A. Cost Sharing in Perspective
B. Package of Reforms
C. Summary and Conclusions
I have argued in this paper that cost sharing in education does not axiomatically result in qualitative improvements in schools and increased finance for education as a whole. It is not, however, my contention that education should (or can) be exclusively the responsibility of the state: individuals will always face costs. Rather it is the application of cost recovery policies for the wrong reasons and on the basis of inadequately understood evidence and theoretical analysis which must be scrutinised.
The conclusion is based on the study of two countries, and experience differs between countries. Another case study could be that of Kenya, which has been held up as a successful pioneer of cost sharing policies. In some sense that is true, but it is also apparent in Kenya that the cost sharing policy has reached its limits; that enrolment ratios are falling; that quality is deteriorating; that accountability for sectoral performance by sectoral managers has been eroded; and that a major reappraisal of how education is financed in the country is required. The Harambee schools and Institutes, while more reliant than many writers appear to think on state funding, played an important part in the development of education in Kenya, but failure to understand the limitations on private finance has led to a crisis of confidence in the education system, and considerable indebtedness of individuals and institutions.
Cost sharing policies have provided a breathing space which have allowed governments to allocate resources to growth which eventually enable them to assume greater responsibility for financing education systems. Cost sharing does not provide a permanent solution to the problems of financing costly education services in countries with weak fiscal management which have education policies which make unrealistic demands on fiscal and household capacity. While in Kenya the policy has run its course and requires major modification, other countries such as Tanzania are embarking on similar policies, and should learn the lessons of those who have gone before.144 Once it is accepted that the attractions of cost sharing as a major financial policy instrument are limited, other reappraisals follow, principally policies related to 'community participation'.145
[144 See Lillis K. and H. Ayot, Community Financing of Education in Kenya, in Bray M. with K. Lillis, Community Financing of Education: Issues and Policy Implications in Less Developed Countries, Pergamon Press, 1988, pp 117 - 129, for a well argued, perceptive and prescient analysis, written 10 years ago, of the Harambee system.145 The best sustained example of scepticism that I know of is Lillis and Ayot op cit. The rest of the book in which the chapter is contained is more ambiguous. However, the message of caution was clear enough.]
One consequence of endorsing cost sharing community based policies is the legitimisation of the cost sharing culture as a response to fiscal stress and inefficient public sector management. By disguising the demands on parents and children in the euphemism of 'participation' the real reasons for cost sharing systems have been obscured. Governments could be bypassed by aid agencies and NGOs, which suited the Bretton Woods institutions' onslaught on government in the late 1980s and first half of the 1990s, and which was consistent with the NGO conception of the world as a place where the rural poor and others were deprived of 'voice'. Government officers could develop community participation policies and thus relinquish responsibility for maintaining large parts of the system.
Another consequence has been the serious deterioration of education infrastructure: in nearly all countries it became - and still is - the norm to expect communities to be responsible for building. Yet infrastructure is expensive in terms of time and money. While communities could reasonably be expected to erect simple temporary or semi-temporary structures, such structures were regarded as permanent by governments, which allocated no resources. Foreign aid agencies frequently financed shells of buildings, to be completed by communities, but in many countries even these were an excessive burden, or simply unsuitable for any number of different reasons. As long as the benefits from education were significant and visible, communities could be mobilised, but when enrolment ratios started to decline it was less likely that communities would take such a strong interest: the reason for declining enrolment ratios were precisely those reasons which reduced enthusiasm for local education development. A vicious spiral of decline set in, and it is now apparent in almost all African countries that the absence of good structures and teachers' houses is a strong factor in enrolment decline. Where teachers have poor living conditions they are absent or late and demoralised, and this affects parental and children's attitudes to school.
One reason given for failure of community participation is that communities did not have sufficient discretion, and there is much to be said for that view. Where communities feel they can exercise some influence over how schools are run, it is possible that education benefits. However, this is a slightly different (though related) issue from that of how schools are financed. The main issue from a financial point of view is that it is unlikely that any component of the education budget can be entirely handed over to local fund raising. Matching grant and other systems which achieve the same end are likely to give added momentum.146 It is time for government budgets to take on responsibility for infrastructure, at least where there are problems in attracting children to school.
[146 For example, in Zimbabwe the government's matching grant to communities for teachers' houses was too low, and few got built. With foreign aid funds the government contribution was increased in 1995, and there was a significant increase in the number of houses as communities then became able to raise the balance of finance required.]
Similarly, it is now almost received wisdom that textbooks should be rented or purchased by children and parents. Yet the cost of providing children with textbooks is not prohibitive within the perspective of total government expenditure. Failure of government managers and aid agencies to understand the necessary sequencing of reforms and the consequent neglect of public sector management has resulted in education system development projects which ignored the root causes of the apparent inability of the state to finance even the cheapest school inputs. Book provision was long of notorious interest to agencies because of the gains that publishers (and paper companies) could make, and because it is a technically easy way of spending aid budgets. Aid provision of books created a culture, which is persistent, of governments believing that they do not need to cover that expense.
People will always face costs, whether a service is 'free' or not. The issues are the level of those costs and the benefits derived from them. The level of cost is largely determined by government policy, whether the costs are faced by private or state schools. Governments set salary levels and determine the curriculum, the number of subjects and the length of the school year. Competition and consumer pressure can have very little impact on cost, except in the narrowest of senses. The essential components of a cost sharing policy should be (a) to aim at the lowest fee possible and (b) to relate benefits from the fee directly to the fee payer. The first component may seem surprising, but without it there are no constraints on how government determines education costs. It means that before any steps to introduce fees are taken, all other steps which are necessary to reduce costs must be taken, whether they are technical, such as control over teacher numbers, institutional, such as decentralisation and autonomisation, or political, such as reducing the scope of the curriculum. The more normal approach has the reverse aim, to maximise fees.
Although the justification of cost sharing policies may be found in their anticipated impact on revenues, efficiency and equity, the public finance and public sector management environment may be hostile. As with other policy reforms, cost sharing should therefore be seen as part of a package of policies, and the pacing and sequencing of implementation should take into account the wide range of factors which influence the supply of education and the demand for it. It is now widely recognised that a key to the success of reform policies is how they are sequenced, especially the order of (a) changes in incentive structures; and (b) changes in institutional structures. Components of cost sharing policy should be as follows, in the order of implementation sequence;
a) public sector reform;b) sectoral finance and management reform;
c) direct linkage of cost recovery to service improvements.
Within all of these a pattern of incentives should be identified. Reform programmes without incentives to reform have less chance of succeeding. The policy should recognise from the outset the fragility of user fees as a base, and the importance of broad-based tax finance, derived from a progressive tax system. Direct and indirect costs should not have a perverse effect and make the overall incidence of compulsory and quasi-compulsory payments regressive.
Public Sector Reform
There are three critical areas of focus in public sector reform: (a) revenue raising; (b) resource allocation and management; and (c) the reform of institutions and management systems. One of the most serious weaknesses of the Bretton Woods led reform programmes is that these three aspects were ignored or compartmentalised: they conformed to the government systems and structures which themselves needed to be changed. Civil service reform (CSR) and public finance reform are usually two separate programmers, with little connection between them: this has certainly been the case in both Ghana and Tanzania. All the three issues have bearings on the budget deficit, and therefore borrowing and public debt, and, by extension, fiscal stress and cost sharing.
One of the more serious mistakes in adjustment policies was the early concentration on expenditure reductions while ignoring revenue raising.147 Badly designed tax systems and inefficient tax collection lead to higher budget deficits or reduced public expenditures, and consequently can have a serious effect on education expenditures. There is little justification for structural changes to education systems in response to fiscal pressure when possibilities for increased revenue are present: even though revenue measures may take some time to come through (though they need not), structural changes can be irreversible and cause permanent damage. The more common case is gradual erosion of infrastructure and of salaries through cumulative expenditure reductions caused by revenue shortfalls.
[147 Partly due to assumption of the presence of the 'Please Effect', which proposes that improved tax revenue leads to government responses which increase expenditure rather than reduce government dissaving.]
Tax finance is the only sustainable source of finance for education. However, innovative possibilities exist in the form of earmarked taxation and forms of local taxation. Such innovations are not viable if the normal system is weak and not enforced. It is sometimes argued that cost recovery is required precisely because taxation systems are weak, but this is a curious argument, in the same category of argument that has resulted in the profusion of parallel project management and aid delivery systems, many established by foreign aid agencies, on the grounds of government failure. For cost recovery systems to work, government systems must work.
In view of the importance of budget reform it is surprising how little serious effort has been made to make it a priority,148 and while large inflows of foreign aid are available, aid fungibility will continue to threaten that budgets will not be improved. Both the countries in the case studies in this paper have weak state machinery, similar to many developing countries, suffering from inefficiency, low morale and principal-agent problems. Overall public expenditure reform is the major condition underlying all sectoral and sub-sectoral financial interventions. There is no justification for a state to pass on the costs of inefficiency to citizens. While tax finance is often costly and inefficient, supplementary costs of inefficiency are even less acceptable when they are compulsory. In that one of the main reasons, if not the main reason, for the introduction of cost recovery policies, is the failure of public finances to cover the costs of education, the tendency is to take the inefficiencies of public finance management as a given and proceed to raise additional revenues. In most economic adjustment programmes cost recovery has been part of a package of measures to achieve fiscal balance through expenditure reduction, not to improve education quality.
[148 'In virtually no country has fiscal reform enabled the budget to become a real tool for effectively managing the development process', Gordon D. F., Debt. Conditionality and Reform, in Callaghy T. M. & J. Ravenhill (eds), Hemmed In: Responses to Africa's Economic Decline, Columbia University Press, 1993, p 122.]
Budgets are the most important policy documents of governments: they are approved by parliaments and therefore have legal status. Yet in many countries they do not reflect stated priorities: there is usually a disparity between government's revealed preferences and the policies they present to their own populations and to foreign aid agencies. Expenditure reform must take place through the budget process. Because of the collapse of the budget process in many countries, foreign aid agencies have established parallel structures and accounting systems, but such practices only paper over the cracks and fail to achieve lasting improvements. Therefore, sectoral reform must also take place through the budget process and the normal systems of sectoral management which centres on budget implementation.
A policy of cost sharing in education or other sectors depends for its effectiveness on the strength of public finance management. In that it is generally recognised that cost sharing policies have a greater chance of success under decentralised conditions, local government fiscal systems take on the responsibility for regulation and accountability. Where these have been weak cost sharing has aroused popular resistance, and fees collected are often not accounted for. Similarly, in order for institutions to have autonomy of financial management strong supervisory and audit systems are needed: in their absence autonomisation of institutional management will not be effective, yet that policy is one of the most potentially important in bringing down institutional costs. In other words, wherever we look for interventions which will facilitate service cost reductions and best use of non-tax finances, we see the essential condition of public sector expenditure management reform.
Financial reform cannot take place without an improved quality of civil service. Most countries, including Ghana and Tanzania, have been trying to tackle public sector management and employment issues for some time, but for various reasons with limited success. Manpower decisions are essentially budgetary decisions, and most civil service reforms (CSRs) did not appreciate that link. In both Ghana and Tanzania the systems of recruitment and of budget are independent of each other, and establishment therefore becomes a fixed cost to budget managers. This is particularly important for education sectors which are the largest single civilian employer in every country. Teacher allocation and reallocation should be effected through the budget (preferably by system managers within a decentralised system where decentralisation is an option) and not by central or parallel planning mechanisms, such as are to be found in Sector Development Programmes and CSRs. Efficient staffing is at the heart of a package of reforms which involves cost sharing, for it is only when personnel are properly and efficiently disposed that parents can rightly be asked to contribute outside the tax system.
Sector Finance and Management
The areas of reform described above are necessary conditions and arguably even sufficient conditions for sectoral reform. Sectoral ministries generally have the responsibility for proposing resource allocation, and for making decisions which affect the costs of education. If ministries do not take action to ensure that the costs of education are fiscally realistic and affordable to government and people, the burden of cost sharing becomes heavy and at the same time the quality of infrastructure and instruction declines.
In both Ghana and Tanzania sectoral management systems require major reform at ministerial and decentralised local government levels, and at institutional levels, but such reforms take time and need to be the subject of consensus. Ministries may have weak financial control systems, and resource allocation may be largely incremental and line item based. There are few incentives to perform within hierarchical management systems in which the main criterion for success is survival, and not good performance defined in some sense as a contribution to improving learning outcomes.
Direct Linkage of Cost-Recovery to Service Improvements
As cost recovery takes place at the institutional level, institutions must be able to demonstrate immediate benefits to those who pay, either directly related to the amount of the charge or more generally through service improvements. This should be a central tenet of a cost sharing policy.
Fees are a form of benefit tax, where payment is made for government services which yield direct and identifiable benefits to the payer. Opinion surveys show that people are more willing to pay if they see tangible benefits. In some cases cost recovery policies may even reduce costs to the user of the service. In the health sector payment for drugs can achieve this, because where there are no drugs patients have to wait for longer periods, bribe or pay higher prices in the market place.149 Similarly, it should be possible for central procurement systems to purchase and distribute textbooks more cheaply than leaving it to the market, and it may be efficient and effective for government to provide such a service on a cost recovery basis, although there will tend to be an overhead subsidy which can be justified on the grounds that more textbooks are in the system than would otherwise have been.
[149 See Abel-Smith B. & P. Rawal, Can the poor afford 'free' hearth services? A case study of Tanzania, Health Policy and Planning, Vol. 7 Nr 4 1992, pp 329-341. There is of course a major problem with the argument that because 'services have deteriorated to such an extent that even the poor have to resort to the private sector to obtain services at much higher cost, charges will be less inequitable than continuing to provide under-financed services..' (p 331). At the time of Abel-Smith's work Tanzania's revenue efforts were very poor indeed.]
Cost sharing policies must therefore be part of wider policy packages designed to address all the issues which make education cost what it does, and which constrain the fiscal capacity of government to support education. There is little sense in imposing cost recovery packages in the absence of wider interventions. They can result in much damage: the decline in enrolments and/or enrolment ratios in many countries confirms the reality of such damage, although there are other reasons for those declines apart from cost.
The lessons for foreign aid agencies are also clear. Donors should be more sceptical of the Bretton Woods prescriptions and their underlying economic arguments.150 They should be more circumspect about cost-benefit data, and about the supposed efficiency and equity effects which accrue from charging fees. They should be more interested in evidence of cost sharing policy success. They should analyse and tackle the underlying reasons for fiscal stress and its effects on education expenditures, and be less preoccupied with developing large-scale projects which fail to fit into a workable sequencing pattern. They should be more interested in genuinely sector-wide reforms rather than fashionable concentration on sub-sectors, at present the primary sector, but sector programmes which are financed in the same way that projects are financed will end up as projects. Their programmes should allow space for 'normal' institutional structures to work, rather than establishing parallel project units, management teams and other parallel systems with enhanced salaries and facilities. Such a change in foreign aid practices would be fundamental, as it would involve reconsideration of their disbursement systems with a move towards direct support to government budgets, and even of the volume of resources they consider necessary.151
[150 Tilak, in Cost Recovery Approaches in Education, op cit, notes how widespread scepticism pushed the World Bank economists into disclaiming their earlier arguments for primary education fees. The World Bank has also gone about in the prevailing wind of scepticism about its earlier health policies. (One notorious example was the case of a study, financed by the World Bank, of cost recovery in health in Kenya, the publication of which was suppressed by the Bank. For the abridged report and a brief account of the affair, see Bloom G. and M. Segall, Expenditure and Financing of the Health Sector in Kenya, IDS, Sussex, 1993).151 It is this issue which causes difficulties for sector development programmes, which are now the New Jerusalem. While their intentions may be laudable (though there is good reason to be sceptical), agency procurement and financing regulations, and aid agencies' inherent desire to exert control, make true budget support difficult to achieve, and sector programmes tend to be large projects in disguise. Sector programmes also seek for perfection, usually with the help of liberal doses of consultants, before starting, which diminishes the opportunity for learning and organic growth within sectoral management cultures and structures. Aid agency driven programmes are also impatient, and subject to the disbursement imperative, with aid agents being more interested in spending money than real improvement. There is evident danger of collusion between donor/lender and beneficiary in seeking perverse objectives. A justification for sectoral programmes which is rarely analysed is that they can make fungibility work positively through the budget. While there has been no analysis of the issue, there is good reason to suppose that endogenous growth as a policy objective applies just as much to public sector as to private sector institutions. Competency based and resource based growth theories of the firm have their potential parallels in public sector organisation, and the danger of sectoral programmes is that they do not take account of all the factors which themselves create new disequilibria while at the same time achieving temporary equilibrium. A dynamic process is started which 'encourages continuous growth but limits the rate of growth' (Penrose, E. T. op cit, p 5).]
In summary, cost sharing should not be seen as a way of overcoming institutional and fiscal failure. Its objectives should be specified; its implementation carefully paced and sequenced; and its effects properly evaluated in order to ensure that evidence and not dogma drives policy development.
The argument advanced in this study is that cost sharing policies have not had the effects anticipated by those who introduced them. Without complementing measures, cost sharing will not increase total resources for education, nor will it enhance the efficiency of resource use, as is often proposed, to give more equitable public finance targeting. Cost sharing has undoubtedly contributed strongly to falling and stagnant enrolment ratios where these have occurred, and the failure of cost sharing to improve the supply of learning inputs and better infrastructure is part of a wider failure to achieve good quality education service provision in many countries. The wider social and economic effects of cost sharing are little understood. This study concludes that there has been little justification to support the introduction of cost sharing policies, and is critical of the linkage which was made between foreign aid entitlement and cost sharing as part of the wider, mainly Bretton Woods Institution, project of rapid public expenditure reductions.
There are six broad and related reasons for failure of cost sharing policies.
First, the fungibility of money from cost sharing and foreign aid has permitted governments to avoid difficult reform decisions. User fees, less obvious compulsory payments, and direct intergovernmental transfers through foreign aid acted as a buffer against poor public sector management. Governments could then maintain higher than necessary spending on external and internal security, loss making parastatals, and other manifestations of extensive government machinery, including excessive teaching forces. Governments, which themselves failed to raise adequate revenues from taxation, required its citizens to make what are essentially tax payments outside their tax systems, consolidating regressive taxation and bypassing public accountability. Moreover, the endorsement of such policies by foreign aid agencies and others gave legitimacy to related policies such as 'community participation' which included the devolution of responsibility to communities for school building and maintenance: this has resulted in the slow but sure deterioration of infrastructure. Payers of user fees and other charges have, in effect, subsidised government failure. Cost sharing may have contributed to the inefficiency and lack of effectiveness of education service delivery more than to its improvement. It might be tempting to qualify this conclusion with the observation that cost sharing is unlikely to have such a significant effect because of the relatively small sums of money involved: while this study has shown that the sums are larger than generally supposed, it is important to realise that cost sharing is an impact at the margin.
Second, the underlying reasons for cost sharing related to fiscal stress and the need to fill a fiscal gap, rather than to augmenting resources within an efficient system to achieve a more equitable public spending pattern, as is argued in the literature. Fiscal stress implies that the education system is probably underfunded and/or inefficient, and that the quality of service delivered is poor. At the level of the household survey evidence suggests that primary education is among the most discretionary of household expenditures and therefore an early expenditure foregone in times of household cash budgetary stress. At the level of the school near total reliance on parental contributions for non-salary expenditures, particularly those on learning inputs, meant that schools fell into debt or had to operate with insufficient resources to provide quality raising inputs. People do not want to pay for a poor service, and as direct and indirect charges did not improve the service, demand for schooling stagnated.
Third, the rigidities of public finance systems were not sufficiently taken into account. The reallocation effects which for many are the underlying rationale for cost sharing did not occur: charging the 'rich' did not result in more resources for the 'poor', and the supposed equity enhancements were not realised. A further reason for the failure to reallocate from higher education to primary education was that any cost reduction measures were welcomed by Ministries of Finance and by the Bretton Woods Institutions as contributions to overall reductions in public expenditures, particularly in the early years of stabilisation and adjustment when rapid expenditure reduction rather than revenue growth were priorities. Moreover, higher education is a very political sub-sector, and difficult to rationalise.
Fourth, the timing of reform was often inappropriate. Policies were introduced during times of growing economic difficulties compounded by squeezes on public expenditure: in some places, such as Zambia, during a time of severe drought and major epidemic. In periods of economic difficulty, employment opportunities are reduced, and the costs of education outweigh perceived benefits, in spite of the misleading arguments derived from cost benefit analysis. People did not want to pay for a poor service which did not lead to employment.
Fifth, during the period of economic adjustment inappropriate policy initiatives pushed up costs, such as a too rapid attempt to introduce Universal Primary Education, '20/20' aid formulae, project based initiatives which expanded curricula and syllabuses, extensions to the basic education cycle, and other measures which could not be afforded from domestic resources, and which there was little likelihood of being affordable in the foreseeable future. These and other initiatives which required external financing to support them were not located in a package of measures necessary to ensure that cost sharing was introduced alongside quality improvement of the education service. Donors and lenders concentrated on narrow projects and were preoccupied with targeted aid, and ignored the connections between policy actions.
Finally, the simplistic assumptions of economic rationality which underlie neo-classical economic theories of demand for a service such as education, and on which much foreign aid policy is based (for foreign aid has for the most part been the main driver of cost recovery policies), could not absorb cultural aspects of demand, and for that and other reasons the orthodox theoretical basis of cost sharing is a poor guide to policy.
This paper has argued that as a consequence of these effects, cost sharing policies have combined with other factors to interrupt enrolment growth (with declines in enrolment ratios and in some cases of absolute enrolment levels taking place in many African countries), and have contributed to the decline in the quality of education services. A properly paced and sequenced package of reform should focus first on bringing the costs of education to manageable levels, with priority given to minimising the costs to households rather than maximising them, as has tended to be the case explicitly or implicitly; and second to structuring education systems so that they can be afforded by governments without excessive dependence on foreign aid, particularly where that aid involves borrowing, no matter how concessional the interest rates.
This study has explored most of the aspects of cost sharing and cost recovery. It argues that the economic basis of many of the arguments for cost sharing policies is simplistic and part of a wider agenda of transferring costs from government to people outside the tax system, often in untransparent and unaccountable ways. Irrespective of how households make choices, the aggregate effect seems to have been that cost sharing has contributed to a stagnation in enrolment ratios and failure to improve the quality of educational provision, and that it has enabled governments to avoid difficult reforms. There is already considerable evidence of the damage that cost sharing policies have had on health care systems and the health of populations, such as rising rates of resistance to antibiotics as people cannot afford to purchase full courses, and it is hoped that this study will raise serious questions about the near and long term effects of those policies on education systems and the opportunities for populations to improve their knowledge and skills.
Perran Penrose
Penquite House
Cornwall 1997
