Argentina - Provincial finances study : selected issues in fiscal federalism (Vol. 2 of 2) : Technical annexes

Report No. 15487-AR Argentina Provincial Finances Study Selected Issues in Fiscal Federalism (In Two Volumes) Volume Il: Technical Annexes July 12, 1996 Public Sector Management & Private Sector Development Division Country Department I Latin America and the Caribbean Regional Office ,,.c,t of the World Bank "  Currency Equivalents Currency Unit: Peso US$1 = ARG$1 Fiscal Year January 1 to December 31 ABBREVIATIONS AND ACRONYMS ATN Aportes Tesoro Nacional, Transfers of National Treasury CAS Country Assistance Strategy CFI Coparticipacion Federal de Impuestos EFCB Emergency Financial Control Board FEDEI National Electric Development Fund FIEL Foundation for Latin Amercian Economic Research FONAVI National Housing Fund GDP Gross Domestic Product GO General Revenue Bond IDB Interamerican Development Bank MAC Municipal Assistance Corporation MCBA Municipalidad de Buenos Aires NBI Unsatisfied Basic Needs PEM Public Expenditure Management  ARGENTINA PROVINCIAL FINANCES STUDY Technical Appendixes' Table of Contents Annex 1: Restructuring Fiscal Federalism In Argentina Introduction ...................................................1 Approach ...................................................3 The Big Questions ...................................................3 Expenditure Assignment ...................................................3 Revenue Assignment ...................................................4 Vertical Imbalance ...................................................4 Equalization ...................................................6 Policy Objectives ...................................................7 Provincial Taxation ...................................................8 Introduction ...................................................8 Criteria for a Good Provincial Tax .................................................. 11 Automobile Taxes ................................................... 13 Property Taxes .................................................. 16 Gross Receipts Tax ................................................... 17 Additional Revenue Possibilities .................................................. 18 Revenue-Sharing and Intergovernmental Transfers .................................................. 21 Introduction .................................................. 21 Intergovernmental Fiscal Transfers: Principles .................................................. 24 The Present Transfer System ................................................... 30 Conclusion .................................................. 39 References .................................................. 40 Boxes: Box 1. Argentina: An "Emerging" Federation? ....................................... ..............2 Box 2. Accountability as the "Bottom Line" . ....................................................5 Box 3. Interpersonal Resdistribution and Intergovernmental Transfers .............. ..............7 Box 4. Municipal Taxes And Fees .................................................... 10 Box 5. What Is A "Provincial" Tax? .................................................... 12 Box 6. The Choice Of Subnational Taxes .................................................... 13 The authors of the Appendixes were as follows: Annex 1, Restructuring Fiscal Federalism, Richard Bird; Annex 2, Walking Through Argentina Fiscal Federalism, based on Juan Sanguinetti's Report; Annex 3- Improving Financial Management Capacity, David Grossman; Annex 4 - Financial Capacity Assessment and Simulation Model - Ben Darche.  Box 7. Limits To Property Taxation ............................................................ 15 Box 8. A Provincial Vat? ............................................................ 19 Box 9. Fiscal Capacity And Fiscal Effort ............................................................ 23 Box 10. Matching (Conditional) Transfers ......................................... 26 Box 11. Changes In Coparticipation Law, 1988-1995 ................................................... 29 Box 12 Determining The "Primary Distribution" ........................................................;... 32 Box 13. Capitation Grants For Education .......................................... 34 Box 14. Determining The Secondary Distribution ......................................................... 35 Tables: Table 1. Tax Collections, by Level of Government and Type of Tax, 1994 ...................... 9 Table 2. Major Provincial Taxes, 1994 ................................................................. 11 Table 3. The Structure of Provincial Revenues,1995 ........................................................ 21 Table 4. Provinces: Selected ndicators ................................................................. 22 Table 5. Transfers to Provinces, 1994 ................................................................. 31 Table 6. Allocation of taxes that make the shared-revenues pool .............. ....................... 31 Annex 2:- Walking Through Fiscal Federalism In Argentina . ...................................... 42 Annex 3: Strengthening The Institutional Framework ............................................... 46 Introduction .................................................................. 46 Assessing Financial Management .................................................................. 48 Financial Management Systems .................................................................. 48 Budgeting .................................................................. 47 Accounting .................................................................. 50 Auditing .................................................................. 50 Performance Measurement .................................................................. 52 Integrating The Components .................................................................. 52 Other Significant Management Systems ................................................................ 53 Relationship Between Responsibility And Resources ............................................... 53 Willingness To Manage .................................................................. 54 Current Status Of Provincial Financial Management ............................. ...................... 54 Strengthening Provincial Fiscal Management ................................................................ 56 Relevant Experience .................................................................. 56 Standard Fiscal Control echanisms .................................................................. 60 Fiscal Emergency Control echanisms .................................................................. 62 Increasing Access To Long Term Finance .................................................................. 64 Relevant Experience .................................................................. 64  Applicability To Argentina ............................................................ 66 Recommended Actions .................................................................. 66 Annex 4: Comparative Creditworthiness Assessment of The Argentine Provinces.. 70 Introduction .................................................................. 70 Creditworthiness Measures in the Context of the Argentine Provinces ....................... 70 Creditworthiness Simulation Model Methodology and Results ................................... 72 Indicator trends .................................................................. 77 M ethodology and Results ............................ ...................................... 78 Conclusions .................................................................. 83 Tables in the Text Table 1: Printout Model to project Main Variables ...................................................... 73 Table 2: Population and Economic Base Indicators and Score ..................................... 86 Table 3: Financial Operations and Debt Indicators .................................................... 87 Table 4: Debt Indicators' scores ........................................................ 88 Table 5: Expenditure Flexibility and Financial Management Score ............................... 89 Table 6: Creditworthiness Matrix ........................................................ 90 Table 7: Simulation of Ranking Provinces ........................................................ 94   Annex 1: Restructuring Fiscal Federalism in Argentina 1 Restructuring Fiscal Federalism in Argentina Introduction 1.1. Intergovernmental fiscal relations are under strain in Argentina. The system of "fiscal federalism" patched together in the last days of the hyperinflationary 1 980s is unsuited to today's circumstances. Over the last decade, Argentina has undergone fundamental restructuring. The restoration of democracy and the curbing of the endemic inflation have made Argentina a very different country than it was in the early 1980s. But the process of structural change is by no means finished. Both democratic politics and sound economic policy demand further major changes in the fiscal relations between levels of government. The present federal fiscal system grew up in an era of non-democratic government and was crystallized in its present form in the middle of hyperinflation. It is no surprise that the fiscal institutions developed in such different circumstances do not serve the new Argentina well. Nor, in the absence of major changes, are they likely to do so in the future. 1.2. An important symptom of the present mismatch of fiscal institutions and fiscal needs has been the emergence of significant deficits in a number of provinces in recent years. Changes are needed both to avoid the emergence of similar problems in the future and to ensure that the "downsized" and restructured state sector provides incentives not only for sound financial management at all levels of government but also for the effective and efficient delivery of public services. 1.3. Fortunately, 1996 offers two important opportunities for change in this direction. First, in completion of the Pacto Fiscal, provincial governments are committed to reform their tax systems in several specific ways. Second, the Constitution mandates that a new "Coparticipation" (Revenue-Sharing) Law is to be introduced by the end of this year. In part for these reasons, Argentine government and researchers have done an enormous amount of work on these subjects, and agreement appears to have been reached on the need for major changes. As yet, however, there is by no means consensus as to the precise nature of the desired changes. 1.4. One reason for the lack of consensus is the complexity of the problems that must be resolved. Some are short-term e.g. the fiscal crisis in a number of provinces. Others are long term e.g. the need to ensure that decentralized services are provided efficiently. Some are not controversial e.g. the need to design and implement a more stable system of transfers. Others are inherently controversial e.g. the extent to which fiscal transfers should be regionally redistributive. 1.5. The complexity of the underlying political-economic situation and the fluid nature of the many proposals for changes in federal fiscal institutions make it difficult to be comprehensive or innovative with respect to fiscal federalism in Argentina. Our approach is to stand back from the  2 Annex 1- Restructuring Fiscal Federalism on-going discussion, raise some fundamental questions about intergovernmental finance and to suggest some of the implications of alternative ways-of answering these questions. Box 1. Argentina: An "Emerging" Federation Although Argentina's constitution has long been fornally federal in nature, its political and fiscal institutions have for many years operated in a centralized fashion. Throughout the world, the range of institutional structures and relations within nominally federal countries is as wide as it is within nominally unitary countries. Indeed, the difference between a tight federation such as Malaysia or Germany and many unitary countries is probably less than that between such federations and looser federations such as India and Canada, in which state governments have more power to act independently with respect to expenditure and taxing patterns. The constitutional label matters less than the reality of how intergovernmental relations work in practice. In the traditional world of fiscal federalism (or multi-level finance), as set out by Oates (1972) in principle everything - boundaries, assignments, the level and nature of transfers, etc. - is up for grabs. In "federal finance," however, jurisdictional boundaries and the assigrnent of functions and finances are generally taken to be fixed at some earlier (constitutional) stage and not open to further amendment in normal circumstances. In the conventional approach to fiscalfederalism, as a rule the central government's policy preferences are clearly dominant (in practice, if not so clearly in theory) as exemplified by the dominance of conditional transfers. Infederalfinance, however, rather than simply assuming consensus on e.g. the degree of fiscal and regulatory harmonization to be attained, or even the degree to which an internal common market should be sought, such matters are in some real sense determined jointly by both levels of government in some appropriate political (constitutional) forum. In addition, both levels of government may properly pursue their own distributive policies, again with no presumption of central dominance. In such a federal setting, intergovernmental transfers are generally both equalizing and unconditional, as, for example, Shah (1995) emphasizes. The appropriate analytical framework in the fiscal federalism setting is thus clearly a principal-agent model in which the principal (the central government) may alter jurisdictional boundaries, local government revenue and expenditure responsibilities, and intergovernmental fiscal arrangements in its attempt to overcome the farniliar agency problems of information asymmetry and differing objectives between principal and agent. The appropriate analytical framework in the federal finance setting, however, is more one of bargaining between principals (who are not necessarily equal) - what one Canadian author (Simeon, 1972) has called, in a useful analogy to the world of international relations, "federal-provincial diplomacy." In some countries - Canada, Switzerland - this may not be a bad description of federal-state relations (Bird, 1986). It is critical to distinguish these two broad approaches in principle because the implications for policy design that follow from them are very different, especially with respect to the design of intergovernmental transfers. Federal finance raises different problems than those considered in the traditional literature on fiscal federalism. For example, unlike the latter, a federal finance perspective must pay close attention to the institutional analysis of reality. Particularly careful attention must be paid to the lessons of public choice analysis with respect to the importance of what has been called the fundamental "Wicksellian connection" (3reton, 1996) between taxing and spending in order to ensure both accountability and the maintenance of a legitimate system of governance in terms of public support of government. As emphasized earlier, the "federal finance" perspective suggests that the appropriate analytical framework is one of negotiation among equals - in the words of Wheare's classical political treatment of federalism (1969, p.10) among federal and state governments that are "...each, within a sphere, co-ordinate and independent." Argentina has by no means as yet gone this far, however, so the basic principal-agent approach has been followed in this report. However, over time the federal finance approach may become increasingly applicable. The present discussion on reforming the coparticipation system perhaps afford an illustration: both sides, center and provinces, must reach an agreement before any significant changes can be made inthe system.. 1.6. Argentina appears to be what may be called an "emerging" federation (Box 1). The decisions to be made with respect to intergovernmental fiscal arrangements over the next few  Annex 1: Restructuring Fiscal Federalism in Argentina 3 years are likely to prove critical influences on the nature of the state in the long run. Such matters should not be decided by a few persons working in isolation under a tight deadline. They are the core of nationhood and will inevitably take time and careful effort to work out and implement. The present report is intended as a first small contribution to this important task. The Approach of this Report The Big Questions 1.7. Four big questions must be answered with respect to intergovernmental finance in any country, whether formally "federal" or not: (a) Who does what? - the question of expenditure assignment. (b) Who levies what taxes? - the question of revenue assignment. (c) How is the (virtually inevitable) imbalance between the revenues and expenditures of subnational governments that results from the answers to the first two questions to be resolved? - the question of vertical imbalance. (d) To what extent should fiscal institutions attempt to adjust for the differences in needs and capacities between different governmental units at the same level of government? - the question of horizontal imbalance, or equalization. 1.8. Ideally, these questions should be approached in the specific circumstances of each country with the aim of achieving the relevant policy objectives - not only the normal public finance trio of efficiency (allocation), equity (distribution) and stabilization but also economic growth as well as such nebulous (but politically resonant) goals as "regional balance." In many instances, of course, there will be conflicts not only between these objectives but also between local and central perceptions of the weights to be attached to them. 1.9 Moreover, intergovernmental fiscal policies must be developed taking into account both political constraints (e.g. the strength of different regions and groups in political decisions) and economic constraints (e.g. the stage of development of financial markets) facing policy-makers. Finally, all policy changes proposed must start from the given set of initial conditions: every country has a history, and the current state of its fiscal institutions in large part reflects the results of the process of policy change over time. Expenditure Assignment 1.10 The question of expenditure assignment, is not discussed extensively in the present report. How government functions should be divided among levels of government is a complex matter, to which each country has its own unique solution. Basically, Argentina has assigned major expenditure responsibilities in the area of education (primary and secondary), health, justice  4 Annex 1- Restructuring Fiscal Federalism (police), and urban services to provincial and local governments. On the whole, this assignment is reasonable, in line with theory and international practice. 1.11 Other important questions could use further study - e.g. the division of functions between provinces and municipalities, variations in the efficiency (or lack of it) with which important services such as education are provided, the degree of national interest in the efficient and effective provision of services (education, some health) related to the level of "human capital" embedded in potentially migratory workers, and so on. Such matters must largely be left aside in the present discussion, however. Revenue Assignment 1.12. Who levies what taxes is one of the major issues in Argentina at present. This question is discussed at more length in the next section. A number of important issues of tax policy design and implementation must be considered in Argentina. The "correct" revenue assignment in a multi-level government structure, particularly in an "emerging" federation like Argentina (Box 1), is by no means clear in principle, and is certainly likely to prove highly controversial in practice. 1.13 The fundamental problems are two. First, the central government can inherently collect most taxes more efficiently than can provincial governments. Second, the potential tax bases that can be reached by the latter vary widely from province to province. The first of these problems gives rise to vertical imbalance; the second produces horizontal imbalance. To some extent these problems may be alleviated by e.g. using provincial surcharges on national taxes, as discussed below. Most of the desired (and desirable) aims of decentralized revenue policy can be achieved solely by allowing variation of the rates of such surcharges, perhaps subject to a constraint on minimum rates to restrict competition for tax base. In addition, since "tax exporting" breaks the critical link between provincial spending and residents' tax burden, care should also be taken to prevent provinces from exporting their tax burdens - e.g., limiting access to the taxation of business. ' Vertical Imbalance 1.14. No matter how revenues and expenditures are reassigned, a problem of vertical imbalance will almost certainly remain, or so worldwide experience suggests. Even if the tax base of the richest province enabled it to balance "own" revenues and expenditures, imbalances would remain for all the rest. Moreover, history suggests that the differential elasticity of expenditures and revenues assigned to different government levels would soon lead to a re-emergence of a vertical imbalance problem even for the richest province. One way or another, structural "gaps" will emerge and have to be dealt with. 'Parenthetically, all of these questions - like those on expenditure assignment - should also be considered with respect to the division of functions and finances between provincial and muncipal governments, although this aspect cannot be further considered here.  Annex 1: Restructuring Fiscal Federalism in Argentina 5 1.15 Tfie most common alternative is to have centrally-collected revenues transferred to provincial governments. Such transfers are of course the major source of provincial revenues in Argentina today, and this situation seems unlikely to change significantly for most provinces in the future. Together with a properly designed provincial (and municipal) tax base, a properly designed transfer system is therefore a key element in the needed rethinking of Argentina's federal finance system. Section 4 report discusses the major issues that need to be considered in this respect. 1.16 It may be useful to emphasize two key points here: First, all transfers (not just "coparticipation") must be taken into account when discussing vertical imbalance: a peso transferred is a peso transferred, no matter what label is attached to it. Second, as many writers on fiscal federalism in Argentina have properly stressed, a critical aim in redesigning the transfer system is to ensure that adequate incentives for "fiscal responsibility" are provided. Accountability is the public sector equivalent of the "bottom line" in the private sector (see Box 2). Without adequate accountability public sector decision-makers are unlikely to make effective and efficient resource allocation decisions (to the extent that any political process can produce such decisions). Box 2. Accountability as the "Bottom Line" If decentralization is to work, those charged with providing local infrastructure and services must be accountable both to those who pay for them and to those who benefit from them. Unfortunately, enforcing accountability at the local level is not always easy. It requires clear incentives from above and the provision of adequate information to local constituents as well as the opportunity for them to exercise some real influence or control over the service delivery system. "Informal" organizations almost by definition must be structured like this or they cannot exist. But it can be a challenge in the political and social circumstances of many developing countries to introduce a similar degree of responsiveness into formal governmental organizations. Accountability is the key to improved public sector performance, and information is the key to accountability. The systematic collection, analysis, and reporting of information that can be used to verify compliance with goals and to assist future decisions is thus a critical element in any decentralization program. Such information is essential both to informed public participation through the political process and to the monitoring of local activity by central agencies responsible for supervising and (usually) partially financing such activity. Unless local "publics" are nmade aware of what is done, how well it is done, how much it cost, and who paid for it, no local constituency for effective government can be created. Unless central agencies monitor and evaluate local performance, there can be no assurance that functions of national importance are adequately performed once they have been decentralized. An important accompaniment of any decentralization program is thus an improvement in national evaluation capacity. Decentralization and evaluation (e.g. cost-benefit analysis) are not substitutes; they are complements. An essential element of the "hard budget constraint" system needed to induce efficient local decisions is adequate central enforcement capacity in the shape of credible infonnation-gathering and evaluation. The "carrot" of central financial support of local efforts must be accompanied by the "stick" of withdrawn support if performance is inadequate, which of course requires both some standard of adequacy and some way of knowing whether performance is satisfactory.  6 Annex 1- Restructwrinf Fiscal Federalism 1.17 The critical point in this respect is accountability at the margin i.e. it is perfectly possible (in principle) for a provincial government to be 90% dependent of federal transfers and still be fully accountable - to its citizens and/or the central government, depending on circumstances. For this reason, the best form of intergovernmental transfer is one the amount of which is fixed in advance, i.e. will not be altered as a result of any (in-period) action by the recipient. Such a transfer by formula implies that at the margin, local actions to raise or lower local revenues or expenditures will directly affect outcomes - which is what is needed to ensure political accountability.2 Equalization. 1.18 Horizontal imbalance, or equalization, is the most controversial, and least amenable to analytical resolution, of the four "big questions" and it is discussed more extensively in section 4 below. At this stage, it may suffice simply to make three comments. (a) Interpersonal and interregional equalization are different matters, and there is no necessary connection between the two (see Box 3). Indeed, it is a serious mistake to attempt to achieve interpersonal distributional goals through interregional transfers: the result is likely to be both redistributional failure and an unduly complex and ineffective transfer system. (b) An equally important distinction must be drawn between equalization and regional development (or "balance") policies: again, there is no necessary connection between the two. In the present paper, to keep matters simple, it is assumed that transfer policies have (and should have) no specifically "regional development" objectives.3 (c) Finally, as discussed later, a properly-designed (and simple) equalization formula can and should provide all the incentives that are needed (or desirable) for "fiscal effort."4 2Two exceptions to this rule, discussed later, are when transfers are explicitly intended either to ensure the provision of specific services at specific levels (i.e. the transfers are essentially payments to provinces acting as agents: see Box 14) or to induce provinces to provide more of certain services than they would otherwise do (see Box 11). 3This is clearly an oversimplification: for more discussion, see Bird (1982, 1984) 4See also the discussion of fiscal effort in Box 10. Incidentally, an equalization formula can also, if desired, achieve the vertical balance goal discussed above. That is, rather than considering the questions of "primary distribution" (federal-provincial) and "secondaxy distribution" (interprovincial) separately, they can be resolved simultaneously (as in e.g. Canada) if desired. This point is not pursued fiuther here, however.  Annex 1: Restructuring Fiscal Federalism in Argenhna 7 BOX 3. Interpersonal Redistribution and Intergovernmental Transfers Some analysts assess intergovernmental transfers in part by how successfully they serve as proxies for direct interpersonal distnbutional measures (Porto and Sanguinetti, 1995). Although most transfers to provinces are unconditional in Argentina, the relationship of transfers and poverty is broadly positive, with poorer provinces receiving more support from the central government, though it is not the very poorest that gain most. The regional redistributive process seldom runs smoothly, however. For instance, the data in Porto and Sanguinetti (1993) indicate that per capita transfers to the poorest provinces (Chaco, Formosa and Santiago del Estero), which have around 40% of their populations under the poverty line, are only slightly higher than the average per capita transfer to all provinces. On the other hand, some wealthier provinces received almost double the average per capita transfer. In many countries, transfers, particularly those for investment purposes, are clearly influenced by politicaUy-related territorial criteria. In Chile, for example, most regional investment is done by the central government (72%) and, as in Argentina, favors frontier regions in the south with low population density (Aisen and Magalanes). Nonetheless, the poorest regions of Chile (Maule and Biobio) received investment per capita close to the average alocated by the central government in 1991 (using data from Espinoza and Marcel, 1993). Of course, mnany poor families live in rich regions. Over half the Chileans with incomes below the poverty line live in the wealthiest regions of the country - the Metropolitan Region of Santiago, Valparaiso and Libertador (all in Central Chile). Tenitorially based transfers are inherently inefficient in reaching these poor families. In principle, the greater the weight that the central government places upon the equitable delivery of essential services to specific groups of poor citizens, the less the weight that can be given to the autonomy of local governments to spend as they see fit. If a principal aim of central policy is to deliver "basic needs" to the poor, either the central government should do it itself, or it will have to develop an elaborate and detailed (and probably not very effective) monitoring system to ensure that local governments perform their role as agents of central policy efficiently and equitably. Decentralization, properly carried out, may have many virtues But doing exactly what the central government wants in terms of income redistribution is not likely to be one of them. International experience suggests that the main guideline for those who would decentralize successfully, while at the same time focusing on poverty-alleviating policies, should be do not conmlicate ntergovenunentalfinance unnecessariy. If the central government wants to deliver specific services to specific (poor) households, it should do so if it can without further complicating intergovernmental finance. Many of the complications, and complaints, characterizing intergovernnental fiscal issues in most countries result from overloading the system with tasks for which it is ill-equipped, such as targeted poverty alleviation. Whenever feasible, direct tansfers to the poor are better than indirect transfers to localities - even poor localities - that are intended primarily to help poor households. In short, if the provision of a nationwide basic unifonn level of health or educational services is an important objective of national policy, the national government should either provide such services itself or directly (e.g. through some variant of vouchers) transfer the needed resources to the target population. Policy Objectives 1.19. Finally, although the point is probably clear enough in what has already been said in this section, it may be useful to emphasize that the discussion of provincial taxes and transfer design in sections 3 and 4 simply assumes that the principal concerns of policy in this area are, essentially,  8 Annex 1- Restructuring Fiscal Federalism the usual public finance triad of equity, efficiency, and stabilization, subject to the usual public choice constraint of political feasibility in the special context of an emerging federal state. 1.20. Equity is considered here only in the limited sense of "equalization" as understood in the usual fiscal federalism literature. Stabilization is taken into account largely in discussing the starting point for the simulation exercise described later (deciding the size of the distributable pool) and how this total might be adjusted over time. The main focus here is on efficiency - and, indeed, even equalization as understood here can be interpreted as an "efficiency" concept.5 1.21. "Efficiency" as used in the discussion of intergovernmental fiscal matters, however, is a label that covers many, concepts - not simply allocative efficiency within the public sector, but also the broader concept of the efficient division of resources between the public and private sectors (political efficiency or accountability) and the narrower concept of administrative efficiency, as well as the real "bottom line" as far as citizens are concerned - the effective delivery of desired public services. The most critical issue in this respect, although not one that can be developed at length here, is undoubtedly the broad concept of political efficiency i.e. who (in an ideal world) determines what should be done - the local political community (and how is efficiency to be secured if resources come from "other people's money"?) or the central government (monitoring, conditionality, and other means of achieving so-called "incentive compatibility)? This debate constitutes an important underlying theme in the following discussion. Provincial Taxation Introduction 1.22 In many ways the key to a sound decentralized fiscal structure lies in the revenue structure of subnational governments. If, as in Argentina, the provinces are responsible for the provision of such important public sector activities as health and education, sound policy would suggest that they should be able to levy sufficient taxes to meet the demands for such services - provided, of course, that those who make the demands are also those who have to pay the taxes. While the needs of national governments and considerations of administrative efficiency and feasibility generally mean that this aim of making each level of government fully responsible for financing its own activities is unattainable, it is sometimes possible to approximate this goal at least for the richest subnational governments. The extreme variation in the economic bases and characteristics of Argentina's provinces (see Table 4) means, however, that transfers are likely to remain the dominant source of revenue for many poorer provinces, no matter how amply taxing powers are bestowed on the provinces. 1.23 At present, most taxes in Argentina are collected by the central government (Table 1). The only tax field where provincial and local taxes dominate is with respect to taxes on assets - the real property tax and motor vehicle taxes. The abolition of the national assets tax in 1994 5For further discussion of this fianework, see Bird (1993); also Boadway and Flatters (1982).  Annex 1: Restructuring Fiscal Federalism in Argentina 9 means that this field has effectively become 100% subnational. The only other field where subnational taxes are important (apart from the social security taxes, which are not discussed here) is with respect to domestic consumption taxes - the gross receipts tax and the stamp tax. Table 1. Tax Collections, by Level of Government and Type of Tax, 1994 Percent Tax National Provincial Municipal Total Income 100.00 0.00 0.00 9.36 Social Security 77.54 22.46 0.00 26.95 Property 14.26 64.61 21.13 6.00 Consumption 74.09 21.83 4.08 47.02 Trade 100.00 0.00 0.00 4.49 Other 74.45 15.80 9.75 6.18 Total 75.03 21.17 3.80 100.00 Source: Instituto de Economia y Finanzas, Universidad Nacional de Cordoba, based on data from Secretaria de Progamacion Economica y Secretaria de Asistencia de Reforma Economica Provincial 1.24. Municipal governments - excluding the Municipality of the City of Buenos Aires, which is included with the provincial level in Table 1 - are unimportant as taxers, and what taxes they do collect are almost entirely in the form of supplementary levies on provincial taxes on gross receipts, real property, and automobiles (see Box 4). For this reason, although there are considerable variations between provinces in the role and importance of municipal governments, the discussion in this section focuses entirely on provincial taxes. 1.25. Unfortunately, the present provincial (and local) revenue structure in Argentina is, and seems likely to remain, inadequate for even the richest provinces to become essentially self- financing. Basically, the provinces have only three important taxes: on automobiles, on real property and on sales (see Box 5). By far the most important of these is the tax on gross receipts (Table 2), 60% of which is collected in the province and municipality of Buenos Aires, with another 20% of the total being collected in the three other "big" provinces - Mendoza, Cordoba, and Santa Fe. The taxes on automobiles and property could certainly be strengthened in design and implementation, but even at their best these two sources of revenue will almost certainly remain both relatively small and relatively inelastic. The only potentially important tax the provinces possess is the gross receipts tax, which (together with the stamp tax) is currently undergoing major reform. This section, after a brief discussion of general criteria for subnational taxes, discusses in turn the automobile tax, the property tax, and the sales tax. The section concludes with a brief review of other possible sources of provincial revenue.  10 Annex 1- Restructuring Fiscal Federalism Box 4: Municipal Taxes and Fees Argentina is unusual in that, stnctly speaking, municipal governments are not allowed to levy any taxes at all. Insad, they are restricted to imposing "tasas" or fees for services. In practice, however, most of these "tasas" appear to resemble local taxes more than fees for specific services rendered. In 1994, 51% of municipal "own" revenue came from (in effect) surcharges on the provincial gross receipts taxes, 23% from taxes on real property, and 11% from taxes on automobiles, with the balance from a vanety of other minor sources. To put this another way, in 1994, although municipalities levied less than 4% of all taxes, they accounted for 18% of total gross receipts taxes, 24% of real property taxes, and 25% of automobile taxes. In some respects, the apparent idea underlying muncipal finance in Argentina makes a good deal of sense. Local governments can and should from an economic point of view be viewed as, in effect, "firms" delivering packages of local public services to residents. Like all firms they should ideally be financially self-sufficient in the sense that people should want what they deliver enough to be willing to pay for it. From this perspective, the first rule of local finance should be: "Wherever possible, charge." For efficiency, charges should be levied on the direct recipients of benefits, whether residents, businesses, or "things" (real property). While user charges are likely to be viewed by local officials solely as a potential additional source of revenue, their main economic value is to promote economic efficiency by providing demand information to public sector suppliers and to ensure that what the local public sector supplies is valued at least at (marginal) cost by citizens. This efficiency objective is particularly important at the local government level since the main economic rationale for local government in the first place is to improve efficiency. Whenever possible, local public services should therefore be charged for - of course, at properly-set prices - rather than given away. Although in most countries much less use is made of charging at the local level than seems desirable, and many of the charges that are levied are poorly designed from an efficiency point of view, at least three types of local "charge" revenue exist almost everywhere: (1) service fees, (2) public prices, and (3) specific benefit charges. "Service fees" include license fees (marriage, business, dog, vehicle) and various small charges levied by local governments essentially for performing specific services - registering this or providing a copy of that - for identifiable individuals. Charging people for something they are reqwured by law to do may not always be sensible - for example, if the benefit of (say) registrating births or deaths is general and the cost is specific - but on the whole there is seldom much harm, or much revenue, in thus recovering the cost of providing the service in question. "Public prices" include the revenues received by local governmuents from the sale of private goods and services (other than the cost-reimbursement just described). In principle, prices of locally-provided services to identifiable private individuals - whether public utility charges or admission charges to recreation facilities - should be set at the competitive private level, with no special tax or subsidy element included. A final category of charge revenue is "specific benefit" taxes. They are (in theory) related in some way to specific benefits supposedly received by specific taxpayers. Examples are: special assessments, land value increment taxes, improvement taxes, front footage levies, supplementary property taxes related to the provision of sewers or streetlighting, development exactions and charges, delineation levies, and so on. Most such charges are imposed either on the assessed value of real property, changes in that value, or on some characteristic of that property - its area, its frontage, its location. Such charges are common in Argenina e.g. in 23 municipalities in the Province of Buenos Aires within the "conurbation" of Buenos Aires, 4 charge for street lightng and cleaning on a frontage basis, 3 on the basis of a fixed charge per residence, and 1 on a combination of the first two bases and 1 on a combination of frontage and fiscal value. On the other hand, the other 14 all charge on the basis of a surcharge on the provincial property tax base (Provincia de Buenos Aires, Noticias de Economia, marzo y abril 1995, pag.29).  Annex 1: Restructuring Fiscal Federalism in Argentina 11 Table 2. Argentina: Major Provincial Taxes, 1994 Tax Perent of Toal Gross Receipts 56.2 Stamp 10.1 Property 17.4 Automobiles 9.9 Other 6.4 Total 100.0 Source: Ministerio del Interior, Secretaria de Asistencia para la Reforna Economica Provincial, Estadisticas de Recaudacion y Coparticipacion. Criteria for a Good Provincial Tax 1.26. Two basic principles of assigning revenues to subnational governments may be suggested. * First, "own-source" revenues should ideally be sufficient to enable at least the richest subnational governments to finance from their own resources all locally-provided services primarily benefiting local residents. o Second, to the extent possible, subnational revenues should be collected from local residents only, preferably in relation to the perceived benefits they receive from local services. * More specifically, among the characteristics that may be sought in an "ideal" subnational tax are the following (Box 6): * The tax base should be relatively immobile, to allow local authorities some leeway in varying rates without losing most of their tax base. * The tax yield should be adequate to meet local needs and sufficiently buoyant over time (i.e., it should expand at least as fast as expenditures). * The tax yield should be relatively stable and predictable over time. * The tax base should be visible, to ensure accountability. * The tax should be perceived to be reasonably fair by taxpayers. * The tax should be relatively easy to administer efficiently and effectively. 1.27 On the whole, is much to be said for allocating taxes on property to lower levels of government, as is done in Argentina. Similarly, a case can also be made for allocating either a "business" tax (like the present tax on gross receipts) or a sales tax (like the proposed retail sales tax) to the provinces. On the other hand, an even stronger case can be made for allocating an  12 Annex 1- Restructuring Fiscal Federalism income tax (in the form of a surcharge) to the provinces. Moreover, both experience in other countries and theoretical considerations suggest that there may be much to be said in favor of giving provinces power to levy a few excise taxes e.g. on alcohol, tobacco, and fuel in addition to the existing tax on autos. These points are discussed briefly at the end of this section. Box 5: What is a 'Provincial Tax? The question posed in the title of this box may seem odd, but in fact it is by no means always clear who has, so to speak, "ownership" of a particular tax source. In principle, a "truly provincial" tax might perhaps be defined as one (i) assessed by provincial governments, (ii) at rates decided by provincial governments, (iii) collected by provincial governments, and (iv) with its proceeds accruing to provincial governments. In the real world, many taxes may possess only one or two of these characteristics, and the "ownership" of the levy may be unclear. In Argentina, for example, through the "coparticpation" system, many taxes accrue in part to the provinces, but their rates (and bases) are determined by the national government, which also assesses and collects the taxes. For most purposes, as is indeed commonly done in Argentina, such taxes might be considered to be really central government taxes accompanied by transfers allocated to the provinces. This interpretation is plausible because (for the most part) there is little connection between the amount transferred and the amount collected locally. On the other hand, what looks to be a central tax and a related transfer program may be considered to be a provincial tax from some perspectives. If, for example, the provincial government determines the tax base and rate and receives all the revenues, then the only role the central government plays is as a collection agent: it may even be reimbursed for its work for instance, by being allowed to retain a small percentage of collections. Presumably, the central government has a comparative advantage in tax collection, and the provincial government has contracted for its services in this respect. In this case, there is no intergovernmental transfer. Although constitutionally some interpret the present Argentine system of "revenue-sharing" along these lines - because in effect the provinces have delegated much of their revenue-raising power to the centre and are, so to speak, "compensated" through coparticipation - this interpretation is not economically meaningful because the revenues are basically not distributed on a derivation basis and all the relevant tax rates and bases are entirely determined by the national government For this reason, the only "provincial" taxes discussed in this report are those for which the provinces have authority to set bases and (within limits) rates, as well as to collect the taxes and spend the proceeds, namely, the taxes on real property, automobiles, and gross receipts. In principle, however, as discussed briefly in the report provinces rmight levy surcharges on national taxes, which would be collected by the national government and the proceeds remitted to the province which imposed the surcharge. Such surcharges would be "provincial" taxes in terms of political accountability even though the national government still determined the tax base and actually collected the tax.  Annex 1: Restructuring Fiscal Federalism in Argentina 13 Box 6: The Choice of Subnational Taxes part from transfers, subnational government taxes around the world may be divided into five groups: property taxes, business taxes (of mnany sorts), sales taxes, excises, and income taxes. The matrix depicts some of the salient characteristics of these revenue sources in terms of a number of possible criteria. While this presentation is obviously impressionistic, it reflects the experience of a wide variety of developing countries. PropertyTax Income Tax Sales Tax Excises Business Tax Mobility + Adequacy + - + ? Buoyancy + + + + Stability + + Exportability +/- +/- + + - Visibility + + + + l Fairness + + ? ? - Acceptability - - ? + Administration ? + ? + + "Mobility" refers to the mobility of the tax base. "Adequacy", "buoyancy", and "stability" to characteristics of the tax revenues. "Fairness" refers to the conventional notion of tax progressivity, and "administration" to the ease with which such a tax can be administered at locally-determined rates. "Acceptability" is a guess as to the political popularity of the different taxes and is clearly related to "visibility" on the one hand and the "exportability" of the tax to non-residents on the other. The perspective taken in assigning the values in the table is that of the benefit model of local government A "+ l" means that the tax is good, a "-" that it is bad, and a "?" that it is indeterminate. A "+/-" means that the tax is good to the extent it falls on residents and bad to the extent it falls on non-residents. The income tax is assumed to be easy to administer locally as a surcharge on the national tax. The property tax is assumed to be costly to administer. The "excise" column assumes that both fuel and vehicle taxes (basically progressive) and taxes on alcohol and tobacco (regressive) are imposed. Automotive Taxes 1.28. The taxation of motor vehicles may be justified for three distinct reasons, and correspondingly divided into three separate categories. First, such taxes may provide significant revenue: in Argentina in 1994, for example, provincial and municipal automobile taxes accounted for 1.6% and national taxes on fuels for 3.3% of all taxes, or approximately 5% in total.6 Second, such taxes may have significant distributive aspects given the presumably high income elasticity of private motor vehicle ownership and usage. Third, since motor vehicles are of little use without 6In addition, an unknown share of collections from the VAT and gross receipts taxes are also of course attributable to automobile sales.  14 Annex 1- Restructuring Fiscal Federalism roads, and roads are basically publicly-provided, there is an obvious strong benefit linkage (see Box 4) between automobile taxation and expenditure on streets and highways.7 1.29 The most important tax from a revenue perspective is the fuel tax, which is also the simplest and cheapest form of automotive taxation from an administrative perspective. Much as the national government appreciates the revenue it receives from this source, there is no reason why fuel taxes could not equally well be levied at the provincial level. Different provinces could impose different taxes, subject the constraint that they would not likely be able to differ much from the rates imposed by their neighbors (e.g. in the Buenos Aires area), given the mobility of the tax base. Administratively, differential provincial fuel taxes could easily be imposed at the refinery or wholesale level, with the refiner or wholesaler acting a collection agent for the provinces and remitting the taxes in accordance with fuel shipments. If an option is to increase provincial "own" revenues (at the expense of national revenues, but presumably with a corresponding drop in national transfers8), a fuel tax would be one of the easiest ways to do so. 1.30 Fuel taxes are related both to road usage and to such external effects of vehicles as accidents, pollution and congestion, although not in any very precise way. To the extent automotive taxation is intended to "price" either the utilization of publicly-provided services or externalities, fuel taxes are at best a crude instrument. Toll roads on the one hand (like the new highway to La Plata) - i.e. "privatization" as it were - and a redesigned set of annual automobile (and driver) license fees can serve this "benefit tax" function much better. Such fees might be based on such features as e.g. age and engine size of vehicle (older and larger cars generally contribute more to pollution), location of vehicle (cars in cities add more to pollution and to congestion), driver records (20% of drivers are responsible for 80% of accidents), and especially the axle-weight of the vehicle (heavier vehicles do exponentially more damage to roads and require roads that are more costly to build). 1.31 Finally, to the extent that it is desired to achieve some redistributional goal through automotive taxation, this can best be done by a national excise tax at the time of initial sale. The present Argentine system, in which every province levies an array of annual taxes that vary with the year and mrodiel of the 'eh,ic)e; a:pare5v!; 'argely in the attempt to levy a "progressive" tax, makes little sernt.&. -is appo-.- is admiinistratively complex and costly; it is not related in any consistent way to any distributional objective; it unduly penalizes newer (and more efficient) vehicles; and it does not "price out" public services or externalities in any meaningful way. 1.32. Provincial taxation of automobiles is a good idea. More specific study of the appropriate design of the automobile tax system in Argentina seems warranted, but better-designed annual license charges could appropriately be set at the provincial level, perhaps within national "guideline" parameters. More res-evue, anG 6etl.er econom'c effec.ts. could be achieved through a revised system of auton;kot;ve ta7: _.r, a: .pro' incial r-venues c(;uk-A be stidl iurther boosted by giving the provinces access to trie iuel tax (through surcharges). Apart from the benefit linkages 7For extensive discussion of this question, see Heggie (1995). 8At present, some fuel tax revenues are channeled into various funds related to provincial investment (see Figure 1). This mnatter is not discussed further here, however.  Annex 1. Restructuring Fiscal Federalism in Argentina 15 already cited, automotive (and fuel) taxation appears to be the only provincial revenue source that, in principle at least, should demonstrate more than unitary income-elasticity, thus matching this aspect of some of the key services (education, health) for which the provinces are responsible. Box 7: Limits to Property Taxation While it is common to note the low level of property taxes in countries such as Argentina and to argue that there is thus obviously much room for increased local "fiscal effort" in this respect. While there is truth in this afirmation, experience everywhere suggests that it is neither quick nor easy to obtain additional revenues from property taxes. The property tax is the most widespread form of local taxation. Unfortunately, experience suggests that such taxes are not easy to administer and that they are never politically popular owing to their visibility and to certain inherent administrative difficulties. Even in the most sophisticated countries, local property taxes seldom yield enough to finance local services. No developed country which depends significantly upon property taxes for local fiscal resources has a local govemment sector that accounts for more than 10 percent of total public spending (Bird and Slack, 1991). Moreover, despite substantial efforts in some countries and considerable foreign assistance, these figures have not changed much over the years (Dillinger, 1991). The property tax may be a useful source of local revenue, but it is unlikely to provide sufficient resources to finance a significant expansion of local public services in any country. Indeed, many countries have been hard-pressed even to maintain the present low relative importance of property tax revenues in the face of varying price levels and political difficulties. A number of conditions must be satisfied for local property taxes to play a more important role in financing local activities (Dillinger, 1991). The political costs of reliance on the property tax are so high that no government with access to "cheaper" sources of finance will willingly do so. Like access to intergovernmental transfers to cover local spending, access to taxes on business which can largely be exported must be curtailed to make property taxes more attractive and, to confront local decision-makers with the true economic (and political) costs of their decisions. Even if this essential structural pre-condition is met. a number of other policy reforns are needed to turn the property tax into a responsive instrument of local fiscal policy. First, and importantly, local governments must be allowed to set their own tax rates: with the Pacto Fiscal's limits, this is no longer the case in Argentina. Secondly, the tax base must be maintained adequately, which is easier to say than it is to do in any country. Finally, a series of procedural reforms is often needed to improve collection efficiency, valuation accuracy, and the coverage of the potential tax base (Kelly, 1994). One reason for the widespread resistance to the property tax is because it is a very visible tax, for several reasons. First, unlike the income tax, the property tax is not deducted at source but generally has to be paid directly by taxpayers in periodic lump sum payments. Taxpayers who pay taxes directly to govermment tend to be more aware of the size of their tax bill than those whose take-home pay is reduced by weekly or monthly tax deductions. The need to make such periodic large payments may well add to the accountability and responsibility of governments, but it also greatly increases the sensitivity of taxpayers to even nominal increases in taxes. Secondlv, the inelasticity of the property tax has a similar effect. Since the base of this tax does not as a rule increase automatically over time, the periodic nominal increases in property tax bills needed to maintain real revenues when price levels rise require increased tax rates. In terms of political accountability, this need to confront the people with the cost of government represents a virtue of the property tax; however, the downside (from the government's point of view at least) is the heightened visibility of nominal tax increases and the accompanying political resistance. Thirdly, property taxes of course finance such services as education, roads, and garbage collection. The quantity and quality of these services (or their absence) is thus readily linked to the property tax. When potholes develop in their set, taxpayers are understandably quick to question the taxes that supposedly finance street repair. Once again, the very feature that makes the property tax a good source of local government revenue in principle makes it especially vulnerable to political resistance in practice.  16 Annex 1- Restructuring Fiscal Federalism Property Taxes 1.33. Argentina is unusual in the extent to which the provinces, rather than the municipalities, collect taxes on real property. Although there appear to be considerable variations in the property tax from province to province, on the whole it is probably fair to conclude that (1) undoubtedly more revenue could be collected from this source in most provinces but that (2) it is almost certainly unrealistic to look to property taxes either for a substantial increase in revenue or as a source for funding all or most of provincial expenditures on such services as education and health (Box 7). 1.34. A rather odd feature of the Pacto Fiscal (Decreto 1907/93) is that it imposes maximum property tax rates (1.2% on rural, 1.35% on suburban (and "subrural"), and 1.5% on urban properties) and further decrees that the tax base in no case can exceed 80% of the market value of urban or suburban, or the value of the land alone for rural, properties.9 That is, the maximum rates that can be established by any province vary from 0.96% of market value for rural to 1.2% for urban properties. This feature is not justified because in principle there is no reason why any maximum needs to be established at all with respect to residential properties - though there may be a case for a maximum (to restrain tax exporting) as well as a minimum (to restrain tax base competition, if desired) with respect to industrial and commercial properties. 1.35 Regardless of the merits of this requirement, however, its existence implies that it would be fair at the present time to assume that the effective property tax rates currently in force in most of Argentina are below these limits. Casual examination of the systems in Santa Fe, Mendoza, and Buenos Aires provinces support this conclusion. In the city of Santa Fe, for example, a recent sample survey found that two-thirds of the lots shown as "vacant" in the tax records actually had buildings on them, while of the lots shown with buildings, 46% actually had more built-up area (30% more) than recorded. In total, 56% of all properties were underrecorded in the records, and the true area of improvements was at least 40% greater than recorded.. Another sample in a nearby municipality (Santo Tome) similarly found 52% of properties underrecorded in one way or another. 1.36. Although there is no information that permits linking these physical facts to values, these figures suggest that it might not be unreasonable to expect at least a 30%-40% increase in the tax base as a result of an improved and updated fiscal cadastre. An increase of around this magnitude appears to be anticipated in Mendoza (where it was said that the tax base was already close to "market values."). Similarly, a 1994 amnesty in Buenos Aires Province reportedly led 400,000 taxpayers (including 170,000 whose land was vacant according to the tax roll) to report 30 million square meters of construction previously unknown to the authorities (Provincia de Buenos Aires, Noticias de Economia, setiembre/octubre 1994, page. 7). 1.37 As this fragmentary evidence suggests, there is undoubtedly considerable room to improve collections from provincial (and municipal) property taxes in Argentina. All the usual techniques 91n addition, the Pacto "recommends" that the municipal surtaxes on the property tax should not exceed 0.4% or 80% of the value, or 0.32%.  Annex 1: Restructuring Fiscal Federalism in Argentina 17 could be used for this purpose10: up-to-date mapping; better and more consistent application of well-established valuation techniques; improved flows of information from Property Registries, local building license authorities, public utilities, etc.; improved collection procedures - none of the offices visited seemed to have the knowledge one would expect of the distribution of past liabilities; and of course improved enforcement against non-payers. All these measures would help and, within reason, no doubt all should be used, albeit with more emphasis than seems evident in Argentina at the moment on improving the "sharp end" - collection and enforcement - rather than the technically more costly (and less immediately productive in terms of revenue) mapping and surveying end. Gross Receipts Tax 1.38 The major revenue sources of the provinces have traditionally been an antiquated gross receipts tax levied at various rates on different activities and an even more antiquated stamp tax. The Pacto Fiscal requires the stamp tax to be abolished and the gross receipts tax changed to a "oretail" sales tax. Considerable steps have already been taken in this direction in most provinces e.g. by exempting primary and some industrial activities as well as suppressing most of the stamp tax. As of 1995, the tax on primary production had been eliminated completely in 6 provinces and reduced to rates of 1% or less in 20 of the 24 (including MCBA) provinces, while the tax on industry had generally been lowered to 1.5%, compared to the general rates of 2.5% on wholesale trade, and 3.5-3.5% on retail trade and services."1 After some delay, the move to a "retail" sales tax is supposed to be completed throughout the country in 1996. 1.39. From the point of view of reducing the economic distortion of the tax system, this change is obviously desirable. From the point of view of strengthening provincial revenues, however, the effects of replacing the existing stamp and gross receipts tax are less clearly desirable. The national government has estimated that a 3.5% rate on retail sales would produce the same revenue as the existing tax, but other estimates suggest that the required replacement rate on the feasible tax base may be as high as 6-7% on average, with the Municipality of Buenos Aires and some provinces (e.g. San Luis) requiring rates as high as 10%.12 Clearly, adding a provincial retail tax (or VAT, which in theory has essentially the same base) to an existing federal VAT of over 20% is not something which is either obviously desirable or likely to be politically or administratively feasible, to say the least. For this reason, there is concern over the possibly deleterious effects on provincial finances of the tax changes mandated by the Pacto Fiscal. 10 '