A. Introduction
B. Background & Definitions
C. Principle Questions
This monograph is concerned with how children in poor countries can gain access to good quality education.1 The basic thesis of the paper is that financial barriers are the main reason for the failure of many countries to provide education to their children. Financial barriers are of two sorts. First, the cost to parents and children is often too high, particularly when economies are in trouble. Second, public finances are in most cases inadequate: however, the financial management of education systems is frequently neither efficient nor effective, so that the state's resources derived from taxes in many cases cannot finance basic learning inputs which they would otherwise be able to do if those resources were managed better.
[1 I am grateful to a number of people for comments and discussion. They include John Mace and an anonymous reviewer, and the numerous officials and others in various countries with whom I have worked in this area. Christopher Colclough kindly permitted me to attend his seminar on cost recovery, and I use the excellent papers extensively in this study.]
That the paper is oriented towards public finance should not be a reason for non economists to be uncomfortable with reading it. Although some of the concepts may be unfamiliar, educators should be able to judge the arguments against their own experience and decide whether they provide a sufficient explanation for the poor quality and falling enrolment ratios which they may be witnessing in their own countries or elsewhere. The paper is critical of the role that economists have played in the formulation of education policy, critical of sectoral management, and critical of the foreign aid agencies' activities, but educators may find that the criticisms come from other directions from those they have customarily employed. 'The purpose and outcome of a great deal of what is said and written about education, however 'scientific' its form, is to persuade and convince. It seeks to establish a basis for agreement on what is, and what should be done.'2 At the heart of the matter is the question of resources and how they are managed, and this paper is intended to further our knowledge of those issues, which affect daily the lives of every teacher in every classroom, and to persuade and convince that some current orthodoxies should be re-examined.
[2 Taylor W., Metaphors of Educational Discourse, in Taylor W (ed.) Metaphors of Education, Heinemann Educational Books & University of London Institute of Education, 1984, p 20.]
The subject of this paper is cost sharing, a term which combines the concepts of direct cost recovery, and thus education pricing policies, and indirect contributions from pupils, their parents and sponsors, which may be voluntary, quasi-compulsory or even compulsory. The costs include opportunity costs: that is, alternatives to being at school (how far opportunity cost is properly a subset of cost sharing is a matter for interpretation, and is discussed further in Chapter 2). The term 'cost sharing' encompasses privately as well as publicly provided services. In this study the term is used when the subject under discussion is not restricted to user fee issues, which are classified under 'cost recovery'. However, the terms are frequently used interchangeably, and, although there is a euphemistic element in 'cost-sharing', its sense is clear enough. Even where families face apparent discretionary costs, closer examination may reveal that those costs are perceived by them to be non-discretionary.
While to some cost sharing is a term which has most significance in specific contexts. such as textbook procurement or school building, and is in various ways linked to concepts of 'community participation', this study, while incorporating and acknowledging those contexts, is in the first place concerned with the complex relation between citizens and the state in the area of financing education and to some extent public services in general. This is the only way the subject can be properly treated because of the linkages between different components of household spending - no one component can be analysed without reference to other components and because cost sharing is frequently treated as a way of earmarking funds for specific purposes, such as textbooks. In some respects this type of hypothecation represents a failure of public finance management systems, although many economists would propose that public bureaucracies are inherently incapable of being 'efficient', a dubious though attractive proposition. In view of that aspect of cost sharing, issues of public finance management are central to the analysis.
The approach of the study is quantitative, but not scientific in the sense that the data can be used to predict future outcomes. There are many different possible approaches to the study of cost sharing. Large scale surveys and associated correlation and probability analysis reveal much that cannot be derived from small scale, qualitative studies, and casual observation techniques can be as illuminating as rigorous statistical techniques. The subject spans government, household and school finance, as each of these requires specialised analysis.
This study tries to cover most aspects of the subject, and such an attempt is bound to be imperfect. While there has been a good deal of qualitative, small scale work, most of it in the form of unpublished reports to be found only in the countries themselves, and a good deal of large scale survey work, unfortunately not all easily available in spite of the fact that they are financed by public money, the focus of this study is on an area which is not commonly analysed, but which for policy purposes is central.
There are also important cultural and sociological aspects to be considered. The economic rationalism which dominates current policy analysis of cost sharing is too simple in its conception. The critical issue overarching cost sharing is its explicit role in a larger project to reduce public expenditures and the role of the state, and that project is driven forward against, in many cases, the weight of evidence that people do not react to the provision of modern education services in the ways they are believed to react. Furthermore, there are many questions to ask about the educationist's, as contrasted with the economist's, assumptions about the functions of education which impinge on consideration of cost sharing, and many of those questions are discussed in this paper. This study is therefore primarily concerned with putting together a modest body of evidence on education expenditures by governments and households and exploring the implications the evidence has for the central questions which need to be asked in order to develop and sustain government expenditure policies.
The structure of the paper is as follows. Chapter 2 considers the principles which underpin cost sharing policies. The following two chapters consist of case studies of specific countries. The case studies are not intended to be merely comparative studies only. They consider various aspects of cost sharing according to the availability of information and data and the nature of the issues facing the country. They are not self contained, and each complements the other: the surveys had different designs and their content has different emphases. The final chapter summarises the issues and draws policy conclusions.
This study approaches cost sharing through six questions:
a) Does cost sharing increase total resources available for education?b) Does cost sharing enhance efficiency of resource use?
c) Does cost sharing affect enrolments and attendance?
d) Does cost sharing improve quality of education?
e) What other effects result from cost sharing in education?
f) Is a policy of cost sharing justified?
The questions are not always easy to answer, and they subsume more detailed questions. The assumption that all of them have positive answers underlies the arguments of those who advocate increased cost recovery and cost sharing.3
[3 These questions may be compared to those asked by Christopher Colclough in Who Should Learn to Pay? An Assessment of Neo-liberal Approaches to Education Policy in Colclough C. & J. Manor, States or Markets? Neo-liberalism and the Development Policy Debate, Clarendon Press, 1991, pp 197-213. He considers the 'neo-liberal' agenda against four elements: (1) user charges at tertiary and secondary levels combined with scholarships to promote both efficiency and equity; (2) loans for tertiary students; (3) encouragement of private schooling; (4) reallocation of 'sayings' to more 'socially profitable' parts of the system. He addresses the 'neo-liberals' on their own ground and on certain technical arguments, particularly those relating to rates of return. Rates of return are so suspect anyway that little is gained by arguing about their levels, though to state that view does not diminish the force of the arguments.For recent overviews which are excellent within their objectives but do not take a public finance orientation, see Bray, M., Counting the Full Cost: Parental and Community Financing of Education in East Asia, World Bank, 1996, and Decentralisation of Education: Community Financing, World Bank, 1996. Those two pamphlets cover an impressive bibliography which confirm the narrowness of the analysis of cost sharing in education.]
The questions are hard to answer because the data are generally not available to determine the effects of cost sharing over time: while there is evidence of falling enrolments and falling utilisation of health facilities over the short term, such evidence would not be sufficient to reject cost sharing policies, partly because of other factors which may affect service utilisation, such as an economic downturn. Another reason for the difficulty in arriving at more than tentative conclusions relates to the counterfactual: what might have happened under different circumstances? Counterfactual analysis can only be indicative, but is an important component of analysis of the effects of stabilisation and adjustment, and of changing policies on public expenditure. Thus, for example, the answer to question (a) above might be negative in relation to a given base year, but nevertheless resources allocated to education from all sources might have been even less without cost sharing.
History matters, and developed countries have arrived at near total support of school education over a century, largely as a result of social and rights pressures as opposed to economic pressures.4 As Table 1 implies, in most countries tax finance accounts for most education spending.5 Many developing countries started their education systems under colonial governments as private systems partly run by missionaries, and their education development has been characterised by the transfer of responsibility for mass education to that state, as occurred in the now developed countries. The variations between countries and the explanations for each country's systems are complex, and simplified versions of history should be avoided. One of the purposes of this study is to suggest that simple explanations have very little use, and can be damaging to policy development because of their origin in the powerful foreign aid agencies on which, unfortunately, many countries have come to depend in the last 15 years. Simplified history combined with orthodox economics is a perilous mixture.
[4 Economics has not always been the dominant discipline in social welfare policy. For example, British social policy was lime influenced by economists in the 1950s and 60s and its designers were unrepentantly collectivist in outlook. See Bulmer M., J. Lewis & D. Piachaud (eds), The Goals of Social Policy, Unwin Hyman, 1990; also the review of the book in the Times Literary Supplement, p 251, March 9-15 1990, by Frances Cairncross. Indeed, an understanding of the development of education in Europe is important to counterbalance a certain ahistorical tendency in much of the literature. See, for example, the excellent account of the rise of public education in England in Gardner P. Schooling, Markets and Public Agency 1833-1944, in Bridges, D. and T. H. McLaughlin (eds), Education and the Market Place, The Falmer Press, 1994.5 The data in the table, although presented confidently in Priorities and Strategies (not including those countries added separately), must be considered with caution, but, because of different sources and measurement criteria, are likely to be indicative.]
Table 1: Total Education Expenditures by Source
|
Group and country |
Public sources |
Private sources |
|
OECD countries | ||
|
Australia |
85 |
15 |
|
Canada |
90 |
10 |
|
Denmark |
99 |
1 |
|
Finland |
92 |
8 |
|
France |
90 |
10 |
|
Germany |
73 |
27 |
|
Ireland |
93 |
7 |
|
Japan |
74 |
26 |
|
Netherlands |
98 |
2 |
|
Spain |
50 |
20 |
|
United States |
79 |
21 |
|
Low and middle-income countries | ||
|
Uganda (1992-1993) (1)(2)(3) |
47 |
53 |
|
Haiti |
20 |
80 |
|
Hungary |
93 |
7 |
|
India |
89 |
11 |
|
Indonesia |
63 |
37 |
|
Kenya (2) (1992/93) |
62 |
38 |
|
Tanzania(1993)(1) |
69 |
31 |
|
Venezuela (1987) |
73 |
27 |
Notes and Sources: From "Priorities and Strategies for Education", World Bank 1995, Table 3.1, p 54.
(1) For Uganda, see World Bank, Access to Education and Health Care in Uganda, June 1996, p 21. For Tanzania see World Bank, Social Sector Review, Draft, April 1995, Table S5, p xxiv. The original table shows data for Uganda for 1989/90 (43%+57%)
(2) Public institutions only. Private sources refer to households only.
(3) Primary and secondary levels only. Private sources refer to households only.
The cost sharing/recovery policies which have been advocated over the last few years for developing (and in many cases developed) countries have a medley of motives behind the arguments. The obvious reason for interest in cost sharing is fiscal stress - the inability of domestic revenues to support education systems - so raising contributions from non-government sources (i.e. outside the tax system) through compulsory charges (cost recovery) and through discretionary charges (cost sharing) increases the total level of expenditures. Less explicitly stated reasons are more ideological, based on assumptions about desirable (often termed 'optimal') levels of public expenditures and taxation and other policies within the macroeconomic frame. That cost sharing enhances equity and efficiency is counter intuitive, yet its proponents regard those attributes as a strong part of their case: many believe that cost sharing will result in increased enrolments, particularly of the poor, and force government to manage resources more efficiently. This is the somewhat narrow approach to the question of equity which is followed in this study: equity is, in fact, a more complex concept concerning compensation, fiscal equalisation and other measures, but for the purpose of this discussion it is hard, I believe, to disassociate equity issues from efficiency issues and to treat them separately.
Furthermore, it is argued that additional resources are made available as a result of cost sharing to increase expenditures on direct learning inputs such as books and to stimulate qualitative improvements. A cycle is created. Enrolment is sensitive to the quality of school experience, which in turn relies on non-government finance. Enrolment is also sensitive to cost. If people don't pay, quality does not improve. If quality does not improve, people don't enrol.
This mixture of necessity and ideology, unsupported by evidence, is confused, all the more so when public expenditure as a whole is taken into account, particularly in countries with a high non-discretionary component in recurrent expenditure.6
[6 'Discretionary' expenditures are those over which the spender has some choice. Governments have no choice (generally) whether they should pay debt, or pensions, for example, and such expenditures are 'non-discretionary'.]
Does cost sharing increase total resources to education?
Total resources for education can be increased through (a) increasing overall public expenditures; (b) reallocating to education within a given level of public expenditure; and (c) increasing non-government contributions with no reduction (or a less than proportionate reduction) in enrolments. Only in the first case is there a net additional claim on public expenditures.
The second and third cases are the most interesting, in that most analysis assumes explicitly or implicitly that the fiscal constraint on government is binding (except for off-budget foreign aid), and that public expenditures should be reduced. Such a generalised response ignores two crucial factors, namely the details of the composition of public expenditures and the fungibility of money.7
[7 Fungibility is the substitutability of money between different uses and is a concept which is central in economics (derived from Latin fungor meaning 'to perform a duty or fulfil an office' which gave rise to a Latin legal term 'such that any unit is substitutable for another'). If somebody gives me a dollar, I can use the dollar I already have for something else, but which dollar do I use, as they were not marked and are indistinguishable from each other? In the same way, if the Ghanaian football team beats Tanzania 3-2, which is the winning goal? Is it the first, second or third? because none could exist without the other. Fungibility is also why foreign aid lending or donation to a project with the highest social return is not what it appears: the loan will always be for the marginal project, irrespective of the ostensible 'purpose' of the money. The whole issue of fungibility in foreign aid has been woefully ignored, or argued not to exist or be malign (e.g. Cassen R. and Associates, Does Aid Work?, Clarendon Press, 1987, p 21). There are relatively few studies of fungibility of aid: see Pack H. and J. R. Pack, Foreign Aid and the Question of Fungibility, The Review of Economics and Statistics, 1993, pp 258 - 265, for one example.]
Most of the countries we are discussing are heavily indebted to foreign creditors, to domestic creditors and to their central banks, and interest costs are shown 'above the line': they are part of the same budget out of which education expenditures are made. The higher the interest costs, other things being equal, the less the finance available for education and other uses, because interest payments are non-discretionary: they have first call on the budget. Although interest costs are temporary, they cannot be ignored, and a counterfactual analysis of a budget without 'excessive' interest costs shows that increasing education (and other) spending is a strong option: more to the point, it shows that reducing spending because of temporary budget problems is counterproductive, because it is easy to cut and very hard to reinstate cuts. As the case study of Ghana illustrates, while expenditures on education do not account for a particularly remarkable proportion of national income, they account for a high (relative to most countries) proportion of the discretionary (after debt cost) budget, implying among other things that debt costs are crowding social expenditures. Consequent public expenditure management thus becomes geared to freeing resources to cope with debt costs. This aspect of the debt problem is rarely brought up in the debates on debt.
There is also the issue of fungibility, considered further in the next chapter. Fungibility - the substitutability of money - is a fundamentally important concept in all analysis of public finance and foreign aid. Its presence turns many (if not most) foreign aid interventions into illusion and gives limited meaning to many (if not most) donor and lender conditionalities relating to budgetary allocations. It is difficult to analyse and its effects are hard to prove, relying heavily on counterfactual speculation The impact of the fungibility of money on cost sharing policies is quite simply that cost sharing is just as likely to achieve the opposite of what is intended as to achieve what is intended, and to reduce expenditures.
Thus, whether cost sharing can increase total resources, as it is often assumed it should do, will depend very much on the underlying reasons. It is more likely to mitigate the effects of reductions in government expenditures rather than to increase total expenditures, or to result in increased expenditures in other parts of the budget.
Does cost sharing enhance efficiency of resource use?
The relation of cost sharing to efficiency is explored theoretically and empirically in this study. Much of the rationale for user charges derives from a belief that they stimulate efficiency and accountability. However, it is equally likely that the reverse may be true, particularly in countries with weak fiscal management, and cost recovery can maintain inefficiency and create problems of lack of accountability. Overcoming resource constraints by charging users can permit inefficient management of resources within the sector and throughout the government budget as a whole, because pressure to find resources through other means is reduced. The fungibility of money can mean that consumers are required to pay for inefficient government sectoral management, which can outweigh any advantages which competition and consumer awareness might bring. Similarly, where fees are raised outside the tax system it has proved hard in many countries to control them.
Another argument which is explicit in the case for cost sharing is that the imposition of charges as well as higher indirect costs at the secondary and tertiary levels of education can permit greater allocative efficiency, creating room for allocations from post-primary budgets to primary education. While the argument has theoretical attractions, it is of little practical validity in the short term because in general governments do not allocate fixed shares of the budget or national income to one sub-sector, and because the sums of money actually raised are small in comparison with what is required. Over the longer term there may be more room to manoeuvre, depending on the overall fiscal position and the political sensitivity of higher education.
Does cost sharing affect enrolments and attendance?
Although enrolment effects from cost sharing are important, attendance effects can be equally important, because reduced attendance ratios affect repetition rates and achievement measures. There is considerable evidence to suggest that attendance ratios are negatively affected by cost sharing as children are sent home for non payment of fees.
Whether total expenditures on education rise as a result of cost sharing measures or not will depend among other things on the relation between enrolments and increased costs. A justification for cost recovery is that it stimulates increases in enrolments largely through the effect of increasing resources and permitting budgetary reallocation. However, cost sharing is more often a response to fiscal stress. Fiscal stress is caused inter alia by general economic difficulties. It is logical to suppose that most people will share in general economic difficulties. Cost sharing is imposed on an already stressed population, and the ratios of food expenditures to total expenditures in a household tend to rise in such times, squeezing capacity to finance items which are not necessities, of which education is one. For example, survey evidence often shows an apparent substitution of health expenditures for education expenditures, and indeed, that primary education is one of the most discretionary of family expenditures: people place priority on expenditures essential for physical survival. Where economies are in trouble, cost sharing policies will affect enrolments, and that is indeed what most evidence suggests. However, it is not easy to distinguish the effect of costs from other effects: for example, most people regard schooling as a route to employment, and in economic bad times employment opportunities are fewer. The economic situation of the people affects their cost-benefit calculus, both from the point of view of opportunity costs (children's labour on farms, for example) and in terms of risk (their perception of the increased probability of unemployment). Similarly, a large number of school children consider public sector employment as the most desirable, and civil service reform may affect that perception.
Thus, while it may appear that the increased cost burdens imposed on households may be responsible for declines in enrolments, the underlying factors are more complex.
Does cost sharing improve the quality of education?
Improved quality of teaching and learning may result from managerial improvements and from better resource allocation. They may also result from improvements in the provision of specific inputs such as textbooks or construction labour: it is reasonable to suppose that availability of inputs enhances quality given that the circumstances are favourable (for example, that there are competent teachers). The history of specific cost recovery schemes for textbooks, for example, has not so far been encouraging, though it is often difficult to see why. Even though specific cost sharing strategies like textbook funds might seem to exert a positive influence on quality, they still have to be considered within the overall menu of alternatives which might be pursued to improve quality.
It is not easy to measure changes in the quality of learning, and less easy to ascribe reasons for quality improvements. Attempts to measure factors which influence learning can result in counter intuitive conclusions, such as that class size or teacher training have no effects: the problem is controlling for all other variables while holding the variable under review constant, and this is extremely difficult in the type of analysis generally employed to measure qualitative changes. At the very least, though, we can reasonably assume that increasing the supply learning inputs such as textbooks has a positive effect.8
[8 Much of the literature is ambiguous on the effects of increased learning inputs, reflecting perhaps the diminishing returns to expenditures on them. However, where the supply of such inputs is very low it is reasonable to suppose increasing returns. See, Hanushek E. A, The Economics of Schooling: Production and Efficiency in Public Schools, Journal of Economic Literature, Vol. XXIV, Sept 1986, pp 1141-1177; and Fuller B, What School Factors Raise Achievement in the Third World?, Review of Educational Research, Vol. 57 nr 3, 1987, pp 255-292. See also Lockheed M. E., Verspoor A. M. and associates, Improving Primary Education in Developing Countries, OUP/World Bank, 1991, chapter 3.]
A textbook fund may also permit governments to avoid structural changes to the budget, which, if undertaken, might result in greater quality gains. In the country case studies in this paper, non-government finance supports a large proportion of, if not nearly all, learning inputs, and that without that finance there would be no inputs. The dependence on non-government finance affects enrolments, which fall because of the costs of schooling. Looked at from that point of view, it might appear that quality enhancement, if it does indeed result from cost sharing, also has a cost in reduced enrolments.
What other effects result from cost sharing in education?
The responses of households to user charges include (a) reallocating from other expenditures to finance the charges; (b) finding additional money; (c) withdrawing from the service; (d) withdrawing from other services; (e) continuing to use the service but refusing to pay.
The first of those choices can have wider effects. Cost sharing is not an exclusive preoccupation of education sector agencies. Costs of health, irrigation and public transport are also significant in many household budgets. Perhaps the most substantial is health charges. Policies in general are sectoral, rather than cross sectoral or programmatic, and the impact of health charges on household ability to pay for education and vice versa is rarely considered. Indeed, it is likely that when faced with competition from other charges, households consider education charges to be the lowest priority.
Finding additional money may mean borrowing, or selling assets. There is evidence in some countries which suggests that people disinvest in physical assets at a more than 'normal' rate in order to pay fees. What is the aggregate effect of more than 'normal' cattle sales in rural areas to raise money to pay user charges? Does it affect the distribution of wealth and poverty? Does it affect economic growth? Where the 'rich' are required to pay for post primary fees, do they forgo alternative investments which might raise economic growth? These and similar questions have not been investigated as fully as they should be.
Ability to pay is a complex concept, and should not be confused with willingness to pay. It is possible for people to pay more than they can afford in certain circumstances, with adverse long run effects, yet simplified analysis can conclude that because they paid, they were both willing and able to pay, which of course in a sense they were: economics has no model for reluctance to pay in the face of absence of real choice whether to pay or not, though many economists would argue that the choice was 'rational'.
While the question of what other effects may result from cost sharing policies is important, this study does no more than acknowledge the issue, as evidence is slight. The question is posed to highlight the need to examine the effect of sectoral policies outside the sector concerned.
Is a policy of cost sharing justified?
The final question, which is the theme of this paper, is whether policies of cost sharing are really justified. The principle intention in considering this question is to highlight the relationship between public sector management and the costs to citizens which result from decisions of governments. Those costs include costs of inefficiency, costs of decisions which make the service more expensive than it needs to be, and costs arising from perverse expenditure priorities. If through better management and more responsive policies the state can reduce the cost of schooling without loss of effectiveness, the case for cost sharing is weak. Yet in most countries increased cost sharing occurs without any significant progress in reducing the cost of learning to pupils and their parents, or in improving services. In other words, parents are forced to pick up the costs of state inefficiency, or of costly state education policies. The argument applies troth to public and private schools.
For example, a cumulative process of curriculum development in response to changing education theories and policies has resulted in overloaded and expensive curricula in many countries. The scope of the curricula influences strongly the level of cost of the system. An alternative to making people pay to support the system is to change the system to fit more with the ability of the state to pay. Similarly, in some countries assessment and examinations push up costs with no visible benefit, with the added effect of perverting education and disrupting local societies. At a more technical level, the failure in most countries to control the allocation of teachers means that salary bills are higher than they need be for any given level of salary. In these instances, the availability to the state of additional indirect and direct finance outside the tax system relaxes a constraint which in other circumstances could force down costs. In this respect, it may also be noted that foreign aid constitutes a form of cost sharing, and also permits states to avoid difficult decisions. Orthodox economic models assume costs as given, and therefore are able to demonstrate axiomatic benefits from cost recovery.
The first step in justifying the introduction of cost sharing policies is to evaluate how far existing provision is compatible with the ability of the state to pay, and how far state provision is efficient. If the education system is too expensive for what it delivers, and if indications are that sector management could be improved, a proper sequencing of reform demands that state provision be rationalised and made more efficient before costs are pushed on to citizens outside the tax system. Similarly, where overall economic management is poor, the state's ability to pay is reduced, but would be greater if economic management were improved.
