THE WORLD BANK Discussion Paper EDUCATION AND TRAINING SERIES Report No. EDT72 Federal Finance, Fiscal Imbalance and Educational Inequality Keith Hinchliffe April 1987 Education and Training Department Operations Policy Staff The views presented here are those of the author(s), and they should not be interpreted as reflecting those of the World Bank Discussion Paper Education and Training Series Report No. EDT72 FEDERAL FINANCE, FISCAL IMBALANCE AND EDUCATIONAL INEQUALITY Keith Hinchliffe (consultant) Education Policy Division Education and Training Department April 1987 The World Bank does not accept responsibility for the views expressed herein, which are those of the author(s) and should not be attributed to the World Bank or its affiliated organizations. The findings, interpretations, and conclusions are the results of research or analysis supported by the Bank; they do not necessarily represent official policy of the Bank. Copyright ( 1987 The International Bank for Reconstruction and Development/ The World Bank ABSTRACT This paper discusses the practice and problems associated with the financing of education in federal political systems as a way of highlighting issues involved in the diversification of the sources of educational financing. The major imbalances between the finances and responsibilities of a federal government and the combined units of a lower level of government, and of the individual units of a lower level of government are concentrated on in general and with regard to education in particular. The experiences of six federal countries in dealing with these imbalances are then described and the lessons learned discussed. TABLE OF CONTENTS Page No. SUMMARY i-iii Section I - INTRODUCTION 1 Section II - EFFICIENCY AND EQUITY IN FEDERATIONS 3 Section III - IMBALANCES IN FEDERAL FINANCING 5 Vertical imbalance 5 United States 7 Canada 7 West Germany 8 Australia 8 Brazil 9 India 11 Nigeria 12 Horizontal imbalance 14 United States 16 Australia 17 Canada 17 West Germany 18 Brazil 18 India 19 Nigeria 21 SECTION IV - FEDERAL FINANCING OF EDUCATION: THE THEORY 23 Vertical imbalance and education 24 Fiscal performance equalisation and education 25 Fiscal capacity equalisation and education 27 SECTION V - FEDERAL FINANCING OF EDUCATION: THE PRACTICE 28 United States 28 Canada 32 Australia 35 Brazil 39 India 43 Nigeria 47 SECTION VI - OVERVIEW AND LESSONS LEARNED 51 General issues of federal financing 51 Vertical imbalance 51 Horizontal imbalance 52 Education finance 52 Conclusion 54 REFERENCES 56 ACKNOWLEDGMENTS Support for this paper was provided by the British Council and the Economic Development Institute of the World Bank. i SUMMARY This paper is concerned with the financing of education in federal political systems. A focus on federal countries is of relevance for at least two reasons. First, several of the most highly populated developing countries are politically organised along these lines and there is a growing concern within them over issues arising from the system of financing. Second, there is a growing interest, generally, in the possibilities and implications of diversifying the sources of educational financing to include individuals, communities and lower levels of government. The experiences of federal countries are particularly relevant for discussions regarding the decentralisation of authority and financial responsibility to this last group. Sets of financial, political and educational issues arising from the methods of financing education in federal countries are described in section I and form the central focus of much of the rest of the paper. Initially, however, in sections II and III the discussion is of the general fiscal issues arising from federalism. Arguments supportive of federal arrangements are largely made on grounds of efficiency and diversity. It is common, however, either initially or over time for two sets of fiscal imbalances to emerge resulting in discrepancies between governments' responsibilities and financial resources thereby reducing efficiency and increasing inequalities. Two types of imbalance potentially exist - vertical imbalance which describes a situation in which: there is a mis-match of responsibilities and resources between levels of government (e.g. between the federal government and all state governments)- and horizontal imbalance which relates to a mis-match between units of a similar level of government (e.g. between state governments). Much effort in all federal countries is focused on devising the most appropriate ways to reduce these imbalances. The countries chosen to illustrate the experiences and changes in federal, fiscal relations are the United States, Canada, Australia, Brazil, Nigeria and India and in section III of the paper the variety of measures which have been taken to alter the automatic or discretionary funding arrangements so as to reduce the imbalances are described. The measures include changes in the constitution to alter both the propor- ii tions in which tax revenues are divided between the federal and state governments and the formula determining the allocation between states, specific purpose grants tied to individual items of expenditure or to educational levels as a whole, shared cost programs and direct compensatory payments between states. An important distinction made concerning the measures to reduce horizontal imbalance is between those which enable all states to provide equal services and those which directly provide the services. In section IV, the paper begins to focus specifically on the financing of education. Again discussion is centered on vertical and horizontal imbalances and the ways, in principle, that federal governments can reduce states' financial burdens by either directly taking over some of their educational responsibilities or increasing transfers which are tied to varying degrees. The underlying principles of federalism are used to discuss the appropriate role for each type of government in the various education levels. In general, the federal government's role is regarded as being limited to instances in which state governments are either unwilling or unable to finance types of education whose effects go beyond the state. Again, measures to increase equality in educational provision between states tend to divide into two groups - those which directly equalise provision and those which increase the capacity to equalise provision. Section V is the central section of the paper and describes the current patterns of the federal financing of education in the six federal countries and trends in these patterns. In the older federations of the United States, Canada and Australia, and also in India, the constitutions give very little responsibility for- education to the federal government. Under ten percent of total educational finance is spent directly by this level of government. In each of these countries, apart from the United States, however, additional transfers which indirectly fund education are also made. This is particularly the case in Australia, where direct and indirect federal funding totals 40 percent of the total and in Canada where it is 22 percent. The level of education most dependent on federal financing in these countries is higher education and this is again the case in Nigeria and Brazil where the federal governments have wider constitutional responsibilities for education. iii Measures aimed at directly equalising educational provision between states (as opposed to those aimed at specific population groups) are few. More in keeping with the principles of federalism, efforts have been made to equalise the provision of states' services as a whole via general revenue grants rather than through sector specific grants. In several countries these efforts remain largely unsuccessful. Local government involvement and responsibility in education varies widely among the federal countries surveyed. Whereas there is a substantial devolution to local school boards in the United States, schooling is highly centralised, administratively and financially, in Australia. The situation in Canada comes between that in the other two countries. In Brazil a quarter of all primary school funding is provided by municipal governments. This is a much higher proportion than in either Nigeria (where local governments have administrative responsi- bility) or India. Federalism is a form of government which essentially guarantees the rights and responsibilities of each level of government. These rights and responsibilities are usually carefully guarded. This paper mainly describes the pattern of educational funding in federal countries and the mechanisms available, and used, to alter the pattern within the limits set. Because of the division of taxing powers, federal governments' revenues normally grow faster in relation to the growth in their constitutional responsibilities than is the case for the state governments. This provides an environment for conflict. As more countries attempt to diversify the sources of educational finance, with the decentralisation of authority which that implies, more analyses aimed at devising ways of controling this conflict, while simultaneously encouraging the growth of an effective and egalitarian educational system, will be required. 1 FEDERAL FINANCE, FISCAL IMBALANCE AND EDUCATIONAL INEQUALITY I INTRODUCTION All central governments face similar questions of how to extend educational provision and increase its quality. Currently, budgetary constraints are formidable in many countries. One option increasingly being advocated is to diversify the sources of educational finance by encouraging parents and pupils, local groups and lower levels of government to contribute, and to decentralise authority and control. It is with the diversification of funding and decentralisation of authority to lower levels of government that this study is concerned. Calls for the diversification of funding and decentralisation of control in education made with the simple expectation that total finances will increase and that the operation of such a system will be non-problematic now need to be augmented by serious consideration of the potential financial and political implications. These, in turn, may have fundamental educational consequences. Among the financial issues which need to be considered are: - does the degree of financial diversification affect the level of total educational expenditures, - in a diversified system, how can overall public spending on the different levels and types of education be centrally influenced, - how can increased spending by one source ensure that this does not simply result in decreased spending by another source, - if existing arrangements for diversification impede optimal expenditure levels, on what criteria and terms can resources be most efficiently transferred? Among the political issues are: - to what extent does decisionmaking power necessarily correspond to financial contribution, - how can a central government which provides a minority proportion of educational finances influence local government policies, - how can governments concerned with equal educational opportunities counter the (often) unequal implications of diversification and decentralisation? 2 Among the educational issues are: - are particular levels of education more appropriate for central government funding than others, - what aspects of education policy most lend themselves to central government control? Diversification of the funding of education has been implicitly linked above to some form of administrative decentralisation. Since the most far-reaching form of decentralisation is federalism, the experiences and responses of nations organised along federal lines may provide useful insight into the questions raised. Analysis of the fiscal arrangements of federal nations is also justified in itself since several of the world's most highly populated countries have a federal, political structure. Some, such as the United States, have existed as federations for centuries, while others , such as India and Nigeria, have operated in this way for only a few decades. In several of the more recent federations the issues of equality, control and the distribution of total resources between different levels and areas of education are high on the educational and political agendas. The remainder of the paper is organised in the following way. In section II, the arguments in favor of federalism are described together with the potential problems or drawbacks. Two of these problems are the existence or development of sets of imbalances regarding revenues and responsibilities of particular levels and individual units of government. Vertical imbalances are those which exist between the central government and all the units of a lower level of government combined. They imply that one or other level of government does not have sufficient resources to carry out its constitutional obligations at an acceptable level. Horizontal imbalances exist when the individual units of a particular level of government each have differing levels of resources relative to their responsibilities. In section III, these imbalances and the general approaches available to reduce them are described. The experiences of fiscal imbalance and the policies directed towards them in the United States, Canada, Australia, Nigeria, Brazil and India are then presented. Following the discussion of federal financing in general, sections IV and V focus solely on education financing. In section IV, the principles behind the different approaches to reducing imbalances are 3 discussed. With regard to reducing the inequalities between the units of a similar level of government, differentiation is made between mechanisms which ensure that the provision of educational services is equal and those which ensure that the capacity to provide similar services is equal. Section V is the data center-piece of the paper. For the six federal countries listed above, the experiences of financing the education system are discussed focusing, where possible, on the time trends in the roles played by the different levels of government and the financial mechanisms to support those roles. A summary of the most important features of these experiences and their implications both for federal countries and for those either operating or considering systems of diversified educational finance is contained in section VI, together with suggestions for further work. II EFFICIENCY AND EQUITY IN FEDERATIONS A federal nation is one which has a constitution covering at least two distinct layers of government and operates at two levels: as a nation and as a collection of related but self standing units. The aim of such arrangements is to either create or preserve a nation while simultaneously maintaining the identity, culture and tradition of the separate units (Hicks, 1978). Essential features of federalism are that no level of government has absolute sovereignty and that the rights of each are defined in the constitution. Procedures generally exist whereby changes to the constitution can be made but these invariably require a wide degree of support. Since federal states and decentralised (or deconcentrated, see Rondinelli et al. 1984, Winkler 1986) unitary states are similar in a number of ways there is a tendency to equate them precisely. This overlooks some very important differences however. In a federation, the existence and rights of the lower levels of government are guaranteed while in all forms of unitary state, lower level governments have no rights independent of the central government and can be abolished at any time. Another difference, and one particularly related to the provision of education services is that within a federal state, local governments, school boards and so on come within the ultimate control of the state 4 governments and central government has no constitutional relationship with them. Any attempt by the center to deal directly with these local bodies can legitimately be regarded by the states as encroaching on their jurisdiction. In a recent paper, Winkler (1986) argues that increased efficiency is the principal argument for federalism. This argument is largely based on the neo-classical premise that greater social welfare gains will accrue the more closely matched are citizen preferences with respect to combinations of taxes and services. This then assumes that the level of authority distributing the services is the one setting the tax rates which fund them. To the extent that transfers from higher level governments fund those services, the efficiency gains claimed are diminished. Increased efficiency of a different sort is also claimed for federalism and, more generally, for decentralisation. This takes the form of increased efficiency in the production of services which may arise as a consequence of the greater likelihood of local administrators adjusting inputs according to their local price differences and from a general reduction in administrative cost. The third argument for potential gains is that the type and quality of service provided is likely to be more appropriate if it is under local control than if it is determined uniformly at the center. Further, the greater degree of experimentation and innovation following on from local control may result in more generally effective practices which may be taken up elsewhere and lead to efficiency gains in the system as a whole. The economic argument for federalism, therefore, largely rests on potential gains in efficiency. There are also possibilities of efficiency losses, however, from three sources. First, there may be a lack of administrative skills locally affecting both the level and quality of services and the collection of revenue. More important there are possibilities of 'interjurisdictional spillovers'. In other words, the benefits of a given state's expenditure may, through migration for instance, not be gained solely within that state leading to a less than optimal provision of the service. Education is a good example, particularly at the tertiary level. A further potential efficiency loss (or problem) arising from federalism can be added. A major aspect of the 5 federal ideal is that the various levels of government have control over resources commensurate with their responsibilities. This does riot necessarily occur automatically. The first (ancient) federations required the parts to contribute to the center in order to maintain the latter's functions. In today's federations, to varying degrees, revenue collection has been centralised and the states depend on various mixtures of revenue-sharing and transfers from the central government. To the extent that a mis-match of resources and responsibilities remains, efficiency is curtailed. Potential inequities are of two types. First, differences in tax bases between states give rise to inequities in that individuals in similar circumstances across states are treated differently in terms of the relationship between benefits and taxes. The second potential inequity arising from federalism, and devolution, is the one most commonly raised - the regional variation in the level of services provided resulting from either wide variations in state revenues or substantial differences in the cost of provision. Two potential forms of imbalance between resources and responsibilities in a federation have been highlighted above - between the center and all the states combined (vertical imbalance) and between the states themselves (horizontal imbalance). The effect of these on educational provision and the mechanisms available to reduce them are the concern of this paper. :111 IMBALANCES IN FEDERAL FINANCING Vertical imbalance A vertical fiscal imbalance between the central government and all lower level governments combined implies that one or other level of government cannot command sufficient resources to carry out its constitutional obligations at an acceptable level. In an economic sense, such a situation is inefficient and welfare would be improved by fiscal adjustment. In general terms, a fiscal imbalance of this sort can be remedied through a variety of mechanisms many of which will be described in greater detail later. These include, 6 apart from different forms of grant, changes in the share of the tax base, changes in the shares of revenues collected centrally and re-allocated, a bypassing of constitutional taxation provisions through the use of public utility charges as taxes and the simple transfer of administrative responsibilities for services of poorer areas within states. The almost inevitable tendency for central governments to possess greater revenues than they require, and vice versa for the states, olus the trend towards an intensification of this situation have been commented on widely. For illustration, the arguments put forward by Hicks (1978) are presented below. Vertical fiscal imbalance, she argues, is almost inevitable due to two forces: on the one hand, central governments' foreign relations and defence duties which encourage them to control customs duties plus the widely perceived desirability of treating income and profits similarly across a country lead to the most productive and rapidly expanding sources of revenue being centralised. On the other hand there is the tendency for the responsibilities of lower level governments for items such as education, public health and housing to be the most rapidly expanding items of expenditure. These two sets of forces result in a built-in imbalance and one likely to increase over time. In many federal constitutions, the imbalance resulting from these pressures has been partly allowed for through the systematic sharing of tax receipts or through responsive institutional arrangements. Over the last four decades, however, the imbalances have increased beyond initial expectations. This has occurred, first, through the general adoption by central government of greater economic responsibility partly via the regulation of public expenditure and tax rates and second as a result of the world economic crises of the 1970s which resulted in the growth of federal government budget deficits controlled partly by increased pressure on all lower level governments. Since the early 1970s the federal governments of Canada, the United States, Australia, West Germany and Switzerland have all introduced legislation to readjust federal-state fiscal relations concentrating mainly on the vertical imbalance. All, according to Hicks, have followed similar policies of: - increasing revenue by raising tax rates or introducing new ones 7 - halting or reversing centralisation by transferring (sometimes back) wider responsibilities and some finance to the states - paying more attention to discriminatory grants and investment - creating more direct relations with local authorities (especially those running the cities). The following paragraphs describe some of the specific measures taken to alter the vertical federal-state fiscal relationship in recent years in these countries. United States - Of all federations the United States is perhaps closest to the ideal of each level of government combining taxing and spending decisions. State revenues depend mainly on taxes levied and collected within the state (83 percent of total state revenues in 1980 came from state income and sales taxes alone). Transfers to localities within states are much more important. However, with the adoption of so-called revenue sharing in 1972, the federal government assumed greater liability for poverty, educational deficiency, crime, pollution, etc., resulting in vertical transfers equal to one sixth of the federal budget in the mid 1970s. Since 1978 these have declined in relative importance. Canada - Despite the differences in size and wealth between states, each is constitutionally powerful vis the federal government. Vertical transfers have been based on tax revenue sharing, shared costs of services and 'contracting out' whereby provinces can contract out of a shared cost program and claim a compensating reduction in federal income tax payments. Legislation in 1977 was designed to reduce the open-endedness of transfers in general and to relate their expansion to only the growth of the economy rather than to provincial expenditure. The changes have been presented as a step towards greater decentralisation and provincial autonomy. 8 West Germany - The main characteristic of West German federalism is its strong local governments which account for two thirds of all public investment. Almost half the states' revenues are transferred to them. Tax sharing is the central feature of fiscal arrangements. Originally the federal government had no rights over income tax revenue but since 1970, the federal-state distribution has been 50:50. Eighty percent of all tax revenues are made a part of vertical redistribution. Another major feature of German fiscal relations, and one which is increasing, is the system of 'joint tasks' whereby different levels of government systematically plan and finance activities together. Australia - In contrast to West Germany, local government in Australia is weak. Reliance for vertical redistribution from federal to state governments has traditionally been on general revenue grants specified by an independent commission. Sixty percent of state revenues result from transfers. Increasingly these grants have been added to by specific purpose grants and there have been moves towards increased tax sharing. After discussing the recent performance of these countries Hicks (1978) argues that they all "... seem to be in some trouble over federal fiscal relations, even to the point of sponsoring new legislation or perhaps constitutional amendments. In fact, the basic problems of fiscal imbalance in a federation have never been solved satisfactorily" (p194). Federations such as the United States, Canada, Australia and Switzerland adopted constitutions at times when their provinces, states and cantons were powerful and when the role of central government was generally perceived to be a slight one. They became federations by aggregation. All federations in the developing world have resulted from disaggregation which has, naturally, tended to result in a relatively 9 strong central government. In the major federations of South America, Africa and Asia this situation is particularly marked in relation to the levying and collection of tax revenues. The result is everywhere a sub- stantial vertical fiscal imbalance. To ensure both adequate levels of funding commensurate with responsibilities and the states' automatic rights to these funds to preserve their legal autonomy, revenue sharing proposals and independent commission reviews have been written into constitutions. Because circumstances relating to both tax revenues and public expenditure needs are not constant, precise formulas to create vertical fiscal balance have not been achieved and arrangements for the distribution of national revenues between all levels of government have been the subject of continual change. Below the experiences of Brazil, India and Nigeria are briefly described. Brazil Brazil has adopted many forms of federalism since 1891 but the present federal fiscal structures date from reforms of 1965-67. The main effects of these were to redefine the tax bases available to each level of government, concentrate control over tax bases and rates at the federal level and vastly expand the importance of federal transfers in the revenues of the state and municipal governments. It has commonly been asserted that these changes have led to a deterioration in state and local finances resulting in additional needs for grants and loans so increasing the fiscal dependence of the states and local governments on the federal government. Mahar and Dillinger (1983) in the most recent treatment of this issue conclude that the dependence on federal funds for investment purposes has led to a greater reflection of national priorities and a loss of control by the states over even their recurrent expenditures. The dominant fiscal position of the federal government can be seen from table 1 which indicates that over 70 percent of public expenditures are made by this level of government. 10 Table 1 Revenue and Expenditure of the Public Sector, Brazil, 1975 (Cr.$ million) Level of Government Federal State Municipal Total Revenue 202,700 70,735 9,767 . (71.6) (25.0) (3.4) Total Expenditure 190,155 81,388 12,353 7 (67.0) (28.7) (4.3) Source: Mahar and Dillinger, 1983, Table 1 Note: Intergovernmental transfers are deducted to avoid double counting. Because of the way transfers are ignored in table 1, the relative importance of the federal government in particular is underplayed. This can be seen from the levels of transfers in table 2. l1 Table 2 State and Municipal Revenues by Source, Brazil, 1978 (percent) States % Municipios Z Exclusive sources 67.8 Exclusive sources 48.5 (Value added tax) (59.2) Transfers 39.7 Transfers 19.6 Federal (9.8) Credit operations 12.6 State (29.9) Credit operations 11.8 100.0 100.0 Source: Mahar and Dillinger, 1983, p9 Transfers are made on the basis of shares of particular taxes levied and collected by the federal government. The most important of these are income tax (12.5 percent to states, 10.5 percent to municipios), manufacturers sales tax (12.5 percent to states, 10.5 percent to municipios) fuel tax (30.7 percent to states, 8 percent to municipios) car tax (45 percent to states) and education salary tax (66.7 percent to states). Up to 1979, earmarking of these funds was common but since then the practice has been much reduced. India The Indian constitution again recognises that the states are not able to meet adequately the functions assigned to them in the absence of access to additional funds. The spirit of the constitution also dictates that those additional funds (in the form of tax sharing and grants in aid) should take place in a manner free from the central government's discretion and be as automatic as possible. In a recent review of 12 center-state resource transfers George and Gulati (1985) suggest that neither the amounts nor the form of the transfers have been appropriate and argue that "... the stresses and strains experienced in federal relations have been largely the result of the difficulties faced by the states in discharging the responsibilities assigned to them on account of the shortage of resources" (p287). The principle of revenue sharing underpins the non-discretionary system of federal financing and the major source of this is specified proportions of particular taxes (mainly income and excise) plus grants-in-aid. The proportions of taxes (and their interstate allocations) are determined by a Finance Commission every five years. The second source of center-state transfers is the Planning Commission. These transfers are not a statutory right and are discretionary and/or conditional. As in Brazil, revenues and expenditures by the central government dominate in India and again like Brazil, the trend is increasing in this direction. Thus the states' share in aggregate budgetary receipts has declined from 38 percent in the First Plan period to 32 percent in the Sixth Plan period. Turning to transfers, these have remained fairly constant at around one third of total central government receipts and in the period 1979-84 almost 42 percent of states' expenditures were financed out of them (George and Galati, 1985, p290). Since the Second Plan transfers have financed around half of the states' Plan expenditure and a quarter of their non-Plan expenditures (Wallich, 1982, Tables 7.8). Transfers made within a clear framework and with no conditions clearly impinge less on the states' autonomy than do discretionary or conditional ones. Between 1951 and 1984, statutory transfers (via the Finance Commission) have accounted for two fifths of the aggregate budgetary transfers and three fifths have had highly discretionary components. An important part of the latter have come from neither the Finance nor Planning Commissions, but from Central ministries and often have had matching conditions attached. Nigeria Revenue sharing is a more dominant method of financing regional/state governments to perform their constitutional responsi- 13 bilities in Nigeria than in either Brazil or India where locally collected taxes are more important. Statutory grants were typically in excess of 60 percent of the total revenues of the regions from 1953-65 and since 1975 have constituted over 80 percent. The formulas used to derive the overall distributions between the federal and state governments as a whole and between the individual state governments have changed many times. Until the late 1970s these two distributions were decided simultaneously. To do this, receipts from individual taxes were wholly assigned to either the state of derivation or to the federal government or a share of the receipts was assigned to one or both of these and the rest to a 'distributable pool' to be allocated between states: e.g. import duties - on tobacco and fuel 100% to state of derivation - on alcohol 100% to federal government - on all other imports 65% to federal government - 35% to distributable pool mining rents - 50% to state of derivation - 35% to federal government - 15/ to distributable pool Over time, the derivation principle has declined. In 1978 major changes were mace to the principle upon which revenues were distributed vertically. A single set of allocations of total revenues was made and not just for federal and state governments but also for local governments. The allocations were fixed at 60, 30 and 10 percent respectively. In a further change in 1981, the percentage allocations were very slightly altered. The result of these moves has been to greatly increase state and local government revenues. The massive increase in oil revenues which accrued to the federal government through the 1970s resulted in excess funds for its constitutional obligations and a consequent growth of non-statutory grants to states. The largest of these was for the universal primary education program from 1977 to 1981. In 1977/78, for instance, total discretionary grants had grown from virtually zero a few years previously to half the value of statutory payments. Since 1981, these have again been sharply pared back. The basing of revenue sharing on virtually all taxes and the incorporation of the procedures for sharing in the constitution result in 14 a reassurance to the states that the terms cannot be changed suddenly at the discretion of the Federal Government. However, the substantial national reliance on petroleum taxes - 75 percent of the 1979-80 federal budget - remains a matter of serious concern, particularly over the last three years as revenue from this source has rapidly declined. Vertical fiscal imbalance can be seen, then, in two ways. The centralisation of taxation together with expanding state responsibilities has led to a mis-match of state own-revenue and expenditures in all federations apart from the United States. The initial response to this situation in many countries has been to establish mechanisms (incorporated in the constitution) which guarantee states part of centrally collected revenues as a right and unconditionally. In most cases this was expected to provide vertical fiscal balance. In the event, state expenditures have often been beyond those which could be financed by state taxation and statutory transfers. The result has been the growth of non-statutory grants, often discriminatory and conditional, including matching elements, and the shifting of administrative responsibility to the center. In each country the response differs as does the resulting relationship between the layers of government. The variety does not suggest that solutions are simple or that those currently adopted are particularly stable. Horizontal imbalance Mechanisms to increase the equality of educational provision across states within a federal system will be the subject of later detailed discussion. Education, however, is only one of the several goods and services commonly supplied by state governments and local authorities under their jurisdiction. In this sub section, the spatial inequalities in resource generation and the issues involved in the allocation of differential levels of transfer between states are discussed. For a number of countries, the experiences with specific allocative procedures and mechanisms will again be described. 15 The countries which have adopted a federal system of government are in general geographically large. Virtually as a consequence, in most there are wide variations in wealth both between states, and within states particularly between the cities and the rural areas. These variations are much wider in the federal developing countries than in the United States, Canada, Australia, Germany or Switzerland. Differences in the economic base of states leads to differences in the ability to raise tax revenues, particularly if similar rates are applied to people in like circumstances. Without redistribution or differing levels of federal transfers, the services provided to citizens in different states will vary considerably. As a result of these considerations, the fiscal issue in a federal system is not simply the division of revenue between the center and all state governments but also the division between the state governments. The major economic argument for federalism is that it produces gains in efficiency as a result of decentralised decisionmaking and diversity. In contrast, the objective of ensuring equality of opportunities and outcomes across states tends to emphasise uniformity. Any consideration of measures to increase spatial equality through differential transfers from the federal to the state governments (and from state to local governments) must take account of this potential conflict. A policy of fiscal equalisation aims to ensure that it is financially possible for citizens to have similar degrees of access to government-provided services irrespective of where they happen to live. In a unitary centralised system of government this is relatively easy to ensure. In a unitary decentralised system (e.g. United Kingdom) it is more difficult since local governments may themselves have different amounts of resources. In a federal system it is most difficult. First, both states and local governments will have different degrees of access to own-resources. Second, local governments, etc. have a constitutional relationship to the state not the central government. And, third, any federal-state transfers have to be seen in the context of the constitutional rights of the states.- 16 There are, in general, two approaches to horizontal fiscal equalisation which have been used in federations. The first, and the one most in line with federal principles, is the fiscal performance equalisation approach. This operates on the expenditure side of the budgets of governments receiving assistance and requires the granting government to specify standards of performance for a particular service. While these standards could be set so as to vary between states, they are more likely to be uniform and must be regarded as a means of centralising control or, at the least, heavily influencing lower level governments. The main tool of the fiscal performance equalisation approach is the special purpose grant. The second approach to fiscal equalisation is labelled fiscal capacity equalisation and operates on the revenue side of the receiving government's budget. The objective of this approach is to make it possible for governments to provide services at a standard level provided they impose taxes at a standard rate. These services may or may not be provided and the standard tax rate may or may not be levied. The decisions are left to the recipient government. This approach is obviously most in line with the federal principle and utilises either the general purpose grant or an appropriate unconditional distribution of tax sharing funds. Most of the federal governments considered in this paper have used a mixture of the two approaches in their attempts to provide what they see is an appropriate balance between the twin objectives of equality and decentralisation. In the descriptions which follow two questions are asked: On what basic criteria have the majority of transfers been allocated between states and in the event of these resulting in significant inequalities, what measures have been additionally taken? United States Horizontal equalisation between states is not given a high priority in the United States, apart from some specific programs. Inter-state inequalities, however, certainly exist. In 1980-81, state-local expenditures per capita ranged from $1284 in Arkansas to $2653 in Wyoming. Fiscal capacity between states is also widening due 17 first to developments in the energy-rich states and then to the recession which has affected states differently. While grants-in-aid from the federal government for specific purposes to the states increased during the 1960s and much of the 1970s, since 1978 they have declined in real terms. As will be described later, intra-state inequalities particularly with respect to education continue to be a major issue and area of substantial state-local transfers. Australia Disparities in economic activity between states are relatively small in Australia. However, as a result of differences in population patterns and the capacity of states to raise tax revenues, there are significant differences between states in their ability to provide comparable services with uniform tax rates. The response to this situation has been the development of a sophisticated system of general revenue equalisation grants to the four financially weaker states based on assessments of both revenue raising capacity and differences in the costs of providing services. Use has also been made of specific purpose payments based on need assessments by independent bodies. Over time, the Grants Commission's approach to states' financial needs has changed from an assessment of minimum financial need to one of full fiscal capacity equalisation defined as the assistance necessary to enable a state to provide services comparable with a standard state without having to impose higher taxes. This has involved, of course, very detailed analysis of state budgets. The assessments of needs, revenues and expenditures depend largely on informed, but subjective, analysis as compared to the use of statistical indicators adopted in some other countries. Specific purpose grants have grown in importance in Australia but these have affected the vertical rather than the horizontal imbalance. Canada Compared to Australia, there is significant economic diversity between the ten Canadian states and ways of providing additional 18 resources for poorer provinces have long been sought. A new Fiscal Equalisation Program was established in 1972 covering (a) general revenue-sharing (b) shared cost services and (c) special arrangements. With respect to (a) the seven provinces with lowest per capita revenues receive compensation based on revenues collected from a specific range of taxes. On the other hand, the expansion of discriminatory grants has been curtailed. This has been done by a levelling adjustment to bring those provinces receiving above average contributions down to the national average and those receiving less up to the average. The whole trend of fiscal arrangements has been to increase provinces' own-revenues and therefore provide greater autonomy but Hicks (1978) argues that the degree of equalisation or redistribution will be decreased. Equalisation procedures in Canada make no allowance for different revenue efforts by the provinces nor do they allow for any differential costs in the provision of services (as is done in Australia). West Germany The major objective of federal fiscal policy in West Germany is to unambiguously equalise standards of living across the federation. Almost half of the federal government's revenue from value added tax is distributed to the weaker states and there is further support for budgetary balancing. All 'financially weak' states have their revenues increased to 92 percent of state average per capita tax revenues (Traber, 1980, 273). In addition to tax sharing, joint tasks and special purpose grants for investment are also features of fiscal equalisation attempts. Brazil The problem of horizontal imbalance is particularly important in Brazil owing to the wide variation in revenue bases of the state governments. In the Southeast, gross regional product per capita is more than twice the national average while in the Northeast it is well under half (Mahar and Dillinger, 1983, p29). Since state revenues are mainly derived from a sales tax, this economic disparity flows through into revenues. The disparities are even wider among municipal governments (table 3). 19 Table 3 State and Municipal Revenues by Region, 1975, Brazil Revenues per capita,, (C/$) Southeast Northeast Frontier State government 1025 312 717 Municipal government Capitals 703 248 244 Interior 243 57 84 Source: Mahar and Dillinger, 1983, p31. As described in the previous section, revenue sharing is the major component of the Brazilian system of intergovernmental transfers. However, of the 12 taxes shared between the federal and state governments only 2 are distributed with any major reference to need. Mahar and Dillinger (1983) conclude that while the transfers favor Northeast provinces the net effect is only a modest reduction in regional revenue disparities - state revenues per capita in the Northeast rising from 37 percent of the national average to 52 percent. With respect to municipal governments, federal transfers result in virtually no net redistribution. The only very small redistributive effects of transfers are partly due to the emphasis placed on population in the allocation formula. India As previously described there are two major bodies which determine federal transfers in India: the Finance Commission and the Planning Commission. Distribution of their resources between states is a key issue in the political economy of the country and has been extensively studied (e.g. Bhattia 1979; Chelliah 1981; Rao 1981; George and Galati 1985). The Finance Commission awards tax sharing revenues and grants-in-aid every five years, to cover non-Plan or maintenance 20 expenditures. Major criteria for distribution have been state population, per capita income and state of tax derivation though the details have changed according to each Commission. Since 1969, transfers from the Planning Commission have been on the basis of the so-called Gadgil formula. Assistance is distributed: 60% on the basis of population 10% to states whose per capita income is below average 10. for above average per capita tax effort 10% for major irrigation and power projects, and 10% is discretionary. In addition to the Finance and Planning Commission transfers, there is also the growing category of discretionary transfers, mainly from federal ministries. It is clear from all the empirical analyses that the transfers have not resulted in a systematic redistribution in favor of low income states. It is also clear that the discretionary transfers have been particularly retrogressive. Table 4 illustrates these points. As Wallich (1982) points out, much of the apparent lack of consistency in the distribution of assistance across states is inherent in the transfer formulas. In the Finance Commission's awards, income and excise taxes comprise the bulk of the transfers and have largely been distributed according to population, which has no explicit bias, and state of collection which has a retrogressive bias. Excise taxes are distributed again according to population, plus per capita income, poverty and revenue equalisation criteria. Grants-in-aid have taken backwardness into account more explicitly. The Plan transfers formula has both progressive elements - per capita income and 'special problems' - and regressive - tax effort, and irrigation and power expenditures. Narabiar (1983) points out that despite all the adjustments to the formula to try to take state backwardness into consideration, the poor states of Bihar, Madhya Pradesh and Uttar Pradesh would have received more transfers if the criterion had simply been one of population. 21 Table 4 Aggregate Budgetary Transfers 1956-81, India State Index Numbers Statutory Plan Discretionary Total Punjab 78 101 159 109 Haryana 75 113 129 103 High Maharashtra 89 66 104 86 Income Gujarat 90 81 105 91 West Bengal 102 71 128 99 Tamil Nadu 86 80 72 80 Kerala 118 101 88 104 Middle Orissa 137 122 125 129 Income Assam 144 153 173 155 Karnataka 90 85 101 92 Andhra Pradesh 98 97 100 98 Uttar Pradesh 86 89 69 82 Low Rajasthan 107 103 193 130 Income Madhya Pradesh 83 99 65 83 Bihar 88 83 84 85 Source: George and Gulati, 1985, p292. Nigeria Revenue sharing is the major source of state (previously regional) revenue in Nigeria (over 80 percent) and over the last three decades the distribution of these revenues across states has been a matter of great political importance and has been subject to continually changing formula. In terms of horizontal fiscal balance, the key decisions have been the extent of the use of the derivation criterion as opposed to the distributable pool and, in turn, the criteria for the 22 allocation of the distributable pool. Up until 1959, the derivation criterion operated to the virtual exclusion of other criteria. With revenues from petroleum increasing at the end of this period, non-oil regions then succeeded in downgrading this criterion and in virtually all the changes which have been made to the procedures of fiscal relations since then (in 1966, 1967, 1975, 1978, 1981) the downgrading has continued. This, in turn has increased the potential for a greater equality in the allocations via the Distributable Pool. From 1969, 50 percent of the Pool was divided into equal parts for each state, while the other 50 percent was divided on the basis of population. This system remained in force until 1978 when a new one was adopted which included five criteria for revenue-sharing each with a different weight. The criteria were (a) equality of access to development opportunities (b) national minimum standards for national integration (c) absorptive capacity (d) independent revenue and minimum tax effort and (e) fiscal efficiency. More recently, in 1981, the criteria of population (40 percent weighting), minimum responsibility (equality) (40 percent), primary school enrollment (15 percent) and internal revenue collection (5 percent) have been adopted. The use of resources other than those from revenue-sharing to increase inter-state equality has been infrequent over the last thirty years, but in the late 1970s as oil revenues exploded they became significant. Mainly, they took the form of specific purpose grants for health, local government, new state administrations, payment of wage increases and, most important, for the universal primary education program between 1977 and 1981. As oil revenues fall, it is likely that any attempt to correct fiscal imbalances between states will again focus solely on the criteria for allocating unconditional block transfers based on revenue sharing. Education expenditures are invariably a high proportion of state and local government total expenditures and therefore discussions of the issues involved and the various mechanisms utilised to correct vertical and horizontal fiscal imbalances in general are highly relevant to the 23 education debate. From this point on, however, the paper focuses more narrowly on education and in the following section the measures available to increase both fiscal performance equality and fiscal capacity equality for educational services across states and localities are presented. This is followed in the succeeding section by a quantitative account of the education financing picture in some of the major federal states. IV FEDERAL FINANCING OF EDUCATION - THE THEORY Previous sections of this paper have discussed the issues and experiences involved in problems of vertical and horizontal fiscal balance in federal countries. In this section the discussion is somewhat more concretised by considering a single area of expenditure, education. Two issues will be concentrated on; the degree of appropriateness of federal government involvement in the provision of particular educational services and secondly the methods by which both vertical and horizontal fiscal imbalance can be reduced. As in the previous section, a distinction is made between those measures which directly affect performance and those which focus on capacity. Three problems of federal finance have earlier been identified (a) the existence of spillover benefits which may reduce expenditures below the optimum level (b) differences in tax bases which give rise to inequities in that individuals in similar circumstances across states are treated differently in terms of the relationship between benefits and taxes. (c) the variations in level of services provided resulting from either wide differences in state revenues or substantial differences in the cost of provision. With respect to all of these, in the area of education it is commonly argued that there is some role for the federal government. In Mathews' (1983) view "If decisions taken at the local level are likely to have significant educational or other effects outside the jurisdiction of the authority making the decisions, the ultimate responsibility for those decisions should.....be transferred to a higher level government." (p6) and again, 24 "...the federal government should have the ultimate responsibility for expenditure decisions which if made by states would have spillover effects beyond state boundaries; and for financing decisions which will permit individual states to be compensated for revenue raising (disparities) and cost disparities as between their school systems." (p6) Vertical imbalance and education With regard to vertical imbalances, the questions of which levels of education are most appropriate for additional, direct or indirect, federal funding and which fiscal mechanisms should be used, have to be answered. Among the factors to consider are - which government is in the best position to interpret the need for the service, - the degree of regional spillover from the service provided, - whether implementation is best carried out by federal or state agencies. (Mathews, 1983, p19) Most of these factors are more likely to suggest federal government involvement at the higher education level including teacher training colleges, technical colleges and universities. In some countries, higher education is a federal function, in some a state function and in others it is concurrent. The status alters the ways in which federal funds for higher education can be used. The factors listed above imply that, in general, federal government involvement at primary and secondary levels should be restricted to broad program strategies. This suggests that general purpose grants focused on objectives such as raising standards or provision of pre-school education are often more appropriate than specific purpose grants. The advantages of such general block grants are that inter-government conflicts are avoided, wasteful administrative overlaps or delays resulting from two bureaucracies are reduced, projects and policies can respond more effectively to local needs and there is no blurring of responsibility for performance, between the two levels of government. The main disadvantage of such grants, of course, is that they preclude federal government influence in areas where it might be appropriate. 25 The other area of 'legitimate' federal financing for state educational responsibilities as identified by Mathews is compensation for limited revenue raising capacity and higher than average costs. The previously described alternative equalisation approaches emphasising Erformance and capacity are discussed below with reference to education. Fiscal performance equalisation and education Any attempt at equalising education expenditures is likely to be based on uniform but arbitrary norms regarding school and class size, student-teacher ratios, teacher-ancilliary staff ratios, library and audio visual budgets and so on. Attempts will largely be partial, equalising to a minimum or an average but allowing some schools or school districts to operate at above standard. At its most basic, expenditure equalisation is essentially concerned with the provision of educational inputs rather than with the level and quality of educational services actually provided. Fiscal performance equalisation (FPE), however, implies more than this with the direct involvement of the giving authority in decisions relating to resource allocation and educational processes. Transfers for FPE may again be of two types - specific purpose grants and sector specific block grants. Specific purpose grants are very much an instrument of centralised control and are transferred to achieve designated resource standards. Payments are likely to be based on federal government assessments of expenditure needs in the context of any differential costs between states in providing services. In addition, conditions may be set in terms of the states' fiscal efforts or expenditure matching, thereby influencing states' behaviour as well as simply complementing it. The chief reason for fiscal effort conditions is to ensure that funds already devoted to education are not diverted to other uses. General revenue conditions commonly specify, for instance, that recipient authorities will maintain or achieve a particular level of expenditure, e.g. a fifth of total expenditure to education. Matching grants may be one for one, or any other proportion. 26 The adoption of a performance equalization approach through specific purpose payments has significant implications for federations. There is no doubt that it does'involve some loss of independence by the state governments within areas for which they have constitutional responsibilities. To many observers this loss of independence results in lower levels of diversity in the amount and quality of educational services provided. Although broadly similar considerations apply to state-local transfers as to federal-state transfers, there is one difference. While responsibilities for education may be devolved to local authorities, constitutional responsibilities remain with the state. The case for state intervention in terms of concentration on performance equalization and specific grants is, therefore, greater. There are five principal mechanisms for determining the amount of specific purpose grants: (a) an ad-hoc consideration on a project by project basis (b) the use of rough formulas based on arbitrary indicators of need formulated by the granting government (c) a detailed and direct assessment of need by the granting government (d) a detailed and direct assessment of need by independent commissions (e) an intergovernmental agency involved in bargaining and negotiations. In practice, assessments for specific purpose grants have been based largely on independent commissions' recommendations in Australia, on formulas in the United States and Canada developed by the federal agencies in the former with a great deal of provincial input in the latter, and through a broad system of 'responsibility sharing' in West Germany. In several developing countries, specific purpose grants for performance equalization in education have undergone periods of popularity and unpopularity. In India up to 1969, all Plan allocations to states could be termed specific purpose and many had matching conditions. Since then block grants and loans have been tied to the Gadgil formula. Revenue conditions are attached such that if a state fails to reach the stipulated level of outlay on, say, education due either to a diversion of funds or a- general lack of resources, a proportionate cut is applied to its share of central assistance under the Plan (Grewal 1975 p. 134). Some specific purpose payments, however, 27 continue to exist for 'centrally sponsored schemes' relating mainly to those of an experimental or spillover nature. Specific purpose payments in Nigeria have been for relatively few programs and were concentrated in the late 1970s. In education, the main one was for the universal primary education program. Grants directly aimed at performance equalisation, therefore, have been relatively unimportant. In Brazil, the same is true. Fiscal capacity equalisation and education The alternative approach to fiscal equalisation works through the revenue side of the recipient government's budget and aims to equalise the capacity of governments, school systems or schools to provide services rather than the services themselves. In other words it makes possible comparable services provided the state and local authorities impose comparable taxes or charges. Differences in fiscal capacity reflect differences in both the relative costs of providing services as well as relative revenue - raising capacity. The approach then involves two parts. First, an assessment of additional costs to provide a standard service resulting from (in educational terms), say, a relatively high proportion of school age population or high costs resulting from a dispersed population. Second, an assessment of the shortfall in revenue due to a low revenue base. If schooling is provided by a single purpose authority then capacity equalisation for education alone is possible. When provided by a multi-purpose authority the approach is not directed exclusively to education. Compared to attempts at performance equalisation, capacity equalisation permits a greater degree of decentralisation. General purpose payments to equalise capacity are made for both vertical and horizontal fiscal adjustment. For the latter purpose they are common in Canada, West Germany and Australia where state inequalities have been approached wholly by the equalisation capacity approach. The Canadian system is restricted to revenue equalisation thereby implicity assuming equal costs of providing services across provinces. Assessment of differential cost is also restricted in West Germany and only in Australia is there an attempt to derive a comprehensive measure of both 28 sides. Again, in Australia, a commission is used for assessment purposes as compared to the use of formula in the other two countries, and the two states with the highest fiscal capacity are taken as the standard as opposed to the use of an average. In the United States, India and Switzerland, on the other hand, to the extent that fiscal capacity equalisation criteria are used at all they have usually been modified by taking fiscal effort of the states or cantons into consideration. Having focused on the principles of federal financing, measures to reduce vertical and horizontal fiscal imbalance in general and for the purposes of educational provision in particular, the following section concentrates on country examples. V FEDERAL FINANCING OF EDUCATION - THE PRACTICE To illustrate in more detail the major points raised in the earlier sections of this paper, data on educational financing have been collected for three developed and three less developed federal states: the United States, Canada, Australia, Brazil, India and Nigeria. These data are used below to examine the problems of federalism as they relate to educational provision (horizontal inequity and disparities in spending: interjurisdictional spillovers) and the responses which have been made by federal governments. First the relevant data and descriptions are presented country by country. These are then followed by a consideration of themes, similarities, dissimilarities and general lessons to be drawn. United States The United States provides the supreme example of a decentralised educational system. In the Constitution there is no direct role for the federal government in education and the states are ultimately responsible. For elementary and secondary schooling there is a substantial degree of further devolution to local school districts, but the states retain overall constitutional responsibility for the results of that delegation. Constitutional responsibility for higher education is also in the hands of the states. 29 The financing of public elementary and secondary education currently involves all three levels of government - federal, state and local - but the relative importance of each has changed over time. Until this century, the roles of both federal and state governments in financing were minimal and schools were funded almost entirely by contributions from within school districts based on local property taxes. Over the last eighty years, however, the states' role in particular has gradually increased to a point where, since 1979, they are the major contributor. Federal government funding has also increased from virtually zero to almost 10 percent of the total in 1979, though it has fallen back since then. Table 5 describes the trends in school financing since 1919. Table 5 Sources of Revenue for Public Elementary and Secondary Schools, United States. Selected Years (percent) School Year Federal State Local 1919-20 0.3 16.5 83.2 1929-30 0.4 16.9 82.7 1939-40 1.8 30.3 68.0 1949-50 2.9 39.8 57.3 1959-60 4.4 39.1 56.5 1969-70 8.0 39.9 52.1 1979-80 9.8 46.8 43.4 1982-83 7.1 48.3 44.6 Source: National Center for Education Statistics, 1983, Table 6.2 National Center for Education Statistics, 1985, Table 1.11 30 In addition to the gradual reversal of state and local dominance in financing, three relatively dramatic changes can be noted: 1. the rapid increase in state financing in the 1930s and 1940s 2. an almost doubling of the (small) federal share in the 1960s 3. the further surge in the states' share in the 1970s. These overall shares of financing by level of government hide wide variations in the financing mix across states. Table 6 demonstrates some of the diversity. The increases in the states' share during the 1930s, 1940s and 1970s resulted from concerns over inequalities in educational provision resulting from a system of local financing based on property taxes. Since the income from such taxes varies considerably within states, it has been argued that state government expenditures are required to at least reduce the educational impact of such variations and provide 'foundation Table 6 Sources of Revenue for Public Elementary and Secondary Schools for Selected States, United States, 1980/81 (percent) State Federal State Local Mississippi 24.3 56.7 19.0 Kentucky 16.0 64.0 20.0 North Carolina 14.0 62.8 23.2 Vermont 7.4 28.4 64.2 Connecticut 6.2 29.4 64.4 New Hampshire 5.3 6.1 88.6 United States 9.3 47.4 43.3 Source: National Center for Education Statistics, 1983, Table 6.3 31 grants'. State government revenues, in turn, are derived from taxes levied and collected within the states - 83 percent of revenues in 1980 came from income and sales taxes alone. Revenue sharing is not a part of fiscal federalism in the United States. The growing dependence of the school system on state government revenues has a potentially serious significance in that these revenues are far more cyclical and sensitive to economic change than are revenues derived from property taxes. In addition to widespread intrastate differences in local revenues for educational expenditures, there are also similar differences between states - higher expenditures in urban industrial states and lower expenditures in the south. This has not, however, led to federal government financing, matching or otherwise, in an effort to reduce inequalities. Federal aid for elementary and secondary schooling has been special purpose and tied, aimed at providing equal opportunities for those with special needs who have been underserved by state and local governments. Over the past fifty years, programs have included: - school milk and lunches (1930s) - educational provision for children of armed forces (1950s) - science, maths, foreign language programs (1950s) - equality programs (1960s) - bilingual programs (1970s) - handicapped children (1970s) Such grants, as Winkler (1986) points out, are not directly focused on any specific problem resulting from federalism and if they do so it is a coincidence. Approximately, two thirds of the total expenditure on higher education is consumed by public institutions. Once again, all three levels of government are involved in funding though proportions differ significantly from those in elementary and secondary schooling. Table 7 presents the appropriate data. There is no general recognition of interjurisdictional spillovers beyond the state in higher education and consequently no general federal support. Federal expenditure is mainly in the form of research contracts and is almost equally distributed between public and private institutions. State expenditures are virtually all directed to public institutions and local sources of revenue support two year junior colleges only. As in the case of 32 elementary and secondary schooling, state revenues are the principal form of government financing of higher education and the share is increasing. The federal government share fell by one third during the 1970s. Table 7 Sources of Current Revenues for Institutions of Higher Education, United States, 1970/1 and 1981/2 (percent) Source 1970/1 1981/2 Government Federal' 19.3 13.3 State 27.6 30.3 Local 3.8 2.7 Private 5.1 4.9 Students Tuition 21.0 21.9 Other 13.1 11.3 Institutional3 10.1 15.7 100.0 100.0 Source: National Center for Educational Statistics (1985), Table 2.14 Notes: 1 includes research grants and contracts, excludes student aid 2 includes federal aid to students endowment income, sales and services, etc. Canada The Dominion of Canada Act 1867 provided for a federal government and four provinces. Six more provinces were added later. The Act gave the provinces exclusive powers to pass laws on education. Today the federal government provides some vocational training plus educational 33 services for only native people, the population of federal territories and the children of armed forces personnel on military bases. There is no federal Ministry of Education. The share of total education expenditures by direct source over the period 1978 to 1982 is described in table 8. Direct federal Table 8 Expenditures on Education by Direct Source of Funds, Canada, (percent) Year Federal Provincial Municipal Fees Other 1978-79 9.0 65.0 19.1 3.8 3.1 1979-80 8.4 65.6 18.5 3.8 3.7 1980-81 8.6 66.5 17.2 3.9 3.8 1981-82 8.6 66.8 16.9 4.1 3.6 1982-83 8.9 67.7 15.3 4.3 3.8 Source: Ministry of Supply and Services, 1984a, Table 39. expenditures as a share of total expenditures on education are very similar to those found in the United States - 8 to 9 percent. Again these are largely categorical and not linked to considerations of horizontal equity or spillovers. Another feature of table 8 is the decline in the share of municipal funding over the four years from 19.1 to 15.3 percent. Of the expenditures on education, 66 percent are devoted to primary and secondary schooling, 20 percent to the universities, 8 percent to post secondary community colleges and 6 percent to vocational training. The direct sources of funding for each education level are shown in table 9 for 1981/82. Almost half of all direct federal expen- 34 ditures on education are for vocational training. Taking the universities alone, the federal government contribution is 12.4 percent, the provincial governments' 68.9 percent with the rest being raised from Table 9 Direct Source of Funding by Education Level, 1981/82, Canada, (percent) Source of Funding Elementary Tertiary Vocational and Secondary Federal 2.7 9.7 68.6 Provincial 67.7 73.5 25.9 Municipal 25.6 0.0 0.0 Fees 2.2 ) ) Other 1.8 16.0 5.5 Source: Ministry of Supply and Services, 1984b, Table 5. fees and other institutional income. Again the direct federal contributions are channelled mainly into research. The financing picture changes substantially if federal government transfers to provincial governments are considered. These are for post secondary education alone. Until 1966, the federal government made grants directly to the universities and colleges. These were then seen as encroaching on the provinces' rights and beginning in 1967 they were substituted by transfers to provincial governments covering 50 percent of operating costs or. $15 per capita of population. This system was in turn replaced in 1977 by one mixing cash grants with a reduction of personal and corporate income tax payments to the center. In 1981/82, these 35 amounts were equal to 68 percent of provincial government expenditures on post secondary education. Inclusion of these grants and tax concessions raises the federal government's contribution to the financing of education from 8.9 to 22.0 percent for 1982/83. Australia The federal system in Australia comprises the Commonwealth Government and six state governments. At the time of federation at the end of the last century, no education powers were assigned to the central government. Since 1946, however, the center has taken virtually total financial responsibility for post secondary education and has increasingly, since the 1960s, provided grants to states for primary and secondary schooling. These were initially tied to libraries and science laboratories in secondary schools but the much larger grants beginning in 1973 have been unconditional block grants. Within the states, the control over education is highly centralised and even regional educational directors have no cash at their disposal. An important feature of the Australian education system is the place of non-government schools. Out of 9400 schools, 2100 are non-government. The Commonwealth Government obtains most of its revenue from income, profit and sales taxes. State government financing comes from state taxes (40 percent) and transfers from the Commonwealth (60 percent). Transfers are of two types: a) the financial assistance grant (or general purpose) based on a formula reviewed every five years and which is equal to 35 percent of all state revenues and b) specific purpose grants. Eighty percent of these grants for education have no further conditions attached. As noted above, state governments in Australia are constitutionally responsible for education. In practice, central government has taken financial responsibility for tertiary education and state governments have the major financial responsibility for primary and secondary education. 'The early 1970s, a period of substantial educational expansion, saw some change in the latter with the Commonwealth increasing its involvement on the premise, not that spatial inequalities existed in the system, but rather that education was generally underfunded and a vertical fiscal imbalance existed. 36 Table 10 Public Expenditure on Education and in Total, Australia, 1971-1976 ($m) Level of Government 1971 1973 1975 1976 Commonwealth Expenditure on schools 99 240 636 730 Total expenditure 8628 11,901 21,600 24,050 % 1.1 2.0 2.9 3.0 State Expenditure on schools 1023 1500 2706 3120 Total expenditure 6370 8436 13,775 16,700 7a 16.1 17.8 19.6 18.7 Source: Grewal, 1981, Table 8. A more comprehensive and contemporary account of all Commonwealth, state and local authority educational transfers and expenditure is provided in table 11. This table demonstrates the importance of the Commonwealth government transfers to states for education. Taking account of such transfers, the Commonwealth funds around 40 percent of the educational system. The distribution of Commonwealth funding by level (for 1982/3) is universities 31 percent; other tertiary 28 percent; primary and secondary 38 percent. The full complex picture of education financing including intergovernmental grants by level of education is provided in table 12. Essentially, this table shows direct expenditures on each level of education plus transfers and expenditures from independent or untied revenues by state and Commonwealth governments. 37 Table 11 Expenditure on Education of all Public Authorities, Australia, 1976-81 ($m) Level of Government 1976 1978 1980 1981 Commonwealth Total expenditure 2212 2526 2931 3352 Grants to states 1719 1960 2372 2738 State Total expenditure 4638 5616 7023 7976 Grant from Comm. 1719 1956 2372 2738 Local Total expenditure 13 13 15 16 Grant from states 13 10 12 14 Expenditure on education as % of total expenditure 16.1 16.0 15.2 15.0 'Own funds' state expenditure as % of total expenditure 57 59 61 61 Source: Bureau of Statistics, 1984, p242. 38 Table 12 Direct and Indirect Expenditures on Education by Level of Government and Level of Education, Australia, 1979/80 ($m) Level of Education Direct Intergovern Own Revenue and Government Expenditure mental Grants Expenditure (1) (2) (3) (4) Primary and Secondary Commonwealth 111 706 818 State and local 3872 - 706 3165 Total 3983 3983 University Commonwealth 170 705 875 State and local 715 - 705 10 Total 885 885 Other Higher Education Commonwealth 85 497 583 State and local 584 - 497 86 Total 669 669 Vocational Training Commonwealth 52 148 200 State and local 463 - 148 315 Total 515 515 Total Education Commonwealth 511 2096 2608 State and local 6194 - 2096 4098 Total 6705 6706 Z State 92 61 % Commonwealth 8 39 Source: Bureau of Statistics, 1982, Table 3. 39 The situation in column (2) reflecting direct expenditures is very different to that in column (4) which presents expenditures once transfers have been taken into account. Only 8 percent of direct expenditures are made by the Commonwealth government, but once transfers are accounted for this figure rises to 39 percent. In effect, though nearly all resources flow through the hands of the states, the Commonwealth funds 20 percent of primary and secondary, 99 percent of universities, 87 percent of the rest of higher education and 39 percent of vocational training. Reasons for these transfers and their pattern have been a vertical fiscal imbalance and the recognition that tertiary level education is the most appropriate level of education to be funded by a federal government. Even so, funds are channelled to the states, not to the institutions directly. Brazil The government structure of Brazil comprises the federal government, 3 federal territories, 23 states and 4000 municipios. Between 1965 and 1967 the government enacted a comprehensive set of federal fiscal reforms, the principal aims of which were to increase the revenues of state governments in poorer regions and to channel increased federal funds to municipios (local governments) with limited tax bases and administrative capacities. To this effect, the rights to collect specific taxes were reallocated among the federal, state and local governments and a more regionally distributive system of inter- governmental revenue sharing was put into place. The main effects of the reforms were to redefine the tax bases available to each level of government, concentrate control over tax bases and rates at the federal level and expand the importance of transfers from revenue sharing to the state and municipal governments. As has previously been shown, only 2 of the 12 taxes used for revenue-sharing purposes include assessments of need in the allocations between provinces. As a consequence despite the revenue sharing and some attempts to lessen regional inequalities, the revenues of individual states and municipal governments vary substantially. This has major consequences for educational provision. 40 Data on both issues are presented in table 13. Expenditures per pupil varied substantially between regions - Southeast Cr$ 1196, Northeast 418, Frontier 642. Table 13 State and Municipal Revenues and Education Enrollments, Brazil, Mid 1970s Southeast Northeast Frontier Revenues per capita (Cr$) State governments 1025 312 717 Municipal governments Capitals 703 248 244 Interior 243 57 84 Enrollment ratios (Z) Primary - urban 69 57 63 - rural 52 27 n.a. Secondary - urban 20 12 12 - rural 3 1 n.a. Source: Maher and Dillinger, 1983, Table 11. Real education expenditures increased by 11 percent a year between 1970 and 1980 and by 1.3 percent a year up to 1983. The greatest increases have been by the municipios and the federal government, resulting in a reduced share of state expenditure from 61.3 to 55.7 percent. Table 14 documents the changes in these shares. Each level of government allocates its educational expenditures in different ways. The 41 Table 14 Share of Total Education Expenditure by Level of Government, Brazil, 1970-83 (percent) Level of Government 1970 1975 1983 Federal (net of transfers) 27.1 25.3 29.5 State 61.3 60.0 55.7 Municipal 11.6 14.7 14.8 Source: From Winkler, 1985, Appendix Tables 13, 14, 15. federal government allocates its highest share to higher education (46 percent) while both state and municipal governments allocate most to primary schooling (63 and 72 percent respectively in 1983). Funding of the three levels of education in 1980 is shared between governments as shown in table 15. For both secondary and higher education, the states' share declined somewhat between 1975 and 1980 (by at least ten percentage points), but this level of government's pivotal position in the financing of both primary and secondary schooling is clear. Municipios have a substantial role only in primary schooling, while the federal government once again dominates the funding of higher education and has been increasing its share of secondary financing. Table 15 underplays the federal government's position in education financing in that the figures refer to direct expenditures by the levels of government and thus any inter-governmental transfers are not accounted for. In table 16 the share of each educational levels' financing by type of government using 'own source' data is shown for 1983. 42 Table 15 Source of Direct Funding for Each Education Level, Brazil, 1980 (percent) Primary Secondary Higher Federal 8.7 28.8 76.9 State 65.8 69.5 23.1 Municipio 25.5 1.7 Source: Winkler, 1985, Appendix Tables 13, 14, 15. Table 16 'Own Source' Share of Education Funding by Level, Brazil, 1983 Primary Secondary Higher Federal 18.9 32.7 75.2 State 57.4 64.3 24.8 Municipio 23.7 - 3.0 Source: See Table 15. Comparing tables 15 and 16, the main effect of taking account of transfers is to increase the federal government's share of primary financing from 8.7 to 18.9 percent. These transfers result from the 43 federal government's one third share of revenues from the education salary tax. As Winkler (1980) argues, the size of these transfers is too small to significantly reduce spending differences and may in fact partly increase inequalities through distributing more to state governments for use in state schools than to local governments for use in lower spending local schools. Matching grants to take account of spillovers at the primary and secondary level are not used. India Responsibility for education in India was transferred from the central government to the states through the Government of India Acts of 1919 and 1935. At Independence, in 1949, the Constitution further clarified the division of powers with the center having direct responsibility for the central universities, standards throughout the higher education sector and professional, vocational and technical training institutions, and a shared responsibility with the states for the vocational and technical training of labor. The rest of education was listed as a state subject. The emphasis on comprehensive economic planning and five year plans adopted since Independence and covering both central and state activities has inevitably led to a larger central government role than the listing of union, state and concurrent functions suggests. The Constitutional Amendment of 1976 which transferred education to the concurrent list can be seen as simply legalising a situation which in practice already existed. The amount of finance required to fulfil the states' responsibilities in education and in other areas is well beyond that available through local levels of taxation. The principle of revenue sharing together with grants-in-aid, therefore, underpins the system of federal financing and, as explained in section III, involves both the Finance Commission and the Planning Commission. After defence, education is the sector of highest budgeted activity. Over the country as a whole, budgetary allocations to education have not kept pace with the growth in enrollments and the rise 44 in prices. As a result, total expenditure per student per year by the center and state governments combined has fallen in real terms. Further, the recent policy paper from the Ministry of Education (1985) notes that as a percentage of total plan expenditure of the center and states together, the share for education has declined from one plan to the next apart from the Third Plan period. In the First Plan the share for education was 7.2 percent and in the Sixth Plan, 2.6 percent. Plan allocations to primary schooling in particular have experienced sharp proportional reductions. Apart from central and state government expenditures, there are few other resources allocated to education. While a third, local, tier of government exists in India it has no resource raising competence defined in the Constitution and receives most of its revenues from the state governments. Similarly, according to the Ministry of Education (1985) "there is hardly any effort on the part of a village or the community leadership or at the level of block or district, to set up or even to help the proper running of schools" (p37). The distribution of education expenditure between Plan and non-Plan activities is shown in table 17 for selected years. Table 17 Distribution of Plan and Non-Plan Expenditure on Education in India. Selected years (percent) Year Plan Non-Plan 1970-1 14 86 1978-9 16 84 1980-1 14 86 1982-3 16 84 Source: Tilak, 1984, p11 Ministry of Education, 1985, p24 45 The Plan expenditures can in turn be divided between those supported by the central government and those by the state governments (irrespective of the source of revenues). This is done in table 18. Table 18 Contribution of the Center and the States to Plan Expenditure on Education in India (percent) Plans Central Government State Governments Total Rs 1 25 75 153 2 25 75 273 3 26 74 589 4 33 67 823 5 32 68 1285 6 30 70 2524 Source: Tilak (1984) p16 Note : Rupees in 10 millions Adding in non-Plan expenditures to the Plan ones shown above, demonstrates the relatively small direct financial role played in education by the central government (table 19). Interestingly, the 1976 amendment placing education on the concurrent list of functions does not appear to have affected center-state government proportional expenditures. 46 Table 19 Center and State Financing of Education, India (Plan and Non-Plan Expenditure) (percent) Period Central Govt State Govts 1st Plan 6.8 93.2 2nd Plan 17.5 82.5 3rd Plan 20.1 79.9 4th Plan * 8.0 92.0 5th Plan ** 8.5 91.5 1976-77 9.0 91.0 1977-78 8.6 91.4 1978-79 9.3 90.7 Source: Tilak, 1984, Table 3 Note * Onwards revenue account only ** 4 year period up to 1977-78 The preceding documentation indicates that direct involvement in education by the federal government is limited to a small set of institutions such as regional colleges of education and centers for advanced university studies, to activities such as the promotion of Sanskrit and Hindi in non-Hindi speaking states and to a few activities such as education of the handicapped. These are mainly areas of spillover and national interest. In addition to the direct expenditures there are the transfers. These, for the education sector as well as in general have not decreased inter state inequality. In a review of this subject, Tilak (1984) argues that, "On the whole one may have to conclude that neither the Planning Commission nor the Finance Commission have been able to introduce progressivity in their transfers of resources to the states for the educational sector." (p23). The result is high levels of 47 inequality in educational spending. Of the nine states categorised as educationally backward in 1982/83, only one was spending at least the national average per capita on education. Examples of the variations in these allocations were R40.5 in Uttar Pradesh, R49.4 in Madhya Pradesh and R51.2 in Bihar compared to R100 in Punjab and R119.5 in Kerala. The central federal issue of horizontal fiscal imbalance has not yet been solved. Nigeria Nigeria has a federal government and nineteen constituent state governments. As has been previously described, the system of federal financing is based largely on mechanisms of revenue sharing rather than on the allocation of specific taxation powers to each level of government. Transfers of resources to the states, however, are not made at the discretion of the federal government. Rather, the mechanisms and the criteria used to divide gross revenues between and among levels of government, though subject to alteration, take the form of constitutional guarantees. Over time, the distributional criteria have been given increasing importance. In addition to these grants which are unconditional and unmatched, conditional grants have also occasionally been made. For most of the period since self government in 1955, primary and secondary schooling have largely been deemed a function of regional, then state, governments. In 1979 constitutional liability for primary schooling was devolved to local authorities. The position of higher education has varied more over time with various mixtures of federal and state involvement, particularly in the universities. Today, the federal government has a direct financial and administrative responsibility for 96 percent of enrollments in universities 23 "" " "t polytechnics 7 " " " " secondary teacher training 1 " " " " academic secondary schools. The remainder of secondary schooling, teacher training and technical education is the direct responsibility of state governments. Primary schooling, though devolved to local authorities is largely funded, at 48 least for the recurrent costs, by the states. Though both state and local governments are recipients of resources from higher levels of government, no funds are earmarked for education. Figures for actual and approved capital and recurrent education spending by federal and state governments for 1981 to 1985 are pieced together and analysed as far as is permitted in tables 20 and 21. Table 20 Actual and Planned Total Federal and State Expenditures on Education, Nigeria, 1981-85 (Naira million) Level of Government and Type of Expenditure 1981 1982 1983 1984 1985 Federal Capital 602 463 297 87 85 % of Total Federal Capital Expenditure 7 6 4 3 2 Recurrent 781 699 663 657 699 Z of Total Federal Recurrent Expenditure 21 20 16 17 19 State Capital and Recurrent 2562 2156 2447 1745 N.A. Federal Educ.Exp. as % of Total Educ.Exp. 35 35 26 30 N.A. Source: World Bank unpublished sources Note: Federal expenditures are all actuals. State expenditures are actuals for 1981 and 1982 and approved budget estimates for 1983 and 1984. 49 Table 21 Actual and Planned Federal and State Education Expenditure by Level, Nigeria, 1981-84 (Naira million) Level of Government 1981 1982 1983 1984 and Education Primary Federal State 824 835 880 948 % Federal 0 0 0 0 Secondary Federal 105 133 109 55 State 844 910 1119 610 % Federal 11.1 12.8 8.9 8.3 Teacher Training Federal 49 64 61 34 State 282 279 318 120 7. Federal 14.8 19.2 16.1 22.1 Polytechnics Federal 72 88 72 66 State 85 88 89 51 % Federal 45.8 50.0 42.1 56.4 Universities Federal 657 532 485 442 State 27 43 40 17 X Federal 96.1 92.5 92.4 96.3 Total Federal 883 817 727 597 State 2562 2155 2446 1746 % Federal 24.2 27.5 22.9 25.5 Source: World Bank unpublished sources Note: Figures for 1981 and 1982 are actual and for 1983, 1984 and 1985 are approved Estimates. 50 The main points which emerge from these tables with respect to the roles and behavior of federal and state governments in the funding of education in Nigeria between 1981 and 1985 are: - Federal capital and recurrent education expenditures as proportions of total Federal capital and recurrent expenditures have fallen during the 1980s, from 7 percent to 2 percent and from 21 percent to 19 percent respectively. - Federal education expenditure as a proportion of total public education expenditure has fallen from 35 percent to 30 percent. Since fees have been increasing, federal expenditures have fallen even further as a proportion of total education expenditure. - Federal expenditures dominate the public funding of the universities (92-96 percent), fund half the polytechnics' costs, 15-20 percent of teacher training and around 10 percent of secondary schooling. The rest is provided by the states and also by local governments which have particular responsibility for capital expenditures in primary schooling. - As the share of federal funding of the education sector as a whole has fallen, it has held up in post-secondary education (96 percent of the total higher education expenditure in both 1981 and 1984). The fall has occurred in the lower levels. The federal government share has apparently fallen mainly because primary school expenditure, in which the federal government does not now directly participate, has grown faster than other levels. The pattern of federal financing in Nigeria goes some way to face the inherent problems of spillovers and spatial inequalities. In some senses, however, it goes beyond the federal spirit by, for instance, funding most of the universities directly rather than through the state governments. Apart from the universal primary education program in the late 1970s, the issue of horizontal inequalities has not been approached by specific purpose grants. Instead, the system of revenue sharing with the gradual reduction in the weight given to derivation and increase in the weight given to distributional factors has been relied on. 51 VI OVERVIEW AND LESSONS LEARNED General issues of federal financing Vertical imbalance In most federal countries, states in general have not been provided with sufficient own-source revenue or automatic (constitutionally guaranteed) transfers from the center to finance their responsibilities. As a result additional non-automatic, discretionary transfers or (particularly in the case of Nigeria and Canada) changes to the constitution have had to be made. Of the additional transfers, shared cost programs have been common particularly in developed countries. However, in the United States these have been significantly reduced over the past eight years and in Canada the concern over these programs for state autonomy has led in some cases to states opting out. In Germany, on the other hand, shared cost programs are extensive precisely because the states are relatively powerful and have the major say in the programs. Of the federal developing countries surveyed, only in India are shared cost programs, run by some individual ministries, common though Plan expenditures in these countries invariably require subsequent state recurrent expenditures. The specific earmarking of transfers to reduce vertical imbalances has been reduced. Such programs initiated by the United States and Nigeria in the 1970s have been cut back or abolished. Similarly, non-statutory transfers in Brazil and India-are less earmarked than they once were. Again, in Canada and Australia the main source of funds to reduce imbalance between states' revenues and responsibilities is the general revenue grant and those specific purpose grants which have been made have mainly been limited to expenditure within whole sectors or within very broad programs such as university education. 52 Horizontal imbalance In the United States, the states directly collect their revenues and federal government measures aimed at fiscal equalisation are few. In contrast, in Australia and Canada resources from federal/state revenue sharing arrangements are an important source of state funds and in both countries arrangements include strong measures to reduce fiscal inequalities between states. Inequalities between states' own-source revenues per capita are wide in the developing countries. However, in Brazil, only a very few of the revenue-sharing sources are distributed according to criteria of need and in India transfers have not been systematically distributed in favor of the poorest states. Again in contrast, in Nigeria, the highly populated and poorer northern states have been able to reduce the influence in the determination of transfers of the derivation criteria and extend first the impact of the population criterion and more recently the additional 'need' factor. In those countries where significant equalisation attempts have been implemented (Australia, Canada, Nigeria) emphasis has been mainly on non-specific block or general revenue grants rather than on sector or project specific grants. In other words, in keeping with federal principles successful equalisation attempts have been based on measures aimed at allowing all states to offer similar services rather than on measures which directly ensure that those services will be provided equally. Educational finance In the developed countries with well-established federal systems of government, no constitutional role is assigned to the federal government for education. Direct federal expenditures are, consequently, largely limited to categorical grants for specific types of students (ethnic minorities, the handicapped, armed forces children, etc.) and for research. In each country these are equal to only 8-9 percent of total educational expenditures. Of the federal developing countries, the situation in India most closely resembles that in the developed countries 53 in that the states largely have responsibility for each educational level. In Brazil and Nigeria however, the federal governments do have overriding constitutional responsibility for higher education and particularly in Brazil are also more involved at secondary school level. Federal educational finance in these two countries accounts for around 30 percent of the total. The degree of local government involvement and responsibility in education varies widely among federal countries. Whereas schooling is highly centralised, administratively and financially, at the state level in Australia, there is substantial devolution to local school boards in the United States including substantial measures to reduce intra-state inequalities. The extent of municipal government enrollment in primary and secondary schooling in Canada comes between that in the other two countries. Differences also exist in this respect between the federal developing countries. In Brazil a quarter of all primary school funding is provided by the municipal governments. This is much higher than in either Nigeria, where local authorities run schools but provide little of the finance, or India. As explained above, efforts to reduce horizontal or inter-state inequalities in the provision of services such as education have largely taken the form of general revenue grants rather than being sector, eg. education, specific. Central governments have been more direct over transfers aimed at reducing vertical imbalances with respect to educational provision. The most common measures to reduce vertical imbalances have been block grants for specific education sectors, particularly higher education. These have been channelled either via the states as in the case of Australia and Canada, or directly to the institutions as in Nigeria and Brazil. In India, only the so-called central universities are funded by the federal government and in the United States no university teaching at all is funded by the federal government. Block grants to institutions or to state governments for other education levels have been far fewer, though the Australian federal government provides 20 percent of primary and secondary schooling costs. Elsewhere, on the whole, specific purpose and earmarked grants to states for education are not common and where they have been used - for example, the universal 54 primary education program in Nigeria - they have been relatively shortlived. Conclusion In the federal countries which have been focused on in this paper, there has been little attempt to ensure equality of educational provision directly. In line with the principles of federalism, questions of spatial inequalities have been addressed more through measures to equalise overall resources than through measures to equalise services in any particular sector. With regard to vertical imbalance between the federal and state governments different approaches have been followed. In the United States, no imbalance is seen to exist in general terms and federal funds are directed towards specific population groups, irrespective of where they live. The low level of federal government funding of education in India again suggests that the states are in general regarded as having sufficient resources to fund education. The response in Nigeria to the increasing demands for education has been to alter the formula for the federal-states division of tax revenues. Another response in Nigeria was to take over full financial responsibilities for existing state-supported universities. This policy has also been followed to different degrees and taking different forms, in Brazil, Australia and Canada. The widespread move towards the federal funding of higher education is not surprising. Because of large scale labor migration between states, considerations of national interest and wide geographical intake of students the spillover effects are considerable. The implications of this situation for the relative financial backing given to the different levels of education may be significant. As has been widely demonstrated, federal government revenues tend to grow at a faster rate than those of state governments. In the introduction to this paper, a list of questions was posed concerning the financial, political and educational implications of diversifying the sources of educational finance. In an attempt to throw some light on these questions, experiences within the most extreme form of financial diversification - fiscal federalism - have been discussed. Though answers to all the questions have not been forthcoming, the data 55 given for six federal countries at different levels of economic development provides a base. Three of the questions currently attracting attention in countries which have already significantly diversified sources of educational finance, or are planning to do so are: - how to ensure resources transferred to states for educational purposes do not simply displace states' existing expenditures, - how to most effectively influence educational development through the addition of only small transfers relative to total expenditure, - how to devise mechanisms for fiscal transfers which, while being linked to central government policy requirements, do not overly threaten state autonomy. These issues merit an extension of the comparative focus of this study. 56 REFERENCES Bureau of Statistics (1982) Expenditure on Education, Australia, 1979-80, Canberra. Bureau of Statistics (1984) Australian Yearbook 1984, Canberra. L. Cantor (1980) The Growing Role of the States in American Education. Comparative Education 16, 1. R. Chelliah (1981) Trends and Issues in Indian Federal Finance, Allied Publishers, New Delhi. Federal Republic of Nigeria (1981) Fourth National Development Plan, Lagos. K. George and I. Gulati (1983) Centre-State Resource Transfers 1951-84. Economic and Political Weekly, Feb.16. B. Grewal (1975) Specific Purpose Grants in a Federal System: Overseas Experience in R. Mathews (ed.) Responsibility Sharing in a Federal System. Research Monograph 8. Centre for Research on Federal Financial Relations, Australian National University, Canberra. B. Grewal (1981) Equalisation Techniques for School Finance: an Evaluation. Occasional Paper 20. Centre for Research on Federal Financial Relations, Australian National University, Canberra. U. Hicks (1978) Federation: Failure and Success, MacMillan, London. D. Mahar and W. Dillinger (1983) Financing State and Local Government in Brazil. World Bank Staff Working Paper No.612, Washington DC. R. Mathews (1983) Fiscal Equalisation in Education. Centre for Research on Federal Financial Relations, Australian National University, Canberra. K. McKinnon and L. Hancock (1979) Australia, in OECD, Educational Financing and Policy Goals for Primary Schools, Paris. R. McMeekin and C. Dede (1980) American Education in the 1980s. Comparative Education 16,3. Ministry of Education (1985) Challenge of Education - a policy perspective, New Delhi. Ministry of Supply and Services (1984a) Education in Canada. A Statistical Review for 1983/84, Ottawa. Ministry of Supply and Services (1984b) Financial Statistics of Education 1981/82, Ottawa. 57 Ministry of Supply and Services (1985) Canada Yearbook 1985, Ottawa. K. Nambiar (1983) Criteria for Federal Resource Transfers. Economic and Political Weekly, May 28. National Center for Education Statistics (1983) Digest of Education Statistics 1983-84, Washington, DC. National Center for Education Statistics (1985) The Condition of Education, Washington DC. National Education Association (annual) Estimates of School Statistics, Washington DC. H. Noah and J. Sherman (1979) Educational Financing and Policy Goals for Primary Schools, OECD, Paris. A. Odden (1984) Education Finance in the States: 1984. Education Commission of the States, Denver. H. Rao (1981) Centre-State Financial Relations, Allied Publishers, New Delhi. D. Rondinelli, J. Nellis, G. Cheema (1984) Decentralisation in Developing Countries: A Review of Recent Experience. World Bank Staff Working Paper, No.581, Washington DC. L. Rupley (1981) Revenue Sharing in the Nigerian Federation. Journal of Modern African Studies 19.2. G. Sawer (1976) Modern Federalism, Pitman, Victoria. J. Tilak (1984) Centre-State Relations in Financing Education in India. National Institute of Educational Planning and Administration Occasional Paper No.5, New Delhi. T. Traber (1980) Receipts and Burden Sharing, the Taxation System and Fiscal Equalisation in the Federal Republic of Germany. in R. Mathews (ed.) Federalism in Australia and the Federal Republic of Germany. Australian National University Press, Canberra. United States Secretary of Education (1982) Prospects for Financing Elementary/Secondary Education in the States, Washington DC. C. Wallich (1982) State Finances in India, Vol.1. World Bank Staff Working Paper No.523, Washington DC. D. Winkler (1985) Finance of Brazilian Primary Education, World Bank, Washington DC (Mimeo). 58 D. Winkler (1986) Federation and Education Finance: the Case of Elementary Education in Brazil and the USA. Paper presented at the University of Dijon, June 23-25. R. Yadau (1980) Tasks Ahead for Indian Education. Comparative Education 16.3.