenditure reviews meet most of the criteria for examples of good policy recommendations: clear, prioritized, specific, implementable, and ideally, costed. In the Albania example, the recommendations are clear, specific, and prioritized (have a time frame for short- and medium-term policy recommendations), and at least part of these recommendations are costed. In the Ethiopia example, most recommendations are clear, quite specific, and somewhat implementable, but not prioritized, which is a drawback, given their number. Democratic Republic of Congo PER (2015) also offers good policy recommendations, but it is not included in this section because of its length. It provides a separate table that lays out a proposed time frame and identifies the ministry (ies) within whose jurisdiction the policy recommendation falls. Bad examples of policy recommendations include: • Recommendations that are not prioritized or sequenced • Vague imperatives that provide policymakers with little or no guidance about available options for fixing the problem • Vague imperatives that fail to acknowledge the fiscal costs, technical complexity, political costs, or time frame required to try to fix the problem • Policy implications or policy directions masquerading as policy recommendations For example: • “In a context of a rapid fall in learning outcomes, there is a need for additional public spending targeted directly at improving quality.” • “Improve the efficiency of public funding in the education sector.” • “Invest more in education, in line with other countries.” • “Increase the allocation of the public expenditures in Technical and vocational education and training (TEVET) as a share of total government education expenditure because the (private) rate of returns is high and the supply of the TEVET graduates is welcomed by the market but remains limited.” 89 Albania PER (2014) Expand Access to Pre-primary Education One priority is the expansion of access to and quality of preprimary education; the cost can be offset by savings from declining student numbers in higher grades (short-term). By 2015, Albania’s preprimary population is likely to stabilize at current levels but the numbers of students in basic, upper secondary, and higher education will drop significantly. Assuming there will be a 95 percent enrollment rate for preprimary and universal enrollment in basic and upper secondary by 2025, fewer basic and upper secondary education places will be needed relative to 2013. On the other hand, Albania’s investments in preprimary and primary education are far behind international benchmarks, especially given studies that confirm robust returns to investment in children’s early years compared to equivalent investments later in life. Preprimary education will therefore need to be expanded by approximately 36,000 places. Based on current unit costs, creating the 36,000 places preprimary education will need in 2025 would cost about 5 percent of the education budget. Using the same criteria to quantify savings, about 9 percent of the education budget can be saved from the decline in numbers of basic and upper secondary students, which can redeployed to finance the expansion in preprimary. Furthermore, since preprimary classes are mainly housed within primary schools, the places freed by basic education could be used to expand access to preprimary. In particular, the school mapping data collected by the Council of Europe Development Bank project (ALB-IPF-TA-10) should be better utilized. Rural to urban migration and expansion of private education should also be taken into account. For these interventions to produce improvements in quality of preprimary education, continued teacher training and capacity building is needed. Improve Governance, Efficiency and Equity Improve management and governance in the education sector (short-term). Sector planning and coordination across different levels of government could be improved to increase transparency and accountability in the use of resources and also enhance policy making and education delivery. Current financing mechanisms and accountability arrangements for pre-university education do not create incentives for schools and local governments to rationalize spending. Limited capacity and lack of a clear monitoring framework exacerbate sector inefficiency. To bring more transparency and equity to education financing, Albania could consider introducing per capita financing (PCF) to fund pre-university education (short-term). Currently, expenditures per pupil vary substantially by region; and it does not seem that regions with poorer results or a higher incidence of poverty receive more funding. Furthermore, schools depend on the REDs and local governments for financing major inputs and meeting maintenance needs, which impairs their ability to sustain a high-quality teaching environment. The introduction of a well-designed PCF mechanism could not only improve transparency around what regions receive and how that relates to student population, but also efficiency and accountability. It could also help provide additional funding to the disadvantaged and hard-to-reach populations. As PCF is phased in, in the medium term both school autonomy and accountability should be heightened (short-term). Currently, schools have only minimal financial autonomy. The PISA 2012 90 Education Public Expenditure Review Guidelines school questionnaire shows that only 20 percent of principals reported formulating their school budgets and less than 10 percent said they had autonomy over teacher hiring, firing, and salaries. Albania should consider giving schools more autonomy on both HR and financial management; but this shift will need to be accompanied by investment in building capacity at the local level and greater accountability for results. While this shift is unlikely to have significant budgetary implications, it will require a cultural shift whereby schools, REDs, and EOs will have to perform new roles. In a more decentralized setting, for example, REDs and EOs should act as advisors to schools. Cost-effectiveness of sub-sector investments could be examined as a basis for reallocating funds. Vocational education students cost the government three times as much per year as general education students, but there is no evidence that there is a concomitant return on the investment in terms of greater learning. Given the tight fiscal space, investment decisions should be driven by benefits that can accrue to both individual children and to society as a whole. For example, preprimary investments can be leveraged to produce positive effects on female labor participation, reach marginalized populations, and reduce the intergenerational transfer of poverty. Consider Increasing Education Spending over the Medium Term In the short term, given the lack of fiscal space and the fiscal consolidation plan in place for the period 2014-16, Albania should carry out reforms to make the sector more efficient and do more with the same level of budgetary resources. In the medium to long term, particularly after 2016, Albania should consider increasing public spending on education. Albania’s small budget envelope for the education sector should be raised gradually from about 3 percent of GDP to 4.0 percent of GDP, closer to the Eastern European average of 4.6 percent of GDP. While there is agreement that more public spending will not guarantee better education quality, Albania ranks at the bottom with respect to both learning outcomes and public spending on education as a share of per capita GDP. Even after exploiting all efficiency gains, there is likely an additional need for budgetary resources to increase the quality of education and learning outcomes over the long term. The additional public spending could be channeled to several areas in which Albania still needs investment, such as: (i) teachers’ professional development; (ii) learning materials and school supplies; (iii) quality of school facilities; and (iv) more time on tasks and activities in schools. While the cost of basic investment in pre-primary education can to some extent be offset by the savings from the declining student population, it also requires well prepared educators, learning materials and school supplies for which additional resources are needed. Improve Quality of Education The government could ensure that tertiary education is not expanded at the expense of quality (short-term). The government should strengthen the regulation guiding the expansion of higher education—largely through the implementation of the 2010 law on tertiary education— both to assure quality and to guarantee alignment with population trends. In the last decade, Albania has seen much higher enrollments in tertiary education, both public and private, in line with the government’s commitment to providing access to all who wish to continue beyond secondary education. Now it is necessary to direct attention to quality—with a view to ensuring that standards articulated in the 2010 Tertiary Education Law, itself aligned with the Bologna Process and overall 91 principles of Europe 2020 Strategy—are complied with. It also means that financing has to keep pace with the increase in the number of students, so quality is not jeopardized. Quality data must be generated and disseminated to the wider public to build evidence and inform policy making (short-term). Currently, there is little evidence on which policies might improve student learning in basic education. Unfortunately, the unreliable household data from PISA does not allow for analysis of determinants of learning outcomes, and causal links have not yet been established between various policy reforms and student achievement. To avoid making policy decisions in the dark, Albania should immediately finalize the design of the EMIS and, once it is in place, use the data in making decisions. This will also help strengthen school accountability. Albania should also seek evidence on policy reforms that have improved student learning in other countries, such as investments in preprimary education, improving teacher and principal effectiveness, and increasing school autonomy and accountability. Ethiopia PER (2015) Resource reallocation to improve equity If the role of public finance is to counteract the inequality in access to education, resulting from unequal income distribution, the government ought to be directing more than 20 percent of public resources towards the lowest income quintile, whereas it is currently directing only 13 percent towards the poorest quintile, less than even the 14 percent share of out-of- pocket expenditure on education contributed by this quintile. At the other pole, 39 percent of the benefits of public education spending goes to the highest income quintile, which contributes 31 percent of out-of-pocket spending. Correcting for this anomaly calls for a reallocation of public education spending, from higher education (which caters largely to the top quintile) to lower primary education (which caters to all classes and disproportionately more to the poorest quintile). The case for such a reallocation of resources from the highest to the lowest level in the education ladder is made stronger by the following facts: (i) more than half the recurring public expenditure on higher education is on provision of free food and lodging to all students in residence, many of who can afford to bear at least part of this cost; (ii) serious classroom and teacher shortages exist in primary schools in particular regions and districts, addressing which will have high marginal impact on efficiency and school performance; and (iii) improved efficiency at lower primary level will increase the access of pupils from poorer households to higher levels, thereby also contributing to improved equity in the distribution of benefits. The provision of free higher education services is based on the rationale that the beneficiaries will pay back after graduation and finding employment, through the “graduate tax”. However, actual level of cost recovery through this instrument is negligible and is not even being reported in any official publication. Moreover, while the graduate tax is in theory a suitable instrument for recovering the academic cost of providing higher education, the non-academic recurring costs need to be, and can be, recovered more quickly. While it may not be politically feasible to withdraw or cut down on subsidies being provided to university students, it should be possible to at least freeze the aggregate amount of subsidy in nominal Birr, and gradually shift a part of the non-academic recurring cost onto the students, complemented by financial assistance to the few students from lower income households who gain admission in higher educational institutions. 92 Education Public Expenditure Review Guidelines Resource reallocation to improve efficiency Additional resources for non-salary recurring inputs, provided through GEQIP, have not had a visible impact on average efficiency in primary education, even though efficiency has improved in better equipped woredas and schools. The average has been pulled down by a lagging quarter of the woredas, where efficiency has declined in spite of additional non-salary funding, in the absence of addressing the binding constraint, i.e., teacher and classroom shortages. The persistence of acute shortages of teachers and classrooms in lower primary education in the case of the lagging quartile of woredas shows that the existing structure of intergovernmental grants is unable to address such shortages, which are concentrated in a few regions and woredas. Formula-based general purpose grants from federal to regional and from regional to woreda governments cannot address such concentrated shortages. Nor can the existing GEQIP school grant in its present design, being distributed to all schools proportional to enrolment. Some mechanism for transferring teachers from one woreda to another has to be worked out. It should ideally be done in a way that makes it gainful for both parties involved in the transfer, and at minimum additional cost to the public exchequer per transfer. An exchange system could be created wherein woredas that have large teacher shortfalls can post their needs/demands—and those that have extra teachers can choose to offer some on transfer. Suppose a rule is established whereby the donor woreda is required to transfer 90 percent of the budget for Remuneration of Transferred Teachers to the recipient woreda (it could be a time bound contract), on the condition that the latter allocates 10 percent and commits to take on all future increases in salary and allowances. For the Donor, the benefit is the 10 percent immediate saving, plus further saving in the future since the continuing commitment towards the lent-out teachers remains fixed in nominal Birr. For the Recipient, the benefit is the ability to hire additional teachers with minimal initial cost and longer time to mobilise internal financing. For the nation as a whole, the benefit is better utilization of available teacher capacity and hence improvement in efficiency at lower cost (than if all woredas simply keep going after desired PTR targets, as fast as their financing capacity permits). Ensuring essential inputs and processes to improve quality Minimum school resources including infrastructure (electricity, water, sanitation), learning resources (textbooks and reference materials) and discretionary funds should be ensured at every school as these are factors that have positive impacts on student learning. Schools should monitor the teaching/ contact time and reduce the time teachers being in school but not teaching. Teachers’ increased efforts in monitoring student attendance will reduce their absenteeism. Teacher professional development should address teachers’ pedagogical challenges and increase their knowledge. Strengthening the system quality assurance capacity will likely bring long-term improvement in learning. Leveraging additional resources for education Ethiopia’s ambitions to become a middle-income country (MIC) by 2025 and its education sector development targets (ESDPV) are admirable, but the achievement of these targets depends on the commitment of all stakeholders: government, families, communities, schools, 93 educators, and administrators. The resources required are very large, necessitating that all key financiers of the system: the government, families and development partners to step up their efforts. Given that a steady 20 percent share of the government budget has been allocated to education over the past decade and the medium-term fiscal framework implies a constrained envelope for government spending as a whole, public resources for education are unlikely to rise above the current level of around 4 percent of GDP. However, the relatively low share of education in private household expenditures indicates that additional resources could be leveraged from private sources. Such potential could be tapped in the case of secondary, TVET and higher education. Analysis of the results of the 2011 Household Income, Consumption and Expenditure Survey shows that a majority of households who send their children to public secondary school have the same spending power as households that send their children to private secondary schools. Yet non-government providers account for less than 5 percent of enrolment in grade 9, and the growth of aggregate supply has failed to keep pace with the growth in demand for general secondary (grades 9-10) education. These facts suggest that it is worthwhile to tap the unused potential for expanding private provision of general secondary education, so as to fulfil the unmet demand with least additional public spending. Higher enrolment of pupils from higher and middle income households in private secondary schools will enable public resources to benefit larger numbers from lower income families. A subsector where demand has apparently declined due to unfulfilled expectations is TVET, where employment and earnings prospects do not seem worth the investment to many households. Low external efficiency, meaning too few among TVET graduates finding the jobs they aspired for, has been the main reason for the recent decline in enrolment. Developing a partnership with the potential employers could perhaps be a more effective way to address technical training needs and at the same time leverage additional resources from the private sector. A public-private partnership approach is also an option for further expansion of higher education in the future, having already created a significant number of publicly funded universities in the country. Strengthening credibility of EMIS data The grade-specific enrolment, repeater and readmitted numbers reported by many of the woredas do not fulfil even a simple consistency criterion, namely, that the drop-out rates implied by the reported data must not be negative. Of a total of over 800 woredas for which EMIS data is available for the five successive years, 2008/09 to 2012/13 (EC01 to EC05), only 37 percent of all woredas in the country are credible and can be used for analysing output efficiency of schools and its determinants. Data submitted by the remaining 63 percent of woredas are not suited for such analysis. Data submitted by schools and woredas need to be checked for consistency and reliability. Realistic planning and target setting Education sector goals and targets need to take into account not only the constraints on supply but also on the demand for education. For instance, the fact that enrolment drops as one moves from lower to upper primary education is, to a significant extent, influenced by the opportunity cost of sending 10-14 year old children to school when they could be doing some work and contributing 94 Education Public Expenditure Review Guidelines significantly to family income. While the first milestone of getting all children (or at least 95 percent) to enter primary education in time is within reach in the majority of regions, this is not the case with respect to the second milestone of getting more than 95 percent to complete the primary cycle. Achieving the latter requires not only improvements in service delivery but also easing of the economic constraints on poor households. Additional primary teacher recruitment needs to be carefully regulated to ensure that it is targeted at those schools/woredas where teacher shortage is clearly a binding constraint to improving efficiency and effectiveness. It is possible to use EMIS data to generate simple indicators of adequacy/inadequacy of (i) teachers and (ii) classrooms, and establish simple guidelines that woreda councils could use while allocating scarce resources. The ratio of Teachers per Section (TPS) = PSR / PTR is a useful indicator. If PTR is much worse than desired target (in a school or woreda) AND TPS is way below 1, then it clearly means that teacher shortage is a binding constraint and of HIGH priority to address. On the other hand, if PSR is much worse than target and TPS is way higher than 1, classrooms/ sections are the most binding constraint and of top priority. Reform options The following are some reform options for addressing the problems highlighted above: • Safeguard the financing for the educationally disadvantaged areas and groups to improve their access to education. This should cover: ▷ ECD/pre-schools (to improve the school readiness) ▷ Grade 1-4 (increasing enrolment of the last 10% and reducing drop-outs) ▷ Grade 5-8 (increasing the supply and addressing the demand constraints for the bottom half ) ▷ Grade 9-10 (increasing the supply and addressing the demand constraints in rural areas) ▷ Preparatory schools (Grade 11-12), universities and TVET programs: ensure that financial aids are available and used to increase their access to these types and levels of education • Introduce a cap on the amount of subsidy to cover non-academic costs (on food and lodging) incurred by higher educational institutions • Reallocate savings on higher education recurring expenditure to initiate a new Special Grant from federal via regions to woredas, targeted at those woredas with most acute shortages of primary teachers and classrooms’ • Establish an appropriate mechanism for transfer of teachers across woredas to reduce the wide inter-woreda variance in PTR • Establish and publicize guidelines for regions/woredas to ensure that additional teacher recruitment is targeted at those woredas/schools where teacher shortage is clearly a binding constraint • Within secondary education, shift the emphasis of additional investments from teacher recruitment to the creation of additional classroom space • Conduct an in-depth study of the demand and supply of TVET training and explore the possibility of a public-private partnership approach to leverage additional resources as well as to improve external efficiency • Encourage families to invest in learning materials (in addition to their current investment in uniforms, bags and transport) 95 • Develop a conducive policy framework for expansion of non-government provision in secondary and higher education, thereby leveraging additional resources • Ensure that data reported by schools and woredas through the Education Management Information System (EMIS) are subjected to adequate and effective consistency checks • Monitor and analyse the sector financing using unit cost approach. Items within the recurrent expenditures that are deemed essential for the efficiency/quality of delivery should gradually be funded domestically to ensure sustainability • Build capacity building to analyse several sources of data (administrative reports, census, surveys and in-depth studies) to analyse, triangulate and interpret the data and identify system challenges and potential solutions • Strengthen the system quality assurance including ▷ assessing students regularly (classroom assessment) and use examination and national (and international) assessment data to provide feedback to schools and policy makers ▷ strengthening teacher training programs (both pre- and in-service) and developing teacher competency validation processes (accreditation, licensing and career development) ▷ using school inspection data to inform school improvement planning, teacher and school leaders’ development; and ▷ conducting school effectiveness analysis using student/teacher/school characteristics 96 Education Public Expenditure Review Guidelines Example 2: Data sources Overall comment: Myanmar (2015) flags challenges associated with data availability and quality for different types of data. Few public expenditure reviews address such data challenges in depth. Samoa PER Notes (2014) has an annex dedicated to detailed information on budget-data sources. Myanmar PER (2015) The government would benefit from investing early in better data and analysis on coverage, quality and equity to prioritize spending towards the 10 Points Policy. Data gaps pose challenges for effective policy-making, especially because most important decisions are made at the Union level rather than at the subnational level where officials would be better informed of local needs and issues. Establishing a solid framework for data compilation and analysis that feeds into policy-making is a long term endeavor (Box 4.2). In the short to medium-term however, the MoE could enhance the capacity for policy analysis by maintaining three critical databases: (i) Education Management Information System to electronically record administrative data, which is currently collected in paper format; (ii) student learning outcomes, and (iii) budget and expenditure data. Box 4.2: Data Challenges • Education Management Information System: MoE [Myanmar] currently collects monthly and annual administrative data (e.g. number and type of schools, teachers and students), compiled in a Statistical Yearbook published with 12 months delay. An MIS would improve quality, timeliness, and accessibility of this important information. • Household surveys: These are important sources of estimates on enrollment, drop-out rates, household spending on education, and returns to education among other things. Existing surveys are dated (2009/10) and do not cover the entire country. Census data and new household survey data will become available in 2015. These data provide invaluable information on enrollments and dropouts. To be useful, though, staff in MOE need training to be able interpret and use these data for policy making. • Accounting for different education levels: Many schools that integrate multiple levels (i.e. primary, middle and high school) do not account for spending at each tier (e.g. nearly all middle schools are also primary). Going forward it will be important to separate cost of each tier to accurately assess costs per student at each level. 97 • Assessment of student learning outcomes: There are currently no regular reviews of whether schooling translates into learning at different stages of the education cycle. The Early Grade Reading Assessments (EGRA) and an Early Grade Mathematics Assessments (EGMA) could be instituted to start the process for grades 1-3. • Government spending: Key to monitoring the effectiveness of links between policy and government spending is timely, relevant, and accurate fiscal data. This will involve review of budget classification and broader FMIS. As part of the writing of this PER, data were compiled in a consolidated BOOST database to facilitate analysis. However, this database need to be updated and actively used by budget staff in MOE – a process which will require training and experience. Samoa PER Notes (2014) Data Annex The PER is primarily based on analysis of disaggregated public financial data covering the seven year period between FY06 and FY12, taking advantage of the FinanceOne FMIS system that has been in place since FY06. This analysis represents the first time that the FinanceOne data has been used to produce a consistent public expenditure analysis. To facilitate the analysis of expenditure based on consistent functional and economic classifications some adjustment to the data were made for the purposes of the analysis. While in more recent years, FinanceOne has full coverage based on GFS classifications, these were not fully in place in FY06 and FY07, so the team remapped expenditure by line item using GFS86 as a guide. While the data from FinanceOne has full coverage of domestic expenditure, externally-funded expenditure was recorded most consistently in other systems held by the Ministry of Finance, so data on estimated grant and loan utilization by ministry was added separately to the analysis. Because this information was only available at a higher level, an analysis of donor-funded expenditure by economic classification or detailed functional classification was not possible. Some further adjustments were made to the FinanceOne data to treat expenditure classifications consistently over the period, to help to illustrate the underlying trends. Firstly, tax expenditure incurred by government ministries that was subsequently rebated was netted off both expenditure and revenue, as it’s treatment in the financial accounts changed over the period which would otherwise give the impression both expenditure and revenue has risen. Secondly, some capital items that were recorded above the line were moved into financing in the budget frame. Thirdly, and most significantly, newly incorporated public beneficial bodies, the National Health Service and the Land Transport Authority were reincorporated into the public accounts. This was necessarily since their new status as corporate bodies meant that halfway through the period, their accounts were no longer consolidated in the public accounts ledger, but instead expenditure was in the form of a public grant. The team used the bodies’ corporate accounts to reintegrate disaggregate expenditure trends for the period since they became corporatized. The South Pacific Games Authority was also reincorporated into the public accounts to smooth out a temporary spike in transfers to that agency relating to the hosting of the South Pacific games. 98 Education Public Expenditure Review Guidelines The analysis of the public wage bill combines FMIS data with data from the payroll system. It also incorporates payroll data from the accounts of the largest public agencies to present approximate estimates of payroll trends for the whole government for the first time. Similarly, the health analysis represents the first time that unified data for the public healthcare sector has been presented since the creation of the National Health Service as an autonomous agency. For the education analysis, the data was adjusted for the spike in trends relating to the South Pacific Games in FY07 and for lumpy tax expenditures to establish the underlying structural expenditure trends. The outcome data are mostly based on the Ministry of Education’s statistical database. Changes are presented in real terms taking FY06 as the base unless otherwise stated. In some cases, additional data has been collected from official government sources including the annual Budget Statement, the World Bank’s World Development Indicators are used for international comparisons and outcome data and other documents. 99 Propoor Spending 45 Degree Line – Equality Example 3: Analysis of source funds Progressive Lorenz Curve of Income or Consumption Nigeria PER (2015) Regressive (Poorly Targeted) Overall comment: The Nigeria example provides a comprehensive breakdown of education-sector financing by source, 100% including public sources by level of government, private sources, and donor funding. In Nigeria, 40 percent of the education sector is funded by private households’ out of pocket contributions while local government constitutes the second highest share (25 percent). Figure 8 presents the sources of finance by origin. Overall, in 2013, the total cost of the education sector (all levels of education) in Nigeria amounted to 2,329.4 billion Naira (14.6 billion US$). The breakdown of the education sector finance was as follows: federal government (18 percent), state government (13 percent), Local Government Authority (LGA) (25 percent), household out-of-pocket payment (40 percent), Universal Basic Education Commission (UBEC) initiative (3 percent), and donors: the remaining 0.4 percent. FINANCE, 2013 Figure 8:FIGURE 8:education Sources of SOURCES finance, 2013 SECTOR FINANCE, 2013 OF EDUCATION sector 2/13 Source: Authors’ estimate from CBN, OECD, Nigeria, State Budget, Federal Government Budget, and General Household Survey Panel 2012/13 9 | 1% 426 | 18% LGA Federal Government GA Household State Household 936 | 40% UBEC LGA 295 | 13% UBEC State UBEC tate Household Federal Gov’t Donors Donors ederal Gov’t 587 | 25% Donors 76 | 3% Source: Authors’ estimate from CBN, OECD, Nigeria, State Budget, Federal Government Budget, and General Household Survey Panel 2012/13 ON FINANCING dependent Devp’t venue Partners Federal Ministry of Finance Federal 0% 20% 40% 60% 80% 100% CUMULATIVE SHARE OF POPULATION 100 Education Public Expenditure Review Guidelines Example 4: Analysis of revenue sources FIGURE 8: SOURCES OF EDUCATION SECTOR FINANCE, 2013 Nigeria PER (2015) Source: Authors’ estimate from CBN, OECD, Nigeria, State Budget, Federal Government Budget, and General Household Survey Panel 2012/13 9 | 1% Overall comment: This example provides a detailed and comprehensive summary of the sources of revenues at all 426 | 18% 936 | 40% administrative levels in Nigeria and explains the revenue-sharing formula. LGA Household 13% the enactment of the 2004 295 |since Figure 12 shows the structure of basic education financing UBEC Universal Basic Education Act (UBE). The law preserves the constitutional responsibility State of states and local governments in Nigeria to provide basic education and expands the federal government’s responsibility in ensuring it is free and compulsory. Federal Gov’t 587 | 25% Donors The proceeds of the Federation Account are shared among the federal, state, and local governments, 76 | 3% in accordance with a revenue-sharing formula and the funding is tracked from the source to the service delivery point. The current formula for dividing up total revenues to government allocates 52.68 percent to the federal government, 26.72 percent to state governments, and 20.6 percent to local governments (Figure 13). The UBE Act was developed based on constitutional mandates, and clearly demarked the financing sources between salary and non-salary. FIGURE 12: THE STRUCTURE OF BASIC EDUCATION FINANCING Figure 12: The structure of basic education financing Internally Federal VAT Independent Devp’t Fund Generated Account & Sources Revenue Derivatives Pool Revenue Partners Federal Ministry of Finance LGA SJGA Federal State Con. Rev. Responsibility Federal Budget UBFC (2%) FMF Salary Non-Salary Salary Non-Salary Salary Non-Salary Salary Non-Salary Education Finances Primary Lower Lower Unity School Secondary Secondary College SUBEB Source: Author’s sketch following funding allocation arrangements in UBE following UBE Act of 2004 Source: Authors’ sketch following funding allocation arrangements in UBE following UBE Act of 2004 Note: Each Note: Each StateState shallamaintain shall maintain special accountatospecial account be called “State toGovernment Joint Local be called “State Account Joint Local Government Account (SJLGA) “into which shall (SJLGA) be whichall paid “into allocations shall to the be paid all allocations local to the government local government councils councils of ofthe the state from the state from the Federation Account and from the Government of Federation Account and from the Government of the State” (Section 162 [6], 1999 Constitution of Nigeria). the State” (Section 162 [6], 1999 Constitution of Nigeria). 101 Education finance depends to a large extent on federal revenues and the ability of the states to finance education expenditures is directly linked to the availability of federal revenues. Figure 13 shows a summary of the sources of revenue for the three tiers of government. Internally generated revenues for LGAs stand at 1.6 percent while allocation from the state government to LGAs only represent about 0.7 percent of their total revenue. This implies that about 95.3 percent of the LGAs’ total revenue comes from statutory allocations, excluding the grants and revenues from the stabilization fund which accounts for the remaining 2.3 percent. Given that salaries are automatically deducted from the statutory transfers at source, this implies that LGAs have, in reality, very little say in education finance, and their role is more symbolic than anything else, given that there is no financial planning or budgeting on their part for the basic education level. At the state level, internally generated revenue (IGR) represents 19 percent of total revenues indicating some potential fiscal space for spending on education based on IGR. In particular, given that the states are responsible for capital and non-salary spending, states’ ability to generate more revenue may suggest variations in resource availability across states for basic education spending. Figure 13: Sources of overall revenues at all administrative levels, 2013 Federal Government State Governments Local Grand Source Governments Total FG’s Share PCT Sub-Total States 13% Sub-Total Statutory Allocation 2,777.4 53.5 2,830.8 1,435.8 615.0 2,050.9 1,107.0 5,988.7 Augmentation 1/ 351.3 6.8 358.1 181.6 101.6 283.2 140.0 781.3 Share from Excess Crude 208.7 4.0 212.7 107.9 60.3 168.2 83.2 464.2 NNPC Refunds - - 0.0 44.9 11.9 56.8 34.6 91.4 SURE-P 191.8 3.7 195.5 99.2 55.5 154.6 76.5 426.6 Share of VAT 106.9 7.6 114.6 381.9 - 381.9 267.3 763.8 FG Independent Revenue 274.4 - 274.4 - - 0.0 - 274.4 Internally-Generated Revenue - 11.0 11.0 574.9 - 574.9 29.3 615.2 Less State Allocation to LG - - 0.0 12.8 - 12.8 - 12.8 Net Internally-Generated Revenue - 11.0 11.0 562.2 - 562.2 29.3 602.4 Grants - - 0.0 69.7 69.7 43.0 112.7 Share of Stabilization Fund - - 0.0 1.3 1.3 16.4 17.7 State Allocation to LG - - 0.0 - - 0.0 12.8 12.8 Others 45.7 - 45.7 8.7 8.7 - 54.4 Total 3,956.2 86.6 4,042.8 2,893.2 844.3 3,737.5 1,810.0 9,590.3 Includes share of the difference between provisional distribution and actual budget. Note IGR is noted for 12 billion naira 1/ and FCT included in state level which make some difference. Source: Cited from CBN annual report with source from “Federal Ministry of Finance (FMF), Office of the Account-General of the Federation (OAGF), and Fiscal returns from state and lovely governments Survey UBFC (2%) FMF Salary Non-Salary Salary Non-Salary Salary Non-Salary Salary Non-Salary Education Finances Lower 102 Primary Education Public Expenditure Review Guidelines School Secondary SUBEB Lower Secondary Unity College Source: Authors’ sketch following funding allocation arrangements in UBE following UBE Act of 2004 Note: Each State shall maintain a special account to be called “State Joint Local Government Account (SJLGA) “into which shall be paid all allocations to the local government councils of the state from the Federation Account and from the Government of the State” (Section 162 [6], 1999 Constitution of Nigeria). Given that the vast majority of basic education salaries come from the federal allocation, which in turn heavily depends on oil, factors that affect oil revenue also directly affect basic education finance. The share of IGR varies greatly by state, and some tend to depend entirely on federal allocation due to low IGR levels. Figure 14 shows (i) total revenue breakdown by IGR; (ii) revenue other than IGR; (iii) the share of internally generated revenues out of total revenue; and (iv) the per capita allocation of non-IGR. The figure shows that share of IGR revenue varies from a low of 1 percent in Benue state (2 % in Borne state) to a high of 41 percent in Lagos. Overall, only four states including Lagos have an IGR share of revenue more than 15 percent of their total revenue—Rivers (22%), Ogun (21%) and Kano (20%). Edo ranks a distant fifth with 14 percent. In addition to Lagos and Kano, five of the 9 Niger Delta states have higher revenue. In general, revenues across states hover around 100 billion Naira, except for the 6 states where it is substantially higher (Lagos, Kano, Rivers, Delta, Akwa-Ibom and Bayelsa). However, since Nigeria has developed an allocation formula justified by rights enshrined in the constitution such as the right of the Niger Delta states to receive 13 percent of oil revenue prior to allocation; resource availability at the state level clearly depends not only on IGR but also on what is being allocated from the federal level. Figure FIGURE14: Sources 14: of OF SOURCES educational EDUCATIONsector finance, SECTOR 2013 FINANCE, 2013 Source: Calculated from CBN Annual report 2013 and per capital allocation from monthly shares of distribution from The Federation Account by State, Nigeria Economic Report, The World Bank Group (2013) Federal Allocation From All Sources 450 100 Internal Generated Revenue (IGR) 400 Share of IGR (right axis) 90 Per Capita Revenue (right axis) 80 350 70 300 60 250 50 200 40 150 30 100 20 50 10 0 0 Borno Taraba Adamawa Bauchi Yobe Gombe Jigawa Zamfara Kaduna Katsina Kebbi Sokoto Kano Benue Niger Kogi Nasarawa Plateau FTC Abuja Kwara Enugu Imo Anambra Ebonyi Abia Bayelsa Akwa Ibom Delta Cross River Edo Rivers Ondo Ekiti Osun Oyo Ogun Lagos North East North West North Central South East South South South West Source: Calculated from CBN Annual report 2013 and Per capital allocation from Monthly Shares of Distribution from The Federation Account By State, Nigeria Economic Report, World Bank Group (2013) FIGURE 3: AN OVERVIEW OF THE COMPLEXITY OF TRANSFERS & FUND FLOWS, 2012 LEVEL BUDGET NATIONAL MINISTRY OF FINANCE MINISTRY OF EDUCATION CULTURE MOF MOEC Adjustment Fund Tugas Central DAU DAK BOS Dekon Pembantuan Functions Own PROVINCIAL Source District Budget Revenue DAU/SDA Provincial Education Central Government O ce Agencies in the Regions PAD 103 Example 5: Analysis of decentralized financing Sudan PER (2014) Sudan State-level Public Expenditure Review Meeting the Challenges of Poverty Reduction and Basic Service Delivery Synthesis Report, Vol. 1, Summary for Policymakers, May 2014 Sudan State-level Public Expenditure Review Meeting the Challenges of Poverty Reduction and Basic Service Delivery, Vol. 2, Background Papers, May 2014. Overall comment: These volumes focus on the strengths and challenges of Sudan’s decentralization arrangements. They do not focus on education per se, although this function is decentralized to subnational units and enters into analyses of decentralization processes. As is often the case in studies of decentralization, these studies had to be based on new data collection in the form of case studies of four of Sudan’s 17 states. Sudan highlights several lessons to which education PERs need to be alert. 1. PEFA issues become particularly salient under decentralization because subnational governments tend to vary significantly in the quality of their public financial management. The opportunities for corruption expand accordingly. 2. Subnational governments usually vary, sometimes significantly, in their capacities to raise own revenues. This reality leads to varying gaps between financing responsibilities and resources and the potential for substantial horizontal imbalances between subnational governments that central government should but may not address. 3. Subnational governments often try to collect cost-ineffective taxes—ones that raise little or no revenue, are costly in terms of tax administration, or that are virtually impossible to enforce. 104 Education Public Expenditure Review Guidelines Example 6: Analysis of education financing by level of government and intergovernmental fiscal transfers Indonesia PER (2015) Overall comment: The Indonesia example provides a focused and detailed analysis of the intergovernmental transfer system and an overview of the education-funds flows in the country. In Indonesia, district governments are responsible for managing the two main assets at the primary and secondary education levels: schools and teachers. Legally, primary and secondary schools are owned by district governments. In fact, when it comes to budgets, the school’s legal status is similar to that of a district government department. Similarly, civil service teachers are legally district government employees, although the hiring process, like that of other civil servants, depends on a number of central government ministries. Provincial governments have very limited authority when it comes to schools, mostly coordinating districts at the basic and secondary levels of education, including with regard to staff development and the provision for education facilities. The central government formulates policy, issues regulations/guidelines and standards at the national level, and still directly controls higher education. Schools have considerable autonomy over operational, budgetary and programmatic decisions. Since 2003, School Based Management (SBM) has applied to all stages of formal education. A degree of decision-making power and management have thereby devolved to the school level, taking account of local norms and encouraging community involvement. The funding system for the education sector is complex, involving multiple sources and transfers across various levels of government. Expenditures for education come from central government funds, transfers to subnational governments, subnational governments’ own-source revenues, and central government spending at the subnational level that is not recorded in subnational budgets. Currently, schools receive funds from eight different sources and four different budgets, including the national, provincial, district and school budgets (Figure 3). 105 Figure 3: An overview of the complexity of transfers and fund flows, 2012 Fund Originated from MoD Fund Originated from MoEC Sectoral Budget Fund Originated from Provincial Budget Fund Originated from District Budget Source: Elaboration based on Permendagri 13/2006 on Guidelines of Subnational Financial Management, World Bank (2009) and Law 22/2011 on AP3N 2012 Note: Adjustment Fund also includes the local incentive grant (Dana Insentif Daerah, or DID) Central government transfers are the main source of revenue for district government budgets (APBD). Central government transfers to subnational governments have more than doubled in real terms since decentralization, accounting for 88 percent of district budgets and 44 percent of provincial budgets in 2009. While the majority of transfers are not earmarked – making it impossible to determine exactly what they are spent on – transfers are estimated to finance about 90 percent of subnational spending on education, and 60 percent of the total national education budget. Subnational governments receive many types of transfers for education spending. The main transfer to subnational governments is the General Allocation Fund (Dana Alokasi Umum, DAU) block grant, which provides funding for the salaries of district civil servants, including civil service (PNS) teachers. DAU transfers represented about 60 percent of district and 20 percent of provincial budgets in 2009. The DAU is allocated through a two-part formula consisting of the “Basic Allocation” and the “Fiscal Gap” (See Box 1 for details on each transfer type). The Basic Allocation, which is calculated largely based on the salary bill for civil servants in the district or province, implicitly incentivizes civil service hiring. Covering about 72 percent of the salary bill, it accounts for about 45 percent of the total DAU. Fund Fund Adjustment Adjustment Tugas Tugas Central Dekon Central DAU DAU DAK DAK BOS BOS Pembantuan Dekon Pembantuan Functions Functions Own Own PROVINCIAL PROVINCIAL SourceSource BudgetBudget DistrictDistrict Revenue Provincial Provincial Education Education Central Central Government Government Revenue DAU/SDA DAU/SDA O ce O ce Agencies Agencies in the Regions in the Regions PAD PAD 106 Education Public Expenditure Review Guidelines Own Own BudgetBudget DistrictDistrict DISTRICT Source DISTRICT Source DAU/SDA DAU/SDA Revenue PAD PAD District District Education Education Revenue O ce O ce DAK DAK Fund Fund from from Originated Originated MoF MoF Fund Fund from from Originated Originated MoEC MoEC Sectoral Sectoral Budget Budget Fund Fund from from Originated Originated Provincial Provincial Budget Budget Fund Fund from from Originated Originated District District BudgetBudget In 2009, districts accounted for 50 percent of total PRIVATE national PRIVATE SCHOOLS PUBLIC SCHOOLS education PUBLIC SCHOOLS expenditures, while provincial SCHOOLS governments accounted for only 8 percent. Salaries are a major component of district spending. When salaries are excluded, education spending is still largely centralized. The central government based on based Source: Elaboration Source: Elaboration Fund alsoFund Note: Adjustment Note: Adjustment on Permendagri Permendagri alsothe includes 13/2006 on includes 13/2006 local incentive on Guidelines Guidelines the local incentive grant grant (Dana (Dana Insentif Financial Financial of Sub-National of Sub-National Insentifor Daerah, Daerah, Management, Management, DID) or DID) World World Bank Bank (2009) (2009) and and Lawon Law 22/2011 22/2011 on AP3N 2012 AP3N 2012 controls the majority of the non-salary budget at all levels of education, from the 70 percent in ECD to 99 percent at the university level. Even in basic education, almost 90 percent of non-salary spending still occurs at the central level. What brings down the overall average to 67 percent is unclassified spending at the district and provincial levels. Figure FIGURE FIGURE 12:12: 12: BYby Spending SPENDING SPENDING BYlevel LEVEL LEVEL OFof OFgovernment, GOVERNMENT, GOVERNMENT, 2001-2009 2001 2001 2009 2009 250 250 4 47 76 66 65 55 54 46 68 8 100 100 90 90 200 200 80 80 70 70 IDR TRILLIONS 2009=100 IDR TRILLIONS 2009=100 65 65 66 66 59 59 65 65 55 55 48 48 55 55 59 59 50 50 150 150 60 60 PERCENT PERCENT 50 50 100 100 40 40 30 30 50 50 20 20 10 10 31 28 35 29 40 47 41 34 41 31 28 35 29 40 47 41 34 41 0 0 0 0 2003 2003 2001 2001 2002 2002 2005 2005 2004 2004 2007 2007 2006 2006 2009 2009 2008 2008 2002 2002 2001 2001 2003 2003 2004 2004 2005 2005 2006 2006 2008 2008 2007 2007 2009 2009 CentralCentral District Source: Source: World World Bank sta Bank based on based sta estimates estimates on state state budget and data budget data Central Regional Financial and Financial Regional information information system data system data District Province District (Sistem Informasi (Sistem Informasi Keuangan Keuangan Daerah, Daerah, SIKD), SIKD),of Ministry Ministry Financeof Finance Province Province Source: World Bank staff estimates based on state budget data and Regional Financial information system data (Sistern Informasi Keuangan Daerah, SIKD), Ministry of Finance Figure FIGURE FIGURE 13: 13:13: NON NON SALARY SALARY Non-salary EDUCATION EDUCATION education EXPENDITURE EXPENDITURE expenditure BY PROGRAMS BY PROGRAMS by programs and of AND level AND OF OF LEVEL LEVEL GOVERNMENT, GOVERNMENT, 2009 2009 government, 100 100 2009 100 100 2 2 10 10 11 11 12 12 90 90 23 23 9 9 80 80 15 15 80 80 6 6 70 70 60 44 44 60 60 60 PERCENT PERCENT PERCENT PERCENT 100 100 51 51 50 50 40 40 75 75 40 40 73 73 30 30 20 20 47 47 20 20 25 25 10 10 0 0 0 0 EARLY EARLY SECONDARY BASIC BASIC SECONDARY UNIVERSITIESOTHERS UNIVERSITIES OTHERS 2008 2008 2009 2009 2008 2008 2009 2009 CHILDHOOD CHILDHOOD EDUCATION EDUCATION EDUCATION EDUCATION EDUCATION EDUCATION TOTALTOTAL NON SALARY NON SALARY Central Central Central Central Central-80S Central-80S District District District District Province Province Province Province Central Central-80S Central District Province Sources: MoF District Province 107 Box 1: An Overview of Current Transfer Mechanisms in the Indonesian Education System From the Central to Subnational Governments This box provides a brief description of the objectives and means by which the various transfer mechanisms from the central government to subnational governments within Indonesia are determined. These transfers represent the major source of financing for subnational governments and thus, to a large extent, explain the level and composition of their spending. General Allocation Fund (Dana Alokasi Umum, DAU) The DAU, according to Law No. 33/2004 Article 1 (21), is a discretionary block grant sourced from the Central Budget (APBN) and aims to equalize the fiscal capacities of subnational governments. It is transferred monthly and directly from central to subnational governments. The DAU is allocated based on a national formula and is the sum of a basic allocation (a portion of the subnational budget for public servant salaries) and the”fiscal gap” (the difference between the estimated fiscal needs and fiscal capacity) of the subnational government. The basic allocation accounted for about 45.5 percent of the DAU in 2010. Fiscal needs are based on regional variables such as population, area, GDP per capita, and the human development index. Fiscal capacity is measured by a region’s own-source revenue and a fraction of total revenue-sharing. Based on Government Regulation No.55/2005, provinces only receive 10 percent of the total DAU, while districts receive 90 percent. Specific Allocation Fund (Dana Alokasi Khusus, DAK) DAK is an earmarked grant allocated to finance specific investment expenditures that are aligned with national priorities and carried out under the jurisdiction of subnational governments. The DAK cannot be used for research, training, administration, or official travel. In 2011, 19 economic sectors received DAK allocations including education, health, agriculture, forestry, trade and various infrastructure sectors (road, irrigation, water, sanitation, rural electricity, housing and local government and remote areas infrastructure). Education is a key priority for DAK spending, with about 40 percent of DAK transfers allocated for education and used primarily for school rehabilitation and quality improvement. The DAK allocation has a formula component that takes into account the fiscal gap and has a 10 percent matching requirement. DAK is transferred in three tranches: the first is allocated after the budget is submitted to the central government; the next two depend on the depletion of the previous tranche. Although DAK is earmarked to fund capital spending, the government allowed some routine maintenance expenditures. Revenue Sharing Fund (Dana Bagi Hasil, DBH) Unlike DAU, which is a horizontal equalization grant, DBH is a vertical equalization grant which consists of revenue sharing from natural resources and taxes. Local governments are obliged to use 0.5 percent of their receipts from the natural resources part of DBH on basic education.” DBH represented approximately 20 percent of total subnational government revenues in 2009. Special Autonomy and Adjustment Funds Special Autonomy Funds include specific grants for Papua, Papua Barat and Aceh (Dana Otsus) and Special Adjustment Funds (Dana Penyesuaian) which include additional allowances for teachers, such as professional benefits for certified teachers and for uncertified civil service teachers, a School Operational Assistance program (Bantuan Operasional Sekolah, or BOS), and local incentive grants (Dana Insentif Daerah, or DID) for education. Central government spending at the subnational level not recorded in subnational budgets (APBD) De-concentration (Dekon) and Co-Administered Tasks (Tugas Pembantuan, TP) Dekon and TP funds originate from the central government’s budget (APBN), and are administered by the provincial Dinas. The funds cover a variety of projects and activities, including school and classroom reconstruction and school quality improvements, social assistance programs (which included BOS until 2011) and capacity building programs for civil servants. Sources: MoF and various laws 108 Education Public Expenditure Review Guidelines Example 7: Analysis of school budget by financing source Philippines PER (2015) Overall comment: This PER investigated the effects of school-based management (SBM) in basic education in the Philippines. Among other questions, it used survey data from three provinces to assess whether SBM alters the resources available at the school level. Using the school-level financing data generated by the survey, the analysis addressed these questions: How has the resource situation at the school level changed in recent years? What is the financial resource situation of schools and what use is made by schools of SBM grants? What resources are schools able to mobilize in addition to Department of Education (DepED) transfers? Does the SBM grant act as a catalyst for the school to access other resources? What uses are made by the school from existing sources? What is the variation in pattern of uses from different sources? What is the relationship between resource allocation and school performance, and what can be done to improve the efficiency and equity of resource allocation? What has been the resource allocation trend in the past few years? What do we know about the equity of resource allocation across schools? What factors contribute to inequality in resource distribution? What is the role of Local Government Units (LGUs) in supplementing the financial resources available to schools? Conclusions assembled from the PER The analysis of school survey data collected in three provinces shows several interesting observations related to the improved availability of financing at the school level. First, the proportion of budget raised by LGUs and communities has been increasing in our sample schools, which is consistent with the expansion of SBMs. There does appear to be some evidence of the so- called ‘fly – paper effect’ where central grants stick to the recipient and overall resource position of the school improves. Parent-Teacher-Community Association (PTCA) funds appear to be higher in the year following a school receiving a SBM grant. Second, the gap between planned and actual budgets looks smaller among schools that perform well in SBM. Financial decentralization to schools has doubled, but at PHP 450 per pupil per year for elementary schools and PHP 965 per year for high schools, school level funds account for only about 5% of overall basic education spending. The sample survey data from 2010 indicates that high schools received about PHP 500,000 from various sources and elementary schools received an average of PHP 134,000 pesos. About 60-70% of these resources come from DepED in the shape of annual capitation grants for maintenance and operational expenses and occasional SBM grants. DepED should consider accelerating the pace of resources transferred to the school level. 109 Analysis behind the conclusions Survey data reveals that school level resources have increased indicating growing school financial empowerment but not substantial school level financial decentralization. An increased share of resources is managed at the school level compared to total national government spending on basic education in the last five years. Funds managed at the school level have grown in absolute and per pupil terms during the last five years. School level resources were almost double among the sample high schools at PHP 965 per student in 2010 when compared with only PHP 449 per student among the sample elementary schools (in nominal terms). Average real per student school-level managed funds doubled from just under PHP 200 in 2007 to nearly 400 pesos in 2010 in the survey sample schools (in constant 2005 prices). Nevertheless, compared to the national average spending per pupil, school level managed resources remain a small proportion of 5.4 percent in 2010. In the sample schools, the overall size of school managed funds has grown from 2006 to 2010 even as Department of Education (DepED) share of school level resources has not increased substantively in the survey schools. It is important to understand how much of these increases in school level managed resources represent a deliberate trend towards financial decentralization by DepED versus a voluntary increase in contributions towards capital expenditures, teacher salaries and various Maintenance and Other Operating Expenses (MOOE) by local government units (LGUs), Parent Teacher Community Association (PTCA) and the community. If this trend were to represent the results of a deliberate strategy by DepED to give increased financial autonomy to schools, the role of DepED grants including school MOOE, School Based Management (SBM) and School Based Repair and Maintenance Scheme (SBRMS) grants would increase over time. Table 3.3 illustrates that DepED resources have hovered around 70% of total school level resources. While the share of resources the school is able to mobilize from local and community sources as compared to what it receives from DepED in terms of MOOE, SBM and SBRMS grants has fluctuated slightly in the last five years, it has always been between 25 to 30 percent of total school resources (except 2010). Table 3.3 Sources of school level funds (mean values in constant 2005 PHP) 2007 2008 2009 2010 2011* DepED 53,000 70% 71,000 76% 98,000 73% 124,000 59% 119,000 73% PTCA 15,000 20% 15,000 16% 21,000 16% 31,000 15% 21,000 13% LGU 2,000 3% 3,000 3% 6,000 4% 28,000 13% 12,000 7% Community - - 1,000 1% 2,000 1% 7,000 3% 6,000 4% Others 5,000 7% 2,000 2% 9,000 7% 20,000 10% 4,000 2% Total (%) 76,000 (100) 93,000 (100) 135,000 (100) 209,000 (100) 162,000 (100) Source: 3D-SFSD. *2011 data not complete. ** Figures have been rounded off to the nearest ‘000 110 Education Public Expenditure Review Guidelines The financial role of LGUs in basic education has been growing in keeping with DepED policies to enable deeper partnerships between DepED and local government units. Table 3.3 indicates that in 2010, the share of non-DepED school level funds rose to a high of 41 percent of total school managed funds. This was driven by LGUs contributing a substantially greater amount of 8.9 percent of per pupil school level funds (up from 6.6 percent in 2007). A possible hypothesis to explain this is that 2010 was an election year, causing local governments to spend more on basic education as a strategy to win votes. It remains to be seen whether LGU funding for basic education as a whole and as a share of total basic education funding continues to increase in real terms post 2010. Of the four major sources of funds that the school can manage, funds contributed by the parent-teacher association (PTCA) form the highest share every year, although the percentage varies between 33 to almost 75 percent of school mobilized non-DepED funds. Nevertheless, about one-third of the sample schools either did not receive any PTCA funds or did not maintain records of these funds in 2010. Further investigation is required to determine if this is an effect of the no- collection policy issued in 2009 which was strictly enforced by DepED from 2010. The surveyed schools did receive some in kind resources in 2011 but the estimated value of these resources was very small at only 0.7% of total funds received by the school. Data on in-kind resources were collected only for the year 2011 and the analysis shows that the only source from which in kind resources form a bigger portion of their contribution was donors, where in kind resources were 18% of total donor contributions. When data about municipal per capita income is included, we find that the richest municipalities received significantly more at 4.5 times the amount in terms of the value of in kind resources as compared with the schools in the poorest municipalities. However, the value of in kind resources received by schools with higher SBM levels of implementation was not significantly different from that received by schools with lower SBM levels of implementation. The proportion of schools that receive the SBM grant annually is growing but still remains small. In 2011 only 20 of the 150 survey schools received the grant (only 2 schools received it in 2007). Three of the 150 schools received the SBM grant in two years. No sample school received the grant more than twice. The percentage of funds represented by SBM grants for those schools that did receive the grant has remained at about one-third of total school level funds in the last three years. The number of schools that received school MOOE grants has increased substantially in the last five years from merely 13 in 2007 to 115 schools in 2011. The percentage of funds represented by school MOOE grants for those schools that did receive the grant has remained at about 60 percent of total school level funds in the last three years. All the schools have school MOOE allocation, and the allocation per school is posted in the DepED website. However, during school visits in the implementation support for National Program Support for Basic Education (NPSBE2-SPHERE), it was observed that there were still schools which: a) opted not to get their MOOE allocation but instead requested their Division Office to provide the Division-procured supplies based on their list of requirements or requested Division to pay their utility bills directly; b) did not request their allocation because of large unliquidated cash advance; or c) the Division decided not to provide the allocation for the same reason. Central-80S District District Province Province 111 An analysis of the sources and uses of funds data for 2011 shows that a greater proportion of school level resources for survey schools that ever received a SBM grant was from LGUs and PTCAs. We can see from Figure 3.2 below that schools that ever received the SBM grant received an average of 26,331 pesos from LGUs in 2011. The average LGU funds in 2011 received by schools that were never SBM grant recipients were much lower at about 7,000 pesos. SBM grant recipient schools were also significantly more likely to raise higher resources from PTCAs in 2011- an average of 49,700 pesos compared with an average of 17,000 pesos for schools that never received the SBM grant. However, SBM grantees mobilized fewer resources from community and other sources compared with non-SBM grantees. FIGURE 3.2: SOURCES OF FUNDS: 2010 Figures 3.2: Sources of funds: 2010 100 6.9 16 10 80 15 29 60 20 40 20 48 54 0 Non SBM Grantee SBM Grantee PTCA PTCA LGU Source: 3D-SFSD LGU Others Community Others Community APPENDIX FIGURE 29: SHARE OF TOTAL HOUSEHOLD EXPENDITURES Share of total household expenditures, by region Source: EMOP 2014 2% 15% Bamako Koulkoro Koyos 11% Gilo 90 80 70 60 PERCENT 50 40 30 20 10 112 Education Public Expenditure Review Guidelines 0 2008 2009 2008 2009 TOTAL NON SALARY Central District Province Example 8: Analysis of household surveys on private spending Mali PER (2016) Overall comment: This example provides a comprehensive and broad overview of household spending on education. It includes information on the share of household spending as part of total education expenditure in Mali; international comparisons on private spending in education (as a share of gross domestic product) based on UNESCO Institute of Statistics data; share of household budget spent on education by region; and the composition of household spending on education by region and socioeconomic 6.9 group. The analysis at subnational level reveals significant variance in how much households spend 10 on education and how they allocate this budget. 29 Household data suggests that in 2014, households in Mali spent approximately CFA 28 to 31 billion on education expenditures, or approximately 8 to 10 percent of all education expenditures.a Private spending on education, estimated in this manner, is roughly equivalent to 0.6 percent of Malian GDP 54 (and this estimate is identical to UIS estimate from 2009). To put this number in context, UIS reports private sources account for 2.4 percent of GDP in Benin, 2 percent in Burkina Faso, 1 percent in Burundi, antee 0.3 to 0.4 percent in Malawi and Niger and 1 percent in Guinea. Households generally spend about 1 percent of their budgets on education, and 65 percent of this expenditure goes to pay for tuition and school fees. How much households spend on education and how they allocate this budget varies greatly from region to region, reflecting both income disparities and disparities in access. Households in Bamako collectively account for half the private spending across entire Mali; households in Mopti, Gao, and Tombouctou incur only 6 percent of all private expenditure, when combined together (Appendix Figure 29). This skewed spending is the result of many disparities including more children, more schools and more income in Bamako. The L HOUSEHOLD EXPENDITURES disparities are also apparent APPENDIX in the share 29: FIGURE of household spending SHARE OF TOTAL on education. HOUSEHOLD In Bamako, households EXPENDITURES allocate 1.75 percent of their budgets to education (twice the national average) whereas in Mopti and Share of total household Source: EMOP 2014 expenditures, by region Tombouctou, this share is only about one tenth of a percent. 2% Appendix Figure 29: Share of total household Bamako 15% Koulkoro expenditures, by region 50% Bamako Bamako 2% Mopti Koyos Koulkoro Gilo 2% Gao Koyos 11% Segou 50% 11% Segou 50% Gilo 10% Kayes 15% Sikasso Mepti Segou Silcasco 2% 10% Koulikoro Mepti 2% Tombouctou Tombouctou 10% Silcasco Tombouctou 10% Source: EMOP 2014 2% USEHOLD BUDGET SPENT ON EDUCATION, BY REGIONS, 2014 10% 2% 113 FIGURE 52: SHARE OF HOUSEHOLD BUDGET SPENT ON EDUCATION, BY REGIONS, 2014 Source: EMOP 2014 Figure 52: Share of household budget spent on education, by regions, 2014 TOMBOUCIOU .13% GAO .41% KOULIKORO .51% MOPTI .11% KAYES .50% SEGOU .63% BAMAKO 1.75% SIKASSO .79% Source: EMOP 2014 Tuition and fees constitute the largest share of household education expenditures. In 2014, this category accounted for 62 percent of all education expenditure. Books follow at 29 percent. All other education expenditures (tutoring, supplies, transportation, etc.) account for less than 10 percent. Books and supplies are a much higher share of household expenditures in rural areas FIGURE (where school fees are lower and private schools are rare. Across rural Mali, 53: DISTRIBUTION households allocate OF HOUSEHOLD nearly half of their budgets to books and supplies whereas the similar share in urban areas is only Source: EMOP 2014 26 percent (Figure 53). Not only urban Malians spent nearly twice as much on education (in nominal terms) they also allocated a much larger share of their income on tuition and fees. Once again, thisMAIL is an indicator of not just household capabilities, but also availability of schooling. The urban households include 4% 29% 5% 44% 62% SIKASSO .79% GAO .41%Koyos 5% Koulkoro 50 IDR TRILLIONS 2 10% PERCEN 10% 11% Gilo FIGURE 12: SPENDING BY LEVEL OF GOVERNMENT, 2001 2009 100 Koyos 40 50% Segou GAO .41% 62% 11% 250 100 50% 450 7 6 6 5 5 Gilo 4 6 8 30 20 71% 10 Mepti 90 Segou 31 28 35 29 40 47 41 34 41 2% 200 80 0 0 Silcasco 2% 2001 2002 2003 2004 Mepti 2005 2006 2007 2008 2009 2001 2002 2003 2004 2005 2006 2007 2008 2009 2% 70 IDR TRILLIONS 2009=100 10% Tombouctou 150 10% 60 65 66 59 65 55 48 55 59 Silcasco 50 Central 2% PERCENT KOULIKORO .51% 50 FIGURE 52: SHARE OF HOUSEHOLD BUDGET SPENT ON EDUCATION, BY REGIONS, 2014 Tombouctou District Source: World Bank sta estimates based on state budget data and Regional Financial information system data 100 40 (Sistem Informasi Keuangan Daerah, SIKD), Ministry of Finance KOULIKORO .51% Province MOPTI .11% 10%Source: EMOP 2014 50 10% 30 20 4% MOPTI .11% 10 31 28 35 29 40 47 41 34 41 FIGURE 67: SHARE OF TEACH 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: own calculations using NUPTIK Data, 2010 2% 2001 2002 2003 2004 2005 2006 2007 2008 2009 FIGURE 13: NON SALARY EDUCATION EXPENDITURE BY PROGRAMS AND LEVEL OF GOVERNMENT, 2009 2% Central 2 District Source: World Bank sta estimates based on state budget data and Regional Financial information system data 100 (Sistem Informasi Keuangan Daerah, SIKD), Ministry of Finance 100 80 SEGOU .63% Province % OF TEACHERS WITH SENIOR SECONDARY 10 11 OR BELOW AS HIGHEST EDUCATION LEVEL 12 90 BAMAKO KAYES .50% SEGOU .63% 80 15 9 23 80 BAMAKO 6 60 1.75% 70 1.75% 60 44 60 FIGURE 52: SHARE OF HOUSEHOLD BUDGET SPENT ON EDUCATION, BYFIGURE REGIONS, 2014 EXPENDITURE BY PROGRAMS AND LEVEL OF GOVERNMENT, 2009 PERCENT PERCENT Tuition 100 ON & Fees EDUCATION 13: NON SALARY EDUCATION 51 50 ITEMS & Other Tutoring 40 114 FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING VARIOUS 40 75 40 FIGURE 52: SHARE OF HOUSEHOLDTOMBOUCIOU Education Public Expenditure Review Guidelines BUDGET SPENT ON EDUCATION, BY REGIONS, 2014 73 2 100 30 Source: EMOP 2014 .13% 100 10 11 20 47 20 Source: EMOP 2014 12 23 90 25 20 Books 9 10 80 15 80 Supplies 6 0 Source: EMOP 2014 70 0 EARLY BASIC SECONDARY UNIVERSITIES OTHERS 60 44 60 CHILDHOOD EDUCATION EDUCATION 2008 2009 2008 2009 0 PERCENT PERCENT 100 EDUCATION TOTAL NON SALARY 51 50 40 75 40 73 30 Central Central 20 Central-80S District 47 20 SIKASSO .79% SIKASSO .79% GAO .41% 0 25 10 District Province Province 0 F EARLY BASIC SECONDARY UNIVERSITIES OTHERS 2008 2009 2008 2009 MAIL RURAL URBAN TOMBOUCIOU .13% CHILDHOOD EDUCATION EDUCATION EDUCATION TOTAL NON SALARY S Central Central KOULIKORO .51% Central-80S District District Province MOPTI .11% Province GAO .41% 4% 3% KAYES .50% SEGOU .63% .13% TOMBOUCIOU KOULIKORO BAMAKO .51% 1.75% MOPTI .11% 8% FIGURE 67: SHARE OF TEA KAYES .50% BAMAKO SEGOU .63% GAO .41% 29% 21% Source: own calculations using NUPTIK Data, 201 private schools as well as upper secondary and tertiary schools that tend to have higher fees. The 1.75% 80 % OF TEACHERS WITH SENIOR SECONDARY FIGURE 3.2: SOURCES OF FUNDS: 2010 OR BELOW AS HIGHEST EDUCATION LEVEL KOULIKORO .51% 44% MOPTI .11% 5% 100 60 FIGURE same thing holds true 53: DISTRIBUTION for households OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS under the poverty line (they spend half their education 6.9 SIKASSO .79% TOMBOUCIOU .13% 16 10 KAYES .50% BAMAKO SEGOU .63% 44% SIKASSO .79% 80 40 1.75% 5% FIGURE 3.2: SOURCES OF FUNDS: 2010 15 62% 71% 60 29 20 budget on tuition, and other half on books and supplies) compared to households above poverty Source: EMOP 2014 100 20 6.9 16 SIKASSO .79% 80 10 40 0 GAO .41% 15 line (for 53:tuition which, and feesOF are 64 percent, see AppendixON Figure 30). 4% 60 29 20 48 54 FIGURE FIGURE 53: DISTRIBUTION DISTRIBUTION OF HOUSEHOLD HOUSEHOLD SPENDINGSPENDING ON VARIOUS VARIOUS EDUCATION ITEMS ITEMS EDUCATION 20 40 0 KOULIKORO .51% Non SBM Grantee SBM Grantee MOPTI .11% MAIL RURAL URBAN 20 48 54 PTCA LGU Source: EMOP 2014 Source: EMOP 2014 Others FIGURE 53: DISTRIBUTIONTuition & Fees SPENDING Tutoring & Other 0 Figure 53: Distribution of household Supplieson various education items spending Community KAYES .50% BAMAKO SEGOU .63% OF HOUSEHOLD Non SBM Grantee ON VARIOUS EDUCATION ITEMS SBM Grantee 1.75% Books PTCA FIGURE 53: DISTRIBUTION OF LGU Source: EMOP 2014 HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Others FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Source: EMOP 2014 Community APPENDIX FIGURE 29: SHARE OF TOTAL HOUSEHOLD EXPENDITURES APPENDIX FIGURE 29: SHARE OF TOTAL HOUSEHOLD EXPENDITURES FIGURE Source: EMOP 2014 MAIL RURAL URBAN Share of total household expenditures, by region Share of total household expenditures, by region EDUCA MAIL RURAL Urban URBAN SIKASSO .79% MAIL Mali RURAL Rural URBAN Source: EMOP 2014 Source: EMOP 2014 MAIL RURAL URBAN Source: own 2% 2% FIGURE 67: APPENDIX FIGURE 29: SHARE OF TOTAL HOUSEHOLD EXPENDITURES APPENDIX FIGURE 29: SHARE OF TOTAL HOUSEHOLD EXPENDITURES Source: own calculati Share of total household expenditures, by region Share of total household expenditures, by region 4% 4% Source: EMOP 2014 3% 15% Source: EMOP 2014 3% Bamako 15% 8% 8% TOTAL SALARY SPENDING FOR BASIC RP. BILLIONS 2% Koulkoro2% Bamako MAIL RURAL URBAN Koyos Koulkoro 29% 29% 11% 21% Gilo 21% Koyos 50% 11% 15% Segou 50% Gilo TEACHER RATIO Bamako Mepti Segou 3% 2% 15% 4% 44% 5% 3% Koulkoro Silcasco Bamako 2% Mepti 4% FIGURE 53: 44% 5% TEACHERS DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Koyos 10% Tombouctou Koulkoro 10% Silcasco 11% 44% Gilo 11% Koyos Tombouctou 5% 50% 8% Segou 50% Gilo 62% 44% 71% 10% 10% 5% 8% 2% Mepti Segou 3% STUDENT 4% Source: EMOP 2014 62% Silcasco 2% Mepti EDUCATION 71% 10% Tombouctou 2% 10% Silcasco 2% Tombouctou 8% 4% 29% 29% 10% 10% 21% 21% 4% 29% 21% FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS 2% 2% FIGURE 52: SHARE OF HOUSEHOLD BUDGET SPENT ON EDUCATION, BY REGIONS, 2014 Tuition & Fees Tutoring & Other MAIL RURAL URBAN Source: EMOP 2014 Source: EMOP 2014 Books Supplies 44% 44% 5% 5% 5% Tuition & Fees FIGURE 52: SHARE OF HOUSEHOLD BUDGET SPENT ON EDUCATION, BY 44% & Other REGIONS, Tutoring 2014 Source: EMOP 2014 Books 44% Supplies 5% 44% 62% 5% 44% 62% 5% MAIL RURAL 71% URBAN FIGURE 67: 62% 71% 71% Source: own calculatio 4%& Fees Tuition TOMBOUCIOU .13% FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Source: EMOP 2014 Tutoring & Other GAO .41% Books 3% TOMBOUCIOU .13% 4% 4% Supplies TOTAL SALARY SPENDING FOR BASIC STUDENT TEACHER RATIO KOULIKORO .51% FIG 8% 4% MOPTI .11% GAO .41% FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING EDU URBAN ON VARIOUS EDUCATION ITEMS KAYES .50% SEGOU .63% BAMAKO MAIL RURAL 1.75% 29% 21% KOULIKORO .51% MOPTI .11% Sourc Source: EMOP 2014 Tuition & Fees Tutoring & Other Books & Fees Tuition Tutoring & Other KAYES .50% SEGOU .63% BAMAKO 1.75% SIKASSO .79% SIKASSO .79% Supplies 44% 5% MAIL 5% Books Tuition 44% RURAL & Fees Supplies URBAN Tutoring & Other Source: EMOP 2014 62% Tuition & Fees Tutoring & Other 71% Books FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Books Tuition & Fees Supplies Source: EMOP 2014 Supplies Tutoring & Other FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS Source: EMOP 2014 4% Books Supplies MAIL RURAL URBAN FIGURE 68: ESTIM EDUCATION TEAC MAIL RURAL 4% URBAN 3% Source: own calculations based o 8% Tuition & Fees Tutoring & Other 29% 21% Appendix Figure 30: Household expenditures in education, size and share, different 4% 3% Tuition & Fees 80k Tutoring & Other Books 8% TOTAL SALARY SPENDING FOR BASIC EDUCATION TEACHERS RP. BILLIONS Supplies 44% 5% APPENDIX 30: HOUSEHOLD FIGUREFIGURE 70k EXPENDITURES IN EDUCATION, 29% 21% 53: DISTRIBUTION OF HOUSEHOLD Books SPENDING ON VARIOUS EDUCATION ITEMS 5% 44% 60k Supplies 44% 62% 5% 71% 50k socioeconomic SIZE,AND groups SHARE, DIFFERENT SOCIOECONOMIC GROUPS 44% 5% 40k Source: EMOP 2014 62% 4% 71% 30k 4% 20k 10k Tuition & Fees Tutoring & Other 0 Books Supplies Tuition & Fees 2009 ratio = 18 Tutoring & Other Tuition & Fees Tutoring & Other Books Supplies Books Supplies Below Poverty APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES APPENDIX FIGURE SIZE,AND SHARE, Level DIFFERENT MAIL 30: SOCIOECONOMIC HOUSEHOLD Above Poverty Level IN EDUCATION, EXPENDITURES IN EDUCATION, GROUPS RURAL URBAN FIGURE SIZE,AND SHARE,53: DISTRIBUTION DIFFERENT OF SOCIOECONOMIC HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS GROUPS FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS FIGURE BELOW 53: DISTRIBUTION POVERTY LEVEL OF HOUSEHOLD ABOVE POVERTYSPENDING LEVEL ON VARIOUS EDUCATION ITEMS Source: EMOP 2014 FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ITEMS 2% Source: EMOP 2014 Source: POVERTY BELOW LEVEL EMOP 2014 FIGURE 53: DISTRIBUTION OF HOUSEHOLD SPENDING ON VARIOUS EDUCATION ABOVE POVERTY LEVEL MAILITEMS RURAL URBAN FIGURE 68: ES EDUCATION T Source: EMOP 2014 43% Source: EMOP 2014 5% 5% MAIL RURAL URBAN Source: own calculations b BELOW POVERTY LEVEL ABOVE POVERTYEXPENDITURES LEVEL 26% Tuition & Fees Tuition & Fees 80k TOTAL SALARY SPENDING FOR BASIC EDUCATION TEACHERS RP. BILLIONS APPENDIX FIGURE 30: HOUSEHOLD Books IN EDUCATION, 70k 26% Other Tutoring &GROUPS Tuition & Fees FIGURE 45: IMPORTANCE SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC Books 5% 60k 64% International Donor Funds, 2015 CFA 50k 5% Supplies MAIL MAIL RURAL RURAL URBAN URBAN 40k 52% Books 30k Tutoring & Other 2006 2 MAIL 4% RURAL URBAN 20k 26% Tuition & Fees 10k BELOW POVERTY LEVEL 64% POVERTY LEVEL ABOVE 5% Books Tuition & Fees Tutoring & Other Tutoring & Other Tuition & Fees Books Tutoring & Other Supplies CFA 53B 0 CFA 61B Supplies Supplies Books Supplies CFA 158B NO EDUCATION COMPLETE PRIMARY 64% 5%5% COMPLETE SECONDARY COMPLETE POST SECONDARY Tutoring & Other 26% Supplies Tuition & Fees Domestic Funding External Funding Books APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 64% 5% APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, Tutoring & Other SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS Tuition & Fees SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS Supplies Tutoring & Other Books BELOW POVERTY LEVEL ABOVE POVERTY LEVEL Supplies BELOW POVERTY LEVEL ABOVE POVERTY LEVEL No Education Complete Primary Complete Secondary Complete Post-Secondary 5% 5% 26% Tuition & Fees 26% Tuition & Fees Books Books NO EDUCATION COMPLETE PRIMARY COMPLETE SECONDARY 5% Tutoring & Other COMPLETE POST SECONDARY Tutoring & Other 64% 5% FIGURE 14: PUBLIC SPENDING ON EDUCATION LEVEL, SELECTED FIGURE COUNTRIES SORTED 45: IMPORTANCE OF DONOR 64% Supplies International Donor Funds, 2015 CFA Supplies Honduras Social Expenditure and Institutional Review NO EDUCATION 6% NO EDUCATION COMPLETE PRIMARY COMPLETE PRIMARY COMPLETE SECONDARY COMPLETE SECONDARY COMPLETE POST SECONDARY COMPLETE POST SECONDARY Source: World Bank/ICEFI social spending database for CA countries; EdStats for the rest of the countries 2006 2007 APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 3% 100 CFA 53B 1 CFA 61B C SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS 33% 26% Tuition & Fees 2% 1% 90 13 7 CFA 158B CFA 172B NO EDUCATION Tutoring & Other NO EDUCATION COMPLETE PRIMARY COMPLETE SECONDARY COMPLETE COMPLETE PRIMARY POST SECONDARY COMPLETE SECONDARY COMPLETE POST SECONDARY 16 19 19 12 21 Domestic Funding 19 15 BELOW POVERTY LEVEL ABOVE POVERTY LEVEL Books Supplies 12% 809% 12 10 External Funding 39 31 9 70 3% 5% 27 12 32 26 5% 60 26 PERCENT 36 APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 50 8 71% 19 GROUPS SIZE,AND SHARE, DIFFERENT SOCIOECONOMICTuition 77% 40 22 FIGURE 14: PUBLIC SPENDING ON EDUCATION LEVEL, SELECTED COUNTRIES SORTED BY GDP P 66% FIGURE 14: PUBLIC SPEN & Fees Honduras Social Expenditure and Institutional Review 69 Honduras Social Expenditure and Institutiona 56% 5% 30 61 18 Source: World Bank/ICEFI social spending database for CA countries; EdStats for the rest of the countries Source: World Bank/ICEFI social spending data Books 51 Tuition & Fees APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 48 100 48 46 1 FIGURE 100 14: PU7 Tutoring & Other Tutoring & Other APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 42 SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS 20 90 13 7 16 12 33 13 Supplies 19 19 21 90 19 22 Honduras Social Expenditur 10 23 15 80 10 15 Tuition & Fees 12 80 Books Tutoring & Other 39 BELOW POVERTY LEVEL ABOVE POVERTY LEVEL Source: World Bank/ICEFI so Supplies BELOW POVERTY LEVEL ABOVE POVERTY LEVEL 31 9 70 70 31 0 26 12 27 32 60 26 32 This estimate is based a BELOW POVERTY LEVEL on estimated the ABOVE POVERTYaverage LEVEL household expenditure on education from the most recent Enquête Modulaire PERCENT 27 36 60 PERCENT 100 Books Nicaragua Honduras Guatemala Georgia Paraguay El Salvador Colombia Costa Rica Panama APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 50 COMPLETE POST Supplies 8 APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 19 50 8 NO EDUCATION SIZE,AND SHARE, DIFFERENT COMPLETE PRIMARY SOCIOECONOMIC GROUPS COMPLETE SECONDARY SECONDARY 40 22 Permanente etSIZE,AND SHARE, Auprès DIFFERENT Ménages 5% desSOCIOECONOMIC (EMOP) data (2014 last quarter), which covers all parts of the country except for Kidal. GROUPS Tuition & Fees Books Tutoring & Other Tuition & Fees Books 30 48 51 61 48 46 69 42 40 18 30 90 46 51 20 APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, 80 Tutoring & Other 33 48 Supplies 20 23 Supplies 10 BELOW POVERTY LEVEL ABOVE POVERTY LEVEL 10 SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS 26% Tuition & Fees 0 0 70 Nicaragua Honduras Guatemala Georgia Paraguay El Salvador Colombia Costa Rica Panama Chile Books Pre-primary and Primary Nicaragua NO EDUCATION COMPLETE PRIMARY COMPLETE SECONDARY COMPLETE POST SECONDARY 60 PERCENT NO EDUCATION COMPLETE PRIMARY COMPLETE SECONDARY COMPLETE POST SECONDARY Secondary BELOW POVERTY LEVEL ABOVE 64% 5% POVERTY LEVEL Tutoring & Other Tertiary and Post-Secondary 50 BELOW POVERTY LEVEL ABOVE POVERTY LEVEL Other 5% Supplies Pre-primary and Primary Secondary 40 Pre-primary and Tuition & Fees Tertiary and Post-Secondary Secondary Books Other 30 Tertiary and Post APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES 26% IN EDUCATION, Tuition Tutoring & Fees & Other Other 20 SIZE,AND SHARE, DIFFERENT SOCIOECONOMIC GROUPS Books Supplies 10 5% Tutoring & Other APPENDIX FIGURE 30: HOUSEHOLD EXPENDITURES IN EDUCATION, Tuition & Fees Supplies 64% 0 115 Example 9: Analysis of donor funding Mali PER (2016) Overall comment: This provides a good example of how to address financial sustainability issues related to donor funding fluctuations. It provides analysis of donor funding before and after the political crisis in the country, examines the role of donor funding (donor activities), and mentions issues related to finding good international or regional comparisons on this source of funding, due to lack of a consistent definition of official development aid (ODA). Mali relies on funds provided by international donors, especially for capital investments. Through 2010, international funding had accounted for nearly a quarter of all budgeted education expenditures. Several international and bilateral donors and NGOs supported the education sector, and the focus of support was largely on school construction, school canteens, teacher training and quality improvements in general, with a strong focus on the south (Table 9). Table 9: Summary of main pre-crisis donor activities Agency Domain of Intervention Geographic Area The Netherlands, CIDA Quality improvement (reading Kati, Koulikoro, Segou, Mopti (Through NGOs and Firms) and writing; textbooks) UNICEF Teacher training, early childhood Bamako, Segou, Mopti, San, development (ECD) support, Kati, Kayes School construction Save the Children, Plan Mali, School canteens, support to Countrywide but inadequate Handicap International, Aga Khan CGS, school construction, ECD to meet all the needs Foundation, Right to Play, Islamic Relief, CRS, BIT, GARDL, JICA World Food Program (WFP) School health and feeding Bamako, Segou, Mopti, San, program Kati, Koulikoro USAID Scholarships Tombouctou, Gao, Kidal Source: Emergency Basic Education Program PAD 116 Education Public Expenditure Review Guidelines Since 2004, donor funding, on average, equaled 0.8 percent of Mali’s GDP. Because there is no consistent definition of official development aid (ODA), it is hard to provide international comparisons. One study puts the average official development aid for education at 5.6 percent of total government expenditures in education across Sub-Saharan Africa but the proportion of ODA in public education resources varies greatly across the region (UNESCO, 2013). For example, in 2008, 72 percent of total education funding in Liberia was financed externally; the comparable share was 42 percent in Ethiopia, 35 percent in Eritrea and Burundi, 26 percent in Niger, 24 percent in Central African Republic, 9 percent in Togo, and 3 percent in Democratic Republic of Congo (all very low income countries like Mali). Donor funding could be unreliable. It completely disappeared in 2012 due the political crisis. Before the crisis, donors had committed upwards of CFA 350 million to PISE III, but according to budget records, only CFA 262 billion, or 85 percent of these funds were received (See Box 4 on the importance of external funds in Mali’s capital improvement plan). After the coup d’état and suspension of aid, donors support to the sector was channeled mainly through UNICEF, NGOs and other direct financing interventions for short-term humanitarian and recovery response. Donor support in education has resumed since the end of 2013, but is not back to normal yet. 2014 and 2015 budgets once again include donor funding, which is expected to pay for approximately 3 to 6 percent of budgeted expenditures (Figure 45). Figure 45: Importance of donor finding, over time Domestic Funding External Funding Source: International Donor Funds, in 2015 CFA 117 Example 10: Analysis of total public education spending Albania PER (2014) Overall comment: This analysis of education expenditure as a share of gross domestic product, total government spending, and per capita spending, provides good international benchmarks based on similar demographics, such as similar shares of school-age populations. Public spending on education is particularly low considering the composition of Albania’s population. About one-third of Albanians are 19 or younger. Table 2.7 shows that Albania compares unfavorably with most other countries with similar shares of school-age population. Even poorer countries like Morocco and Vietnam invest more of their income on education. Public spending per student is among the lowest in Europe even taking into account Albania’s income level. Table 2.7: School-age population and public spending on education Share of Public expenditures on Public expenditures on education GDP per Country population below education as share of as share of total Government capita (PPP) 19 years old (%) GDP (%) expenditures (%) Albania 32.5 2.9 9.5 9,403 Europe and Central Asia Kazakhstan 33.5 3.1 - 13,667 Azerbaijan 32.7 2.4 7.2 10,125 Armenia 29.1 3.1 13.7 8,417 FYR Macedonia 25.0 - - 11,834 Serbia 23.7 4.4 10.6 11,801 Croatia 21.0 4.9 9.9 20,964 Mexico 39.7 5.3 19.4 16,734 Chile 30.8 4.1 19.4 22,363 New Zealand 27.8 7.3 18.7 32,219 OECD Ireland 27.5 6.5 9.8 43,683 United States 26.9 5.6 12.7 51,749 Peru 39.8 2.6 18.1 10,765 Morocco 38.3 5.4 - 5,220 Colombia 38.2 4.4 15.8 10,436 Countries Other Brazil 33.9 5.8 18.1 11,716 Vietnam 33.6 6.8 20.9 3,787 Source: World Development Indicators, UN DESA Population Projections and UNESCO Institute for Statistics Note: Population figure are 2010 estimates. Expenditures are 2010 or latest available and GDP per capita is from 2012. BELOW POVERTY LEVEL ABOVE POVERTY LEVEL 2% 5% Tuition & Fees 26% 43% Books Tutoring & Other 118 Education Public Expenditure Review Guidelines 52% 64% 5% Supplies 4% NO EDUCATION COMPLETE PRIMARY COMPLETE SECONDARY COMPLETE POST SECO 2% 1% Example 11: Analysis of functional classification 6% 3% 9% 12% 33% 26% 5% 5% Honduras PER (2015) 56% 5% 66% 71% 77% Overall comment: This analysis examines education expenditures by functional classification--i.e., level of education or subsector. It provides comparisons with countries at similar stages of economic development and provides international benchmarking for changes in the composition of spending by level of education as the country develops. In Honduras, primary education takes up the of public bulk 45: FIGURE spending IMPORTANCE onFINDING, OF DONOR education, but this share OVERTIME is expected to fall as the country develops. Pre-primary and primary educationa alone represents 51 International Donor Funds, 2015 CFA percent of all educational expenditures, while, on average, a quarter of the budget is devoted 2006 2007 2008 2009 2010 2011 to secondary education and around 15 percent to tertiary education. This distribution is in line with CFA 53B CFA 61B CFA 49B CFA 56B CFA 21B CFA 39B countries at similar stages of economic development, such as Guatemala and Nicaragua (Figure 14). CFA 158B CFA 172B CFA 168B CFA 192B CFA 213B CFA 248B However, countries with a higher level of economic development tend to show more balanced Domestic Funding spending across levels, with their spending on primary education ranging between a minimum External Funding of 20 percent and a maximum of 35 percent. Therefore, as Honduras develops, international benchmarking suggests that public spending on primary education should decrease in relative importance compared to other educational levels, even if public spending on primary education increases in absolute terms. FIGURE 45: IMPOR International Donor Funds, 2015 CFA Figure FIGURE 14: 14: Public PUBLIC spending SPENDING onLEVEL, ON EDUCATION education by level, SELECTED COUNTRIES selected SORTED countries BY GDP PER CAPITA sorted by GDP per capita (circa 2013) Honduras Social Expenditure and Institutional Review Source: World Bank/ICEFI social spending database for CA countries; EdStats for the rest of the countries 2006 100 1 CFA 53B 13 7 10 Pre-primary and 90 16 19 19 12 21 19 22 Primary CFA 1 15 80 12 10 39 17 32 Secondary 31 9 70 26 Tertiary and Post- 2011 12 60 27 32 26 32 Secondary PERCENT 36 CFA 39B 50 8 19 38 Other 42 CFA 40 69 22 61 18 30 51 48 48 46 46 42 Domestic Fu 20 33 35 26 External Fun 10 23 0 a p al Re nd a ile ua Co bia ica a as ay Co or m gi m , ad ur Ch g gu la te ea aR or m na ra Fin nd ua or lv ra lo Ge Pa ca st Sa Pa K Ho G Ni El Source: World Bank/ICEFI social spending database for CA countries; EdStats for the rest of the countries. Pre-primary and Primary Secondary Tertiary and Post-Secondary Other 119 Kenya PER (2014) Overall comment: A functional analysis shows that Kenya’s university subsector receives a disproportionally large share of the current sectoral budget in contrast to primary education. These relative shares raise serious questions about equity and almost certainly about efficiency. Intra-sectoral composition in the education sector can benefit from rationalization in order to enhance efficiency and equity. The education budget allocation in 2014/15 is skewed in favor of tertiary education at over 40 percent of the total sector budget, which compares unfavorably to 26 percent allocated to primary education. A functional analysis shows that Kenya’s university sub-sector receives 39 percent of the current sectoral budget in contrast to 26 percent for primary education (grades 1-8). These relative shares raise serious questions about equity and almost certainly about efficiency. Figure 1.21: Current expenditure composition in education could undermine efficiency and equity Total education sector budget, 2014/15 Primary Education Secondary Education University Education TIVET Other Source: Staff computation based on the National Treasury data a In this report, due to reasons of consistency and comparability, primary education includes the first two cycles of basic education, and secondary education includes the third cycle of basic education (lower secondary education) and high school (upper secondary education). The Honduran educational system is structured according to the following levels: pre-primary education (ages 3 to 5), basic education (1st cycle, ages 6 to 8; 2nd cycle, ages 9 to 11; 3rd cycle, ages 12 to 14), high school education(ages 15 to 17) and higher education (ages 18 to 22). 120 Education Public Expenditure Review Guidelines Example 12: Analysis of education spending by economic classification Honduras PER (2015) Overall comment: This example shows analyses of education expenditure by economic classification (with a particular focus on spending on wages) and provides regional and international benchmarking on the wage bill. The wage bill, accounting for almost 90 percent of the total public spending on education, is strikingly high in Honduras when compared to similar countries. A large wage bill is partly attributed to the high average level of teachers’ salaries, especially after the significant increase in the minimum wage in 2009 (by 63 percent). Figure 16 shows that Honduras spends considerably more on the wage bill than its neighboring CA countries or even other LAC countries with higher income, such as Chile or Colombia. The share of expenditures going to salaries is also much higher than countries with top-class education systems such as Finland and Korea. In 2012, only 2 percent of the total public spending on education went to construction, renovation, rehabilitation and/or non-routine maintenance of the facilities. Other recurrent expenditures accounted for the remaining 8 percent. This picture is quite similar for higher education. Between 2008 and 2011, the share of wages averaged 83 percent of total higher education expenditures. Nevertheless, universities devote a larger share of its budget to capital expenditures, averaging 7 percent for the same years. FIGURE 16: WAGE BILL AS A PERCENTAGE OF PUBLIC EDUCATION SPENDING, CIRCA 2013 Figure 16: Wage bill as a percentage of public education spending, circa 2013 Source: SEDUC (2012). O cial data for El Salvador (2011). EdStats for the rest of the countries. 100 90 88 87 82 81 80 80 79 79 % PUBLIC EDUCATION SPENDING 73 70 68 60 57 50 49 49 40 30 20 10 0 Honduras Costa Rica Guatemala Chile Colombia Nicaragua Jordan El Salvador Paraguay Panama Finland Korea, rep Source: SEDUC (2012). Official data for El Salvador (2011). EdStats for the rest of the countries. 121 Example 13: Issues related to budget formation and execution Solomon Islands PER (2011) Overall comment: This example describes and analyzes issues of budget formation and execution in the Solomon Islands. It assesses accountability for the use of resources and results achieved, and recommends options for addressing the identified problems. The PER team found three main issues related to budget formation: a) policy priorities, plans, and budgets are not well linked or integrated; b) budgetary allocations are made with little consultation with the line ministries or citizens and without a feed-back process to facilitate corrections; and c) there is a general lack of consistency over time in allocations to public services. It is not clear how budgets support SIG policy and plans. The Ministry of Planning requests that each ministry develop a four year corporate plan. Ministries are also asked to prepare an annual or operational plan for each year of the corporate plan. These plans lay out in more detail, what will be achieved from the implementation of each activity. These plans typically include a large number of goals, often far more than could feasibly be achieved under present human and budget constraints. National plans set targets for outputs and/or outcomes but these are not generally accompanied by cost estimates. By contrast, the national budgets are organized in terms of inputs and do not include clear expectations for service delivery targets. National plans and budgets are not linked: the budget does not show clearly and consistently what activities are to be delivered and how spending contributes to government policy priorities. Consultations with line ministries have been poor. The only portions of a ministry budget that are systematically deliberated upon are a small fraction of the recurrent budget set aside for new activities,a and the development budget.b Decisions regarding most of each ministry budget are largely incremental, meaning that the bulk of the allocation decisions are percentage increases over the previous year’s baseline allocations – decisions are made with insufficient regard for what ministries should achieve or how much their goals would cost. There are few mechanisms for consultations with citizens. Consultations with citizens and provincial governments are not used well in forming and prioritizing national plans and line ministry corporate plans. The budget process includes no formal processes for hearing citizen input and encouraging debate regarding allocations of resources. Unlike in many other poor countries, there is very limited discussion within media or civil society regarding resource allocation decisions included in the budget. There may be a need to better align budgetary allocations with ministry requirements. There is a general lack of consistency over time in allocations to large, priority public services such as policing, health, or education. 122 Education Public Expenditure Review Guidelines Actual spending does not resemble approved allocations. While execution across some Ministries has been relatively close to what was approved by Parliament, there are more ministries where actual expenditure has very substantially exceeded or fallen short of budgets. This is shown in Table 3 below. In 2009 for example, 6 of 30 budget heads were spent within plus or minus 10 percent of their original approved recurrent allocations. Execution of the development budget is very weak. While SIG contributes only a fraction of the funding for the development budget, Table 3 shows that fraction has been underspent in most ministries. In 2006-09, for example, no more than two ministries spent within plus or minus 10 percent of their approved allocation from the consolidated portion of the development budget. The majority of ministry allocations were under-spent. This is partially the consequence of SIG placing a higher priority on recurrent obligations (notably the wage bill) but it may also a reflection of generally weak capacity for project implementation throughout SIG. Table 3: Budget execution, 2006-09 2006 2007 2008 2009 Deviation in SIG Controlled Resources (% of Budgeted Amount) 0.3 4.8 -11.9 -13.6 Recurrent Spending a Good Execution (Number of budget heads withing ±10%) 10 14 16 6 Poor Execution (Number of budget heads) 5 15 14 24 Over +10% of approved allocation 2 8 2 4 Under -10% of approved allocation 3 7 12 20 Spent without allocation 0 0 0 0 Total Budget Heads with Allocations or Expenditures 15 29 30 30 Development Spending, Consolidated Good Execution (Number of budget heads withing ±10%) 0 2 1 0 Poor Execution (Number of budget heads) 9 13 25 26 Over +10% of approved allocation 1 1 0 1 Under -10% of approved allocation 8 12 25 25 Spent without allocation 0 2 0 0 Total Budget Heads with Allocations or Expenditures 9 17 26 26 Development Spending, Non-appropriatedb Good Execution (Number of budget heads withing ±10%) 3 10 4 5 Poor Execution (Number of budget heads) 14 13 23 20 Over +10% of approved allocation 6 6 10 3 Under -10% of approved allocation 8 7 13 17 Spent without allocation 7 2 1 2 Total Budget Heads with Allocations or Expenditures 24 25 28 27 a Excludes sector budget support for health and education. b The outcome in 2009 reflects changes in the accounting of support for police and justice programs as well as some under- counting in other ministry programs rather than actual under-spending. Sources: MOFT and World Bank staff calculations using most recent database from January 2011. 123 There has been little accountability for the use of resources and results achieved. The public has not been able to see what activities and programs the government was using money for, nor the results achieved, because of the way the budget is presented and because of delays in reporting actual outcomes. Line ministries have not been held to account by Ministers, in part because there is little information about what activities they will deliver or results they will achieve. It is important to deal with the formation, execution, and accountability problems together because they are mutually reinforcing. The incentive to budget well increases when three conditions are met: i) line ministry officials have confidence that they will be consulted; ii) ministry officials have confidence that allocations provided will match what was agreed during consultations; and iii) the approved budget is the final word -- ministers cannot press for changes except in clear cases of emergency. Similarly, the incentive to execute each budget faithfully improves when line ministry officials believe it meets their needs and when they are held accountable for their decisions. a The process also makes it is difficult to know with certainty whether bids for alleged new activities represent genuine new activities or instead represent inflated costs for ongoing activities. b The public expenditure review team received mixed reports as to how much scrutiny development budget expenditures get. 124 Education Public Expenditure Review Guidelines Example 14: Analysis of per student spending Armenia PER (2011) Overall comment: This offers a good example of examining the distribution of educational inputs (including per student spending, average class size, average school size, and student-teacher ratio) by school location and community type. The analysis points to inefficiency in the utilization of educational inputs across schools. The main contrast in educational efficiency within Armenia is between small rural schools and large urban ones. In 2009/10, rural schools had an average of 174 students, with an average class of 13.6 and a student-teacher ratio of 8.0. In contrast, urban schools educated an average of 450 students, grouping them 21.8 per class and 11.4 per teacher. Rural schools have been able to flourish with smaller classes because they have ample funding (rural schools receive 53 percent more per student than urban ones) and do not run into infrastructure constraints (nationwide, rural schools utilize less than half of their building capacity). Similar trends are evident when comparing schools in mountainous/highly mountainous locations to non-mountainous ones, as well as the only schools in small communities versus all other ones (Figure 4.5). The pattern of potential inefficiency is even more evident when viewed through the lens of school size. Nearly two-thirds of Armenia‘s schools have 300 or fewer students. These institutions receive 42 percent of all PCF funding and employ 43 percent of staff in general secondary education. Yet the utilization of educational inputs in these schools is strikingly inefficient. For example, among the 369 schools with 100 or fewer students (27 percent of all schools in Armenia), the average class size is 5.6 with 3.7 students per teacher. These schools utilize only 27 percent of their available building capacity, compared to 56 percent nationwide. 0 E Hond N N Pan Korea,Korea Fin K K Jo El Salv Costa Costa Guate Para Nicar Colo El El Honduras Rica Guatemala Chile Colombia Nicaragua Jordan El Salvador Paraguay Panama Finland rep FIGURE 4.5: DISTRIBUTION OF EDUCATIONAL IMPACTS BY COMMUNITY TYPE, 2009/10 125 Source: World Bank calculations based on data from the National Statistics Service (NSS), The National Center for Education Technology (NaCET), the Assessment and Testing Center (ATC) and MOF MPACTS ACTS BY COMMUNITY BY COMMUNITY TYPE,TYPE, 2009/10 FIGURE 4.5: DISTRIBUTION2009/10 OF EDUCATIONAL IMPACTS BY COMMUNITY TYPE, 2009/10 Republic of Ameri Source: World Bank calculations based on (a) Average data PCF from the National Allocation Statistics Service (NSS), per Student (b) Average School Size S), The National Center for Education Technology (NaCET), the Assessment and Testing Center (ATC) and MOF (000s AMD) (Number of Students) nter TC) and (ATC) MOFand MOF FIGURE 4.5: DISTRIBUTION OF EDUCATIONAL IMPACTS BY COMMUNITY TYPE, 2009/10 Source: World Bank calculations 300 Figure 4.5: Distribution of educational inputs by community type, 2009/10 (a) Average based on PCF data from the National Allocation Statistics Service (NSS), per Student 500 (b) Average School Size Republic of Ameri 250 The National Center for Education Technology (000s AMD)(NaCET), the Assessment and Testing Center (ATC) 400MOF and (Number of Students) (b) Average (b) Average SchoolSchool 200 Size Size Republic of Armenia 300 Republic Republic of America of America 150 Republic of Ameri (Number (Number of Students) of Students) 300 (a) Average PCF Allocation per Student 100 250 500 200 (b) Average School Size 50 (000s AMD) 400 100 (Number of Students) 200 0 300 0 150 l s ra ou l us 500 500 n ra ta s 300 0 m y s en ity 200 500 n nt us ou 0 m y us en ity Ru no ou ba 40 om nit 40 Com nit 40 Com nit 100 in Ru ba 40 om nit 40 Com nit 40 Com nit ud n ts ai ou ino a no m tain ud n ts Ur in u St u t Ur u St u 250 t m n ai m n n- nta 50 ou 100 400 n w ol i om w ol i om w ol i om ou n w ol i om w ol i om w ol i om No ou ou 400 400 No ou 200 M ho n C M M ho n C ly M 0 y m m ith n a 300 l m ith n a h 0 - Sc s l i h n Sc s l i l ig ly ou oo 150 s ig u < ly nou oo ra H s u < 300 300 ou alH Ontai Sch ou n s 0 m y en ity Ontai Sch ur an n ly us 200 u y en ity Ru ba in ou n o 100 in ud n ts ou n o b a R m t O ain ud n ts Ur n y St u t a m O in Ur St u t l n n 0 m n- Not nta t ou No n 50 100 ou n 200 200 No ou No ou yM ho n C M M ho n C l M y m ith n a 0 l m ith n a gh - Sc s l i 0 h n Sc s l i i ly ou oo ig < lynou oo lH s 100 100 < ra ou lH us in h n ra s ai ch 0 m y en ity ta S c an nt ly Sus un nlyou y en ity Ru no ba in Ru ud n ts i ou n o b ta in ud n ts Ur u St u ta m O in Ur u St u un s y mo t Oa un 0 m n- Notnta t On 0 0 on Noun On o o No ou yM o < aC o nC yM M < aC ho in C us us l M al al gh hl - Sct ol i ith n ur ur n n ith n s nt us Hi ty s ho 4 n Co uni s ty oy o o g l Hi oo Mou nou nlSc ol in ou nou ly itholoin mm nou ba ba niiN t t in in nle ch i R R o On Sch < a C de un Sn hs ta ta (c) Average Class Size (d) Student-Teacher Ratio t Ur Ur in ith n tu mu St nlyu n yn Sc i i i un un w l i S mm 0 t Om a a y nl cho n-m nta a nl nt ly ount Od m om tO o o u ou 40 No o o yM yM No Sc < l i a C C M m l l 0 12 0 gh gh n- Hi Hi No o 25 10 (c) Average Class Size (d) Student-Teacher Ratio yS N On w cho o y Sho 20 8 6 nl 15 y tO tO 12 NoO 10 4 No 25 10 5 20 (c) Average Class Size 2 8 (d) Student-Teacher Ratio 0 0 15 6 s l us ra ou n l nt us 12 0 m y us en ity ra Ru