A Three Part Hands-On Training Module
FISCAL DECENTRALIZATION IN EDUCATION
A Three Part Hands-on Training Module
By Taryn Rounds Parry
Department of Political Science
University of Georgia
October 1996
CONTENTS
Introduction: Why Use Intergovernmental Grants?
Part A: Types of Grants and their Impacts
A.1 Matching Rate and Local and Central Government Share of Spending
A.2 Local Government Response to a Change in Income
A.3 Local Government Response to a Change in Central Government Transfers
A.4 Local Government Response to a Change in Price
A.5 Estimating Elasticities (optional exercise)
Part B: Lump Sum Grants
Simulation 1: Foundation Grant
B1.1 Foundation Simulations
B1.2 Simulating Alternatives (optional exercise)
B1.3 Evaluating Alternatives: Efficiency
B1.4 Evaluating Alternatives: Equity
B1.5 Lorenz Curve (optional exercise)
Simulation 2: Poverty Formula
B2.1 Increase Capitation Transfer (same amount for all municipalities)
B2.2 Increase Distributed According to Relative Poverty Level
B2.3 Comparing Alternatives
Part C: Matching Grants
C.1 Simulation 1: Simple Matching Grant
and Compare to Lump-sum Grant (of same amount)
C.2 Simulation 2: Variable Matching Grant
FISCAL DECENTRALIZATION IN EDUCATION
INTRODUCTION
Suppose you are assigned to a country that is considering decentralization of education. What strategy should you recommend? What data need to be collected in order to evaluate various approaches to decentralization? This training module presents some important considerations in the decision to decentralize education (or other social services) and provides some simulation exercises to show the impact of various intergovernmental grant designs on the distribution of benefits and on stimulating local education expenditures.
This session begins by explaining why intergovernmental transfers are usually an essential part of decentralization in developing countries. Part A reviews the different types of grants and how their impacts may differ. In particular, the concepts of income and price elasticity are important for simulating the response of local governments to different grant formulae. Short exercises in this section teach techniques and concepts that participants will later apply in simulations using actual data in Parts B and C. Part B gives some examples of simulations of lump-sum type grants. The simulations reveal that providing resources may not be sufficient to stimulate higher education expenditures so that minimum required expenditures may be necessary. Part C gives some examples of simulations of matching grants. These grants are more successful in stimulating education expenditures than lump-sum grants.
This training package contains this written guide and a diskette with the following Excel files:
Part Aedndec.xls
estelast.xls (optional exercise)
Part Bfoundatn.xls
lorenz.xls (optional exercise)
poverty.xls
Part Cmatching.xls
INTRODUCTION
The Logic of Decentralization
Should education be decentralized?
1. To review briefly, some of the benefits of decentralization include the following:
greater diversity and increased responsiveness to the needs of the people (greater social efficiency)
more efficient administration as decisions can be made more quickly and with better information (technical efficiency)
additional cost savings if lower cost areas can pay lower wages and prices than when the system is centrally administered
Decentralization can have other benefits such as promoting greater participation, and generating more local support. It may generate greater local funding for services which can relieve fiscal stress experienced by central governments.
2. On the other hand, decentralization clearly has some problems:
The localities often lack the expertise and technical resources to perform the newly decentralized tasks. However, this is only a short-term problem, and a reason to implement decentralization slowly rather than abandoning it completely. The central government must support decentralization by providing support and helping the local governments develop the expertise required to perform the decentralized tasks.
Some services have economies of scale as well as diseconomies of scale. Standardization has some advantages, and centralized administration of education my require fewer resources than decentralized administration. For example, decentralized administration of education in Chile requires more personnel (per student) than when education was centrally provided.
One main concern with decentralization is inequity and inadequacy of resources. Centrally provided services often (not always) redistribute revenue to resource poor areas. If decentralization requires local communities to rely much more on their own resources, significant disparities will result. Especially in developing countries, many communities may not have adequate resources to provide even a minimally acceptable standard of education.
Another concern that is particular to education is spillovers across local jurisdictions that may lead to less than optimal investment in education. Literacy and basic skills produce better informed citizens who have better health and who tend to have smaller families. Education has a positive impact on economic growth and a negative impact on crime. For these reasons a good education system benefits the entire nation. Education is a national public good, not a local public good, and pure local finance of education would lead to under-investment in education.
3. Intergovernmental transfers address some of the concerns with decentralized provision of education. They permit the central government to continue to be a major source of funds for education while at the same time, local officials can make expenditure decisions and see to the day to day operation of schools. This can help to reduce inequity by providing adequate resources, and it permits some national funding for a national public good.
4. However, intergovernmental grants are not without their own set of problems:
Rent-seeking Powerful local government officials will attempt to influence central government finance decisions in order to accrue more resources for their community. This will interfere with the development of grant distribution formula as they will want the formula to favor their characteristics. More than likely, their influence will have a detrimental impact of any equity improving aspects of the transfer mechanism.
Poor Decision Making The recipients of intergovernmental transfers tend to look upon grants as "free money" since they do not bear the costs of raising the funds. When local governments do not have to finance their own expenditures, they tend to spend inefficiently since spending brings local benefits while the costs are borne by others. Therefore, some incentives may be necessary to achieve greater efficiency.
Transfers Do Not Guarantee Higher Expenditures If demand for education is usually income inelastic and price inelastic, transfers may not overcome equity or spillover problems. Grants also tend to reduce local effort so that grants merely supplant local expenditures: If you spend $100 on education and I offer to give $100 that must be spent on education, you may continue to spend $100 on education, but now the $100 you used to spend on education can now be used for other things. The $100 can also be used to give local citizens a tax break.
5. POINT: The first question to address is whether decentralization is worthwhile. This means identifying how the present system is working (collection of baseline data) and determining what can be gained through decentralization. Another important consideration is what decisions are being decentralized? This module assumes finance and expenditure decisions are the decisions that are being decentralized, but decentralization may be more narrowly or more broadly defined. Box 1 shows list of questions and types of data that need to be collected in order to answer these questions.
0.
Box 1: Assessing the Situation
1. How does the current system work and what changes are proposed with decentralization?
School Organization
How is the schooling system organized? (pre-school-primary-secondary-vocational etc.)
The structure of a school system is almost always set by the central government. Will this situation change with decentralization?
Curriculum and Teaching Methods
Is the curriculum a national curriculum (centrally deter mined)? Are there any adjustments for regional or cultural differences? Will decentralization permit localities to determine their own curriculum and teaching methods?
Examinations and Supervision
How are students and teachers evaluated? Are there national examinations? Which level of government performs the evaluations and sets examinations? Will this be changed under proposed decentralization plans?
Teacher Recruitment and Compensation
Which level of government sets accreditation standards? hires teachers? sets promotion rules and rules for dismissals? How are teachers' salaries determined? Which of these things are expected to change under proposed decentralization plans?
Finance of Operating Expenditures
How is school operation financed and in particular who pays teachers' salaries? Do intergovernmental transfers exist at present? How are they distributed? Are new transfer mechanisms proposed with the decentralization plans? What percent of this finance comes from local sources? Is this share expected to increase under decentralization? What are local government sources of revenue? How are localities expected to raise additional funds for education? What legal restrictions exist on their ability to raise local revenues?
School Construction and Finance
Who determines where schools are built? Who sets construction standards and are they uniform nationwide? How is school construction financed? Do any intergovernmental transfers exist specifically for financing school construction? How are these transfers distributed? How much of school construction costs are financed by localities at present? Is this share expected to change with decentralization?
2. What are the country's attributes?
What is the distribution of income? Is there a large degree of variation in income and resources across the country? Do climate and natural resources vary greatly? What are the different cultures in this country and where are their cultural centers? Is there one common language or are many languages spoken? What is the political system and in particular, how are local government officials chosen? Who are the minorities and are they represented by the political system? Do any channels exist for citizen participation in national or local decision making?
3. What are the potential gains from decentralization?
Social Efficiency
The greater the uniformity of the education system and the greater the disparities in cultures, climate, and resources, the greater the potential for decentralization to improve "social efficiency" or improving the match between citizen preferences and the quality, type, curriculum, methods, and quantity of education provided. However, if minorities or low-income groups are shut out of the local decision making process or if local officials are chosen by the central government, these gains are minimized.
Technical Efficiency
In theory decentralization is expected to improve technical efficiency and lower costs. In practice it is difficult to assess. First some data are needed regarding costs under the current system and how these costs compare to local costs. Are centrally determined teacher salaries much higher than other local salaries? Will decentralization permit localities to determine teacher and staff salaries? Will decentralization permit localities to use cheaper local resources? Is central administration slow and encumbered by red tape?
Equity
Many expect decentralization to worsen inequities. However, if the present centralized system is already inequitable in the way resources are distributed through intergovernmental transfers, legal restrictions or other policies, decentralization may not be any worse.
1
PART A: TYPES OF INTERGOVERNMENTAL GRANTS AND THEIR IMPACTS
6. Usually intergovernmental transfers are divided into the following three categories:
General Revenue Sharing--the central or state government shares revenue (from a sales tax for example) with lower levels of government. Usually the distribution is on a per-capita basis, but the distribution formula can be more complicated.
Block Grants--are specified for a specific type of expenditure, but have few or no other restrictions attached. Their benefit is lower administrative costs, and few compliance costs for recipients.
Categorical Grants--are not only for a specific type of expenditures, but have many strings attached. They are expensive to administer for both the grantor and the recipient of the grant.
7. The first two types of transfers are mainly to equalize or redistribute resources. Categorical grants are given in order to induce local governments to spend more on education. Despite all their restrictions, they may not be very successful. This occurs if the grant funds are used to substitute for locally funded education expenditures, and the money the locality used to spend on education is used for other purposes. The net effect is the same as if no requirements were given. For this reason, categorical grants often have revenue effort or minimum expenditure requirements.
Distribution of Grants
8. Both block and categorical grants can be distributed as project grants (by application) or by formula (by specified criteria). The simplest type of formula grant is a capitation grant which is based solely on population or number of clients (students in the case of education). Both types of distribution methods are used for education. Usually capital investment in new schools or extensions of schools are funded by project grants, while operating expenditures are funded by capitation grants.
9. This training module will focus on two main types of formula grants: lump-sum and matching grants. Both project and formula grants can be lump-sum or matching grants. Matching grants require the local government to spend money (on education) in order to receive any money. Lump-sum grants do not have this requirement and hence, are unrelated to local expenditures on education.
10. With a matching grant, the central government will match each local dollar spent on education with a specific amount of centrally-raised funds for local education expenditures. For example, if the central government matches every local dollar spent on education with a dollar from the central government, half of total education expenditures would be funded by the central government and half would be financed by local revenues.
11. Matching grants lower the "price" of education from the local government's point of view--one dollar increase in expenditures costs less than one dollar of local revenue. Lump-sum grants work by increasing the local government's total income by giving them more income to spend. Because of this difference in design, matching grants stimulate spending more than lump-sum grants--assuming no expenditures requirements. This principle will be illustrated in the simulation in part C.1 below.
12. Since the matching amount in the example above is one dollar for one dollar of local expenditures, the matching rate (R) is equal to one. However, the matching rate can be any amount. If the matching rate is less than one, the central government will fund less than half of total education expenditures, and if the matching rate is greater than one, the central government will fund more than half of total education expenditures. Since local governments in developing countries often do not have access to major sources of revenue, the matching rate should typically be higher than one; otherwise the transfer payment will not be very large. Matching grants can be open-ended (no limit of grant received) or closed-ended (grant amount is limited so that after a certain level of local expenditures, the grant received will not increase).
13. The purpose of the following exercise is to show how to calculate the share of education spending financed by the central government (C) and the share of spending financed by the local government (L) when the matching rate (R) is known. Alternatively, sometimes the central and local government shares are known, and then the matching rate (R) must be calculated. These calculations will be used in the matching grant simulations in part C below.
Exercise A.1:Determining the Matching Rate, Central Government Share, and Local Government Share of Spending
14. Open the file edndec.xls and go to the worksheet entitled "Matching Rates." Table 1 gives you the opportunity to calculate the central government share of education spending (C), the local government share (L), and the matching rate (R). The variables have the following relationships:
C = R / (1+R)
C = 1 - L
R = (1/L) -1
Use the above equations to complete the colored blanks in table 1.
15. To interpret your results you can think about the central government share as the percentage decrease in the localities price of education. For example, if R=1, then each local dollar generates one grant dollar so the grant finances 1/2 of total education expenditures. For the locality the price of education decreases by 50%. If R=.5, then each local dollar generate fifty cents in grant funds, and the grant finances 1/3 of total expenditures. For the locality the price of education decreases by 33%.
A.2 LOCAL GOVERNMENT RESPONSE TO A CHANGE IN INCOME