PRACTICAL GUIDELINES for Preparing a Public Expenditure Review for Education at the District Level

Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized 53286 PRACTICAL GUIDELINES for Preparing a Public Expenditure Review for Education at the District Level Notes Examples Data Key questions Data

Table of Contents TABLE OF CONTENTS 2 1. INTRODUCTION 3 2. PLANNING AND BUDGETING 4 3. EDUCATION EXPENDITURE 5 1. Source of district education budget 5 2. District education spending picture 6 3. Composition spending analysis 6 4. Household out-of-pocket spending analysis 9 5. Comparative analysis 9 4. EDUCATION PERFORMANCE 12 1. Input analysis 12 2. Output analysis 13 3. Outcome analysis 15 5. EQUITY AND EFFICIENCY ANALYSIS 16 1. Equity analysis 16 2. Efficiency analysis 17 3. Best practice frontier 18 6. APPENDIX 20 DATA REQUIREMENT S 20 DATA ANALYSIS 22 2

1. Introduction In recent years district education expenditure has grown rapidly both in terms of level and as a share of national education expenditure. Total district education expenditure increased from Rp 26 trillion in 2001 to Rp 52 trillion in 2006 and constituted 50 percent of total national public education expenditure in 2006. For many districts, education is a priority in local government budgets and on average absorbs almost one third of spending at the local government level. Moreover, education at the district level has been prioritized with the enactment of Law No. 20/2003 on the National Education System, which requires central and sub-national governments to allocate a minimum of 20 percent of their budgets to the sector. However, outputs and outcomes still vary despite the increase in education spending. Some districts lag behind, while others have made significant progress in meeting their education goals. The variations in teacher distribution, the supply of schools, the quality of infrastructure facilities, and other input resources may be factors behind the difference in outcomes. The lack of synchronization between planning and budgeting, as well as inefficiencies in budget allocation, may also be hampering outcomes. Understanding the spending patterns of district governments and how these correlate with other inputs and outputs is one of the tools that can help to translate the significant resources in education into improved outcomes. By conducting a public expenditure analysis on education, a proper assessment of the effectiveness and efficiency of district government spending can be made. This manual is intended as a guide for stakeholders at the district level to conduct their own assessments of public expenditure on education in order to achieve the above objectives. The preparation of this manual is intended to be part of SISWA capacity building program for local governments. The development of the manual is based on the methodology that has been applied in the PEACH (Public Expenditure Analysis and Capacity Harmonization) program. 3

2. Planning and Budgeting Objective The objective of this chapter is to: (i) analyze the consistency between planning, budgeting, and actual expenditure patterns; (ii) analyze the consistency between sectoral and regional plans; (iii) analyze whether the education sector plan reflects the actual issues and challenges faced by the education sector in the concerned region; and (iv) analyze the strengths and weaknesses of the planning and budgeting process. The basic documents required to conduct the analysis are budget documents (APBD) that can be obtained from the budget offices and the planning documents from the Bappeda office. Type of analysis required 1. Development priorities analysis. (i) Describe regional development priorities and discuss priorities for the education sector as mentioned in regional RPJMD (medium-term development plan). (ii) Assess whether the priorities reflect actual issues and challenges of the education sector in the concerned region based on current educational outcomes and outputs. (iii) Assess the level of linkages and synchronization between various planning documents. 2. Planning and budgeting linkages. Look at the budget as a manifestation of development planning and establish development priorities. Assess whether the existing budget (APBD) actually links with the planning documents. Assess whether the public budget reveals the development goals stated in the planning documents. 3. Good practice for planning and budgeting: Review whether the existing planning and budgeting documents take into account the Medium Term Expenditure Framework (MTEF) and performance-based budgeting (PBB) principles. See how the concepts of the MTEF and PBB are applied in the current budget system. 4. Participative development planning and budgeting. See if the concerned region has applied a participative process in development planning. Understand the benefits and drawbacks of such a process. 5. Regulatory framework analysis. Examine whether the process of planning and budgeting is in accordance with existing regulations and meets the required timeframe. 6. Budget performance system. Check the availability of performance indicators and analyze whether the performance indicators have been correctly defined. Ensure that the performance indicators are realistic and measurable. 4

3. Education Expenditure Objective The objective of this chapter is to understand how local governments spend their education budgets by looking at the source of education spending, trends over time, the composition of education spending by economic classification and program, and by comparing education spending with other sectors, and by comparing education spending with other districts or with the national average. The basic documents required to conduct the analysis are budget documents (APBD) that can be obtained from the budget offices or Bappeda, Education Dinas budget documents (DASK), and planning documents in the sector. Type of analysis required 1. Source of district education budget. The aim is to obtain an overall picture of the sources of the education budget by looking at central government education spending on the districts including BOS funds, district education spending, and household expenditure within the district. Sample 1. Source of district education expenditure, 2005 The sample chart above shows relative shares of the major contributors towards education spending in districts. The chart suggests that spending on education is largely from local governments, followed by households and then the central government. This chart also provides an overview of the current level of fiscal decentralization in the education sector. To 5

create the chart, data are required on local government budgets from district budget reports, central government deconcentrated funds from provincial governments or treasury offices, and household data from BPS. 2. District education spending picture. Here the objective is to better understand how trends over time in total district education spending compare with overall district spending. The analysis is conducted by looking at absolute spending over time, the share of education spending to total district spending, and education spending per capita of the district compared with other neighboring or similar characteristic districts. For trend analysis, constant price values are preferable. Sample 2. District education expenditure trends, 2001-06 The sample chart above provides two sets of information within a single chart: first, it shows the trend in absolute education spending and, second, it shows the share allocated towards education from the total district government budget over time. The chart prepares the ground for further analysis into reasons behind the decline in the share of education spending despite an increasing trend in absolute education spending. 3. Composition spending analysis. Here the objective is to look at the composition of education spending categorized into routine versus development, or capital versus non-capital. Spending can be further disaggregated into economic and program classifications. Analysis of economic classifications can be explored by breaking down spending into personnel, goods, operational/maintenance, travel and others. For education spending by program, the information is normally available in a separate budget book that is held by the working unit of the district education office. A disaggregation of spending by education level can also be conducted if data are available. It is worth noting that prior to analyzing the spending composition, it is important to provide background on the current budget format (direct and indirect spending) and the evolution of the local government s budget format, particularly in 6

terms of how spending has been classified over a 5-6 year time span. A mapping between various budget formats, particularly on the spending side, is necessary in order to obtain a consistent trend analysis of spending by composition. Sample 3. Composition of routine and development education expenditure, Percent 2001 2002 2003 2004 2005 District Rtn Devt Rtn Devt Rtn Dev Rtn Devt Rtn Devt Kab. Asahan 93.0 7.0 92.6 7.4 92.6 7.4 93.1 6.9 96.5 3.5 Kota Binjai 89.0 11.0 91.0 9.0 84.1 15.9 91.1 8.9 90.3 9.7 Kab. Wonosobo 96.7 3.3 92.5 7.5 88.5 11.5 84.2 15.8 90.7 9.3 Kota Magelang 97.2 2.8 92.7 7.3 71.2 28.8 87.7 12.3 88.2 11.8 Kab. Minahasa 98.2 1.8 99.9 0.02 98.9 1.1 96.8 3.2 88.6 11.4 Kota Manado 83.2 16.8 98.8 1.2 91.8 8.2 96.4 3.6 96.0 4.0 Kab. Timtengsel 87.1 12.9 90.5 9.5 89.5 10.5 84.4 15.6 82.2 17.8 Kab. Belu 95.4 4.6 94.5 5.5 86.1 13.9 87.2 12.8 87.8 12.2 Kab. Jayawijaya 68.5 31.5 76.9 23.1 86.9 13.1 97.9 2.1 90.7 9.3 Kab. Jayapura - - 51.3 48.7 45.8 54.2 92.1 7.9 80.1 19.9 All districts Indonesia 89.3 10.7 86.9 13.1 87.1 12.9 88.9 11.1 86.1 13.9 The table above gives a breakdown of expenditure into routine and development spending. In order to obtain a classification of routine and development spending, mapping between the budget format that consists of apparatus and public, and the previous budget format that consists of routine and development, needs to be conducted. This mapping is particularly important given that sub-national budget formats underwent changes in 2003 and more recently in 2007. The mapping is intended to give a consistent historical trend of spending. 7

Sample 4. Nias economic composition of routine expenditure, 2001-05 Routine spending based on economic classification can be further disaggregated, as in the chart above, into personnel, good and services, operational and maintenance, official travel, and others. These economic classifications make it possible to see which component receives the largest allocation and how this affects other components. In the sample chart, spending on personnel absorbs the largest share of the district education budget over time, leaving only minor shares for the other components. Analysis can also be done by looking at the share going towards maintenance and linking this to the condition of education infrastructure and facilities across districts. Sample 5. Program composition of the visited districts, 2006 8

In addition to the economic classification, analysis can also be conducted by looking at the breakdown of spending by program classification. The sample chart above shows the share of spending that is allocated to each program. However, it should be noted that the spending classifications by program may not always be the same from one district to another. A further analysis based on this chart can be conducted by comparing the spending allocated towards each program with the outputs/outcomes related to the program. 4. Household out-of-pocket spending analysis. In this analysis, the contribution from households in terms of their out-of-pocket spending on education is analyzed. For example, this analysis looks at: the size of the contribution of households towards education; the trend of these contributions over time; and the composition of spending within different income levels. The impact of new education financing programs, such as the BOS program, scholarship programs, etc. on household spending patterns can also be analyzed. Sample 6. Household out-of-pocket expenditure in some districts, 2004-06 5. Comparative analysis. Here a relativity analysis across districts and provinces, as well as across strategic sectors, is conducted. Compa rison between the amount of spending allocated to strategic sectors in specific years provides a picture of whether allocations for education spending are sufficient or not. Some key questions include: what share is allocated to education compared with other strategic sectors such as health and infrastructure; how does the district performance on education spending compare with other districts or nationally; and which districts have the highest level of per capita education spending. The reasons for significant differences should then be elaborated. 9

Sample 7. Overall sectoral expenditure in Kab. Nias and Kab. Nias Selatan, 2001-05 Sample 7 provides a comparison of spending on education and other major sectors. The aim of the chart is to show the priorities of the local government in allocating its resources and to obtain an overview of the relative level of education spending compared with other strategic sectors. Analysis can also compare spending in terms of the share (percentage) that is allocated to education with other sectors. The sample chart above suggests that education has been prioritized in the local government budget over several years and current trends could be expected to continue in the absence of any significant change of policy. Sample 8. Per capita education expenditure, 2005 10

In addition to an assessment across sectors, a comparison across districts and provinces is also helpful in providing a picture of how a district has been performing compared with other districts. The comparator districts can be selected based on indicators such as districts within similar geographical group, economic groups, or other relevant characteristics. A national average figure can be used as a reference or a basis point to assess how far ahead or behind a district is on a specific indicator/target compared with other districts in the country. 11

4. Education Performance Objective The objective of this chapter is to look in more detail at the links between spending and some of the key performance indicators in the education sector. A discussion of performance is informed by looking at some of the indicators on the input side (e.g., facilities, human resources, etc.), as well as outputs (e.g., enrollment rates, literacy rates, etc.), and indicators on the outcome side (e.g., test scores). A comparison between sub-districts, districts, and national average figures is useful in providing some insights. The basic documents required to conduct the analysis are: annual district education statistics; planning documents at the Dinas level; BPS surveys (Susenas, Podes, etc); and District in Figures reports. Type of analysis required 1. Input analysis: Analysis in this section looks at the condition of facilities, learning and teaching equipment, and human resources in the education sector. Some key questions include: Are there enough schools in the district? What is the condition of the classrooms? Are there adequate books for students? Are there adequate numbers of qualified teachers for the district (student-teacher ratio, classroom size), etc.? Links with education spending can be seen for example by comparing the condition of classrooms with the maintenance and operational budget, or by comparing the number qualified teachers with the budget for teacher/staff capacity-building programs. Sample 9. Primary schools/1,000 primary school-aged children in Kab. Wonosobo, 2006 12

The chart above shows the ratio of primary schools per 1,000 primary school-aged children in Kabupaten Wonosobo. The ratio attempts to measure whether the area has an adequate number of primary schools. In the above illustration, the ratio varies from a high of 14 schools in a sub-district to a low of three schools in a sub-district. In the highest sub-district 14 schools serve 1,000 school children or about 71 children per school, indicating a surplus number of schools. In contrast, in the sub-district with the lowest ratio, one school serves about 333 school children, or an average 55 children per grade, indicating that the number of schools is slightly below the general recommended level. The chart can be used to indicate which areas still require new schools and which areas have already been saturated. Sample 10. Teacher qualifications at the primary school in some districts, 2006 This chart shows the wide variation between districts in terms of teacher qualifications. The regulatory framework mandates that teachers should have at least a D-4 diploma (equivalent to S1) or bachelor degree (S1). Most districts have only a small proportion of bachelor graduate teachers at primary level, while most teachers in lagging districts are only high school (SLTA) graduates. This chart can be used as a guide by the district government in allocating more resources for upgrading the qualifications of teachers in the district. 2. Output analysis: The analysis in this section assesses the achievement of outputs in the education sector by analyzing various output indicators over time and comparing these with other districts or national averages. Some examples of output indicators that are used in this analysis are: net and gross enrollment rates, drop-out rates, literacy rates, percent of population aged 15 and over that has never attended school (this can be linked to the literacy rate), and the mean number of years of schooling. The link to education spending can be seen by comparing spending on the sector over time (5 to 6 years) with changes in output performance indicators. 13

Sample 11. Junior secondary gross enrollment rate in Kab. Nias, 2001-05 This chart shows gross enrollment rates at the junior secondary level over the period 2001-05. There are several ways to interpret this chart: first, the trend over time reveals that Kab. Nias has improved its gross enrollment rate at the junior secondary level. Second, the gross enrollment rate can be compared with provincial and national averages, revealing that Nias is lagging behind both. Sample 12. Mean years of schooling in some districts, 2001-05 The mean number of years of schooling indicates the average number of years that the population in a certain district spends at school and therefore indicates the quality of human capital in the area. A mean number of years of schooling of about 9 years, as seen in Kota Manado and Kota Magelang, implies that the average population in these cities has completed junior secondary education. Note that this indicator of human capital will be low where the older population is less educated than the current school-age population, reducing the average for the entire population. 14

Sample 13. Literacy rates in some districts, 2001-05 The above chart shows the adult literacy rate in Indonesia, defined as the ability to read, write, and do simple calculations. In order to be able to function in a modern society, an adult should have these basic skills. A breakdown of literacy rates by age group indicates whether a low literacy rate occurs in the school-age population and can therefore be addressed through the school system, or whether it occurs in the population beyond school age, which may indicate the need for an expansion of informal adult education programs. 3. Outcome analysis: This section assesses education outcomes through learning achievement parameters such as average test scores in the main study course. The analysis can be conducted through the use of time series and district comparisons. The analysis in this section uses a similar approach to that used in the output analysis. Often, the output and outcomes analyses are combined into one if outcomes indicators are only a few. 15

Objective 5. Equity and Efficiency Analysis This chapter looks in more detail at the efficiency of spending on education and the equity of that spending. Analysis of equity in the sector looks at the geographical distribution of inputs, outputs and outcomes, as well as the distribution of spending across income levels. The level of efficiency is found by estimating the cost effectiveness of given inputs with attained outputs, as well as by comparing these with the education goals of the district. A comparison between subdistricts and districts, as well as with the national figures, is also useful in providing analytical insights. The basic documents required to conduct the analysis are district budget (APBD), annual district education statistics, planning documents at the Dinas level, BPS surveys (Susenas, Podes, etc), and the District in Figures. Type of analysis required 1. Equity analysis: This analysis looks at variations by gender, household income, residential location (rural vs. urban), and ethnic or religious group. Variations are analyzed based on input, output and outcome indicators. Some key questions are: How do school enrollment rates, drop-out rates and learning achievements, vary by sub-group? What is the average spending by households in different income quintiles on education? What is the distribution of teachers in urban vs. rural areas? What are the government programs or policies providing financial incentives in education? Sample 14. Net enrollment rate by income quintile, 2004 16

The chart above illustrates the distribution of enrollment rates across household income groups. Based on the information provided in the chart, one can analyze whether enrollments are similar across income quintiles for each level of schooling or whether variations exist. The chart above suggests that enrollments in those districts analyzed at the primary level are close to universal for all income quintiles. This is not the case, however, for the secondary level, where wide gaps in enrollment still occur between the poorest and richest quintiles. District quintile Sample 15. Aggregated district expenditure per poverty quintile, 2005 Per capita district expenditure (Rp) Education expenditure per school student (Rp) Non-personnel education % tot exp Private HH out-of-pocket (Rp million) Poorest 858,296 1,097,255 6.99 26,067 2 807,164 941,367 5.02 43,121 3 595,605 1,048,785 5.38 62,138 4 896,367 1,402,839 5.01 66,642 Richest 912,590 1,574,280 4.68 114,132 The table above provides a similar approach by using poverty quintiles at the district level. The table compares education spending of aggregated districts and households in five district quintiles, from the poorest (quintile 1) to the richest (quintile 5). 2. Efficiency analysis: Analyzing whether the budget has been spent efficiently and effectively depends on information such as the availability of performance systems, annual evaluation systems, and minimum service standards, etc. Some key questions include: Does education spending match targeted goals and objectives? What is the link between achievement in each level of education and employment and wages? What is the basis for employing and locating teachers: geographical need, merit or seniority? What criteria are used in deciding to build a new school and its location? What do the trends in school-age population imply about the inputs required in the future? Has the district been able to meet its minimum service standard targets? Sample 16. Over- and under-supply of elementary teachers, 2006 17

The chart above shows the share of sub-districts within a district with an excess or shortage of teachers. The data used consist of existing numbers of teachers in the sub-districts and the estimated required numbers of teachers for schools in each sub-district. The difference between the two sets provides an estimation of the excess or shortage of teachers. Analysis of the over- and under-supply of teachers provides a picture of teacher distribution in districts and how the country overall can improve its efficiency in teacher distribution, particularly given that such a large share of education spending goes towards personnel. Sample 17. Student-teacher ratio (STR) for the primary level in Kab. Timtengsel, 2006 The chart above is a sample of a simple graph on the variation of teacher distribution across sub-districts. This kind of graph can also be applied to other indicators, such as the number of schools per 1,000 primary school-age students, literacy rates, expenditure per capita, etc, and is intended to help analysis of disparities at the district and provincial levels. 3. Best practice frontier: The efficiency frontier analysis show variations in education sector performance across sub-district/district/province. This analysis attempts to illustrate best practice districts that maximize their outputs given a certain level of inputs. Hence, it can also be interpreted as measuring the level of efficiency of the education system in a district. The efficiency frontier analysis is developed using a set of input data (such as the number of schools, student-teacher ratio, local government spending on education, etc) and a set of output data (such as enrollment rates, literacy rates, test scores, etc.). These two datasets are then processed using a factor analysis method with the objective of creating an index of inputs and outputs with which to assess efficiency in the sector. 18

Sample 18. Best practice frontier of education sector performance at the district level 19

Data Requirements 6. Appendix The tables below provide researchers with data requirements and sources of data. The time span of each dataset will depend on the requirements of each analysis. Quantitative data No Type Name Description Source Indicators 1 Fiscal APBN National Expenditure by economic classification and function. MoF Deconcentrated and Assistance Task Expenditure APBD Regional budget (Kab/Kota & Province Level) consists of revenue (by items), direct and indirect expenditures to be mapped with the previous spending format. BPS ; MoF Revenue, expenditures by economic classification and sectors 2 Non Fiscal Social Indicators DAU DAK Population Census DAU Allocation and Basic Data used for DAU calculation (Kab/Kota & Province Level) DAK Dana Reboisasi and DAK Non Dana Reboisasi (Kab/Kota & Province Level) National population census, conducted once every 10 years MoF MoF BPS DAU Allocation, population, poverty incidence, area, IHBG, etc. DAK Allocation (Infrastructure, health, education sectors) Population SUSENAS SAKERNAS SUSENAS consists of CORE (annual) and MODULES (once in three years). It covers characteristics of household and members of household within sampled household. The Annual National Labour Force Survey (SAKERNAS) covers national labor market characteristics of all working age BPS BPS Education attainment, literacy rate, % of urban population, household income and expenditure. Labor force (by sector), employment rate, 20

individuals within sampled households. unemployment rate, etc. Economic Indicators PODES GRDP Village Potential Survey (PODES) provides information about villages/desa characteristics and infrastructures. Regional Products (kab/kota and provincial level) by current & constant price BPS BPS No. of school facilities, distance to school, road and infrastructure condition, etc. Sectoral products Local Gov't Characteris tics GDS Governance Decentralization Survey 1, GDS 1+, GDS 2 WBOJ & PSKK- UGM Governance indicators and decentralization (transparency, accountability, service quality) Qualitative documents Type Details Period Source Information needed Planning Budget Planning process Budget Preparation Budget Publication Budget Implementati on Latest Bappeda, Education What plans are in place? Dinas Does the public participate in the planning process? If yes, how? Is there a plan/strategy for improvement in education sector? How is the achievement of the plans monitored? Latest Financial Bureau, Is there a formal mechanism for the Bappeda, Education public to participate in the budgeting Dinas process? If yes, how effective is it? How is quantitative information used to make budget decisions? How are decisions made regarding government intervention in a particular sector? Which institution/body makes decisions on the final budget allocations? Latest Bappeda, Sekda Is the budget made available to the public? If yes, how (newspaper, government gazette etc.)? Latest Biro Keuangan, Which unit is responsible for Bappeda disbursement? Has the local treasury (BUD) been established? What are the payment mechanisms (SPP)? 21

Are there problems in with cash management? If yes, which? Has the budget been revised during the fiscal year? Performance Budgeting Latest Bappeda Has performance budgeting been introduced yet? If yes, how is performance monitored? Transfers Transfers implementati on Latest Biro Keuangan What is the size of allocation from transfers (DAU, Revenue Sharing, and DAK) that are allocated toward education? How is the allocation process being done? **Format -- electronic wherever possible Data Analysis Agreeing on data sources and analysis It is important that data sources are agreed upon in order to have a consistent dataset throughout the analysis, especially in cases where the analyses in different sections are conducted by different researchers. What needs to be agreed? a. Data to be used in the analysis. For example, budget data are available at the both central and local levels. It is important to agree on which datasets are to be used at the beginning of the process and before the data collection process starts. Ideally, fiscal data come from data collected at the district level, as these data are based on official documents that have already been audited. However, data that come from the central level is useful for a comparison with other districts or national level. b. Data on population and regional CPI (Consumer Price Index). Different population data are available, with sources ranging from the central BPS, the regional BPS, to different agencies within sub-national governments. Agreement on which data are to be used is important because the data are used to generate per capita figures. Sometimes, CPI for a region is unavailable. Agreement is needed on which regional CPI to use in order to normalize the figures against inflation. Moreover, agreement is needed on the base year for the real figures. Usually, the base year is the earliest year in the dataset. However, in order to make the analysis forward-looking, it is recommended to use the latest year as the base year. The advantage of this is that, when comparing trend figures and cross section figures (which are the latest), numbers will be consistent. c. The required time span. This is an important factor that influences the quality of the analysis. Ideally, the longer the times span the better, although there is data limitation at some point. A 22

province established after decentralization will not have data prior to 2000. Ideally, a minimum of five years of data is needed. d. Data limitations. When making regional comparisons, the team will face data limitations, especially data from regional sources. Usually, data collected in the region do not include interregion data. Data from central sources cover inter-region data but usually do not contain the level of detail as found in the regions. Therefore, it is recommended to use regional sourced data for analysis and comparison within a region, while central sourced data is used to conduct inter-region analysis and comparison at the higher aggregated government level. This is also something that needs to be agreed upon. Note on the analysis: 1. All analysis that involves data from different years should be converted into real figures (constant price figures). 2. When presenting quantitative analysis, avoid using details to the last digit. Round it up to the nearest thousand, million, billion, or trillion, although this is not necessary for per capita figures. 23