II Principles of Cost sharing and Key Issues

A. Introduction
B. Rights
C. Economics and Finance
D. Culture
E. Foreign Aid
F. Conclusions

A. Introduction

This chapter sets out a framework within which the principles underlying cost sharing need to be considered. The principles may be divided into four related and interdependent categories:

a) rights of citizens;
b) economics and finance;
c) cultural aspects;
d) in the case of developing countries, foreign aid and national autonomy.

B. Rights

The framework in which children's rights are normally considered is that provided by the UN Convention on the Rights of the Child. Most governments are signatories to the convention, which lays certain obligations on them. As far as education is concerned, states are obliged to 'make primary education compulsory and available free to all' (Article 28). Article 4 of the Convention recognises the possibility of resource constraints: 'with regard to economic, social and cultural rights, States Parties shall undertake such measures to the maximum extent of their available resources and, where needed, within the framework of international cooperation'.

The Convention begs many questions, some of which will emerge in this paper. Nevertheless, its intentions are clear. The child has rights to services, to protection, and to participation and self-determination. The child must be recognised as an individual with volition, which may be contrasted to some extent with instrumentalist views of children virtually as the property of parents as a resource worth investing in, exemplified in the human capital literature.⁹ The significance of rights based approaches and their related legal obligations is that they may imply policies which are not consistent with economic efficiency (however defined), and that they have costs attached to them which should not be discretionary costs because of legal enforceability.

[⁹ See de Vylder S. Children's Rights, Development Strategies and Macroeconomic Policies, Radda Barnen, January 1995, pp 2 ff]

Opportunity Cost

One interesting aspect of rights approaches and their effect on economic analysis is that of opportunity cost, which in some sense serves also as a metaphor for the adult view of the world embodied in most policy concerned with children. The concept of opportunity (or alternative) cost expresses the basic relationship between and choice. Opportunity cost is therefore the evaluation placed on the most highly valued of the rejected alternatives in the presence of scarcity. It is critical to understand that the alternative is that which might be and not that which might have been without the qualifying reference to choice. If no choice was present, although it is possible to value what might have been, it would not be correct to refer to these values as opportunity costs, since they did not represent a lost opportunity. If education is compulsory (as basic education is in the countries used in this study as case studies), no choice (legally) exists, and therefore there is no opportunity cost: although in the absence of choice it may sometimes be useful to consider alternative values, those values cannot be considered as opportunity costs.

Where education is voluntary opportunity costs, to be present, require someone to do the choosing, and only the chooser can know what the best alternative is: the value of the alternative exists in the mind of the chooser, and nowhere else. The cost must be borne by the chooser and cannot be shifted to anyone else.

As opportunity cost exists within the mind of the chooser, and cannot be objectified or measured by anyone else, it cannot be readily translated into money, although this is common practice in education economics. The cost only exists at the moment of decision. If a child decides between his/her alternative employment and going to school in favour of the latter, once the choice is made, the cost vanishes, for that which is rejected can never be enjoyed, nor can it be recovered. They become sunk costs, which are irrelevant, except in so far as we are interested in quantifying sunk costs. Opportunity cost is forward-looking. It is 'choice-influencing' rather than 'choice influenced'.¹⁰ Economists tend to value a child's or a household's opportunity cost of schooling in terms of earnings foregone, for example, related to an imputed market agricultural wage, and the opportunity cost of not attending school in terms of the returns foregone. Both these sets of calculations, while of interest, can only be speculative.

[¹⁰ See Buchanan J, Cost and Choice, University of Chicago Press, 1977, and Professor Buchanan's entry on opportunity cost in the Palgrave Dictionary of Economics.]

It is likely that many, if not most, children want to go to school, and it is equally likely that many are not able to attend because of family pressures to work. Family income foregone is a cost to the family of a child attending school, but if both the family and the child choose to attend school, the relevance of the cost becomes less clear. If the child chooses to attend but the family does not wish it, the cost is ascribed to the family and not to the child, who, it is generally agreed, has a right to go to school. This is an example of the importance of defining a position on children's rights, and compulsory education is partly a response to the danger of parents not wishing children to attend school. It might be argued therefore that family income foregone is irrelevant, though nonetheless real. In terms of economic theory, the concept of opportunity cost is relatively straightforward in market settings (where opportunity cost at the margin equates to cost), but once it is extended beyond market settings its interpretation becomes complex.

C. Economics and Finance

Mark Blaug in his classic monograph published in 1973 reflected a change in thought on how education policy should be made. 'Suffice it to say', he wrote, 'that the concept of education planning for economic objectives is an untidy mess, but it is a paragon of order compared to educational planning for social, political and educational objectives. Is it perhaps that sociologists, political scientists, psychologists and educationists have lacked a framework of decision making in which their positive findings may be fitted? If so, cost-effectiveness analysis is such a framework, which would permit social scientists other than economists to make their contribution to the subject'.¹¹

[¹¹ Blaug M, Education and the Employment Problem in Developing Countries, ILO, 1973, p 25.]

Since Professor Blaug wrote those words economists have come to dominate education policy making (as opposed to education practice), and neo-classical¹² economics with its emphasis on market solutions and maximisation principles has developed the analytical tools with which are most familiar. This section reviews some attributes of those tools, with particular reference to partial equilibrium analysis; to the propositions that people ('economic agents') are 'rational' (often loosely used as short hand for economic rationality); and to the economics of public pricing. What are the theoretical and practical criteria on which we can draw to determine public policy for the finance of education and therefore the scope and level of cost sharing? How far are the widely accepted approaches which dominate theoretical discussions of user charges useful as a foundation for policy formulation?

[¹² In this discussion the terms 'orthodox' and 'neo-classical' are used interchangeably. This is perhaps too loose, but this is not the place for an extended definition and justification of terms. An important theoretical work which sets out the definitional issues is Hodgson G. Economics and institutions, Polity Press, 1988. There is a powerful body of literature which sets out to question both equilibrium theoretical systems and that part of orthodox economics which has been critical of equilibrium (e.g. the Austrian School).]

Economic Justification for State Provision of Education: 'Market Failure'

The role of government and tax finance the provision of education is generally justified by the presence of conditions for 'market failure'. In other words, the market alone would not supply sufficient education services to be economically 'optimal', and the state must intervene in order to compensate for the shortcomings of the market. For example, in the cases where people do not fully appreciate the social benefits of education (i.e. the benefits to others rather than themselves, as in the apparent case of girls' education and fertility) or cannot afford basic education, society, represented by the state, will 'demand' more education than the aggregation of individuals will demand - the social demand will exceed the private demand. Individual consumers will not take into account the benefits to society which would accrue should they 'demand' more education, and because they do not take those benefits into account, their demand (when added together) for education is less than the 'optimum', meaning the level of demand which benefits society most. In such cases the state should finance education so that price does not act as a barrier to participation. Similarly, where education and training are considered to be necessary conditions for economic growth, society and therefore the state (in principle at least) have a common interest in increasing the educational level of the population.

The presence of conditions for 'market failure' implies a recognition that some 'subsidy' to education services is required, but the presence of fiscal constraints suggests that 'market clearing' solutions will depend on the extent to which users of the service are able and willing to pay for them. This is a difficult judgement, as is the judgement of what the socially 'optimal' desired level of provision might be.¹³ In orthodox economics, willingness to pay is the maximum price that can be charged without reducing an individual's welfare or utilisation of services. The policy issue is therefore one of determining the appropriate mix of tax and non-tax financing.

[¹³ One curious result of 'market' policies based on orthodox economic analysis is the power given to policy makers, who have assumed some of the functions of the central planners who are now so vilified. Thus, Jimenez (Jimenez E, Pricing Policy in the Social Sectors Cost Recovery in for Education and Health in Developing Countries, Johns Hopkins University Press, 1987) suggests that 'policy makers must agree on the level of consumption at which the marginal social benefit of another unit of education service is equal to its marginal social cost' (p 68, fn 3) and that 'the basic needs level of consumption can be interpreted as a minimum amount that society or the policy maker considers desirable' (p 133). Similarly, another influential writer equates 'society' with 'government' in considering social returns (Psacharapoulos G. and M. Woodhall, Education for Development: An Analysis of investment Choices, OUP, 1985, pp 35 - 37). There are many reasons why such 'decisions' should not be left to 'policy makers' acting on behalf of society: planners and 'policy-makers' have their own preference functions. Geoffrey Hodgson (Economics and Institutions, Polity Press, 1988, p 246) writes that there is 'an internal contradiction in much New Right analysis. If it were as absolutely rationalistic or subjectivist as claimed then it would have to abandon any rational argument to persuade others of the policy proposals to which it is attached. This inner contradiction caused by combining extreme subjectivism with rational argument was identified by Michael Oakeshott ... when he wittily described Hayek's work as 'a plan to resist all planning'.]

Financing of Education: Tax or Fees?

Tax finance is the most stable source of finance for services. Most, but not all, people would accept that tax systems should be progressive, that is, the better off should contribute a larger proportion of their income at the margin than the less well off. User fees, on the other hand, are a comparatively fragile base for social service finance. Moreover, fee finance relates to the individual consumption of services and can interfere with efficient and equitable service provision because of the effect on distribution: one function of the state and tax systems is to redistribute resources. A broad based tax system should not be hijacked by the better off in their pursuit of benefit from services, but to a greater or lesser degree there will always tend to be a distribution in favour of the better off Indeed, it may even be desirable. ¹⁴

[¹⁴ 'Higher supply strongly biased towards more privileged groups may be worse than lower supply'. Stern N, Comment on Social Sector Pricing Policy Revisited, Proceedings of the World Bank Annual Conference on Development Economics, 1989, pp 139 - 142, quoted in Tilak J. B. F. Cost Recovery in Education, op cit. p 15.]

Government expenditures derive from taxes which include income and transaction related taxes, borrowing and in the cases of the countries we are discussing, foreign aid. A distinction between direct compulsory or quasi-compulsory payments made by students and parents for schooling and payments made by governments is to some extent artificial. Bird defines a tax as 'a compulsory contribution from an individual to the government without reference to any special benefits conferred on him as an individual...', and a fee as 'a compulsory payment intended to defray the cost of a service undertaken by the government primarily in the public interest but conferring a measurable advantage on the fee payer'.¹⁵ He acknowledges the difficulties posed by education. He does not take into account the fungibility of money: defraying the cost of a service is the same as supporting the cost of an alternative object of expenditure for any given level of total expenditure. Fees permit governments to have large armies or large debt just as much as they 'pay for' education. In other words, fee income can allow governments to maintain an overall pattern of expenditure which may be inefficient, inequitable or indeed repressive.¹⁶

[¹⁵ Bird R. M, Charging for Public Services: A New Look at an Old idea, Canadian Tax Foundation, 1976, pp 16-22.

¹⁶ This proposition might be rejected on the grounds that the finance involved is small relative to total spending. However, that is the significance of the margin: in cases where most of the education budget is salaries the smallest reduction in expenditure has an effect on employment and therefore political consequences. The straw that broke the camel's back is an example of impact at the margin.]

The attraction of fee based systems is that fees can be earmarked (allocated to specific purposes) more easily than general taxation.¹⁷ Graduate taxes or payroll taxes can be earmarked taxes, and strong arguments have been made in their favour.¹⁸ However, earmarked taxation, if acceptable to Treasuries, which it rarely is because it constrains financial policy, requires relatively sophisticated systems to administer, systems which are lacking in most developing countries. Fees which do not pass through consolidated revenues are not threatened by Treasury policy. User charges, fees and other types of cost recovery, when not centrally administered, bypass the normal budgetary and accountability processes, and therefore offer none of the constraints that tax finance offers. They often bypass the political process, as well as government accounting processes: several countries are beginning to experience problems from the unaccountability of the user charge process.

[¹⁷ For a review of the arguments for and against earmarking, see Teja R. S. and B. Bracewell-Milnes, The Case for Earmarked Taxes, IEA Research Monograph Nr 46, 1991.

¹⁸ Colclough C, Raising Additional Resources for Education in Developing Countries: Are Graduate Payroll Taxes Preferable to Student Loans?, International Journal of Educational Development, Vol. 10 Nr 2, pp 169 - 180.]

Education Pricing: Efficiency and Distribution

Pricing theory is principally concerned with efficiency. In its purest form, the price of the service should be set at its marginal cost of production, as that is the only 'correct' price. However, the marginal cost of providing education is not easy to determine: it may be the additional pupil, which may be low, or the additional school, which may be high. In the absence of any pricing rule for education, the principle of 'excess demand' is adduced.

An early formulation of the rules for the mix between tax finance and fees was by Thobani in 1983 in the case of Malawi, and is sometimes known as the 'Thobani Rule'.¹⁹ The Thobani Rule requires a market clearing price of education services, and subsidies are only made in cases where socially desired service provision solutions exceed market solutions, where less education is provided by the market than is socially desirable. If the constraint on resources is such that the supply of services is insufficient meet the demand, some of those who demand the service in vain will, it is argued, be willing to pay for it. There would be a lower aggregate supply of services in the case of full 'subsidy' than in the case where people paid, because the available resources cannot meet the costs of the full subsidy. If fees are charged, the total level of resources available increases, all other things being equal. Therefore, the argument runs, charging fees augments the total level of resources for education, so that all demand can be satisfied, even though the subsidy does not cover the full costs. In this case, the rule is to raise the user charge so long as there is excess demand for the service.

[¹⁹ Thobani M, Charging User Fees for Social Services. The Case of Education in Malawi, World Bank Staff Working Paper Nr 572, 1983. More specific analysis was undertaken in Tan, J-P, K. H. Lee and A. Mingat in User Charges for Education: The Ability and Willingness to Pay in Malawi, World Bank Staff Working Paper Nr 661, 1984.]

The converse of the rule can also apply. Where there is a constraint on fee payment, it would follow that fees should be reduced and/or subsidies increased in order to attain the socially desired level of service supply.²⁰ In reality there are always constraints on both public finance and on household finance.

[²⁰ For an extended discussion of these theoretical issues see Bertrand T. and R. Griffin, The Economics of Financing Education: A Case Study of Kenya, World sank Staff Working Paper Nr 402, 1983, Annex 1.]

The rules can be formally demonstrated by the construction of demand functions, showing how the demand for education services is influenced by their price. However, the analysis is much more complicated than it appears to be at first sight in the standard texts, simply because of the difficulty of specifying the demand curves. In applying supply-demand principles the product must be defined. Education cannot be easily categorised as one product, yet the notion of private excess demand implies that there is one product with one demand curve.²¹ The question must be asked, excess demand for what? For example, if there is excess demand for poor quality schooling, what does this mean? Poor quality schooling is a different product from good quality schooling, and must therefore have a different demand curve.²² Private excess demand may have little meaning under such circumstances, though some writers see a virtue in the private provision of low quality education.²³ In many countries education cost benefit analysis appears to suggest 'high' rates of return, while at the same time the low quality of service provision is widely recognised, leading to the logical conclusion that governments do not need to invest in improving quality (particularly as such expenditures may reduce net social benefits). A proper specification of demand curves is by no means an academic point, and can provide a conceptual framework within which to consider some of the paradoxical results which seem to emerge from much education policy analysis.

[²¹ Throughout this discussion 'education' is synonymous with schooling. Such a synonymy is not always acceptable. we may be talking about a demand for a certificate, or a demand for an unemployment substitute, as well as a demand for learning. These considerations highlight the difficulty of analysing education within a simple demand-supply framework (at least as far as economic principles are concerned). Furthermore, perceptions of the product change, and a different thing might be demanded at different points in time.

²² Indeed, the conclusions reached by Armitage and Sabot in Kenya suggest that there is an effective demand for low quality education. see Armitage J. and R. Sabot, 'Efficiency and Equity Implications of Subsidies to Secondary Education in Kenya', in Newberry D. and N. Stern (eds), The Theory of Taxation for Developing Countries, OUP, 1987.

²³ Such as James E, Public Policies Towards Private Education: An International Comparison, International Journal of Education Research, Vol. 15, 1987, and Armitage and Sabot op cit.]

Moreover, the simple models we are discussing assume that the costs of education are given. If costs can be reduced, fees and/or government expenditures can be reduced. Where costs are in some sense unnecessarily high, households are then required to pick up the bill for government or institutional inefficiency. In this case it would follow that any fee policy would be specifically concerned with minimising fees and reducing costs as a matter of principle.

While Thobani and others argued that income from user charges should yield additional finance for education, an alternative school of thought holds that user charges should be a substitute for government expenditure, and that the burden of service financing should shift from centrally or locally collected taxation to direct payment. In other words, reducing excess demand through increased subsidies is itself not socially 'optimal' because the marginal costs incurred through additional taxation exceed the marginal benefits to society.²⁴ In the real world of public finance management it cannot be assumed that fees will augment total resources.

[²⁴ Ie finance raised through taxation uses up resources in its collection and is itself costly. Resources are used up in administration and in waste. There are also social costs. For example, where tax systems are regressive and disproportionately favour the better off, any services paid for out of tax finance have by definition social costs to the poor. Where taxes have high opportunity costs such as might be found in a country with high marginal rates, the social benefits from government spending on education would have to be adjusted accordingly: this is rarely considered in the social cost-benefit calculus. The additional resources needed to increase subsidies are usually not possible to calculate and are hidden overheads, but nevertheless they would result in an inefficient solution in this context.]

Efficiency

User price theory suggests that charges enhance both efficiency and equity. The attraction of the argument for some is therefore that they can save money and, at the same time and with one policy instrument, promote a fairer and more efficient use of resources. Cost recovery is seen as a 'multipurpose remedy'.²⁵

[²⁵ Hall P. H. Land, R. Parker & A. Webb, Change, Choice and Conflict in Social Policy, Heinemann, 1975, quoted in Tilak J. B. G. Cost Recovery Approaches in Education, National Institute of Educational Planning and Administration, New Delhi, 1994, p 13, paper prepared for the Workshop on the Social and Economic Effects of Alternative Methods of Financing Education and Health Services in Developing Countries, Institute of Development Studies, Sussex, March 1994. This paper, and others from the same workshop, appear in Colclough C. (ed), 'Marketizing Education and Health in Developing Countries: Miracle or Mirage?', Clarendon Press, Oxford, 1997 (in press at the time of writing).]

User charges can in principle affect efficiency different ways. Additional resources can permit increased technical efficiency as well as enhanced cost effectiveness. A combination of additional fee finance and reduced tax finance can put pressure for a better and more accountable allocation of public finance.²⁶ Fees, it is argued, can force consumers to appreciate the value of services and not to use them unnecessarily. Because the costs of higher education to individuals is low in relation to the returns, more people demand it than would do so if they faced the 'true' costs.²⁷ The argument is analogous to the concept of 'frivolous' demand for health services, and it is a difficult argument to apply to education: there is little evidence to suggest that the argument has substance even in relation to health services.²⁸

[²⁶ There are several (related) dimensions of efficiency, including cost-effectiveness, technical, allocative, and incremental output efficiency, and they relate both to the ways in which education policy is formulated and to the way in which public finances are managed. Cost-effective provision of education involves increasing the education and skill status of the population for a given budget; technical efficiency involves attaining a given level of education provision for minimum cost; and allocative efficiency involves the least cost combination of inputs for a given output. Incremental output efficiency relates the additional outcomes associated with each efficiency measure with each other: do we want major improvements in reading while numeracy declines?

²⁷ 'If households faced the true social costs of obtaining more education, they would confront a lower rate of return; instead they are induced to obtain more schooling.' Jimenez E, Pricing Policy in the Social Sectors Cost Recovery in for Education and Health in Developing Countries, Johns Hopkins University Press, 1987, p 40. But the intention behind subsidising higher education is precisely to induce more demand. The issue is whether it does or not, and how that demand is distributed across income groups.

²⁸ See Creese A, User Charges for Health Care: A Review of Recent Experience, in Health Policy and Planning, Vol. 6 Nr. 4, 1991, pp 309 - 319.]

The argument for efficiency effects of user charges and private schooling also holds that competition is created by consumer choice, and that this drives down costs. Apart from the fact that most consumers face little choice in what school their children can attend, the role of competition itself as it is defined by orthodox economists is troublesome. It is not axiomatic that competition, which can be 'at once the god and the devil',²⁹ is always desirable, and that it inevitably enhances welfare.

[²⁹ Penrose E. T. The Theory of the Growth of the Firm, John Wiley and Sons, New York, 1959, p 265. The most advanced area of economics in the sense of its rejection of simple neo-classicism is the economics of the firm, where much interesting theoretical and empirical work is now taking place. For the role of competition and economic theory of competition, see Best M, The New Competition, Polity Press, 1990, and Hodgson G. Economies and Evolution, Polity Press, 1993. There are many reasons why (neo) liberal dogma about competition should be considered sceptically, including reasons relating to efficiency and reasons relating to equity and ethics (e.g., see Hirsch F. Social Limits to Growth, Routledge and Kegan Paul, 1977).]

Equity

To many, the main argument in favour of cost sharing policies, particularly at post basic education levels, is that they induce a more equitable distribution of opportunities and resources. To others, such a concept is counter intuitive. Equity, it is argued, is achieved by increasing resources through fees charged to those who can afford them through present or future resources, and thereby more children are enabled to receive education. Greater allocative efficiency allows resources to be redirected from higher education to primary education, and this permits poorer people to send their children to school, or, alternatively, the funds can be used to give scholarships to poorer students.

Equity arguments are deceptively simple. There are few if any examples of obvious reductions in public spending at higher levels in favour of lower levels of education, especially where costs per pupil are concerned, partly because in most countries the management of resources for education is weak and the systems are inefficient. The effect of reallocations is just as likely to be to redirect finances towards inefficient uses, which is hardly fair on those who pay, whether they be 'rich' or not. Moreover, in a broader context, the fungibility of money implies that fee income can perpetuate rather than obviate inequitable resource allocation.

Nevertheless, equity of provision of education opportunities across income groups should be an important social goal. The question is how it is to be achieved, and the history of earmarked and targeted expenditures in countries with weak fiscal management suggests that the answer is not straightforward. For example, measurement of how far government expenditures are redistributive, neutral or regressive (in the sense that the better off benefit from them more than proportionately to their share of income), can to some extent be made on the basis of household data and information on government expenditures.

Household finance combines with public finance to make up the total package of expenditures on education. More affluent households have a higher level of discretionary resources to allocate to education. Relative affluence is often concentrated in particular geographic locations, but in most communities there is a spread of incomes. One function of public finance is to compensate for income diversity through the tax system and through the distribution of benefits financed by taxes. Analysis of the geographical distribution of resource allocations can show how far public budgets are equitably shared between the population, and at the household level it is possible to determine to some degree how equitably public resources are distributed between income/expenditure groups, irrespective of location.

The incidence of the benefits from public spending can be derived from data on the number of people in a population in given income (expenditure) brackets and the average expenditure per enrolled pupil. Enrolments of different consumption groups can be identified, and on the assumption of a single national average expenditure per pupil the amount of public expenditure 'captured' by each income group can be plotted on a Lorenz curve to show the progressiveness or regressiveness of public spending. The case studies of Ghana and Tanzania show the results of benefit incidence calculations, and are intuitively acceptable, in that they show progressive distribution of resources at the primary levels, regressing at higher levels.³⁰

[³⁰ The technique is a useful one but not without flaws, and tells us nothing about the quality of expenditures on the different income groups. As with much poverty oriented analysis, it is also not always clear that the information adds a good deal to what is already apparent, and when the costs of data collection and processing are taken into account there must be some question about the use of scarce resources on expensive surveys: Living Standards Measurement Surveys cost up to $1 million, not including staff costs.]

In Tanzania and Ghana (and in many other countries), the incidence of the benefits of public expenditure on primary education is broadly progressive. The issue is how to increase and sustain enrolments at the basic level at the same time as making the incidence of benefits from post basic education expenditure more progressive. Simply charging fees for post basic education will not achieve that objective without strong complementary actions: it merely increases the marginal rate of tax on 'rich' families with few redistributive benefits.

The Assumption of Rationality: Government and Household Reaction Functions

Analysis of policy requires assessment of how agents react to interventions. How do governments and households react to cost sharing and cost recovery? What are the effects of the fungibility of money? These are the central questions, from which the questions we ask in this paper are derived.

The Government Reaction Function

A paradox in the theoretical system we have been discussing is that on the one hand governments are frequently assumed to be inefficient, so the less that government manages the better; while on the other hand the success of market prescriptions depends on an efficient bureaucracy, as well as a set of complementary interventions which would ensure success. There is a set of 'implementation conditions' upon which the success of any reform depends, and it is the failure of economic reform packages to reflect that concept at the sectoral level which is the cause of much of the decline in enrolments in many countries.³¹

[³¹ Creese A. and J. Kutzin, Lessons from Cost Recovery in Health, World Health Organisation, 1994, paper prepared for the Workshop on the Social and Economic Effects of Alternative Methods of Financing Education and Health Services in Developing Countries, Institute of Development Studies, Sussex, March 1994. See Chapter 5 for an outline of a possible package.]

The network of assumptions we have been considering extends beyond individual behaviour and equilibrium, which we discuss below, to assumptions about perfect flexibility in public finance mechanisms: it is almost as though the abstract conception of the rational individual is mirrored by the abstract conception of the rational bureaucracy. The capacity of administrative systems to respond to calls for the reallocation of finance, targeted scholarship systems, fee exemption mechanisms, and so on, is notoriously weak: all of these require resources which are usually not available. Welfare economic analysis of user charges involves assumptions about public finance management, in particular that reallocation of finance to those areas yielding the highest social returns is in fact possible, and that government expenditure can be efficient and 'optimal'.

The reaction of government to fee income and cost sharing is the central fiscal question we are addressing, and the functional relationship between a government and an additional resource injection may be expressed via a reaction function.³² Whether explicitly or implicitly, the effect of fiscal squeeze has been to regard cost recovery as a way of reducing the burden of public expenditure. Were other things to be equal, total ('private' + public) expenditures on education would rise as a result of increased cost recovery, but in reality cost recovery policies are usually intended to mitigate the effect of falling public expenditures. This can be looked at in one of two ways:

a) without cost recovery expenditures on education would fall by more than they would with it; and

b) without cost recovery measures government would have been forced to reallocate in favour of education away from, say, defence, or to improve revenue collection. Fungibility in the budget permits government to charge parents fees which effectively finance less socially desirable outcomes: without the option of cost recovery expenditures on education might be maintained and a better composition and balance of public expenditure be achieved at the same time.

[³² See Heller P. 'A Model of Fiscal Behaviour in Developing Countries: Aid, Investment and Taxation', American Economic Review, Vol. LXV Nr 3, June 1975, pp 429-445.]

(a) and (b) above represent two possible counterfactuals, and, as with all counterfactuals, it is not possible to say whether or not they are 'true' for any given case. The questions raised are central to cost sharing policy, because recovery of the costs incurred (out of tax finance) by the state for education and health permits the state to maintain levels of inefficiency and misallocation. Even where there is an explicit policy of substitution of non tax compulsory financing for tax financing, the gains which the proponents of such policies might expect will not be forthcoming if the overall level of public expenditure is not reduced or if there is no progressive reallocation.

The Household Reaction Function

The reaction of families and individuals to cost sharing policies and patterns of government expenditure is equally complex. How do people respond to better services? How much is the demand for education or certification affected by the quality of provision of service? Do sectoral interventions have a greater effect on household responses to education policies than do the employment effects of economic policies? There are two broad sets of views of the how and why decisions are made in households. On the one hand many models are based on the assumption that households are homogenous groups with one set of preferences, while others, more in keeping with anthropological studies, are based on more complex assumptions about how individuals within households negotiate their interests.³³ There is increasing interest in the specific role of children in decisions affecting their future. In many cultures children do not get a substantial share in decisions about schooling, although in some, such as Ghana, with more complex family patterns, children may not only decide but have responsibility for raising money. In some societies women prefer to spend money on their children's education, whereas men have other preferences.

[³³ For a review see Alderman H. et al, 'Unitary Versus Collective Models of the Household: Is it time to shift the burden of proof ?', The World Bank Research Observer, Vol. 10, nr 1, Feb 1995, pp 1 - 19.]

There is a wide range of economic and cultural influences on household reactions to education policy. For example, Tilak has shown how, in the case of India, there was a positive relationship between government and household absolute expenditures on education. Families appeared from his data to spend more on education as government spent more.³⁴ If there is a causal relationship, one explanation might that the level of public expenditure has a strong enough effect on quality, perhaps in the form of school inputs, for people to want to spend more. This would be consistent with survey experience and with experience in the health sector.³⁵

[³⁴ Tilak J. B. G. Family and Government Investments in Education, International Journal of Educational Development, Vol. 11, Nr 2, pp 91-106, 1991.

³⁵ Bennett S, The Relationship Between Public and Private Systems in Health Care, Health Policy unit, London School of Hygiene and Tropical Medicine, 1994, p 16, paper prepared for the Workshop on the Social and Economic Effects of Alternative Methods of Financing Education and Health services in Developing countries, Institute of Development Studies, Sussex, March 1994.]

Few countries have time series of data for household education expenditures which would allow the hypothesis of complementarily between government and family expenditures on education to be explored further.³⁶ Intuitively, it makes a good deal of sense. For example, survey results indicate that parents will pay more if they feel it is worth it: if there is indeed a link between total expenditures and quality of output, then increased government expenditures could trigger increased family expenditures. In order for cost recovery to become a successful policy, the consumers of the service must see benefits. One of the underlying reasons for resistance to cost recovery in higher education is that there is little apparent benefit to students, who have little incentive to accept it. Were their conditions to be improved, it is possible that they would show more willingness to pay.

[³⁶ For example, using regional cross-sectional data, there was no systematic relation between household and government spending in Viet Nam (Penrose P. Review of Public Expenditures on Education in Viet Nam, UNDP/Ministry of Finance, Hanoi, 1995). But cross sectional data would not be strong evidence either way.]

The contrary hypothesis is that government expenditures crowd out private expenditures (West's 'public/private displacement mechanism'³⁷), particularly those on private schools. High levels of government expenditure on education in developed countries are, it is argued, a principal reason for relatively low private expenditures when compared with less developed countries, where 'excess demand' prompts parents to seek their own solutions.³⁸ The absence of time series data makes it difficult to take these conjectures further, though West's study of education finance in Britain through the 1 9th and early 20th centuries is food for thought. An interesting aspect of the same issue is found in Chile. One study shows that after the privatisation of education in Chile total household expenditures on education increased, but profit-taking by school owners resulted in lower average expenditures on direct and indirect inputs than previously.³⁹

[³⁷ West E. G. Education and the Industrial Revolution, Batsford, 1975. He argues that there was a significant 'public/private displacement mechanism' in Britain between 1833 and 1945 as public (state) schools took over from private schools and total average expenditures on education expenditure declined (Chap 15).

³⁸ See for example Johnes G. The Economics of Education, MacMillan, 1993, pp 81 ff, and James E, Public Policies towards Private Education, op cit.

³⁹ Schiefelbein E, Restructuring Education through Economic competition: The case of Chile, Journal of Educational Administration, Vol. 29, Nr 4, 1991, pp 17-29.]

One way in which we find out how people react to cost sharing policies is through household surveys.⁴⁰ Household income is made up of cash income through work, sales, borrowing and transfers, and non-cash income, such as the imputed value of food and own production. Household surveys have different ways of measuring income, and they are not always clearly specified in analysis based on them. It is notoriously difficult to determine income in questionnaires, and a little less difficult to determine expenditures: often therefore household expenditures and cash savings are used as a proxy for (cash) income and other non-cash values are added. There is a strong relation between income and expenditure, but there is a less strong relation between cash expenditure and total income (cash plus imputed income) among poorer people.

[⁴⁰ see Behrman J. R. 'Human Capital Formation, Returns and Policies: Analytical Approaches and Research Questions', Journal of International Development, Vol. 8, Nr 3, 1996, pp 341-373 for an overview of issues of technical analysis of household data in a neo-classical empirical economic framework. The use of the term 'household' is in many ways unsatisfactory, but there is no easy shorthand alternative. 'Household' has come to mean the opposite of government, so that total expenditures = government + household. The pervasiveness of the technology of planning now far surpasses what was available to the central planners of the command economies, and as market economies are increasingly susceptive to the preferences of planners as survey techniques become more sophisticated.]

Fees and other schooling costs affect household savings and investment, and by extension there must also be an effect on national economies. One reaction to fees which is not well documented is that the poor may sell significant proportions of their assets (i.e. more than that which is normally set aside for financing their children) in order to pay fees and other charges: they switch investments from physical investments to human capital. Fee obligations may be one cause of declining household abilities to sustain their basic needs as well as reduce their capacity to generate wealth. Part of the problem is that although the returns to schooling appear to be 'high', those who do not earn are not included in the calculations, and families take increasing risks in deciding to 'invest' in schooling, even assuming that their expenditures are voluntary.

Wealthier families can afford more risk⁴¹ and are likely to spend more on education. Their opportunity costs include investment in business: the alternative 'investments' to education are not restricted to capital or equipment investments with which returns may be compared, but also in recurrent costs of entrepreneurship and other wealth creating activities. User charges will necessarily have a macroeconomic cost, the issue being that of the size of the effect and how the costs and benefits balance against each other.

[⁴¹ Levhari D. and Y. Weiss, 'The Effect of Risk on the Investment in Human Capital', American Economic Review, Vol. 64, Nr 6, December 1974, pp 950-963.]

Risks are greater and less affordable for poorer families. Poor households have more children, and the ratio of total education costs to total household cash income will be higher in the case of large households than in the case of smaller households on the same cash income. One consequence of this is the additional burden placed on families by extending basic education without compensation measures. It is also likely that in poor areas the number of households represented in a school will be less than in better off areas, because of larger families. The average expenditure per child would thus be lower, and the poverty of the school would be worsened by reduced income from parents.

Economic Rationality

Over the last few years the economic concept of 'human capital' has gained ascendancy in the social sector literature and dominated foreign aid discussions on education.⁴² 'Rational' human beings are assumed to make investments in themselves and their children on economic grounds, and society also makes investments in individuals on economic grounds. Investments by definition yield streams of financial and economic returns. In the case of human beings these are expressed in the form of lifetime earnings and other financial and non-financial returns, which benefit both society and individuals. These streams of returns can be estimated, and internal rates of return both to individuals and to society as whole derived as guides to policy.

[⁴² Compare Mark Blaug's assessment in 1976: 'In all likelihood, the human capital research programme will never die, but it will gradually fade away to be swallowed up by the new theory of signalling, the theory of how teachers and students, employers and employees, and indeed all buyers and sellers select each other when their attributes matter but when information about these attributes is subject to uncertainty'. (Blaug M, 'The Empirical Status of Human Capital Theory: A Slightly Jaundiced Survey', in Journal of Economic Literature, Vol. 14, Nr 3, Sept 1976, pp 827-855.]

One assumption which is hard to accept as a basis for serious policy formulation is the view that economic coordination is simply a matter of price signalling in markets to which individuals react predictably (a crude definition of 'rationality'). We need to take account of a wide variety of factors relating to culture, institutions and other components of individuals' environments. The problem with simplified propositions is that while they are useful tools of analysis, they can discourage deeper understanding of the connecting principles relating to relevant phenomena.⁴³

[⁴³ In his important book, Equilibrium and Evolution: An Exploration of the Connecting Principles in Economics, Manchester university Press, 1991, Brian Loasby suggests, following Adam Smith, that 'we try to make sense of the world by imposing patterns on it, and then sticking to them as long as they are tolerably successful in allowing us to feel that we understand what we observe and what we experience.' He continues: 'Smith argued that people like to feel comfortable, and that they do not feel comfortable unless they can link together in their own minds the phenomena to which they are exposed. People prefer not to have to think; but what they like even less is the feeling that they do not understand, and in such a situation they are driven to seek an explanation.... The motivation of science, therefore, according to Smith, is the psychological need to invent a set of connecting principles which will make sense of experience, and thereafter leave the brain in peace.' (pp 6 -7). The drive to define connecting principles has resulted in the rational choice equilibrium method of economics, a method which gives tranquillity to the economist's imagination. The world of the economist's imagination is not the world of policy: Herbert Simon wrote 'the decision maker's ... perceived world is fantastically different from the real world' (Simon H. A, Models of Bounded Rationality, MIT Press, Cambridge MA, 1982, Vol. 2 p 306), and the consensus of economists in the subject matter we are discussing is very much responsible for many of the problems faced by education systems in poor countries now.]

The importance of the price mechanism for education policy making is illustrated by the current hegemony (at least within the Bretton Woods Institutions) of the analysis of internal rates of return to education. Rates of return are seen by many as an indicator of how individuals perceive the financial benefits to themselves derived from additional education, and as an indicator of the economic benefits to 'society'. Where there is a high 'private' benefit relative to the estimated 'social' benefit, it is implied that there is excess demand and that therefore fees must be charged. Because the 'social' internal rates of return to primary education are generally estimated to be higher than those to secondary or tertiary education, and because the divergence between 'private' and 'social' benefits is usually calculated to be less at the lower levels, policy advice based on rate of return analysis is invariably that public expenditures should be directed at those lower levels of education while cost recovery must be pursued at higher levels, principally at the university levels.⁴⁴

[⁴⁴ Social rates of return are computed simply by adjusting estimated private rates to include costs to the state, with little corresponding adjustment on the benefit side of the cost-bereft calculus, particularly those which might be caused by externalities.]

There are two kinds of question which can be posed about the value of education cost-benefit analysis. The first relates to the rational choice basis of the assumptions derived from it, and the second relates to reasons of a more technical and methodological nature which cast doubt on the reliability of the solutions offered. In many respects the robustness of rate of return analysis is surprising, and tells us more about the absence of tools which enable us to make education policy than the usefulness of the technique itself.⁴⁵ Although there have been a number of criticisms of the foreign aid agencies' reliance on rates of return in its policy 'dialogues' with governments, criticisms do not appear to have affected the widespread acceptance of the results offered, principally the policy focus on expanding basic services and rationing access to post-basic services through price.⁴⁶ The range of objections to the technique as a main basis for policy decisions is too wide to discuss fully in this paper, but I return to the subject where it is relevant, particularly in regard to post primary policy: there are in any case very strong reasons for not placing too much credence on education rates of return. Moreover, even for those who do have faith in such analysis, the problem should lie more in how the findings are used in relation to public expenditure allocations as a whole (especially when they are driven by foreign aid conditionalities), and whether the correct response is to charge fees and raise costs for post-primary education rather than to eliminate expenditures elsewhere in the budget with even lower 'social' rates of return, as theory would suggest.

[⁴⁵ But the need for such a tool derives not from 'market' approaches but rather from central planning requirements. A point which was often missed from the rate of return versus manpower planning debate is that both techniques are used by central planners, but one appears to be more sensitive to market signals than the other.

⁴⁶ The criticisms of the reliance on internal rates of return for education policy purposes are strong enough to require the results to be used with circumspection, if at all. Where they are used well, such as by Knight and Sabot in Knight J. B. and R. H. Sabot, Education, Productivity and Equality: The East African Natural Experiment, OUP, 1990, the authors caution readers that they did not 'wish to perpetuate the illusion of precision created by oversimplification' (p 51). Knight and Sabot, rightly, do not dismiss either the methods or the results, but suggest that the reliance on survey data for cross section averages does not take into account the changing conditions facing new entrants to employment, and that rates of return may be weak guides to policy. A strong (and sometimes intemperate) attack with many well argued points which I have not seen answered is by Curtin in Curtin T, The Economics of Public Investment in Education in Papua New Guinea, University of Papua New Guinea Press, 1991. See also Bennell P. Using and Abusing Rates of Return: A Critique of the World Bank's 1995 Education Sector Review, IDS Working Paper Nr 22, Sept 1995, and two other Working Papers (Nrs 23 & 24).

One interesting development is how more recent calculations (most of the cited IRRs are a decade or more old) show high private relative to social IRRs to primary education, and more or less equalised rates for secondary education (as in the recent calculations for Ghana). This may well be a trend, and it will be interesting to see how aid policy is affected. See Psacharopoulos G. 'Returns to Investment in Education: A Global Update', World Development Vol. 22, Nr 9, pp 1325-1343.

It is often suggested that nobody pays much attention in practice to rate of return findings, but experience suggests otherwise. Not only is lip service paid to them, but Treasury officials, often frustrated by the absence of reasoned justifications for expenditures in education sectors, latch on to them as at least one way of making sense of education spending.]

D. Culture

Although any discussion of cost sharing will have an economic bias in that the justifications for cost sharing policies relate to the financing of education, the cultural and social environment in which cost sharing activities take place will to a great extent determine how they work. There is considerable cultural diversity between countries which directly affects people's reactions to cost sharing, and the insights from anthropological and sociological research are important in balancing the cruder assumptions of economic rationality.⁴⁷

[⁴⁷ There have been a number of important studies over the years which bring cultural issues to the fore. Perhaps One of the finest is Philip Foster's Education and Social Change in Ghana, Kegan Paul, 1965. Another very important study is Serpell R. The Significance of Schooling. Life journeys in an African society, Cambridge University Press, 1996. Also Musgrove F. Education and Anthropology: Other Cultures and the Teacher, John Wiley, 1982.]

In 1932 T. S. Eliot wrote 'Questions of education are frequently discussed as if they bore no relation to the social system in which and for which the education is carried on... Education cannot be carried on in a void: our questions raise other questions, social, economic, financial, political. And the bearings are on more ultimate problems even than these: to know what we want in education we must know what we want in general, we must derive our theory of education from our philosophy of life'.⁴⁸ Western education can have a different significance in many traditional societies, and the influence of foreign aid can be, in Denis Goulet's phrase, 'shockingly ambiguous'.⁴⁹ One reason for the ambiguity is that modern education systems in non-Western cultures have not arisen in the same way out of a 'philosophy of life' in a natural fashion. Sociological and anthropological literature is replete with accounts of the reactions of 'traditional' peoples to modern education, and dissenting economic literature has also emphasised the consequent problems of transporting foreign models via the medium of foreign aid.⁵⁰

[⁴⁸ Eliot T. S. Modern Education and the Classics, in Selected Essays, Faber & Faber, 1966, pp 507-516.

⁴⁹ Goulet D, The Cruel Choice: A New Concept in the Theory of Development, New York: Athenaeum, 1971.

⁵⁰ 'External doles ... tend to bias the development process in directions based on external prototypes which are often inappropriate and therefore damaging. Such a sequence retards development rather than promotes it... Adverse results are all the more likely when the expenditure within the country is undertaken by people who do not themselves bear the cost.' Bauer P T. Dissent on Development, Wiedenfield and Nicolson, 1971, p 103.]

In a sociological echo of the language of economics, the concept of capital investment may be extended to that of investment in 'cultural capital'. The range of meanings of the concept of cultural capital is wide, but modern education systems may stimulate capital growth where it may be less desirable and destroy cultural investment without creating new capital.

In an example of the first phenomenon, as governments seek means of targeting public finance to overcome inequalities and inequities in the system they come up against the self-replicating nature of social status and provision where parents with education are able to use the publicly financed system to perpetuate and consolidate their advantages.⁵¹ At the same time, the pressure from the foreign aid agencies to 'liberalise' education provision and encourage the development of private schools will tend to create a 'cycle of privilege' if accompanied by a reduction in public finances for state schools.⁵² Cost sharing will always tend to create inequalities simply because schools in better off areas or with better off pupils are better financed than others.⁵³ If schools rely on cost recovery for non-salary items, and cost recovery is unevenly distributed between schools, underfunding is built into the system, even where the incidence of public expenditure may be progressive. Such a tendency would work to offset the supposed equity effects of cost recovery.

[⁵¹ See, for example, the discussion and analysis in Halsey A. H, A. F. Heath and J. M. Ridge, Origins and Destinations: Family, Class and Education in Modern Britain, Clarendon Press, 1980, pp 75 ff and pp 198 ff.

⁵² State education is often viewed from the World Bank as an 'inferior good', meaning something that people cease to purchase as their incomes rise.

⁵³ See, for example, Bray M, Community Financing of Schools in Less Developed Countries: Rationales, Mechanisms and Policy Implications, paper prepared for the Workshop on the Social and Economic Effects of Alternative Methods of Financing Education and Health Services in Developing Countries, Institute of Development Studies, Sussex, March 1994. There are other local phenomena. For example, in some places in the north of Ghana, parents do not wish to contribute anything to schools because of the competition from teachers in farming. Teachers, who earn salaries, are able to finance farming inputs (as well as use children as free labour), and this gives them an advantage over poor farmers with little cash working capital. Throughout every system local effects such as these strongly influence attitudes to cost sharing.]

An example of the second phenomenon is the relation between 'traditional' wisdom and skills and those promoted in modern schools.

A good deal of anthropological debate centres on the assumption and definition of (economic) rationality.⁵⁴ Anthropologists have much to say about the impact of markets and economics on societies to which these phenomena have been relatively recently introduced, particularly by foreign aid. Indeed, anyone who has worked in countries which are the subject of the attentions of foreign aid project designers should have been able to see the various patterns of resistance which are mounted against them.

[⁵⁴ In some instances economists can be treated as a homogenous group, but this should not be the rule: one example of this treatment is in R. L. Stirrat's interesting paper Economics and Culture: Notes Towards an Anthropology of Economics, School of African and Asian Studies, University of Sussex, 1994. This paper was one of many delivered at a conference with the theme of 'tensions between economics and culture' in the University of London Institute of Education.]

People possess economic rationality in the sense that they will respond in some way to economic stimuli. This is not the same as saying that those responses will outweigh all other responses, or that economic rationality is not tempered by other types of rationality. Self-interest is not confined to economic betterment, but includes it. The Mandevillian baker bakes bread for a living and not for altruistic reasons, but he may also consider it important to make good bread (quite apart from the market implications of baking bad bread), and may also take seriously his place in society as a contributor to general welfare. Economic rational man is stimulated exclusively by prices, an assumption as absurd as it is wrong. As surveys illustrate, people desire education for other reasons than economic reasons, but of course for economic reasons as well.⁵⁵

[⁵⁵ Though the Becker position would be that all behaviour is by definition rational to the behaver (Becker G. The Economic Approach to Human Behaviour, Chicago, 1976), a position not universally shared.]

For example, one interesting paper records that in Mozambique 'peasant families were generally unable to ascribe any meaning to what was taught in schools in terms of content... Instead, schooling was basically ascribed a purely functional value, that of opening up the door leading to modern society and its well-being. But even in this respect, the relationship was ambiguous'.⁵⁶ One reason for the ambiguity was that education offered an escape route from the peasant condition, but that at the same time the route was 'long, costly and insecure'. Another reason was that education took a member of the group away from the group, transforming him or her into something different. While this might be acceptable for boys, it represented a great risk for girls in a matrilineal society. Research such as this goes some way towards the explanation of why there is underenrolment in the presence of 'high' private rates of return.⁵⁷

[⁵⁶ Palme M, Being Respected but Teaching Hieroglyphs: addressing the question of the primary school teacher, culture and local community in rural Mozambique, paper presented to conference on 'tensions between economics and culture' in the University of London Institute of Education.

⁵⁷ Although, of course, the returns are only to those in employment.]

Survey literature in other countries shows similar phenomena. In Sierra Leone, for example, parents in some parts of the country see western education as haram (sinful), as breeding ingratitude, and as alienating children from their families and homes.⁵⁸ The general reaction by the agents of the modern education sector - teachers, administrators, foreign aid workers is that parents are 'ignorant', they 'do not appreciate education', and that community attitudes to education are in some sense irrational. This applies also to girls' education, where refusal to send girls to school is seen as an oppressive measure. In reality, parents' attitudes may be very rational, even in the economic sense, as in, for example, Pakistan, where parents who have traditionally been opposed to the education of their daughters may change their views if they see economic benefits deriving from it.⁵⁹ It is also important to note in the consideration of cost sharing that the benefits accruing from education can be more important in people's perceptions than the costs of attaining it, for the benefits are measured in other ways than the material, and are frequently seen as 'social' rather than 'private'.

[⁵⁸ Kroma S. Factors Influencing School Enrolment and Attendance in Rural Communities in Sierra Leone, (mimeo) Njala University College, Sierra Leone, December 1993.

⁵⁹ Bacchus K. & I. Farah, Tensions Between Economics and Culture in the Provision of Education for Girls, Institute for Educational Development, Aga Khan University, Karachi, 1995. The paper was delivered to the London Institute of Education conference.]

'All societies seem to have a somewhat ambivalent attitude towards children. They are accorded the status of humans as far as our basic orientation is concerned. We do not treat them like animals: we talk to them, greet them, expect them to eat, to toilet and to dress like humans; but we do not expect them to take responsibility for their actions or their words. They are thus incomplete humans, in a state of transition, creatures about to be persons. Their actions are promissory, provisional, subject to correction and to being forgotten. It is this ambivalence which motivates the activities we call rearing, socialisation or education through which we attempt to prepare children for adult personhood.'⁶⁰ Parents and communities are fairly clear about what they expect from their children, and there is always a tension between economic betterment and the attributes and behaviour of a person in any given society. In African countries, as in many Asian countries, communities are rural and small, and education must provide skills and attributes for survival both in those communities and in the bigger world of the towns and cities. Yet a key function of school is 'getting children out of their families'⁶¹: schools become the 'third cultural reality' of Malinowski, and they transcend the cultures in which they are situated in order for societies to progress.

[⁶⁰ Serpell, op cit. p 70.

⁶¹ Musgrove, op cit. p 174]

The importance of understanding what constitutes education in any society in relation to policies which require citizens to pay for it lies in the need to understand the product. The economic justification for making citizens pay for the education of their children is partly founded on calculations of economic returns, and therefore in some sense involves repayment of a debt to society (although economic justifications often conceal a simpler truth which is that the state has insufficient resources, for whatever reason). The implication might be that people pay for the perceived entitlement to employment. The product is deemed as one which enables those who purchase it to be employed, and secondary benefits may or may not be learning how to be better people, and so on. Evidence suggests that when employment opportunities do not flow from success in schooling, school is seen as only mildly relevant, which is a sad indictment of any education system.

The intermediation between schooling and the economic benefits which accrue from it is the examination. In some countries the examination has complete dominance over all aspects of the system. Educational considerations are of little relevance. It is widely accepted among teachers and pupils that if an item is not examined there is little point in teaching it. As Ron Dore wrote in 1980,

Schools belonged unmistakably to the modern sector. Timetables were adjusted to the city week, not to the rhythms of the agrarian calendar. Teachers were hired and paid in accordance with the canons of modern bureaucracy, not indigenous custom. Furniture, layout, and curriculum design explicitly mimicked imported models. Schooling was socialisation into modern sector life - in anticipation of a future which only a small minority could enjoy. Hence, an overwhelming importance was attached to examinations, which determined who should belong to that minority.⁶²

[⁶² Dore R, The Future of Formal Education in Developing countries, in Simmons J. (ed) The Education Dilemma: Policy Issues for Developing Countries in the 1980s, Pergamon, 1980, pp 72-73.]

Dore reiterates the difference between the emergence of education systems in industrialised countries and those in developing countries: 'where they lack certain pre-industrial traditions it is harder to sustain the guise that education is aimed at personal development and spiritual enrichment rather than money-earning opportunities, a fiction which in older societies derives some strength from traditional ideals and mitigates the effect of the qualifications disease'.

The extent to which parents and children perceive the relative advantages of schooling in economic or non-economic terms depends on tradition and culture as well as the robustness of employment opportunities, and will vary between and within countries. There is frequently a belief that school can build moral responsibility and discipline, partly in reaction to the breakdown of traditional relationships in communities. In that cost sharing policies rest on assumptions of economic benefits, their success or failure will be sensitive to the attitudes of communities to the benefits of school, which, one would suppose, is related to success of children in examination systems. However, lack of information or the ability to interpret information, or indeed the lack of any choice, parents may be unaware of the poor quality of schooling their children receive as measured by examination results.

Although there are many aspects of the cultural environment which impinge on education policy, perhaps most critically language issues, another aspect which is relevant to cost sharing is the culture of management in the public sectors. The general model of cost recovery supposes among other things a perfect flexibility in public finance allocation. Such a flexibility cannot easily be achieved by any large organisation, public or private, and not at all in hierarchical and unaccountable public sector bureaucracies staffed by underpaid staff who live often in fear of their superiors and where there is no reward for innovation and much reward for caution and inactivity. Many of the reforms proposed for the management of education systems which emanate from foreign aid agencies seem to derive from idealised rationalist perspectives, assuming an underlying willingness of the recipients of these attentions to change. Yet in many countries the simplest of analyses reveals the irrationality of change, and preference for the 'inefficient' status quo, when viewed from the local perspective. Efficiency brings unemployment, and there is in most developing countries a shared goal of employment with widespread unwillingness to be responsible for or collude in other people's unemployment. Also, larger work loads may interfere with the necessary business of making a living from second or third jobs.

Much of the rhetoric about communities and 'ownership' is based on an assumption that measures can be taken to thrust on teachers accountability to the communities in which they serve. Apart from the fact that teachers will be accountable to those who pay their salaries (if they are accountable at all), the foregoing paragraphs suggest that the widest perception of the product which teachers deliver is employment (or movement up the education system) through certification. As Serpell argues well in the case of Zambia, the impediments to teachers perceiving their responsibilities to their communities as opposed to the wider educational bureaucracy are centred on the nature of the product, rather than on what should be the process, for teaching to examinations may not involve much which can be construed as 'learning'.⁶³

[⁶³ Serpell, op Cit. pp 132 - 135]

All this is to emphasise the scope of the 'implementation package' in which cost sharing policies should be included. To implement reallocation and efficiency in public expenditures requires action along an extended chain ranging through different government agencies, including schools. It requires individuals to make arguments and claim more accountability than is strictly necessary, as well as requiring them to relocate accountability in ways which the system in which they operate cannot easily support.

E. Foreign Aid

It is worth also considering the culture of foreign aid agencies, which can be generally characterised as bureaucratic and hierarchical, with inward looking accountability, as evidenced in the Wappenhans report for the World Bank.⁶⁴ The role of foreign aid agencies in the context of the present discussion can hardly be exaggerated: cost recovery was a central part of economic adjustment strategies on which aid was conditional, and there were few dissenters within the agencies. Moreover, most aid policy was directed towards expansion and increasing costs, and it disrupted the natural evolution of education systems within fiscal feasibility.⁶⁵ It had a direct impact through this mechanism on cost recovery imperatives.

[⁶⁴ The 'Wappenhans Report' was prepared by a former Vice-President of the World Bank in 1992 but never published. It attributed much of the high failure rate of World Bank projects to a 'culture of commitment' in which staff were appraised more on the basis of the projects for which they gained approval and committed funds rather than by any success of the projects themselves. This seems to be common to most agencies.

⁶⁵ cf '... aid represents the import of resources not generated within the receiving economy. This may mean that the skills which would have been generating those resources were never in fact called for or learned, and are not available to use the resource when it is provided. In a more complex case, the educational system of a country has to be paid for from its productive resources ..., but if education is rapidly expanded by the use of unearned aid resources (without corresponding growth in the economy) there may not be employment for its graduates....' Bauer op cit, pp 105-106, quoting Guy Hunter.]

Another way foreign aid affected cost was via the high level of debt incurred by countries.⁶⁶ In many countries, up to one third of recurrent budgets is allocated to paying debt costs (domestic and foreign).⁶⁷ Priority is given to paying debt, and the large spenders in the budget such as education are constrained. The point to note is that debt is (or should be) a temporary phenomenon. Yet foreign aid agencies in their drive to impose fiscal rectitude on countries take debt costs as given and ask for substantial and in some cases irreversible changes in education policy in order to reduce expenditures. Were debt not to be such a burden these changes would not be necessary. Foreign aid has therefore promoted contradictory aims of expansion and contraction, which is hardly a sound base for policy.

[⁶⁶ How far should countries borrow to finance their social sectors? The argument is that without education economies will not grow, and that growth provides the fiscal space to repay loans. That has not transpired in many countries with have borrowed heavily for education. Moreover, the benefits of education loans may be doubted: in Tanzania and Ghana it is hard to see the returns to past loans for education.

⁶⁷ For example, external debt costs account for over a third of the Ghanaian development budget for education. Nearly all this debt is to the World Bank. Internal debt cannot be allocated to sectors, though of course it finances them. There is a connection between aid to education and internal debt through the cost push effects of aid (e.g. through the expansion of systems beyond fiscal feasibility) which forces governments to borrow for deficit finance. World Bank officials commonly urge that IDA aid is almost grant aid, a disingenuous argument as the chickens of the 1970s and 80s come home to roost in the 90s.]

To a large extent foreign aid upsets a natural process of evolution in education policy, as countries are subject to swings of education fashion from outside: from the emphasis on university education; to vocational education and 'diversified' secondary curricula; to the current emphasis on 'basic' education. Depending on the capacity of the recipient country, staff from aid agencies have great influence over domestic policy, and in many cases even draft it.⁶⁸

[⁶⁸ As in the case of Tanzania where social sector policy was written and published by World Bank staff.]

Foreign aid projects have played a significant role in expanding education systems beyond a size which can be supported by domestic tax finance. Perhaps one of the most destructive initiatives has been Universal Primary Education (UPE) and the huge costs associated with it. Many countries expanded their systems too rapidly to maintain any quality, creating a large pool of untrained teachers, a stock of poorly built schools and large equipment and materials deficits. They also created raised expectations of employment which, not materialising, have prompted withdrawals from schools rather than increased enrolments, the opposite of what was intended. A current example is the education reform and consequent free compulsory universal basic education policy in Ghana which extended the cycle from six to nine years in a fiscally stressed country.

Paradoxically, in many countries foreign aid agencies, while pressing for basic education policies, allocate the greater part of their donations to post-basic uses: this is certainly the case in Tanzania. However, this should not matter simply because of the fungibility of money. Indeed, there is a reasonable case for foreign aid to be directed at higher levels of education on the grounds that such interventions are much more likely to assist the poor than sub-sectoral interventions at the basic level.

There are two reasons which support such a view. First, post basic education has heavy infrastructural costs which, if they come from the same budget as primary education, will effect primary education spending: this is arguably one reason for the aid fashion of requiring communities to build their own primary schools, while post basic institutions receive capital budgets from government. Secondly, the ratio of private to public expenditures at the primary level are higher than the ratios at post-primary levels simply because costs are higher. The marginal dollar has a greater impact at the primary level than at higher levels. It would follow from this that policies should aim to create conditions and incentives to maximise parental contributions to primary schools rather than to reduce them. This would necessarily involve reducing the private costs of secondary education. Indeed, the argument can be extended: foreign aid should finance heavy infrastructure (roads, water and sewerage) within recurrent maintenance and operational cost constraints, and work through fungibility to push more resources from domestic revenues to social expenditures. The foreign aid fallacy is the assumption that foreign aid finances only achieve what they are directed targeted to achieve, which ignores fungibility.⁶⁹

[⁶⁹ Contrary to the belief of many commentators, governments' natural priorities tend towards higher education rather than basic education: hence the multiplicity of foreign aid conditionalities. If this were not so, why would conditionalities be required? Kenya provides an interesting example. Most aid to education was stopped for several years for political reasons. During the period of no aid government allocations to education investment showed entirely different priorities. Details are given for one section of the education system in the 1994 Kenya Public Expenditure Review, Chap 8.

Similarly, if priorities in the Ghanaian education development budget are ranked with and without foreign aid, higher education receives the highest ranking when only domestically financed investment is included, whereas basic education takes first rank when aid is included. Similarly, over the whole government development budget education expenditures rank fourth at 3.5 per cent of the total when only domestic finances are considered, but rises to the second highest priority over all expenditures at 11 per cent when foreign aid is added. The situation is complicated by 'matching funds', which are domestic revenues required to match foreign aid allocations: see Penrose P. Budgeting and Expenditures in the Education Sector in Ghana, Ministry of Education, April 1996, Chap 5. (There is a deplorable tendency for agencies to demand fixed coefficients for budgetary allocations to education (e.g. the absurd '20/20' formula), something which, apart from all the measurement ambiguities involved, would never be countenanced by public finance managers in any country). In Ghana there has been intense pressure over recent years on the higher education budget with corresponding pressure to raise its share of the total. One argument commonly used by those pushing for reallocation is that basic education benefits from foreign aid.]

The implementation of cost recovery policies is frequently a part of wider conditionalities relating to fiscal balance and a part of institutional development programmes the design of which contains serious flaws of pacing and sequencing.⁷⁰ Throughout the adjustment period the combined reduction of expenditures in real terms⁷¹ and increased costs to families invariably featured highly in Bretton Woods Institution (BWI) conditionality lists. These conditionalities were more attached to ideology than evidence.⁷² The history of foreign aid over the decade 1985-1995 (and perhaps beyond) will be shown to have had an immensely destructive effect on education systems in many countries with hasty 'reforms' imposed without regard to practicality or educational effect.⁷³

[⁷⁰ One such example was cost recovery in Zambia. Bretton Woods conditionalities, supported by donors, imposed cost recovery policies for health and education during a period of drought, and before significant public finance reforms were implemented.

⁷¹ David Sahn's study was more concerned with social sector expenditures as a percentage of national income, rather than in absolute real terms. As a percentage of national income education expenditure was quite robust in many countries, while at the same time falling in real terms. Both measurements are valid as partial measurements of real movements in fiscal effort, but it is a little disingenuous to argue, as is often the case, that real expenditures are maintained on the basis of constant or rising GNP/GDP shares. See Sahn D. E, Public Expenditures in Sub-Saharan Africa During a Period of Economic Reform, World Development, Vol. 20, nr 5, pp 673 - 693.

⁷² See Colclough C. & J. Manor, States or Markets? Neo-liberalism and the Development Policy Debate, Clarendon Press, 1991, pp 197-213

⁷³ This perhaps somewhat strong statement will be countered by 'what choice was there?', and while in relation to certain structural reforms it is probably the case that little choice existed, in the case of education there were many choices, some related to wider reform programmes and some simply to paying more attention to evidence and less to dogma. In both cases reforms paid little heed to initiating medium term measures to tackle the underlying cost structures of education, instead preferring to concentrate on superficial measures to reduce budgets. The failure to understand the essential requirements of civil service reform was another example. One documented example is in Carnoy M. and C. Torres, Educational Change and Structural Adjustment: A Case Study of Costa Rica, UNESCO, September 1992. The authors argue that Costa Rican secondary education never recovered from the effect of 'reform' policies initiated in the 1980s. It is easy to destroy and hard to rebuild.]

F. Conclusions

This chapter has set out issues for inclusion in the analytical framework for the case-studies of cost sharing. The common economic justifications for cost sharing are contradictory and inconsistent, and their rationalist assumptions do not reflect reality, either of how people behave or how public finance systems react. The politics and economics of cost sharing in developing countries are rather founded in fiscal crisis⁷⁴ and in the ideology of 'low' public spending and budget deficits. They are also heavily influenced by foreign aid policies, which are themselves unstable over time and frequently inconsistent between donors and lenders, and which impose additional fiscal burdens on countries. As a result, there has been a general failure to understand the content, pacing and sequencing of financial reforms which improve the resource flows to education.

[⁷⁴ A point generally accepted in Bray M. and K. Lillis (eds), Community Financing of Education: Issues and Policy Implications in Less Developed Countries, Pergamon Press, 1988. The book provides interesting case studies and is written, as the title suggests, with minimal analysis of public finance aspects of community financing. Keith Hinchliffe, in a cogent review article Neo-liberal Prescriptions for Education Finance: Unfortunately Necessary or Inherently Desirable? in International Journal of Educational Development, 1993, Vol. 13, Nr 2, pp 183-187 asks the same question in the title, although he is cautious in his approach to the answer.]

The subsequent chapters consider the evidence from selected countries and attempt to answer the questions posed at the beginning of this paper.