Report No. 16996-AR Argentina The Fiscal Dimension of the Convertibility Plan A Background Report (In Two Volumes) Volume l: The Main Report January 22, 1998 Argentina Chile and Uruguay Country Management Unit Poverty Reduction and Economic Management Unit Latin America and the Caribbean Regional Office Document of the World Bank Currency Equivalents Currency Unit: Peso (As of December 1997) US$1 = ARG$1 Fiscal Year January I - December 31 Glossary of Acronyms ANSeS Social Security Public Program AMPO Minimum Contribution to the Public Pension System ATN Treasury Grants BCRA Central Bank of the Republic of Argentina BHN Public Mortgage Bank BOCON Government Debt Consolidation Bonds BONEX Dollar denominated Treasury Bonds CPI Consumer Price Index EFF Extended Fund Facility EPH National Houseliold Survey FEDEI National Electric Development Fund FIEL Foundation for Latin American Economic Research FONAVI National Housing Fund GDP Gross Domestic Product HUBN Households with Unsatisfied Basic Needs IMF International Monetary Fund INDEC National Statistics & Census institute LIBOR London Interbank Offer Rate MCBA Buenos Aires (city) MERCOSUR Southern Cone Common Market NGO Non-govermmental Organizations OECD Organization for Economic Cooperation & Development PAYG Pay as you go PEP Partial Subsidy for Private Employment PYMES Small and Medium Enterprises USA United States of America VAT Value Added Tax YPF Oil Company Vice President: Shaliid Javed Burki, LAC Director: Myrna Alexander, LCC7C Lead Economist: Jolin Underwood, LCC7A Task Manager: Paul Levy, LCC7A COUNTRY DATA -ARGENTINA AREA POPULATION DENSITY 2766.9 thous. sq.km. 34.7 million (mid-1995) Country density 1991 11.7 hab.per sq.km 1.4% annual growth Rural density al 16.9 hab. per sq.km of arable land POPULATION CHARACTERISTICS at HEALTH bl Crude Birth Rate (per 1000-1995) 19.8 Population per physician (thous.) 0.4 Crude Death Rate (per 1000 - 1994) 7.5 Population per hospital bed (thous.) 0.2 Infant Mortality (per 1000 live births -1994) 22.0 INCOME DISTRIBUTION bt/ DISTRIBUTION OF LAND OWNERSHIP % of national income, highest quintile 51.0% % owned by top 10% of land owners % of national income, lowest quintile 5.0% % owned by smallest 10% of land owners ACCESS TO SAFE WATER (1993) ACCESS TO ELECTRICITY (1989) % of population - urban 73% % of population 95% % of population - rural 17% NUTRITION al EDUCATION Calorie intake as % of requirements 119.2% Adult literacy rate % (1980) 95% Per capita protein intake (grams per day) 99,7 Primary school enrollment % at 100% GNP PER CAPITA IN 1996 di 8,410 GROSS DOMESTIC PRODUCT IN 1996 of ANNUAL GROWTH RATES (% constant prices) USS Bill. % (current prices) of GDP 1965-73 1973-80 1980-95 1996 GDP at market prices 300.5 100.0 4.3 2.2 1.2 4.4 Gross Domestic Investment 55.6 18.5 6.8 4.3 11 5.4 Gross National Savings 51.6 17.2 2.5 2.3 0.4 2.5 Current Account Balance -4.0 -1.3 ExportsofGoods&NFS 27.1 9.0 4.7 14.1 6.2 8.6 Imports of Goods & NFS 27.9 9.3 0.6 13.3 5.3 10.0 OUTPUT, LABOR FORCE AND PRODUCTIVITY IN 1995 Value Added (constant prices) Labor Force fS V.A. Per Worker rg $ Thousand % of Total Thousands % Arg $ Agriculture 917 7.5 2,973 12.0 308 Industry 4,324 35.3 7,780 31.4 556 Services 6,996 57.2 14,023 56.6 499 Total GDP at Factor Cost 12,237 100.0 24,776 100.0 494 GOVERNMENT FINANCE gJ Federal Govemrnment Provincial Govemrnment Million Pesos % of GDP Million Pesos % of GDP 1996 1996 1996 1996 Current Revenues 46,000 15.3 Current Revenues 27,658 9.2 Current Expenditures 48,397 16.1 Current Expenditures 25,462 8.5 Capital Revenues 501 0.2 Capital Revenues 316 0.1 Capital Expenditures 3,564 1.2 Capital Expenditures 4,150 1.4 Surplus -5,514 -1.8 Surplus -1,829 -0.6 ad For the period 1982-1985. bf For the period 1970-1976. ci For the period 1987-1992. dt Current US dollars. Estimated using Bank Atlas methodology. et Current US dollar estimates, calculated from data in constant Arg $ 1986. ft Calculated by applying 1980 census shares to 1994 population. g/ Cash Basis in current Pesos, includes Central Administration, Social Security and net balance of Public Enterprises. Capital revenues includes privatization incomes. COUNTRY DATA - ARGENTINA MONEY, CREDITAND PRICES 1992 1993 1994 199! 1996 (Millions of Pesos; and of period) Money and Quasi MoneY 20,739 29,479 33.246 32.817 38,026 Domestic Bank Credit to Public Sector 9,098 9,020 9,428 8,157 9,137 Domestic Bank Credit to Private Sector 33,593 42.206 50,561 50,566 56,045 MoneyandQuasiMoneyas%ofGOP 9.2 11.4 11.8 11.7 12.7 Wholesale Price lndex (April t991t00) b/cl 104.7 1064 106.9 115.1 119.5 Annual pemcentsg changs in: General Wholesale Price Index dv 6.0 1.6 2.3 6.5 3.2 Bank Credit to Public Sector .. -0.9 4.5 -13.5 12.0 Bsnk Credit to PrrYe Sector ., 25.6 19.8 0.0 10.8 MERCHANDISE EXPORTS (Avearsga 1991 -196) &t BALANCE OF PAYMENTS af 1992 1993 1994 1995 1996 USS Mlin. % of Total Primary products 4073.5 25.0 Manuf. of agricultural ongin 6062.5 37.2 Exports of Goods, NFS 14,727 15,625 18,508 23,889 27,076 Manuf. of industrial origin 4513.3 27.7 Imports ofGoods. NFS 18,386 20,711 25,605 23.826 27,900 Fuels 1667.0 10.2 Resource Balance -3,659 -5,086 -7,097 63 .-24 Total Merchandise Exports 16316.3 100.0 Interest Payments (net) -2.695 -1,109 -1,277 -1,070 -1,417 OtherFactor Payments (net) .845 -1,273 -1.270 .1,871 -2,106 Total Public Debt Outsttnding & Disbured (End 1996)bI US$ Min. Net Current Transfers 749 411 320 432 334 Bdalnce on Curnt Account -6,450 -7.057 -9,324 -2,446 -4,013 Total 97105 IBRD 5316 Capital Account 10,162 11,408 9,868 2,377 7,795 IDB 4756 IMF 6279 Private Sector 8,632 7,065 6,376 -5,417 -1,940 Bilaterals 10162 Public Sector 1,530 4,343 3,492 7,794 9,735 Bonds 68341 Commercial Banks 1751 Changes in Gross Reserves I- = increase) 3,712 4.351 544 -69 3,782 DEBT SERVICE RATIO, 1996 55.3% RATE OF EXCHANGE IBRDIIDA LENDING, DECEMBER 31, 1996 (Mln. USS) bl USS i= Arg.S I IBRD IDA Outstanding & Disbursed 5316 a/ Source: Ministry of Economy br Source: INDEC cl Average index for the year. di Based on the average index for the year. TABLE OF CONTENTS EXECUTIVE SUMMARY ...............................................................i INTRODUCTION ...............................................................1 MAIN REFORMS UNDER THE CONVERTIBILITY PLAN . ...........................................................1 CHAPTER I: RECENT PUBLIC SECTOR PERFORMANCE .................................3 CONSOLIDATED FISCAL ACCOUNTS ..................................................................................3 Consolidated Fiscal Accounts on a Cash Basis .............................................................4 Composition of Federal Revenues and Expenditures .................................................7 Consolidated Fiscal Accounts on an Accrual Basis ........................................................9 Comparative Analysis ofAccounts on Cash and Accrual Basis .................................... 10 COMPARISON OF DEFICITS AND THE CHANGE IN PUBLIC SECTOR DEBT .............................. 12 FISCAL SOLVENCY AND SUSTAINABILITY ........................................................................... 15 SUMMARY AND RECOMMENDATIONS FOR FUTURE WORK .................................................. 16 CHAPTER II: THE FISCAL IMPACT OF STRUCTURAL REFORMS ............... 19 A. FISCAL IMPACT OF ARGENTINA'S PENSION REFORM .................................................... 1 9 Background to Argentine Pension Reform ............................................................ 19 Expected Fiscal Improvements The Basic Pension ................................................... 30 The Basic Pension ................................................................................. 30 The Basic Pension ................................................................................. 30 Labor Taxes and Pension Surpluses .......................................................................... 31 Social Assistance Needs ................................................................................. 31 Agenda for Future Work ................................................................................. 32 B. PRIVATIZING ARGENTINA'S PUBLIC PROVINCIAL BANKS .............................................. 33 Public Provincial Banks Performed Worse Than Either Privatized Provincial Banks or Private Banks ........................................................... 34 Privatizing Public Provincial Banks Forced Provincial Governments to Realize "Costs" Incurred Prior to Privatization ........................................................... 36 The Size of the Residual Entity Depends Upon the Quality of the Public Provincial Bank's Assets ............................................................ 37 The Insolvency Risks Posed by Privatized Provincial Banks Are No Greater Than Insolvency Risks Posed by Other Private Banks ........................................................... 40 Fiscal Pressure and Poor Bank Performance Increased the Likelihood that the Provincial Government Would Decide to Privatize the Public Provincial Bank .......... 41 The main focus of this report is Argentina's fiscal experience during the first years of the Convertibility Plan, its achievements and challenges. Contributors to this report were Messrs./Mmes: P. Levy (task manager and fiscal issues), J. A. Zapata, D. Rosenblatt, M. Freire (fiscal federalism), A. Schwarz, R. Rofman (social security), R. Cull and G. Clarke (provincial bank privatizations), G. Ramirez (fiscal accounts), H. Dinh (fiscal solvency and sustainability), M. Basile (provincial fiscal cash accounting estimates), B. Alberdi, FIEL (FONAVI). The report benefited from substantial contributions by the Superintendency of AFJP in the estimation of social security fiscal projections, the Central Bank (Messrs. J. Bolzico and A. Powell and their staff) for their contributions in the evaluation of provincial bank privatizations, and the Ministry of Economy (Programaci6n Econ6mica) for the collaboration in the preparation of provincial fiscal accounts on a cash basis, and the analytical support in the study of the revenue sharing system (coparticipaci6n). The report benefited from D. Veter's contributions on the subject of fiscal federalism, and S. Albert from his contributions on the privatization of provincial banks. CHAPTER III: CREATING A MORE FLEXIBLE FISCAL SYSTEM .................. 45 A. REFORMING THE REVENUE SHARING SYSTEM WITH THE PROVINCES ............................. 45 Principles ......................................................... 47 Background ......................................................... 47 The Current System of Intergovernmental Transfers . ............................................... 48 Per Capita Federal Transfers and Direct Federal Expenditures by Province .............. 51 The Pattern of Provincial Expenditures . .................................................................... 52 Provincial Potential Tax Revenues ......................................................... 58 Towards a Greater Fiscal Efficiency ......................................................... 59 Conclusions ......................................................... 66 Research Agenda for Policy Reform ......................................................................... 68 B. THE FONAVI PROGRAM .......................................................................... 69 The housing market in Argentina ......................................................... 69 Public intervention in the housing market ......................................................... 70 Cost-benefit analysis of the FONA VIprogram ......................................................... 73 Conclusions and recommendations ......................................................... 77 EXECUTIVE SUMMARY 1. This report is intended as a first step to a process of reviewing various dimensions of the fiscal situation and outlook in Argentina, given the primordial importance of fiscal stability in the Convertibility Plan. In that sense, this report is a follow-up of the 1996 Bank report: "The Convertibility Plan; Assessment and Potential Prospects", which reviewed the macroeconomic impact of that plan. The present report should be seen as a building block that establishes a basic knowledge of fiscal conditions in Argentina, evaluates selective important developments and programs affecting Argentina's fiscal condition, and sets up a framework for future work. 2. Following an impressive improvement in public finances during the first three years of the Convertibility Plan, fiscal performance faltered in 1994, and deteriorated during 1995 and 1996. That deterioration was due, among others factors, to the recession-related fall in revenues, and incremental expenditures/lower revenues related to. structural reforms (including social security reform and reduction in employer contribution rates to labor taxes). Sustaining prudent fiscal performance remains a priority for strengthening and sustaining the Convertibility Plan. While the current economic recovery may address part of the fiscal problem if expenditures are kept under control, it is becoming increasingly clear that the fiscal accounts have been, and continue being affected by the important structural reforms associated with the Convertibility Plan. Additionally, there is a need to flexibilize and improve the efficiency of revenue collection and expenditures. 3. Key themes of this report are a review of the consolidated performance of fiscal accounts over the 1991-96 period, and a review of some of the structural reforms that affected the fiscal accounts, such as the reform of the social security system and the privatization of provincial banks. On the revenue side, this report advances in the assessment of revenue sharing arrangements (coparticipaci6n) between the federal government and provinces. On the expenditure side, this report evaluates the complex housing program (FONAVI), financed with transfers from the federal government but managed by the provinces. CONSOLIDATED PUBLIC SECTOR PERFORMANCE 4. The evidence presented in the first chapter of this report, which reviewed recent consolidated public sector performance, points to a fiscal imbalance in the 1991-96 period that is broader than generally recognized. The main elements leading to that assessment are fiscal expenditures which have been treated in an extra-budgetary manner, principally including expenditures related to structural reforms (particularly in 1995 and 1996), and the treatment of debt reconciliation through the issue of bonds. These adjustments point to a continuous deterioration of consolidated fiscal accounts (on a cash basis) over the 1992-96 period. If privatization proceeds are excluded from these cash accounts, Argentina's consolidated public sector accounts were at best in balance for only one year, 1993. - ii - 5. Early gains in revenue collection, related to the defeat of inflation and better administrative management, plus privatization proceeds, led to expectations of further gains that did not materialize. As a result, on a cash basis, the deficit of the Central Government, more than that of the provinces, continued to grow, if for no other reason than that the Central Government had better access to financing than the provinces (which particularly in 1995 built up arrears). Nevertheless, the deficit of the provinces was large both on an cash and an accrual basis in recent years. Furthermore, the quality of spending in the provinces remained essentially unreformed well until 1995, when the tequila crisis forced the initiation of a process of provincial reform. 6. The above evidence points to the need for fiscal adjustment, both in the conventional sense of securing sustainability of fiscal accounts, but also for sustaining the success of the Convertibility Plan where degrees of freedom for economic policy making are limited by design. Additionally, these constraints are compounded by the lack of flexibility created by the present revenue sharing system with the provinces, and the limited control that the Central Government has over its expenditures. 7. The present work has reconstructed estimates of provincial fiscal accounts on a cash basis, enabling for the first time the presentation of consistent consolidated fiscal accounts both on a cash and accrual basis. It is recommended that the presentation of provincial fiscal accounts on a consistent cash basis be institutionalized, and made public, in order to obtain a consistent picture of the consolidated fiscal accounts. FISCAL IMPACT OF REFORMS 8. The Government of Argentina has undertaken, and is contemplating a number of structural reforns, with potentially significant fiscal impact. It is important that the fiscal impact of structural reforms be assessed to understand the future implications of both ongoing and proposed reforms, to ascertain that appropriate measures are taken in order for these reforms, desirable as they may be, not to undermine the Convertibility Plan. However, the threat to the Convertibility Plan is not structural reforms, but fiscal imbalances which prevented the accommodation of such reforms. As indicated in the first chapter of this report, the fiscal deficit obtained in 1996 is not sustainable, particularly in the framework of the Convertibility Plan, and in the context of Argentina's high external debt to service ratios. 9. Reforms undertaken both by the Central Government and provincial governments should be thus assessed. In that context, the assessment of two structural reforms was undertaken in this report: (i) that of the national social security system; and (ii) the privatization of provincial banks. In shear impact, social security reform is a structural change with major fiscal implications (contributing to the temporary increase in the deficit, and affecting provincial finances). The privatization of provincial banks has been particularly controversial at the local govermnent level, and the better understanding of its costs and benefits could facilitate future decisions in that area. 10. In the future, it would be useful to evaluate the fiscal impact of additional reforms, such as trade reforms, the impact of the proposed reform of the state, and other proposed - iii - sectoral reforms such as in health and education. It would be useful to institutionalize the evaluation of fiscal implications in the future assessment of proposed reforms. Social Security System 11. The review of the national Social Security system suggests that despite the initial costs currently being borne by the Government of Argentina, the pension reform has clearly been worthwhile, both fiscally and from its objectives of achieving more equity and transparency. Even including the reduction of employer contributions, ten years after the reform the deficits in the pension system will be lower than the non-reform case. 12. However, longer term there are three areas of concern which the Government may need to address. These involve the size of the basic pension, the use of labor taxes and choice of financing for this basic pension, and the need for social assistance arising from the declining pension coverage among the elderly. i 3. If evasion does not decline as projected, employer contributions might not be sufficient to cover the basic pension in the long term, and in that case the Government might want to consider reducing the amount of the basic pension to 2 AMPO. 14. In the absence of an earlier and more comprehensive tax reform, or reform of the minimum pension system, eventual surpluses in the social security system might allow the lowering of payroll taxes, followed by the reduction in other taxes currently transferred to the system. 15. In the longer term, due to both the increasing retirement age and the greater number of years of contribution required before collecting a pension, a much smaller percentage of those over age 60 will receive pensions, increasing the demand for social assistance payments. This should be included in the calculation of fiscal cost of the reform. 16. In the future, with evasion of labor contributions remaining a serious problem, the evaluation of its causes, and combating it, remain a priority. Second, additional work needs to be done on the distributional implications of the pension reform. Finally, it is imperative that resources be committed to produce a flexible, user-friendly pension model that all relevant Government institutions can access. Privatization of Provincial Banks 17. The study of privatization of provincial banks documents the poor performance of public provincial banks compared to both large private banks and recently privatized provincial banks. The main finding of this analysis is that the estimated future fiscal cost of recapitalizing the money-losing public provincial banks exceeds privatization costs. While the focus of this analysis is on fiscal benefits--which appear to be substantial--it is quite likely that the most important benefit of these privatizations is the economic growth that should derive from improved credit allocation. Additional findings include: - iv - a) Most costs recognized at the time of privatization were created prior to privatization, and thus privatization merely exposed existing problems. b) The size of the residual entity, compared to the privatized entity, depended upon the quality of provincial assets and upon features of the sales contract. c) Available post-privatization evidence suggests that performance of the privatized provincial banks has been better than that of public provincial banks and similar to that of large private banks. This indicates that the risk of insolvency for privatized provincial banks is no greater than the risk posed by the typical private bank. d) Empirical evidence suggests that fiscal pressures, political incentives, and bank performance all affected a provincial government's decision to privatize. Poor bank performance greatly increases the probability that the provincial government will privatize the bank. Likewise, provincial fiscal difficulties increase the probability that the government will privatize. A governor of the ruling party was found to be approximately five times more likely to privatize than when a governor was a member of an opposition party. CREATING A MORE FLEXIBLE FISCAL SYSTEM 18. The Central Government is facing difficulties in revenue collection that go beyond the impact of structural reforms, likely deriving from problems both in tax structure and administration. Tax evasion remains a serious challenge to the fiscal authorities. IMF estimates indicate that VAT compliance levels increased from 33.9 percent in 1989 to 64.4 percent in 1992, but decreased to 59.5 percent in 1993, and to 55.1 percent in 1994. In 1994, VAT compliance in Uruguay was 72 percent, and in Chile reached 82 percent. Similar difficulties are evident in the collection of the income tax. In 1995, again according to IMF estimates, income tax collections amounted to 2.2 percent of GDP, while in Chile reached 4.3 percent, and in Brazil 4.5 percent of GDP. An evaluation of causes of high tax evasion in Argentina would greatly facilitate the current discussion of tax reform. 19. With high rates of evasion in VAT and income taxes, and high labor tax rates, the evaluation of a potential tax reform would also be advisable, following the assessment of the causes of tax evasion in Argentina. Regarding labor taxes, Argentina is faced with high rates of labor taxation, at a time of urgent need to address a high rate of unemployment (related also to rigid labor legislation at a time of significant structural changes in the economy). Mandatory payroll contributions added to 38 percent of gross wages prior to the 1996 payroll tax reduction. Following the 1996 reduction, that rate declined to 32.5 percent, which is still high by international standards1. Effective payroll taxes are even higher (almost 60 percent) when adding other costs associated with current labor regulations. A reduction in labor tax rates would have fiscal implications that need to be evaluated prior to reform. The negative direct short -termn revenue collections effect See "Estimating the Benefits of Labor Reform in Argentina", The World Bank, 1996. v - of reduced labor taxes should be assessed against the growth in employment, particularly 2 in the formal labor market, increasing thus the rate of compliance 20. The current rate of VAT (21 percent) is on the high side by international standards, and in view of its subpar collections rate mentioned above, should be reassessed for possible reduction. However, any reduction in labor taxes or the VAT should be compensated by other sources or revenue (at least until meaningful expenditure reductions are identified). A possible candidate compensating for these two tax rate reductions is the income tax, which as mentioned above, remains a very inefficient source of tax revenue compared to other countries. The Revenue Sharing System 21. Tax reform cannot be contemplated in Argentina without assessing the revenue sharing system of the Central Government with the Provinces. The evaluation of the revenue sharing system (coparticipaci6n) should be an integral part of an assessment of tax reform, and precede it. 22. The results of the analysis of the revenue sharing system, although preliminary, are important. First, the structure of total flows to the provinces changes significantly when all the federal-financed programs are included. The direction of the change does not reflect such factors as population or poverty; some provinces end up losing part of the cake, while others gain. Second, there is a substantial potential for increasing tax revenues -- provinces' own tax revenues could increase by 45% by emulating the collection record of the most efficient province -- making provincial governments more responsive and fiscally responsible to their constituencies. Third, per capita expenditures vary widely among provinces and such variation has no apparent justification on poverty accounts or population density. Per capita expenditure seem to follow transfers availability which indicates large margins for savings. Fourth, devolution of tax revenues to the provinces on an origin base has a dramatic effect in the way that largest provinces could finance their needs but leaves a clear role for equalization programs due to the concentration of tax bases. Fifth, using piggy backing system on the income tax has a relatively small impact which reflects the limited role of this tax in the present Argentine tax system. Sixth, using a Canadian equalization formula brings very sound results on the way equalization could work taking into account differences in provincial tax bases. The formula could be enriched in the future including factors such as the cost of providing public services. While further analysis is needed to guide the reform agenda, these preliminary results reveal the potential that can be realized due to the new data collected, organized and estimated for this report. 23. The analysis of the revenue sharing system focused on the issues encountered in the distribution across provinces (e.g., secondary distribution) more than the primary distribution (e.g. general division between the federal and aggregate provincial public 2 i.b.i.d. - vi - sectors of the total pie of public resources). Assuming that the primary distribution remains roughly the same, some recommendations are due: a) there should be more sharing of tax bases rather than tax revenues, where provinces could charge a surcharge on the taxes collected by the federal government in their jurisdictions, thus causing local populations to face the cost of the local services provided; b) there should be one revenue-sharing pool, so as to increase the autonomy of federal government tax policy and tax effort; and c) there should be a greater correspondence with federally collected taxes remaining in the federal treasury and a greater role for provincial government tax collections, with those revenues remaining with the provinces that collect them. 24. The transition into a better and more transparent system requires, a priori, that all parts of the federal system agree on the basic strategy and pillars of the new system. Fundamental pieces of the work include: a) Decision on which taxes will be federal and provincial b) Definition of which taxes constitute the revenue pie to be shared between federal and agreement on the future system (as function of GDP growth, constant in real terms, etc.) c) Consolidation of all programs and transfers into a single transfer program (primary distribution) to increase transparency and predictability of the flows. This would help provinces have a better grasp into their own budget constraints. d) Definition of the parameters to be taken into account in an equalization formula: (i) tax capacity and tax basis; (ii) population and social needs; (iii) relative cost differentials. e) Establishment of an analytical body with the function to discuss with all the referent parties -- e.g., provinces and national government -- the elements (rather than the results) of the new system f) Review of the tax system at the national level 25. To plan for this transition, much ground work will be needed. Basic information on economic and social variables need to be collected at provincial level. This includes data on employment, GDP, industrial activity and cost of service delivery -- which are basically nonexistent outside Buenos Aires -- along with studies on tax potential and tax effort. - vii - 26. On the expenditure side, the fact that the Central Government has limited flexibility over its expenditures, as seen in the first chapter of this report, and a significant portion of that expenditure is destined for social program, suggests that any efforts to control the fiscal deficit would require an evaluation of the efficiency of expenditure programs, and in particular the public expenditure review of social programs. The last chapter of this report undertook an assessment of the Argentina's housing program, FONAVI. FONA VI 27. A review of the FONAVI program showed it to be inefficient, by not meeting its intended targets and being expensive. Even taking into consideration a basic needs approach (i.e. justifying for social purposes a degree of subsidy to the program), the program has been shown to remain cost ineffective, raising questions about its desirability in its current structure, and in the context of the current overall fiscal situation in Argentina. More specifically: * Higher subsidies than necessary undermine the program. FONAVI is providing excess space per person, and higher costs result in targeting fewer beneficiaries; * FONAVI construction costs have been above market costs reducing the net social benefit of the program; a FONAVI shows very high default rates resulting in higher net social costs; * FONAVI is providing subsidies to higher income households than the lowest quintile income population. 28. While the efficacy of FONAVI has improved significantly since 1993 with its decentralization to the provinces, important problems remain. Assuming the desirability of continuation of a transitional housing program in Argentina for the lowest income group on account of financial and other market inefficiencies, an alternative worth considering would be focusing the program not on the supply of houses and their finance, but on a direct subsidy on the demand for housing. There is a need to separate subsidies from the financing of houses. Also, subsidies should be one time capital grants or housing allowances that have a finite duration or a built-in review procedure.. They must be well-targeted, measurable and transparent and should avoid distorting housing markets. 29. There is a role for the government to ease access to better housing conditions for the poorest families of Argentina that need not be done through FONAVI. Effective enabling strategies should deal with the causes rather than the symptoms of housing problems. The government could focus on different operational market mechanisms to stimulate housing demand (develop property rights, a competitive mortgage finance market, streamline the subsidy program), to facilitate the process of housing supply (provide infrastructure for residential land development, and secure a competitive land - viii - and housing development), and create an overall institutional framework for managing the housing sector and ensure adequate access to housing by the poor. 30. Restructuring of the revenue sharing system with the provinces should give flexibility to the Government to finance other public sector investments of higher return and allow for a possible decline of the size of the housing program. In that sense, it is recommended that a public expenditure review of government programs be undertaken. 31. A public expenditure review remains in order, both to secure savings, but more importantly for Argentina, in increasing the efficiency of spending. So far, most of the emphasis in addressing the fiscal imbalances of 1995 and 1996, has been on revenue collection. Expenditures have not been sufficiently analyzed, particularly in view of the high percentage of expenditures in social programs and argued to be "difficult" to cut. More fundamentally, focusing nearly exclusively on revenue changes, which has been the emphasis so far, may be a mistake, since recent literature suggests that countries which implemented fiscal reforms through adjustments in expenditures (particularly on transfers and government wages) were more successful in sustaining gains than countries that dealt with fiscal disequilibria by raising tax rates, or cutting public investment. INTRODUCTION 1. Following an impressive improvement in public finances during the first three years of the Convertibility Plan, fiscal performance faltered in 1994 and deteriorated during 1995 and 1996. That deterioration was due, among others factors, to the recession-related fall in revenues, and incremental expenditures/lower revenues related to structural reforms (including social security reform and reduction in employer contribution rates to labor taxes). While the current economic recovery is expected to improve fiscal performance in the near future, it is becoming increasingly clear that the fiscal accounts have been affected by the important structural reforms associated with the Convertibility Plan. Furthermore, there is a growing need to flexibilize and improve the efficiency of both revenue collection and expenditures. 2. Fiscal strength and stability is a major factor in the maintenance of the Convertibility Plan. Under this plan, the economic authorities have limited scope for pro- active policies. Monetary policy is severely circumscribed since the Central Bank is essentially a currency board, and money creation is directly related to the availability of international reserves. Fiscal policy is also circumscribed by the impossibility of financing deficits through monetary creation. Two other factors add to the complexity of fiscal policy in Argentina. First, the revenue sharing system with the provinces limits the Federal Government's ability to use tax policy efficiently, for example at times of economic crisis, as in 1995, since the major revenue sources are shared with the provinces and the leakages to the provinces from tax increases are large. Second, the majority of public expenditures now go to social programs that are difficult, politically, to cut. At the same time, empirical evidence suggests that countries with strong and flexible fiscal systems are in a better position to deal with the consequences of volatile capital flows, to which Argentina remains vulnerable. 3. This report is a follow-up to the 1996 Bank report: "The Convertibility Plan; Assessment and Potential Prospects." That report reviewed the notable achievements and challenges of the Convertibility Plan and its implications, including highlighting the need for fiscal stability. This report is intended as a first step to a multi-year process of reviewing various dimensions of the fiscal situation and outlook in Argentina, given the primordial importance of fiscal stability in the context of the Convertibility Plan. In that context, this report is a building block that establishes a basic knowledge of fiscal conditions in Argentina, both at the federal and provincial levels. MAIN REFORMS UNDER THE CONVERTIBILITY PLAN 4. The main pillars of the Convertibility Plan' have been: ISee "Argentina; The Convertibility Plan: Assessment and Potential Prospects," 1996, the World Bank. -2 - (a) Monetary reform, through the Convertibility Law, subsequently supplemented by the new Charter of the Central Bank2 financial sector reforms ensued; (b) Fiscal reform, initially through a sharp improvement in the administration of the tax system and later through the redefinition of tax instruments and rates; (c) Public sector reform, through debt restructuring, civil service reform, fiscal restructuring and an ambitious and successful plan of divestiture, and deregulation of product markets; (d) Social security reform, allowing for a new capitalization mechanism operated by the private sector; and (e) Trade reform, through the elimination of export taxes and most quantitative restrictions on imports, and the reduction of the level and range of import tariffs. 5. The macroeconomic impact of the above reforms has been positive, resulting in high economic growth in the first four years of the Convertibility Plan (average growth rate of 8.9 percent over the 1991-94 period), accompanied by increases in productivity, and generation of export- and investment-led growth during 1993 and 1994. The 1995 regional crisis reversed some of the gains from the first four years of the plan and contributed to higher unemployment, but by the second half of 1996 the economy resumed economic growth, which continued strong into 1997. Even taking into consideration the recession year 1995 and slow recovery in 1996, average economic growth in the six years of the Convertibility plan reached 5.8 percent. That pace of growth, jointly with the reduction in inflation, contributed to a significant expansion in public revenues and expenditures. 6. The major restructuring of the Argentine economy requires significant reallocation of both labor and capital to obtain the potential for productivity gains. However, lack of flexibility in labor and capital markets (the latter in view of substantial public -- national and provincial -- presence in that sector) generated dislocations, none more important than the creation of high level of structural unemployment. Argentina needs the fiscal ability to respond to important social needs following shocks. That need is yet another reason for a review of the fiscal situation. 2 See "Estimating the Benefits of Labor Market Reform in Argentina," 1996, the World Bank. -3 - CHAPTER I: RECENT PUBLIC SECTOR PERFORMANCE 7. During 1991-94, the performance of the Argentine public sector was commendable in many respects. First, following a decade of public deficits, the Federal Government was able to achieve fiscal balance or surplus in 1992 and 1993, eliminating the need for inflationary finance. Second, a series of difficult to reverse reforms in institutions and policies reduced the size of the public sector3 to 22.4 percent of GDP in 1994, primarily through privatizations of public enterprises, which a decade earlier accounted for approximately 10 percent of GDP and now account for only 0.5 percent of GDP. The current size of the public sector is substantially lower than the 40 percent average rate in OECD, and lower than in other Latin American countries. 8. Fiscal performance started deteriorating in mid-1994, prior to the 1995 elections, and deteriorated further during 1995 and 1996 principally for two reasons: the economic recession and the cost of some key structural measures, such as the social security reform and the reduction in employer social security taxes. In 1995, despite important efforts to rein in expenditures, the shortfall in revenues generated a federal fiscal deficit, on a cash basis, equivalent to 0.5 percent of GDP, or US$1.4 billion (including privatization receipts of US$1.2 billion), and a consolidated fiscal deficit (on a cash basis) of 1.6 percent of GDP (see Table 1). In 1996, continued difficulties in revenue collection, along with higher provincial transfers on account of a revenue sharing guarantee setting a floor on transfers to the provinces, resulted in a deterioration of the federal deficit to US$5.5 billion. Despite the decline in the provincial deficit, to an estimated $1.7 billion, the overall deficit rose to 2.5 percent of GDP. 9. The consolidated fiscal accounts of the Argentine public sector have exhibited a gradual but continued deterioration between 1993-96, raising questions about the trend of fiscal accounts, important as they are for the sustainability of the Convertibility Plan. Early indications suggest all improvement in fiscal accounts in 1997, albeit in the context of a strong economic recovery. The following sections provide a closer look at these accounts, factors influencing their performance, and methodological issues that affect the magnitude of the public deficit. CONSOLIDATED FISCAL ACCOUNTS 10. The quality of statistics on federal and provincial fiscal accounts has been improving significantly, following the debacle of Argentina's inflationary period. One of the important remaining methodological difficulties in the presentation of such consolidated accounts is that provincial accounts (expenditures but not revenues) are officially reported on an accrual basis, while federal accounts are reported on a cash and 3 Including Central Government and provincial operations, but excluding municipalities, operations of the Central bank, public enterprises (except to their net impact on Central Government finances), and certain operations that are recorded below the line, and are discussed later in this report. -4 - on an accrual basis, but with limited historical information for the latter, thus making consistent presentations and analysis difficult. 11. In this work, we have reconstructed estimates of provincial fiscal accounts on a cash basis, in order to present consistent consolidated fiscal accounts. This work presents consolidated fiscal accounts both on a cash and accrual basis4 and points to methodological issues that create discrepancies between fiscal accounts and the growth in public sector indebtedness over the 1991-96 period. Consolidated Fiscal Accounts on a Cash Basis 12. The Table below incorporates official cash accounts for the Central Government, and estimates of provincial fiscal accounts on a cash basis, which until now have been unavailable. The latter estimates were prepared with the cooperation and supervision of the Ministry of Economy and allow for the first time a consistent view of consolidated fiscal accounts on a cash basis in Argentina. The methodology for estimating provincial fiscal accounts on a cash basis is outlined in Annex I. 13. The results, presented in Table 1 , show that the consolidated public sector deficit (including the Central Government, and provinces, but excluding municipalities) increased substantially between 1994 and 1996. It is of interest to note that the provincial fiscal deficit on a cash basis has remained approximately unchanged between 1993 and 1996 at about 0.6 percent of GDP. That is, on a cash basis, the deterioration of the consolidated fiscal accounts in the 1994-96 period was due to Central Government operations, with the deterioration of both the social security accounts and those of the National Administration. 14. The provincial fiscal balance on a cash basis, which came under stress during 1995 in view of the tequila crisis, did not deteriorate significantly as a share of GDP (in contrast to the accounts on an accrual basis which marked a noted deterioration in 1995), because sources of financing were severely constrained, and provinces accumulated arrears. In 1996, changes in the social security system had a major detrimental fiscal impact on the Central Government's accounts, both due to the transfer of affiliates to the private capitalized system (estimated by the IMF as 0.8 percentage points of GDP per year), following the 1994 social security reform and by the decline in social security contributions (aportes patronales), intended to reduce labor costs (IMF estimated cost of 0.8 percentage points of GDP per year). Furthermore, in 1996, despite the economic recovery, the surplus of the National Administration (excluding social security) continued to decline despite a further absolute decline in expenditures. The slowdown in privatization receipts played a role in the apparent deterioration of the national Fiscal accounts on a cash basis take into consideration the movement of funds at the time they occur, irrespective of the fiscal year to which they correspond. Accounting on an accrual basis registers imputed expenditures of a fiscal year, irrespective of the time of actual payment on receipt. Both accounting systems are useful and relevant, since each one has obvious drawbacks. - 5 - administration's fiscal balance, as did the decline in non-tax revenues, lower Central Bank transfers to the Treasury due to the remuneration of liquidity requirements, and higher capital expenditures. Table I Consolidated Public Sector Accounts 1991-1996 (% current GDP Cash basis) 1991 1992 1993 1994 1995 1996 Current Revenues 18.7 21.0 21.5 21.2 21.2 19.3 Central Government 10.5 11.4 11.9 11.6 12.0 10.1 National Administration 5.9 5.9 5.7 5.5 4.5 3.9 Social Security System (1) 4.5 5.5 6.2 6.1 7.5 6.2 Provinces 8.2 9.6 9.7 9.5 9.3 9.2 Current Expenditures 18.2 19.4 19.7 20.4 21.3 20.2 Central Government 11.0 10.8 10.5 11.4 12.5 11.8 National Administration (2) 6.5 5.4 4.1 4.3 3.8 3.8 Social Security System 4.5 5.4 6.4 7.1 8.7 8.0 Provinces 7.2 8.6 9.3 9.0 8.8 8.5 Current Balance 0.5 1.6 2.2 1.2 0.7 (0.1) Central Govemment (0.5) 0.6 1.8 0.7 0.2 (0.8) National Administration (0.6) 0.5 1.9 1.6 1.4 1.0 Social Security System 0.1 0.1 (0.2) (0.9) (1.2) (1.8) Provinces 1.0 1.0 0.4 0.6 0.5 0.7 Capital Receipts 1.1 0.9 0.4 0.4 0.6 0.3 Central Government 1.0 0.8 0.3 0.3 0.4 0.2 National Administration 0.9 0.6 0.2 0.2 0.3 0.1 Social Security System 0.1 0.3 0.0 0.1 0.1 0.0 Provinces 0.0 0.1 0.1 0.1 0.1 0.1 Capital Expenditures 1.9 1.8 1.7 1.8 1.7 1.7 Central Government 1.0 0.8 0.6 0.6 0.4 0.3 National Administration 1.0 0.8 0.6 0.6 0.4 0.3 Social Security System 0.0 0.0 0.0 0.0 0.0 0.0 Provinces 0.9 1.0 1.0 1.3 1.3 1.4 Balance (including privatizations) (0.3) 0.7 0.5 (0.6) (1.2) (2.4) Central Government (0.5) 0.6 1.1 (0.0) (0.5) (1.8) National Administration (0.7) 0.3 1.2 0.8 0.6 (0.1) Social Security System 0.2 0.3 (0.1) (0.8) (1.1) (1.7) Provinces 0.1 0.1 (0.5) (0.6) (0.7) (0.6) Balance (wlo privatizations) (1.5) (0.1) 0.3 (0.9) (1.6) (2.5) Privatization incomes 1.2 0.8 0.2 0.3 0.4 0.1 Trust fund (Banks) 0.0 0.0 0.0 0.0 (0.3) (0.2) Ex-Cajas Provinciales 0.0 0.0 0.0 0.0 0.0 (0.2) Capitalized interest 0.0 0.0 (0.3) (0.4) (0.4) (0.3) Balance (including extra budgetary accounts) (0.3) 0.7 0.2 (1.0) (1.9) (3.1) (1) Includes transfers from Central Government to Social Security Institutions (PAMI), which are excluded from National Administration current revenues. For 1995, revenues include $1.34 billion in promissory notes for payment of social security contribution arrears. (2) Includes net result of public enterprises. Includes: National Administration, Social Security and estimations for Provincial Governments. Source: Ahorro - Inversion, Ministerio de Economla y de Servicios Publicos, World Bank elaboration Bank estimates. -6 - 15. Public enterprises have played an important role in reshaping Argentina's public finances. As can be seen in Table 2 below, the impact of privatizing public enterprises was particularly evident in 1991-93 period. While in 1989 public enterprise expenditures reached 11 percent of GDP5, intervening privatizations reduced the size of Argentine public enterprise expenditure to less than 0.5 percent of GDP by 1995/96. The privatizations eliminated the operational deficits of public enterprises, which in 1989 were 2.5 percent of GDP. Public finances benefited not only from the elimination of this deficit but also from the temporary surge in privatization proceeds. These funds came at the early stages of implementation of the Convertibility Plan, when the establishment of credibility of the Plan, anchored as it was on the viability of fiscal discipline, was crucial. The decline and elimination of federal fiscal deficits in the first three years of the Convertibility Plan were aided significantly by privatization proceeds, as the table shows. (In 1991, direct cash privatization proceeds approached 1 percent of GDP.) These proceeds also contributed to the partial cancellation of public debt. Longer-term, empirical evidence from other countries suggests that fiscal accounts will likely benefit from better allocation of resources in the economy, through the improved management of privatized enterprises and future generation of tax revenues (a subject for further analysis). Table 2 PRIVATIZATION OF PUBLIC ENTERPRISES 1991 1992 1993 1994 1995 1996 (Million current pesos - Cash basis) Public Enterprises operat. balance -982 -93 690 -25 -59 -55 Operational Revenues 8,835 8,980 5,016 1,567 1,091 918 Operational Expenditures 9,817 9,073 4,326 1,592 1,150 973 (as a % of current GDP) Public Enterprises operat. balance (0.5) (0.0) 0.3 (0.0) (0.0) (0.0) Operational Revenues 4.9 4.0 1.9 0.6 0.4 0.3 Operational Expenditures 5.4 4.0 1.7 0.6 0.4 0.3 Central Government Balance (838) 1,402 2,730 (84) (1,373) (5,514) Privatizations 2,194 1,781 523 733 1,171 375 Balance w/o privatizations (3,032) (379) 2,207 (817) (2,545) (5,889) Source: Ahorro - Inversion, Boletin Fiscal, World Bank presentation. 16. Since the Convertibility Plan was put in place in 1991, the total expenditures of the core public sector (excluding public enterprises) has been fairly stable as a percentage of GDP, at a time of very rapid GDP growth. The composition of expenditures however has shiftedfrom the federal central administration toward the social security system and to the provinces. Social security system expenditures increased by 3.5 percentage points of GDP, but much of this change was due to a change of accounting. Notably, PAMI and other social safety net programs like family allowances are now counted as part of the "Argentina: From Insolvency to Growth", The World Bank, 1993. While the 1989 estimates are not strictly comparable with estimates in this document because of the recalculation of Argentina's national accounts, they are useful as an order of magnitude. - 7 - social security administration. Also, provincial expenditures increased by almost one percent, partially attributed to the final phase of decentralization of primary/secondary education and health services. 17. The improved performance of the Federal Government was primarily the result of large increases in current revenues, an increase of 2.8 percentage points of GDP between 1990 and 1994, associated with the reduction in inflation, better institutional revenue collection capabilities, and the resumption of economic growth. VAT collections (with the VAT raised in the early years to a near uniform rate of 18 percent) showed the fastest growth, approaching 6 percent of GDP, more than twice the 2.5 percent average of the 1980s. As noted above, massive privatizations also became a primary source of revenue, averaging about 0.7 percent of GDP annually between 1991-93, when the bulk of the privatizations occurred. Of the $18 billion raised through privatizations between 1990- 93, half was received in cash, and the rest in bonds and transfer of liabilities. Since 1995, however, revenues as a share of GDP have declined, because of faltering tax collections, the decline in labor taxes, and transfer of social security contributions to the private capitalized pension system. 18. Expenditures, as a share of GDP, have declined when compared to the 1980s, but the major cut in public spending took place during the hyperinflation of 1989. The national administration benefited from the growth of the economy, raising its expenditures between 1990 and 1994 by a real 32 percent. Concurrently, by 1994, the share of social expenditures to total federal public expenditures, reached 64 percent. During this period, the Government was able to increase average salaries and partially restore salary differentials.6 As regards consolidated public sector employment levels, evidence suggests that any decline was exclusively the result of the sale of public enterprises, since, excluding public enterprises, the increase in public employment in the provinces more than compensated for any reductions in permanent federal employees (through a shift in expenditure responsibilities). Composition of Federal Revenues and Expenditures 19. Recent fiscal performance (1996, see Table 3 below) is indicative of the constraints facing the federal authorities. From a nominal revenue base of $46.5 billion, the Federal Government, when excluding transfers to the provinces (nearly all automatic), has discretionary revenue of only $30.6 billion (including privatizations). If expenditures on the social security system are taken into consideration ($19.1 billion), plus interest payments on debt ($4.6 billion), the Federal Government is left with revenues of $6.8 billion, or 2.2 percent of GDP, for all other expenditures, including 6The transfer of education and health responsibilities to the provinces accounted, for an important reduction in the number of positions in the Federal Government. However, it is worth underscoring that the cost of the transfer continue to be covered by the federal authorities, as part of the Pacto Federal, which assigned about US -8 - public administration, and capital expenditures. Abstracting from the efficiency of expenditures, it can be argued that there is considerable short-term inflexibility from the expenditure side. Adding to these difficulties is the complexity of raising additional revenue for Federal Government use, since national revenues are automatically shared with the provinces. Table 3 National Administration: Structure of revenues and expenditures in 1996 Current Pesos on a cash basis, Billions (including Social Security) National Social Administration Security Total Revenues Current Revenues 31.9 14.1 46.0 Tax Revenues 29.2 4.0 33.2 Soc. Sec. Contrib. 0.3 10.0 10.3 Other Current 2.4 0.1 2.5 Capital Revenues 0.4 0.1 0.5 Privatizations 0.3 0.1 0.4 Other * 0.1 0.0 0.1 Transfers to Provinces 15.8 0.1 15.9 Current 13.2 0.1 13.3 Capital 2.6 0.0 2.6 Net Revenues (w/o Privatizations) ** 15.8 14.0 30.2 Net Revenues (wI Privatizations) 16.5 14.1 30.6 Expenditures Current Expenditures 15.8 19.3 35.1 Personnel 6.2 0.5 6.8 Goods and Services 1.9 0.2 2.1 Interest 4.5 0.1 4.6 Transfers (non provincial) 2.4 3.8 6.1 Other Current 0.1 0 0.1 Social Security Payments 0.8 14.6 15.4 Capital Expenditures 1.0 0 1.0 Total Expenditures 16.8 19.3 36.1 Operational Balance 3.1 -9.0 -5.9 w/ privatizations 3.4 -8.9 -5.5 *Excluding 250 million pesos in moratoria borrowing. **Net of transfers to provinces. 20. Another breakdown below displays the percentage of net revenues absorbed by each category of expenditures. Aside from the inflexible expenditures mentioned above, another large category of expenditures is non-provincial transfers, which include transfers to the private sector for social welfare programs and transfers to the public universities, which have administrative autonomy, and consume an additional 8 percent of revenue. 21. Finally, we can add the provincial governments and examine the composition of expenditures across economic categories. We see the effects of decentralization in the large size of payroll expenditures and goods and services expenditures at the provincial level compared to the federal level of government. The lion's share of traditional government functions and administration is with the provinces. Excluding Central -9 - Government expenditures on social security, the National Administration spends 5.7 percent of GDP and Provinces 9.9 percent of GDP. Table 4 Federal Government Expenditures, As a Percentage of Net Revenues*, 1996 National Social Security Total Administration Institutions (w/o soc. security) Current Expenditures 52 63 115 Personnel 20 2 22 Goods/Services 6 1 7 Interest 15 0 15 Transfers (non prov'l) 8 12 20 Other Current 0 0 0 Social Security Payments 3 48 50 Capital Expenditures 3 0 3 Total Expenditures 55 63 118 Operational Balance 10 -30 -19 w/ privatizations 11 -29 -18 Total revenues, including privatizations, net of transfers to provinces. 22. Overall, according to Government estimates, social expenditures of the consolidated public sector have risen from 51 percent of total expenditures in the 1980-82 period to 70.7 percent in the 1994-96 period. Over the 1994-96 period, 40 percent of social spending was destined for social security programs and 60 percent to social sector spending. Of the total public sector spending in that period, 54.7 percent was executed by the Federal Government, 38.4 percent by the provinces, and 6.9 percent by municipalities. However, the Federal Government was responsible for 79.4 percent of all social security expenditures, while provinces and municipalities were mainly responsible for executing social programs (49.9 and 11.5 percent respectively). Abstracting from the efficiency of these programs (an issue to be addressed later) the political ability of Central and Provincial Governments to treat those expenditures more flexibly at times of crisis is severely circumscribed. In that sense, it is imperative to undertake a public expenditure review of social programs, since they dominate public expenditures. Consolidated Fiscal Accounts on an Accrual Basis 23. On an accrual basis, the presentation of consolidated accounts is restricted since data for the Central Govermment on that basis are available only for the 1994-96 period (detailed information on federal and provincial accounts can be found in the Annex). Provincial accounts are available on an accrual basis throughout (although revenues are calculated on a cash basis). In the 1994-96 period, there was a deterioration of the consolidated fiscal deficit. -10- Table 5 Federal and Provincial Expenditures, As a Percentage of GDP, 1996 Total National Social National Administration Security Subtotal Provinces and Institutions Provincial Current Expenditures 5.4 6.6 12.0 8.5 20.5 Personnel 2.1 0.2 2.3 5.0 7.3 Goods/Services 0.7 0.1 0.7 0.9 1.6 Interest 1.5 0.0 1.6 0.3 1.9 Transfers (non prov'l) 0.8 1.3 2.1 2.3 4.4 Other Current 0.0 0.0 0.0 0.0 0.0 Social Security Paym's 0.3 5.0 5.3 N/A* 5.3 Capital Expenditures 0.3 0.0 0.3 1.4 1.7 Total Expenditures 5.7 6.6 12.3 9.9 22.2 Operational Balance 1.1 -3.1 -2.0 w/ privatizations 1.2 -3.0 -1.9 -0.6 -2.5 *In the provincial accounts **Provincial accounts are on a accrual basis on the expenditure side, while federal accounts are on a cash basis. 24. According to official data the consolidated deficit of the public sector increased to 1.3 percent of GDP in 1995, and to 1.4 percent in 1996 (up from 0.5 percent in 1994). The increase in the 1995 deficit is due to: (i) the increased deficit in the provinces, when the tequila crisis and the national recession contributed to a loss in coparticipated and own revenues, with little reaction time for meaningful adjustments in expenditures; and (ii) the elimination of the surplus of the national administration when the actual increase in expenditures was not matched by the expected increase in revenues, again due to the recession, but also structural changes such as the reduction in employer social security contribution rates, and the shift of contributors to the new private capitalized social security system. 25. The increase in the 1996 consolidated deficit was totally due to the increase in the deficit of the Central Government, since provincial deficits declined sharply, when reduced availability of other financing sources forced the restructuring of provincial finances. Nevertheless, part of the provincial deficit decline was due to the transfer of some provincial pension funds to the national system, and higher transfers from the Federal Government. Another reason for lower than anticipated revenues was the continued high rate of tax evasion. Comparative Analysis of Accounts on Cash and Accrual Basis 26. The comparison of consolidated fiscal accounts on a cash and accrual basis generate a quite different outcome in the short term, although longer term the two approaches converge. A comparison can be made for the last three years where comparative data exist. Clearly, on a cash basis, between 1994-96 Argentina's consolidated fiscal deficits have been higher than on an accrual basis, indicating that cash payments are being made for prior years' commitments. This phenomenon is most - 11 - Table 6 Consolidated Public Sector Accounts 1994-1996 (Accrual Basis, % current GDP) 1994 1995 1996 Current Revenues 23.1 22.8 20.9 Central Government 13.5 13.5 11.7 National Administration 8.5 8.6 8.3 Social Security System 5.0 4.9 3.4 Provinces 9.5 9.3 9.2 Current Expenditures 21.3 22.0 20.2 Central Government 12.2 12.8 11.8 National Administration 6.4 7.1 6.6 Social Security System 5.9 5.7 5.1 Provinces 9.1 9.2 8.5 Current Balance 1.7 0.8 0.6 Central Government 1.3 0.7 (0.1) National Administration 2.2 1.5 1.6 Social Security System (0.9) (0.8) (1.7) Provinces 0.4 0.1 0.7 Capital Receipts 0.4 0.6 0.3 Central Government 0.3 0.4 0.2 National Administration 0.3 0.4 0.2 Social Security System 0.0 0.0 0.0 Provinces 0.1 0.1 0.1 Capital Expenditures 2.6 2.7 2.6 Central Government 1.2 1.2 1.3 National Administration 1.2 1.2 1.3 Social Security System 0.0 0.0 0.0 Provinces 1.4 1.5 1.3 Balance (0.5) (1.3) (1.6) Central Government 0.4 (0.1) (1.1) National Administration 1.3 0.7 0.6 Social Security System (0.9) (0.8) (1.7) Provinces (0.9) (1.3) (0.5) Extra budgetary expenditures Trust fund (Banks) 0.0 (0.3) (0.2) Ex-Cajas Provinciales 0.0 0.0 (0.2) Capitalized interest (0.4) (0.4) (0.3) Balance (with extra budgetary capital exp.) (0.8) (2.1) (2.3) Includes: National Administration, Social Security and Provincial Governments. Source: Boletin Fiscal. Ministerio de Hacienda y Obras y Servicios Sociales. - 12- pronounced for 1996, where the largest divergence is observed. Payments for the crisis year of 1995 were made in 1996, both at the national and provincial levels. The improved access to financing permitted that adjustment. 27. During the 1994-95 period the provincial cash deficit was lower than the accrual deficit, indicating that they accumulated arrears. In contrast, the Central Government was paying off older debts, particularly to the social security system (1995-96). This is consistent with the need in 1996 for the provinces to make a large adjustment in their fiscal accounts. Despite the paying off of past arrears, at the end of 1996 some year-end expenditures of the Central Government were again passed on to 1997, such as certain transfers to the provinces, and part of the annual salary bonus to public employees. Provinces also delayed payments to 1997. COMPARISON OF DEFICITS AND THE CHANGE IN PUBLIC SECTOR DEBT 28. Reconstructing Argentina's debt obligations has been a difficult process, and accurate official historical data are scarce. Nevertheless, reliable but preliminary information on national public sector debt for the last three years has become available (1994-96), which together with recent estimates of provincial public sector debt for 1995- 96, allows the estimation of the consolidated public sector debt of these last two years (see Table 7). 29. A few observations are in order. Clearly, public sector debt has been increasing, reaching in 1996 32.3 percent of GDP. Together with provincial indebtedness (excluding obligations to the national government), Argentina's consolidated public sector debt reached 36.1 percent of GDP. Most of the national public debt (77 percent in 1996) is external debt both for historical reasons (through amortization of pre-Convertibility domestic debt by inflation), market limitations during the Convertibility Plan in issuing domestic debt, and the low level of fiscal deficits during the first three years of the Plan (1991-93). 30. Even though reliable debt information is concentrated in the last three years, it is of interest to note that the increase in public indebtedness does not correspond to the deficits observed in Tables 1 and 6. Table 8 highlights these discrepancies. In that table, it is shown that the national public deficit on a cash basis in the last two years is less than half the increase in national public indebtedness. This leads to the conclusion that the fiscal accounts, in their current form, do not reflect the full range of public expenditures that should be included in fiscal accounts on a cash basis (such as the reconciliation of arrears, or other expenditures related to structural reforms). Table 8 reconstructs such items that are either treated below the line in fiscal accounts or excluded altogether. The discrepancy between the public deficit identified in Table 1 and the increase in national public debt of the last two years consists of: (i) the financing of trust funds in 1995 and 1996 for the restructuring of provincial and private banks; (ii) the cost of transferring provincial pension funds to the national pension system; (iii) the capitalization of interest of zero coupon debt-consolidation bonds; (iv) for 1996, advanced financing for the - 13 - following year; and (v) the net issue of debt consolidation bonds (included in the residual). Table 7 Consolidated Public debt 1994 - 1996 (US$ million and % of GDP) 1994 1995 1996 I. National Public Sector Debt 80,313 87,091 97,105 (% current GDP) (28.5) (31.0) (32.3) External Public Debt 63,155 70,524 74,588 (% current GDP) (22.4) (25.1) (24.8) Domestic debt 17,384 16,912 18,932 (% current GDP) (6.2) (6.0) (6.3) II. Provincial debt (net*) 10,204 11,380 (% current GDP) (3.6) (3.8) Consolidate Public Sector (I + 1I) 97,295 108,485 (% current GDP) (34.7) (36.1) (% current consolidated public sector revenues) (159.1) (178.5) * World Bank adjustment for double counting of debt between levels of government. Source: Boletin Fiscal and Consolidated Provincial information from the Central Govemment. 31. It can oe argued that these items accounting for the discrepancies should have been included in the annual fiscal accounts (they are currently recorded below the line). If the major below the line items -- trust funds for bank restructuring, the find for the transfer of provincial pension funds, and capitalized interest --were included in the consolidated public sector deficit, the deficit increases significantly to 3.1 percent of GDP in 1996. 32. The proceeds from privatization have been in different forms (cash, bonds, cancellation of liabilities) making their measurement difficult. Here, we present cash privatization proceeds of the Central Government (provincial privatization proceeds are not included in the provincial fiscal accounts) that were ultimately made available in the financing of budgetary expenditures (see Table 7). These are likely a low estimate of privatization proceeds. 33. Although privatizations cannot be seen in isolation, since they influence revenue and expenditure decisions, excluding them as revenue (since they were extraordinary revenues that cannot be repeated) suggests that the fiscal performance during the 1991-96 period is less favorable when seen from a more sustainable perspective. In general, adjusting for budgetary privatization proceeds, the consolidated finances of the Argentine - 14- public sector have been in deficit throughout the 1991-96 period, averaging an annual deficit of 1.4 percent of GDP. Table 8 Public Deficits and Variations in Public Debt (US$ Million) 1994 1995 1996 Change in national public debt n.a. 6,778 10,310 Public deficit (cash basis w/o privatizations) 817 2,545 5,889 Trust funds for banks 0 834 333 Ex Cajas Provinciales 0 0 310 Advance Financing (C.B. deposits) 0 0 1,882 Capitalized interest 1,000 1,188 1,170 Others (a) n.a. 2,211 726 Change in provincial debt n.a. n.a. 880 Provincial deficit 2,448 3,533 803 Others n.a. n.a. 77 (a) Other includes in 1995 the consolidation bonds (US$ 1716) and Bank Capitalization (US$445). Source: Boletin Fiscal. Ministerio de Economia y Obras y Servicios Publicos. 34. An additional adjustment warranted in these accounts is the issue of bonds for the recognition of old debts, which was very significant in the 1992-94 period. There are two alternative ways of incorporating these bonds in fiscal accounts (on a cash basis): include the total amount of the debt issue at nominal value at the moment of issue or include bonds at market value at the moment of issue, and subsequently add market valuation changes plus accrued interest in case of zero coupon bonds. The first approach tends to concentrate the deficit in the year of issue, while the latter spreads the deficit across years (but is more difficult to calculate). Neither of those are currently used in Argentina's fiscal accounts (the bonds are included in the stock of debt), except for interest. Their inclusion would increase significantly the annual deficits over the 1991-96 period. To give a sense of magnitude for recognition bonds, the stock of BOCONs, BOTE and BOTESO outstanding at the end of 1996 amounted to $19 billion, or 6.3 percent of GDP. If this item is added to the cumulative deficit of that period without adjusting for privatizations (and subtracting capitalized interest since those bonds are counted at nominal value) the Argentine consolidated fiscal deficit over the 1991-96 period averages an annual 2.0 percent of GDP (compared to the average of .85 percent from the official national accounts and estimated provincial accounts). Adjusting for privatization proceeds, the average consolidated fiscal deficit rises to an annual 2.5 percent of GDP. 35. The above adjustments in Argentina's fiscal accounts go a long way in accounting for the discrepancies with the increase in the stock of debt. A greater effort to estimate the role of issuance of recognition bonds during that period is needed to clarify further the state of Argentina's fiscal accounts in that crucial period of adjustment. - 15- FISCAL SOLVENCY AND SUSTAINABILITY 36. This section reviews issues concerning solvency and sustainability of Argentina's public sector fiscal accounts. It is well recognized that a public sector is considered solvent when its discounted primary surpluses are sufficient to cover the curTent public 8 sector debt (which includes external and domestic debt) . Knowledge of future economic growth and paths of real interest rates are needed for solvency analysis. For this reason, most empirical studies focus on sustainability analysis which uses only current information on the country. 37. Total debt of the public sector (internal debt and public and publicly guaranteed external debt) is sustainable if it can be serviced each year without incurring new internal or external debt relative to output.9. Sustainability is a more stringent condition than solvency. Maintaining the sustainability condition over all periods ensures solvency, although the reverse is not true, i.e., solvency does not require sustainability for all periods. 38. The central variable for the public sector solvency and sustainability is the primary fiscal balance, i.e., the fiscal balance net of interest payments on domestic and public and publicly guaranteed external debt. For each year in the relevant period, the fiscal balance is calculated that would make the public sector solvent given the public debt stock of the previous year, the future projected real interest rate (10 percent), and GDP growth rate (5.5 percent). The desired fiscal balance is then subtracted from the actual or expected balance to find the adjustment needed to obtain public debt solvency. A positive number indicates that an adjustment needs to be made; the higher the number, the more adjustment needed. This analysis does not distinguish adjustment from revenues or expenditures. All fiscal variables are expressed as ratios to GDP. 39. For public sector sustainability, we first calculate the fiscal balance needed to achieve sustainability, i.e., to keep the ratio of public debt to GDP constant. Again a measure of the required adjustment is calculated by subtracting this desired balance from the actual (or expected) fiscal balance. 7This section is based on a background note prepared by Mr. Hinh Dinh from the Bank's Country Creditworthiness and Risk Division. 8 Solvency is given by the equation: eY -) dt = Bo ,where B is the debt stock, Y is GDP 0 yY and y its growth rate; r is real interest rate. 9 The sustainability condition is given by a simplified expression: - = (y-r) + (X / Y) (x-r ),were D is primary fiscal deficit, B and B are total y y y domestic and external debt stock, respectively; i is the grant component of budgetary, X are exports and x its rate of growth. - 16- Table 9 Fiscal solvency and sustainability 1996 Prmary balance (billion of current pesos) 1.1 Primary balance/GDP 0.4% Real GDP growth 4.4% Public sector solvency adjustment (% of GDP) 0.6% Public sector sustainability adjustment (% of GDP) 0.7% Source: Argentina Country Operations Division data and authors calculation. 40. The analysis of sustainability is simplified in the case of Argentina, since Argentina has instituted a currency board which fixed the peso exchange rate, and inflation tax, which is usually available to other countries, does not exist. Thus, Argentina is not affected by changes in the value of its currency, and does not rely on inflationary financing. Table 9 above shows the results of the analysis. The first three rows indicate fiscal balances and GDP growth. The fourth and fifth rows show the adjustment in the primary balance, expressed as a percentage of GDP, that would make the public sector solvent and sustainable. For 1996 further fiscal tightening would have been desirable. To obtain sustainability and longer-term solvency, Argentina's Central Government needed in 1996 a better fiscal performance amounting to 0.6-0.7 percent of GDP. However, the above results have to be qualified as follows: (i) the fiscal accounts did not incorporate the consolidated fiscal balances discussed above -- or lack of appropriate data -- but only that of the Central Administration; and (ii) were the fiscal performance to include the additional corrections discussed in the previous sections (i.e. the issue of recognition bonds, capitalized interest, fiduciary funds etc.), the required fiscal adjustment presented in the above table would have been higher. Therefore, results presented here are only indicative and likely an underestimate of the needed fiscal adjustment. Furthermore, the strong restrictions of the Convertibility Plan on the financing of public debt, make again those estimates a lower bound of fiscal adjustment. SUMMARY AND RECOMMENDATIONS FOR FUTURE WORK 41. The evidence presented here points to a fiscal imbalance in the 1991-96 period that is broader than generally recognized. The main elements leading to that assessment are fiscal expenditures which have been treated in an extra-budgetary manner, principally including expenditures related to structural reforms (particularly in 1995 and 1996), and the treatment of debt reconciliation through the issue of bonds. These adjustments point to a continuous deterioration of consolidated fiscal accounts (on a cash basis) over the 1992-96 period. If privatization proceeds are excluded from these cash accounts, Argentina's consolidated public sector accounts were at best in balance for only one year, 1993. 42. The above evidence points to the need for fiscal adjustment, both in the conventional sense of securing sustainability of fiscal accounts, but also for sustaining the success of the Convertibility Plan where degrees of freedom for economic policy making are limited by design Additionally, these constraints are compounded by the lack of -17- flexibility created by the present revenue sharing system with the provinces (an issue that will be addressed in a further report), and the limited control that the Central Government has over its expenditures. 43. Early gains in revenue collection, related to the defeat of inflation and better administrative management, plus privatization proceeds, led to expectations of further gains that did not materialize. As a result, on a cash basis, the deficit of the Central Government, more than that of the provinces, continued to grow, if for no other reason than that the Central Government had better access to financing than the provinces (which particularly in 1995 built up arrears). Nevertheless, the deficit of the provinces was large both on an cash and an accrual basis in recent years. Furthermore, the quality of spending in the provinces remained essentially unreformned well until 1995, when the tequila crisis forced the initiation of a process of provincial reform. Addressing the issues emerging from the brief analysis in this report, the purpose of which was primarily to clarify basic fiscal accounts on a consistent basis, will be complex, particularly in view of the nexus of central government and provincial responsibilities both at the revenue and expenditure levels. 44. The Central Government is facing difficulties in revenue collection that go beyond the impact of structural reforms, likely deriving from problems both in tax structure and administration. Tax evasion remains a serious challenge to the fiscal authorities. IMF estimates indicate that VAT compliance levels have increased from 33.9 percent in 1989 to 64.4 percent in 1992, but decreased to 59.5 percent in 1993, and to 55.1 percent in 1994. In 1994, VAT compliance in Uruguay was 72 percent, and in Chile reached 82 percent. Similar difficulties are evident in the collection of the income tax. In 1995, again according to IMF estimates, income tax collections amounted to 2.2 percent of GDP, while in Chile reached 4.3 percent, and in Brazil 4.5 percent of GDP. An evaluation of causes of high tax evasion in Argentina would greatly facilitate the current discussion of tax reform. 45. With high rates of evasion in VAT and income taxes, and high labor tax rates, the evaluation of a potential tax reform would also be advisable, following the assessment of the causes of tax evasion in Argentina. Regarding labor taxes, Argentina is faced with high rates of labor taxation, at a time of urgent need to address a high rate of unemployment (related also to rigid labor legislation at a time of significant structural changes in the economy). Mandatory payroll contributions added to 38 percent of gross wages prior to the 1996 payroll tax reduction. Following the 1996 reduction, that rate declined to 32.5 percent, which is still high by international standards !. Effective payroll taxes are even higher (almost 60 percent) when adding other costs associated with current labor regulations. A reduction in labor tax rates would have fiscal implications that need to be evaluated prior to reform. The negative direct short -term revenue collections effect 10 See "Estimating the Benefits of Labor Reform in Argentina", The World Bank, 1996. - 18- of reduced labor taxes should be assessed against the growth in employment, particularly in the formal labor market, increasing thus the rate of compliance' 1. 46. The current rate of VAT (21 percent) is on the high side by international standards, and in view of its subpar collections rate mentioned above, should be reassessed for possible reduction. However, any reduction in labor taxes or the VAT should be compensated by other sources or revenue (at least until meaningful expenditure reductions are identified). A possible candidate compensating for these two tax rate reductions is the income tax, which as mentioned above, remains a very inefficient source of tax revenue compared to other countries. 17. Tax reform cannot be contemplated in Argentina without assessing the revenue sharing system of the Central Government with the Provinces. The evaluation of the revenue sharing system (coparticipaci6n) should be an integral part of an assessment of tax reform, and precede it. 48. On the expenditure side, the fact that the Central Government has limited flexibility over its expenditures, as seen in this Chapter, and a significant portion of that expenditure is destined for social program, suggests that any efforts to control the fiscal deficit would require an evaluation of the efficiency of expenditure programs, and in particular the public expenditure review of social programs. A public expenditure review remains in order, both to secure savings, but more importantly in Argentina, in increasing the efficiency of spending. So far, most of the emphasis in addressing the fiscal imbalances of 1995 and 1996, has been on revenue collection. Expenditures have not been sufficiently analyzed, particularly in view of the high percentage of expenditures in social programs and argued to be "difficult" to cut. More fundamentally, focusing exclusively on revenue changes, which has been the emphasis so far, may be a mistake, since recent literature indicates that countries which implemented fiscal reforms through adjustments in expenditures were more successful in sustaining gains than countries that dealt with fiscal disequilibria by raising tax rates, or cutting public investment. i.b.i.d. - 19- CHAPTER II: THE FISCAL IMPACT OF STRUCTURAL REFORMS A. FISCAL IMPACT OF ARGENTINA'S PENSION REFORM Background to Argentine Pension Reform 49. Argentina has undertaken a major pension reform effort, the implementation of which began in July 1994. The reform objectives were to make the system fiscally sustainable in the longer term, to make the system more equitable and more transparent, to reduce evasion, and to reduce the government's sole responsibility as pension provider. As a result. the reform had to cover a wide array of areas, from administrative reform to policy reform. 50. Prior to the reform, there existed in Argentina 20 separate pension schemes which were sector-specific and operated at the national level. Then each of the 24 provinces maintained at least two separate pension schemes, one for general provincial civil servants, and at least one for special categories of provincial civil servants, most commonly for teachers, and police. On top of that, several of the larger municipalities also maintained separate pension schemes for civil servants. Each of these schemes had its own administrative structure, its own collection mechanism for contributions, its own set of eligibility conditions, and its own benefit and financing structure. Consequently, the system as a whole was often considered unjust and non-transparent, which led to a loss of confidence on the part of workers and a reluctance to support the system both politically and through active contribution. 51. The reform radically changed this structure. Eighteen of the twenty national systems were consolidated into one, ANSeS. Only the military, national police, and federal judiciary were allowed to maintain separate systems. The Federal Pact of 1993 also allowed provincial and municipal governments to turn their pension schemes over to the national system. Such a proposition was highly attractive to many of the provinces since their pension schemes were running large deficits, and under the terms of the pact, the province would no longer be liable even for pensions of current retirees. On the other hand, all workers would immediately have to abide by the eligibility conditions of the national system which were much stricter than most of the provincial conditions, and would receive benefits calculated at national rates, which were on average half what the provinces were paying. Through the first half of 1997, 10 provinces, and 2 municipalities had chosen to transfer. The only sub-system within the provincial systems which did not transfer was the one for the provincial police, which will eventually be consolidated into the pension scheme for the federal police. 52. From both the efficiency and administrative points of view, the consolidp*# itself has been beneficial. Removing impediments to labor mobility between between provinces improves labor market flexibility and makes any neces- -20 - adjustments in one sector less painful and less costly. Furthermore, the pre-existing systems were more sector specific than occupation specific, and pensions had very limited portability. For example, a secretary in a provincial school system would be eligible for a teacher's retirement scheme, which was generally even more generous than the normal civil servant retirement schemes, and would suffer a loss of pension if she changed jobs even for another secretarial position within the public sector. There would be an even greater loss if she moved to private sector job. This fragmentation of the labor market made necessary public sector downsizing next to impossible. 53. On the administrative side, there are clear economies of scale in administering pension systems. Consolidation of the systems also helps eliminate the possibilities of inidividuals receiving multiple pensions and reduces opportunities for fraud and evasion. Furthermore, the responsibility for payroll tax collection was consolidated within the federal tax administration authorities, to allow economies of scale on the tax collection side as well. 54. From the point of view of improving equity and transparency, having the bulk of the labor force under the same system, with the same eligibility conditions, with the same benefit structure, and with the same contribution structure, is clearly an improvement. In the very short run, some level of inequity remains in that people already receiving pensions were allowed to retain the benefits that had already been granted under the previous inequitable systems. Furthermore, the level of record-keeping in each of these sub-systems varied tremendously, but was generally poor, forcing ANSeS to make guesses about what benefits were due. All of these files are being carefully reviewed by 12 ANSeS, but this process is lengthy and will take time 55. On the policy side, retirement eligibility conditions have been tightened to make the system more fiscally sustainable. The minimum retirement age is rising to 65 for men and 60 for women, fully phased in by the year 2001, from the previous 60 for men and 55 for women. Furthermore, whatever early retirement provisions existed in the individual systems prior to consolidation have been abolished. Finally, the minimum years of contribution required to receive a pension have also been raised from 20 years to 30 years, and the base on which a defined benefit pension is paid has been increased from the average of the 3 highest of the last 10 years to the average of the last 10 years. The first two reforms reduce the inflow of new pensioners, but the increase in required years of contributions also discourages individuals from taking advantage of the pension system with only a few years of service. This both improves equity and discourages evasion. Similarly, lengthening the base period over which the pension is calculated discourages individuals from arranging for large salary increases just prior to retirement, which again makes the system more equitable and limits the ability of individuals to benefit from underreporting of salaries. eontext of this work, ANSeS initiated the review of a stratified sample of its files, which has ,ompleted at the time of completion of this report because of a difficulties in locating files - their contents. -21 - 56. While the above reforms were clearly substantial, the most radical reform has been a redefinition of the role of the State in pension provision. In the long run in Argentina, the State will assume the role of only providing a flat pension to all those who have contributed for 30 years. This pension will be highly redistributive and will provide a basic sustenance pension. Individuals will also contribute to an individual account in a pension fund, where the funds will be accumulated and invested. Each individual will receive an additional pension from the pension fund, a pension based on one's contributions (and therefore earnings) and the rate of return on investments. These pension funds will be chosen by the individual and will for the most part be privately managed, but regulated by the State. This combined system is more fiscally sustainable, is more equitable, more transparent, and reduces incentives to evade in comparison with the previous system. 57. While this was the conceptual design of the new system, at the time of the 1994 law full implementation in this manner was not politically feasible. Individuals, accustomed to the Government providing their entire pension, were not completely comfortable with the notion that the Government would provide only a basic pension and the remainder would have to come from private savings accounts. To accommodate these individuals, the Government allowed those individuals who did not want individual accounts to instead contribute their additional contribution to ANSeS, which would then provide them an earnings-related pension in addition to the basic pension they would have earned anyway. In practice, more than 90% of the new entrants to the labor force are choosing individual accounts and since they cannot change back to the State system later, in the longer term the initial conceptual design will be fulfilled, without a messy political battle. 58. The benefit formulas were also altered to better maintain fiscal sustainability. The basic pension is 27.5% of average wage, with increases of 1% of the basic pension for every year of contributions greater than 30. Individuals also receive 1.5% of the average of their last 10 years' salary for every year worked prior to July 1994. However, individuals who stay with the State system only receive 0.85% for every year worked after July 1994. Thus, a person who worked 35 years in the old system and received average wage throughout, but retires under the new system, receives 81.4% of average wage. This was calibrated to approximate the old system's benefits, so that people close to retirement would not suffer a reduction in pension benefits. A new worker who works 35 years under the new system will receive only 58.6% of average wage. Thus, the average pension payment is being reduced 28% over 35 years. For those who choose the individual accounts, the State pension will be reduced even further to 28% of average wage, which is about one-third of the pension today. 59. However, this long term fiscal gain does not come without short term fiscal pain. Overall contribution rates in Argentina are 27% of wage. Sixteen percentage points automatically go to ANSeS to cover the basic benefit. The remaining eleven percentage points can either go to the individual accounts or to ANSeS to fund the earnings-related pension. But in lieu of ANSeS's receiving 26% of wage as before, which it could then - 22- use to finance current pensions, for contributors who have chosen to contribute to the individual accounts, approximately 70% of total contributors, ANSeS now only receives 16% of wage, with the remainder going to the individual accounts. Thus, in the short run, ANSeS suffered a loss of revenue, without a corresponding reduction in expenditures. Expenditures did not fall immediately because pensioners already receiving pensions continued to receive their pensions based on whatever previous eligibility conditions had applied to them and with the benefit levels based on whatever had applied to them as well. 60. The Government imposed two remedies to deal with the shortfall in ANSeS. First, it earmarked a series of taxes, including a portion of the value-added tax, whose proceeds would go directly to ANSeS. A portion of privatization proceeds were also earmarked for ANSeS, adding further to the revenues of ANSeS. Second, in March 1995, the Government passed a Solidarity Law, which eliminated the automatic wage indexation of pensions which had existed in all the previous systems. Any increase in pensions in progress would be determnined by Congress annually based on what funds were available. Also within the Solidarity Law, a ceiling was imposed on current pensions which would be phased in over a short transition period. Both of these measures in the Solidarity Law were designed to control the expenditures of ANSeS. However, with little recent wage growth and low inflation, the indexation measure, in the short run, has little impact. Ironically, had Argentina been experiencing the inflation of the 1980's, this measure would have had a drastic impact. But under the current fiscal conditions, this measure will only have impact in the longer term. Similarly, evidence from the Municipality of Buenos Aires, which is probably the wealthiest in the nation, shows that only 4% of pensions being paid would be affected by the ceiling. Nationwide, this number would be much lower. 61. Prior to the reform, the social security system was running deficits in the order of 1.5% of GDP. Immediately post-reform, the deficits in ANSeS increased due to the loss of contribution revenue. However, in the future, as the impact of tightened eligibility conditions is felt and as the average pension payment falls relative to average wage, the fiscal situation will improve. Eventually, the aging of the population will result in deterioration, but this will not occur for at least 40 years, and by that time individual accounts will be well established, making further reductions in the basic benefit more politically feasible. These results are shown in the next section. Expected Fiscal Improvements 62. Figure 1 shows the approximate expected path of the finances of the current public social security system both including and excluding the earmarked taxes, from 1996 to 205013. While ANSeS also is responsible for unemployment insurance and the family allowance system in Argentina, these projections cover only the pension system, 13 These estimates were derived using the pension model of the Superintendency of the Pension Funds, SAFJP, by Gustavo Stirparo and Pablo Lattes, under the direction of Rafael Rofinan and Gustavo Demarco. - 23 - benefits for old age retirement, disability, and survivors. The performance of the private pension system is not considered here since it has limited fiscal impact and is well- covered elsewhere.14 63. Table 1O lays out the labor market assumptions behind the projections themselves"5. Labor force participation rates are expected to rise in Argentina from their current 41.8% to 50%, which is approximately the average rate in developed countries today. Table 10 ASSUMPTIONS FOR SOCIAL SECURITY PROJECTIONS Variable 1996 2050 Year in which it reches final value Labor force 41.8% 50.0% 2032 participation rate Unemployment rate 17.4% 7.0% 2016 Contributors 1 35.4% 70.0% 2039 Employed % of contributors who 74.9% 56.0% 2048 are employees % of employees who 61.4% 100.0% 2014 choose funded option % of self-employed 46.2% 100.0% 2018 who choose funded option Real wage growth 0.0% 1.0% 1999 64. Unemployment rates are expected to drop from their current 17.4% to 7%, relatively quickly, as a result of growth in the economy and labor reform. The ratio of contributors to the employed is expected to rise from 35.4% today to 70% in the long term, as record-keeping improves in ANSeS and individuals have greater incentives to contribute to a fairer system. Even the long term numbers are well below the ratios found in countries like the United States and Japan. However, given the very high evasion rates in Argentina today, this increase is projecting a halving of the evasion rate. Given that many of the people who evade are among the self-employed, the percentage of contributors who are employees is expected to fall over time. As noted above, in the longer term, all contributors are expected to choose the funded option. At the current moment, most of the contributors who have not chosen this option are relatively close to retirement. As these individuals retire and new entrants choose the funded option, the 14 See Vittas, "Private Pension Funds in Argentina's New Integrated Pension System," World Bank Working Paper, June 1997. 5 Developed jointly with World Bank staff. - 24 - funded option will cover a greater and greater percentage of the workforce. Finally, the model assumes modest real wage growth. Pension Reform As Initiated 65. Figure 1 shows that the pension system if limited to only revenues from contributions, by employers and employees, is currently running deficits, in the order of 1.7% of GDP, and will continue to run deficits until the year 2009. From that point forward, the system will run surpluses until some time beyond the projection period. The surpluses will begin to fall after 2030 due to the aging of the population, and presumably sometime long into the future, the system will most likely begin to run deficits again. The dotted line on Figure 1 shows the deficits and surpluses of the system if the earmarked taxes specific to ANSeS are included in the revenue projections for ANSeS. The system Figure 1 Evolution of Pension System, 1996-2050 Surplus/GDP 0.03 0.02 - ............ .- ..,,,,,. 0.01 , 0 (0.01) (0.02d i. .. I . .l . t .. 1996 2001 2006 2011 2016 2021 2026 2031 2036 2041 2046 Years Base-Contributions only Base-Contributions+transfers begins to run surpluses much sooner, in 2001. The surpluses do start to fall, but the fall begins a little later, in 2034. These projections include deficits generated by the transfer of the provincial civil servants, which amount to 0.3% of GDP in 1997. These numbers fall steadily as the current stock of pensioners among provincial civil servants falls, causing a deficit of only 0.1% of GDP by 2010. 66. The improvement in the fiscal situation is largely due to the policy impact of the pension reform. However, demographics begin to move against the system in the early 2030's as shown. The-continued surpluses from that point forward are only temporary and due to the increase in the number of contributors as a proportion of the employed. As - 25 - the number of contributors rises, revenues in ANSeS immediately go up. Expenditures remain relatively constant because there is no increase in the number of elderly who qualify for retirement. Expenditures will only start to rise when the increased contributors begin to qualify for pensions, a minimum of 30 years later. Since by 2050 contributors have stopped rising relative to employment, revenues will get no further boosts. However, expenditures only reflect increases in contributors which occurred by 2020, 30 years before. Therefore, expenditures are expected to continue rising further, with no revenue increases to balance them. Table 11 Fiscal balance of the public pension system (% of GDP) Base scenario Current scenario Benchmark (1994) With reduction in contributions Year (Counterfactual) w/o gov. With gov. w/o gov. With gov. transfers Transfers transfers Transfers 1996 -1.24% -1.74% -0.47% -2.26% -0.99% 1997 -1.31% -1.68% -0.67% -2.22% -1.27% 1998 -1.31% -1.45% -0.45% -2.00% -1.04% 1999 -1.26% -1.33% -0.31% -1.88% -0.91% 2000 -1.22% -1.19% -0.16% -1.74% -0.75% 2001 -1.20% -0.93% 0.12% -1.48% -0.47% 2002 -1.21% -0.81% 0.26% -1.36% -0.33% 2003 -1.18% -0.66% 0.41% -1.22% -0.17% 2004 -1.19% -0.55% 0.55% -1.10% -0.03% 2006 -1.16% -0.41% 0.70% -0.96% 0.12% 2006-20 -1.40% 0.23% 1.41% -0.30% 0.88% 2021-35 -1.55% 0.65% 1.91% 0.22% 1.48% 2036-50 -1.41% 0.47% 1.74% 0.17% 1.44% 67. Figure 2 takes a more pessimistic view that the ratio of contributors to the employed never rises beyond 50%. The rate of increase is the same as in the projections shown in Figure 1, but when the ratio reaches 50%, all further improvement ceases. It is difficult to imagine no increases in Argentina in the future, since increasing income levels should increase efficiency of tax collection and make evasion that much harder. As a result, the lower line in Figure 2 can be taken as the worst case scenario with regards to evasion improvement. As shown, even in this case, deficits based on only contribution revenues continue only until 2012. Future surpluses are smaller than in the base case, but they show roughly the same pattern of decline beginning in the early 2030's. - 26 - Figure 2 Minimal Reduction in Evasion Surplus/GDP 0.01 0.005 0 *.. (0.005) - (0.01) - (0.015) (0.02996 2001 2006 2011 2016 2021 2026 2031 2036 2041 2046 Year Base-Contributions only Worst-Contributions only Pension System as it Stands Now 68. However, while both of the above figures show a decidedly good result for the pension system, neither of these figures correspond exactly to the current situation. The Government of Argentina in 1996 instituted a reduction in employer contributions, initially sector-specific and later universal, in an effort to lower labor costs and stimulate employment. Instead of an employer contribution of 16% of wage, employers on average are now contributing only 12% of wage. This is an average since the reductions are region-specific. This has caused a further reduction in the revenues going to ANSeS, and results in greater fiscal deficits within the pension system. Figure 3 shows the case of the reduction in employer contributions. Based on contribution revenues alone, the system will run deficits until 2019. However, if we consider the earmarked taxes, the system will begin to run surpluses in 2004, only 3 years later than the system without the reduction. As Figure 3 shows, though, by the end of the period, there are virtually no surpluses left if only contribution revenues are considered, and since the reduction in employer contribution uses the base case as the starting point, all the issues of further expenditure increases without firther increases in revenues which arise from the contributor expansion hold for this case as well. - 27 - Figure 3 Reduction in Employer Contributions Surplus/GDP 0.03 0.02 .*** 0.01 , 0 (0.034996 2001 2006 2011 2016 2021 2026 2031 2036 2041 2046 Year Base(Contrib only) Base(Contdb+Transfers) Reduction (Contrib only) Reduction (Contrib+Transfers) Benchmark Case 69. How do these results compare with what would have happened had there been no reform? It is always exceedingly difficult to build a counterfactual case, particularly three years after a comprehensive reform has taken place. So much of the economy has changed and so many of the incentives have changed that it is difficult to determine what would have happened had the old rules prevailed. Figure 4 attempts to provide some insight into what would have happened had the old pension system prevailed, in the benchmark case. The benchmark revenue assumptions are derived by assuming the ratio of contributors to the employed remains constant from 1996 onwards. Benchmark expenditures assume that if no contributors would choose the pension fund option in 1996 expenditures of the system are projected to follow the growth in the population of males above age 60 and women above age 55. This expenditure value is clearly just an estimate since in fact the expenditures include pensions for the disabled, and survivors as well. However, this benchmark case probably represents a rosier picture than the true counterfactual since the old system was deteriorating faster than economic indicators would have projected. 70. Even in the rosy counterfactual presented in Figure 4, the reform is currently costing the Government around 0.5% of GDP. By 1999, only five years after the reform, Argentina will be better off with the reform than without the reform. So while the reform has been costly for the Government, the costs are well worth the fiscal improvement in the medium term. Even in the case of the reduction in employer contributions, the reform is costing the Government only 1% of GDP currently, which by international standards is not an excessive cost for a lasting pension reform. By 2004, the Government will be better off having accomplished the reform. - 28 - Figure 4 Comparison to Benchmark Case Fiscal Implications of Reform vs. Benchmark Net Fiscal Cost of Reform Surplus/GDP Surplus/GDP 0.01 0.03 0.005 0 0~~~~~~~~~~00 (0.005) - 0.01 /,f1 (0.01) 0 (0.015) g (0.01) -' (0_02) -;_I.________(0.02)396 2006 2016 2026 2036 2046 (06 2046 Yr996 2006 2016 2026 2036 2046 Year |ase(Contrib only) Benchmark Reduc.(Contrib only) BaseBenchmark Current-Benchmark Other Issues Related to the Reform 71. Another issue in the context of the current reform is the impact of the Solidarity Law of 1995, which removed automatic wage indexation and replaced it with discretionary increases dependent on budget availability. All of the above simulations assume that pensions in progress will receive increases as wages of workers rise, but only covering 75% of the increase in worker wages. That would be roughly equivalent to the expected rise in prices over the period. However, it is possible to calculate the cost of such a policy relative to one where pensions in progress receive full indexation and where they receive less indexation. Figure 5 attempts to make just such a comparison. Because expected real wage growth in Argentina is low, only around 1%, and inflation is very low as well, changes in indexation make very little difference. The early years of the projection period have been omitted from Figure 5 to allow the reader to make the little distinction that exists between the various types of indexation. 72. Another proposed reform is the increase in retirement age for women from the current 60 to the same retirement age as for men, at age 65. This reform also has some impact, but the impact is fairly minimal, largely because of the lower labor force participation of women, particularly at upper ages. These projections are shown in Figure 6. The system shifts from deficit to surplus one year earlier with the rise in retirement age, and the surplus at the end of the period is 0.35% of GDP instead of 0.28% as under the base case. This is not to say that equal retirement ages for men and women is poor policy. Given the greater longevity of women, equalizing retirement ages is good policy. However, from the fiscal point of view, the consequences are not huge. - 29 - Figure 5 Sensitivity to Indexation Surplus/GDP 0.008 0.007 0.006 - 0.003 &-. 0.004 0.003 2011 2016 2021 2026 2031 2036 2041 2046 Year Base-Contributions only Complete Index Less Index Figure 6 Rise in Retirement Age for Women SurpluslGDP 0.01 0.005 - 0 (0.005) - (0.01) (0.02996 2001 2006 2011 2016 2021 2026 2031 2036 2041 2046 Year Base - Contributions only Women Retirement Age Conclusions 73. Despite the initial costs currently being borne by the Government of Argentina, the pension reform has clearly been worthwhile. Even including the reduction of the employer contributions, ten years after the reform the deficits in the pension system will be lower than the non-reform case. -30 - 74. However, in the longer run there are three areas of concern which the Government may need to address. These involve the size of the basic pension, the use of labor taxes and choice of financing for this basic pension, and the need for social assistance arising from the declining pension coverage among the elderly. The Basic Pension 75. First, after the transition period and when no one is any longer entering the public defined benefit scheme, the basic pension, worth 27.5% of average contributory wage, will be financed from the employer contributions. If we assume that all individuals work and contribute for 40 years and retire for 20 years, the 16% original employer contribution will finance a pension worth approximately twice the contribution, 32% of wage. Since some of the contribution will be required to purchase disability and survivors' insurance, the 16% contribution will be just sufficient to cover the pension promise. However, with the reduction in employer contribution, contributions will not be sufficient to cover the basic pension in the long run; twice 12% is much less than 27.5%, even without considering the deductions for insurance. These problems do not show up in the above projections because of the assumed decline in evasion. Should evasion not decline or should the system be projected another 50 years, these problems would appear. 76. One option would be for the Government to reduce the amount of the basic pension to 2 AMPO from the current 2.5 AMPO. An AMPO is the average personal contribution collected by the system. Since the rate is set at 11% of wage, 2 AMPO would lower the basic pension to 22% of average wage, rather than the current 27.5%. Financing purely by contributions would still be tight, but the deficit would be minimal. 77. The 30-year minimum contribution period imply a strong penalty for workers who fail to contribute for that period (although the enactment of the advanced age pension went some way to mitigate this adverse effect). Such a requirement may fight evasion but may still encourage some strategic manipulation (i.e. contributions for small amounits, especially by self employed people, to meet the eligibility requirement). A more equitable approach would be to express the minimum pension guarantee as an accrual rate of the average covered wage per year of contribution. Yet another alternative would be to apply a two-part rule, consisting of a flat pension payable to all old people plus a percentage of the average covered wage for every year of contribution, but to require people with less than 5 years of contributions to be subject to an income and asset test. Such an approach would integrate a social assistance pension with the ordinary retirement pension system. It would avoid poverty traps and would discourage strategic 16 manipulations 16 See "The Argentine Pension Reform and its relevance for Eastern Europe", Dimitri Vittas, mimeo, June 1997 - 31 - Labor Taxes and Pension Surpluses 78. Other issues which the Government needs to address concern labor taxes and other sources of finance for the basic pension. Labor taxes in Argentina are still extraordinarily high, despite the reductions in employer contributions. All of the simulations shown above suggest that the pension system with or without the earmarked taxes could generate surpluses within a relatively short period of time, surpluses which could persist over thirty or more years, but which will be steadily declining over time. Questions arise over the suitability of generating surpluses in the public pension system. How will they be invested? Will the Govermnent be able to avoid the political temptation to dissipate the surpluses by raising benefits? 79. In the absence of an earlier and mnore comprehensive tax reform, discussed in the previous chapter, or previous recommendations for reforming for example the minimum pension system, a better solution for these surpluses might be to lower payroll taxes when the surpluses begin to appear (to address distortions in the labor markets), followed by the reduction in other taxes currently transferred to the system. SocialAssistance Needs 80. Another issue which needs to be raised is the lower level of beneficiary coverage after reform. As Figure 7 shows, currently a high percentage of individuals over the age of 60 are receiving pensions. In the longer term, due to both the increasing retirement age and the greater number of years of contribution required before collecting a pension, a much smaller percentage of those over age 60 will receive pensions. Ultimately, this may lead to increasing demand for social assistance payments. In some respects, the cost of the additional social assistance should be added to the fiscal costs of the pension system or subtracted from its surplus to evaluate the true impact on the aggregate fiscal accounts. Otherwise, we would be comparing a system where a large percentage of elderly receive some form of public support through the pension system to a later system where fewer elderly receive public support through the pension system, ignoring the other forms of public support these elderly would receive. 81. However, despite these caveats and issues for further reform, the conclusion has to be that the Argentine pension reform has been successful, both fiscally and from its objectives of achieving more equity and transparency, despite early costs to the Government. - 32 - Figure 7 Ratio of Beneficiaries to Population Over Age 60 0.8 0.6 - 0.4- 0.2- Agenda for Future Work 82. Three specific issues could facilitate further work in improving the public component of the pension system. First, , evasion of labor contributions remains a serious problem in Argentina, and further evaluation of its causes is necessary in order to propose a strategy for combating it. Evasion of social security contributions reaches a weighted average of 45 percent, according to Governmnent estimates, with evasion of employees at 31 percent, and of self-employed reaching 72 percent. As can be seen from the above projections, reducing evasion in social security contributions could have a significant fiscal impact. 83. Second, additional work needs to be done on the distributional implications of the pension reform. The new system is very redistributive and should reduce poverty among the elderly if all people join and contribute throughout their lives. In practice, a 30 year qualifying period may eliminate pensions for those with incomplete work histories. These individuals are generally poorer than people with complete work histories. 84. Finally it is imperative that resources be committed to producing a flexible, user- friendly pension model that all relevant Government institutions can access. The model used to produce the above simulations is good and with a very small additional investment, the model could serve as the Govermnment pension model and could be used to measure the costs and benefits of all the proposals suggested above and numerous variations on these proposals. - 33 - B. PRIVATIZING ARGENTINA'S PUBLIC PROVINCIAL BANKS 85. Argentina offers a unique opportunity to study a cross-section of bank privatizations. At the beginning of the decade, each Argentine province owned at least one bank. The performance of these publicly-owned provincial banks in the 1990s has been substantially worse than that of private banks, and the losses incurred by the banks have imposed a large fiscal cost upon the provinces. Beginning in 1991, when the provincial government of Corrientes passed a law authorizing the privatization of Banco de Corrientes, provincial governments started to consider privatizing the public banks. The trickle of provincial bank privatizations became a flood after the "Tequila Crisis" of December 1994. Of the nearly thirty provincial banks, almost half had been privatized by the end of 1996, and several other privatizations had been authorized but not completed. 86. This section of the report discusses the bank privatization process and the resulting fiscal implications for the provinces. First, it documents the poor performance of the public provincial banks compared to both large private banks and recently privatized provincial banks. Simulation models, based upon the observed loss rates in net worth, indicate that value of the future fiscal cost of recapitalizing the money-losing public provincial banks exceeds privatization costs. Second, it analyzes the "costs" associated with privatization noting that many of them were incurred prior to privatization, and thus privatization merely exposed existing problems. Third, the section describes how assets and liabilities were divided between the privatized entity and a "'residual entity." It finds that, in general, the size of the residual entity depended upon the quality of provincial assets and upon features of the sales contract. Fourth, available post-privatization evidence suggests that performance of the privatized provincial banks has been better than that of public provincial banks and similar to that of large private banks. This evidence indicates that the risk of insolvency for privatized provincial banks is no greater than the risk posed by the typical private bank. Importantly, although we cannot quantify the social welfare benefits, the data in this section give strong indications that credit allocation is improving. 87. While the focus of this paper is on fiscal benefits -- which appear to be substantial17 -- it is quite likely that the most important benefit of these privatizations is the economic growth that should derive from improved credit allocation. Indeed, these benefits probably would make it wise to privatize even if the fiscal calculus did not favor privatization. The section closes with a study of the political economy of bank privatization in Argentina. It finds, consistent with Bureaucrats in Business, that fiscal pressures, political incentives and bank performance all affect a provincial government's decision to privatize.'8 17 The fiscal benefits may even be slightly greater than those estimated here because we ignore the any additional tax revenue that will be collected from the privatized provincial banks. 18 World Bank (1996), Bureaucrats in Business, (World Bank, Washington DC). - 34 - Public Provincial Banks Performed Worse Than Either Privatized Provincial Banks or Private Banks. 88. Balance sheet data from 1991-96 for provincial, private, and privatized banks indicate that the rate at which the bank's net worth relative to total liabilities declined was, on average, much higher for provincial banks than for either privatized or private banks. In fact, the private and privatized banks nearly maintained their ratio of net worth to liabilities throughout the period 1992-96 (see Figure 8).19 Figure 8 Ratio of Net Worth to Total Liabilities 38%, 33% ._28%- -23% Z18%, 8% 1992 1993 1994 1995 1996 | Private Privatizing *Prov w/o BA XProv BA WProv w/o BA, COR Notes: Private includes the ten large banks in the private banking sample. Privatizing includes those had privatized by 1996, and five banks who were in the early stages of privatization in 1996, but had not as yet sold their banks. Prov BA is Buenos Aires. Prov w/o BA is for all public provincial banks except Buenos Aires. Prov w/o BA, COR is for all public provincial banks except Buenos Aires and the Banco de La Provincia de Cordoba. The figure presents data for only 1992-96 to foster better visual comparisons. 1991 was a year of severe financial disinternediation borne of extremely high inflation. The result was a very high ratio of net worth to liabilities, which declined dramatically in 1992 as inflation and disintermediation subsided. The 1993-96 data should give the most accurate depiction of the steady-state situation. 89. These balance sheet data also allow us to construct a simple simulation model of the re-capitalization cycle. Results from the simulations are presented in Table 12.20 Using the estimated loss rate and assuming that every three years provincial banks I From 1991 to 1992, however, there was a large drop in the ratio of net worth to liabilities for all banks. That decline was largely the product of the exceptionally high ratio in 1991, which reflected severe financial disintermediation (i.e., low liabilities) borne of high inflation. As disintermediation subsided, ratios gradually returned to steady state levels. Throughout the period, however, provincial banks had the steepest declines. 20 The simulations are discussed in greater detail in a background paper which is presented as an annex to the main report. See Clarke, George and Robert Cull," Why Privatize? The Case of Argentina's Public Provincial Banks." - 35 - increase their capital to bring their ratios of net worth to assets in line with Argentina's prudential standards, we calculate the future costs of re-capitalization.21 The cost estimates indicate that the present value of future re-capitalization far exceeds the costs associated with privatization. For a public provincial bank with 25 million pesos in net worth, whose net worth as a percentage of total liabilities declines at 9% per year (the average loss rate for public provincial banks during the 1 990s), the present value of future re-capitalization payments would be nearly 400 million pesos. The costs associated with privatizing that same bank would not exceed 130 million pesos.22 With some recovery of residual assets, the privatization costs would, in all likelihood, be even smaller than that. These large re-capitalization costs, moreover, represent a substantial fiscal burden relative to provincial deficits and expenditures. Table 12 Estimates of the Costs of Re-capitalizing a Typical Public Provincial Bank Ratio of Net Worth to Assets in Year 1: .115 (Arg. Requirement) Net Worth in Year 1: 25,000,000 Loss Rate in Ratio of Net Worth to Liabilities: .090 (Sample Mean, Prov)* Rate of Growth in Nominal Liabilities: .113 (Sample Mean, 1996)* Discount Rate is 10% for Present Value Calculations* Year Net Worth Total Total Implied Nominal PV Assets Liabilities Asset Re-Cap. Re-Cap. Growth Payment Payment During millions millions millions Cycle millions millions (1) (2) (3) (4) (5) (6) (7) 1 25 217 192 .020 0 0 3 -11.9 226 238 .020 42.9 29.' 6 -14.8 280 295 .020 53.1 27.2 9 -18.3 347 366 .020 65.8 25.3 12 -22.6 430 453 .020 81.5 23.6 96 -9092.0 172747 181839 .020 32731.0 3.2 99 -11262.8 213994 225257 .020 35881.6 2.6 Total _ 205515 383 * These assumptions and further results using a broad range of discount and loss rates are discussed in detail in an annex to the report. 21 The qualitative results of the analysis are similar when we assume that public provincial banks re- capitalize in each year, every five years, or every seven . Though arbitrary, the three-year re- capitalization assumption seems plausible, and, most importantly, is not an important determinant of the outcomes presented here. 22 For further details regarding these calculations and for a discussion of the assumptions made in the estimation the reader is referred to . The annex also presents additional results for a range of parameters. - 36 - Privatizing Public Provincial Banks Forced Provincial Governments to Realize "Costs" Incurred Prior to Privatization. 90. An interesting feature of the loss rate data is that, in the first year after a privatization was announced, the net worth of the bank in question always dropped dramatically. These declines coincided with stringent audits of the privatizing bank's assets, which suggests that, prior to privatization, net worth data was somewhat less reliable than for the post-privatization period.23 This indicates that the loss rates used in the simulation models, which were based on pre-privatization balance sheet data, may understate the severity of the public provincial banks' situation. Lacking other data, however, we had little choice but to rely on these balance sheets. We recognize that the simulations in Table may understate substantially the costs associated re-capitalizing a typical public provincial bank. 91. How a province recovers its residual entity's assets and how it dispenses with its liabilities will have a substantial short-term fiscal impact. In almost all of the privatizations to date, division of assets and liabilities between the privatized and the residual entity was dictated by political and banking realities. Provinces managed to sell off only the perforrning assets in their banks' portfolios. From an economic perspective, banks with negative net worth and little hope of future profits should sell at negative prices. Politicians may, however, find it difficult to explain to constituents why the government has to pay a private bank to acquire its public provincial bank. From the perspective of a private purchaser concerned about its banking reputation, moreover, being saddled with a public provincial bank's liabilities and non-performing assets is an unattractive alternative. Borrowers that once found it in their interest to default on public provincial bank loans may now find it their interest to behave as responsible debtors with respect to a private bank. Better to let the privatized entity start afresh with such a borrower rather than force it to collect on past provincial loans that neither the state- owned bank nor the borrowers truly expected to be fully repaid. 23 This is not to suggest that these banks had not undergone external audits in the past. Rather, the audits undertaken prior to privatization were more rigorous, presumably so as to attract private buyers. Whatever the motivation, the post-audit write-down of assets and the increased provisioning which it implied were substantial. Throughout the section, these audits are referred to as pre- privatization audits. Not surprisingly, the losses realized in the course of these audits were often attributed to privatization by those who benefited from the old arrangement. Privatization or no, however, these losses had been incurred, and would have had to be treated in some way. To confront them as overtly as did the privatizations may have entailed additional costs, but these were political not fiscal costs. - 37 - The Size of the Residual Entity Depends Upon the Quality of the Public Provincial Bank's Assets. 92. In most cases assets and liabilities were apportioned between the privatized provincial bank and the "residual" entity by matching performing assets with nearly an equal amount of liabilities. Therefore, the sales prices of the privatized entities were quite small, especially in comparison with the size of the residual entity. Additionally, most provinces a reed to jointly capitalize the privatized provincial bank with the winning bidder. The sum of these capitalization costs and the eventual losses associated with liquidating the residual entity will be the realized costs of privatization. To some small extent, they were defrayed by the sales price of the privatized entity. Table 13 Size of Residual Entities Bank % of Pre-Priv. % of Pre-Priv. % non- % Normal Physical Assets Liabil. performing Pre- Pre-Pnv. Assets as % of Priv. Pre-Priv. (1994) (1994) Assets (1) (2) (3) (4) (5) (6) Chaco 54.0 51.3 **32.6 **52.6 8.9 Formosa 36.7 66.2 4.0 79.4 2.5 Mendoza 61.2 61.2 20.2 49.5 0.3 Misiones 28.6 65.8 12.5 71.6 5.3 Prv Soc Men 49.7 49.7 34.2 43.4 1.0 Rio Negro 24.6 23.7 47.3 28.4 1.2 Salta 73.3 58.7 75.9 7.5 4.5 San Juan 34.2 76.4 23.1 68.8 1.5 San Luis 31.6 87.1 12.9 80.9 35.5 Sant del Est *81.1 *84.4 ***71.8 ***13.8 3.7 Tucuman 66.7 67.0 53.4 43.8 4.6 Sources: The % non-performing and the % normal loans data comes from B.C.R.A. balance sheets. Data on physical assets comes from Fondo Fiduciario balance sheets for residual entities at the time of privatization. Prv Soc Men is the former Banco de Prevision Social de La Provincia de Mendoza; Sant del Est is the former Banco de La Provincia de Santiago del Estero. The asset data for Santiago del Estero implied that their residual entity was 135% of pre-privatization assets, an obviously unrealistic estimate. The figures reported here for Santiago del Estero in columns (1) and (2) are, therefore, based on their reported assets in the year of their privatization (1996). 1993 data were used. By 1994, the effects of privatization were evident in Chaco's portfolio quality data. ... 1993 data were used. No data were available for 1994. 24 Put another way, many provinces decided to maintain ownership of some fraction of the shares of the privatized entity. Because they could have presumably sold these shares and used the proceeds for other government projects, the retained shares should be thought of as a fiscal cost associated with privatization. 25 The Santiago del Estero privatization was finalized in September, 1996. The Fondo Fiduciario data for the residual entity are dated July 31, 1996, and should, therefore, be an accurate reflection of assets and liabilities at the close of the sale. The asset and liability figures from the Central Bank for 1995 for Santiago del Estero are much smaller than those in the residual entity. Inflation in 1995 was relatively low, so it seems implausible that nominal assets and liabilities could have increased so much as to permit a residual entity larger than the old public provincial bank. - 38 - 93. While capitalization costs and sales prices are fixed, the realized costs associated with liquidating the residual entities are still quite uncertain. A province that negotiates its way out of substantial liabilities while at the same time recovering substantial non- performing assets may sharply curtail the short-term cost of privatization. Unfortunately, there is multi-period data on the liquidation of the residual entity for only three provinces. In many cases, residual entity asset recovery was left to the privatized entity who received a percentage of the recovered loan payments. Percentages were typically decreasing in the quality of the asset recovered; the worst loans -- which were presumably the most difficult to recover -- generated the highest percentage payment for the manager of the residual entity. We should emphasize that with only three cases, it is too early to analyze the links between liquidation contracts and asset recovery. We also emphasize, however, that even if no residual entity assets were recovered, the re-capitalization simulations indicate that the typical province would generate large fiscal savings from privatizing its bank. 94. On the liabilities side, the Federal Government and the World Bank employed an interesting strategy to ease the fiscal burden of provinces facing massive short-term liabilities that heretofore had gone unrealized. With Bank assistance, the Federal Government created the Fondo Fiduciario, a fund designed to convert short-term provincial liabilities into long-term loans. Proceeds from the Fondo's long-term loans went to pay the residual entity's creditors; the province will pay off the Fondo over time. The fiscal impact of converting short-term obligations to long-term loans is difficult to estimate as the counter-factual is difficult to specify. It is not clear how provincial governments would have financed these liabilities without access to the Fondo Fiduciario.26 In our fiscal estimates, therefore, we ignore the benefits of meeting provincial obligations on beneficial financial terms. The political feasibility of the privatization program, however, depended crucially on the ability to convert obligations from short to long term. A province in a desperate fiscal situation would have found it hard (if not impossible) to finance its obligations privately; meeting them all at once would have made a terrible fiscal situation worse, if it could have been done at all. Indeed, given the tax base of many provinces, meeting all public provincial bank obligations immediately was probably unaffordable. The extent and the terms of Fondo Fiduciario assistance to each privatizing province in greater detail in an annex. 95. The sales prices of public provincial banks can also be related to characteristics of the sale. Again, the proceeds from the entire privatization transaction will often be negative because capitalization costs and the eventual losses of the residual entities may 27 far outweigh the small positive sums received for the privatized provincial bank. How negative depends, to some extent, on features of the sale and the nature of assets. For 26 These provinces probably had little, if any, access to private credit at this time. Obviously, this does not imply that the price paid for the privatized entity was ever negative. What was privatized was the so-called "good bank," which represented a fraction of the former bank. Good banks were designed as viable businesses with equity sufficient to meet regulatory capital requirements. - 39 - example, in some cases, the public provincial ban-k had a substantial amount of fixed assets (buildings, in particular) that may have been difficult to liquidate and therefore may have had a dampening effect on price.28 We regress residual liabilities on features of the old public provincial bank including its ratio of performing loans to total loans (just prior to privatization) and its fixed assets. Although the number of observations available is quite small, the preliminary results confirm economic intuition -- burdensome features of the sales contract and low quality provincial assets are reflected in the net returns to privatization (i.e., a large residual entity).29 Future privatizers may benefit from these estimates when contemplating the short-term fiscal impact of privatization. 96. Table 14 summarizes the total fiscal costs associated with privatization for individual provinces. The total potential short-term costs associated with privatization are equal to the sum of any capitalization costs bome by the province and residual entity liabilities, minus the price paid for the privatized entity (column 5). There were, however, other incremental costs associated with privatization such as fees paid to advisors, auditors, and investment bankers. Those costs were presumably quite small in comparison with the others and are, therefore, left out of the analysis. Final privatization costs will depend on the extent to which provinces recover residual assets. Column 6 lists total privatization costs assuming that 20% of residual assets are recovered; Column 7 assumes a 50% recovery rate. Clearly, total costs will be quite sensitive to the success of the recovery effort. The total cost estimates in Table 14 assume that the province pays off all of the residual liabilities. To the extent that they are able to negotiate their way out of some liabilities, the total cost figures in columns 5-7 should be reduced. 28 The total potential short-termn costs associated with privatization are equal to the sum of any capitalization costs borne by the province and residual entity liabilities, minus the price paid for the privatized entity. There were, however, other incremental costs associated with privatization such as fees paid to advisors, auditors, and investment bankers. Those costs were presumably quite small in comparison with others costs and are, therefore, left out of the analysis that follows. 29 The simple regression that displays the relation between the variables is: % Assets = 27.8 + .548 (% non-performing) + .798 (% physical assets) (t-stat) (2.53) (2.49) (0.39) N=9, Adj R-Squared .352 (% non-performing) is the percentage of total loans in the worst two B.C.R.A. loan classifications. The physical asset variable is the percentage of pre-privatization assets that ended up in residual entity balance sheets under the heading "bienes de uso." Many of the public provincial banks had an abundance of branches and buildings that purchasers might have preferred not to own (so they would not have to re-sell them later). The dependent variable is the percentage of pre-privatization assets shifted to the residual entity from Table 13. - 40 - Table 14 Total Fiscal Costs Associated With Privatization Under Various Residual Asset Recovery Scenarios Province Amount Value of Total Total Total Fiscal Total Fiscal Total Fiscal Paid for Privat. Liabil. of Assets of Costs of Costs of Costs of Privat. Shares Residual Residual Privatization Privati- Privati- Entity Retained Entity Entity 0% Asset zation: 20% zation: 50% by Prov. Recovery Asset Asset (millions (millions (millions (millions Recovery Recovery pesos) pesos) pesos) pesos) (5) (6) (7) (1) (2) (3 4 Chaco 6.3 2.9 233.1 245.3 229.7 180.6 107.0 Ent Rios 15.1 n.a. n.a. n.a. n.a. n.a. Formosa 9.3 4.0 244.9 135.7 239.6 212.4 171.7 Mendoza 20.1 0.6 666.6 666.6 647.1 513.8 313.8 PS Mend 8.2 0.2 292.1 292.1 284.1 225.7 138.0 Misiones 9.1 0.0 331.8 144.2 322.7 293.9 250.6 R Negro 10.2 1.8 47.4 49.2 39.0 29.2 14.4 Salta 4.4 1.1 73.4 91.6 70.1 51.8 24.3 San Juan 11.3 3.8 175.3 78.6 167.8 152.1 128.5 San Luis 5.4 0.0 81.8 29.7 76.5 70.5 61.6 Sant Est 6.7 0.0 227.3 199.6 220.7 180.7 120.9 Tucuman 10.3 2.6 262.9 261.7 255.2 202.9 124.4 Notes: n.a.- not applicable. In the Entre Rios pnvatization, no residual entity was created. As computed, here, the total cost estimates here are not meaningful for that case. PS Mend is the former Prevision social de Mendoza; R Negro is Rio Negro; and Sant Est is Santiago del Estero. The Insolvency Risks Posed by Privatized Provincial Banks Are No Greater Than Insolvency Risks Posed by Other Private Banks 97. Another potentially important fiscal issue is the future solvency of the privatized provincial banks. If privatized provincial banks continue to operate as poorly as public provincial banks, they will go bankrupt. Although predicting the future solvency of any bank, public or private, is a speculative endeavor, the available post-privatization data indicate strongly that privatized provincial banks have become quite distinct from provincial ones in their operations. In fact, data presented in Table 15 on operating income and costs indicate that privatized provincial banks now differ little from the ten largest private banks in operation in 1996. Similar results obtain with respect to portfolio quality and credit allocation by sector.30 A strong case can be made, therefore, that the solvency risks posed by privatized provincial banks are no worse than those posed by the typical private bank. Given the incentives faced by owners of the privatized provincial bank, this may come as little surprise, especially to economists. Empirical confirmation of simple economic intuition, however, may be especially welcomed by policy makers -- although it should be emphasized that the post-privatization experience has not been long, and future data are required before ironclad conclusions can be drawn. 30 See annex for details. - 41 - Table 15 Ratios of Operating Income to Costs Financial Income from Total Income/Administrative Services/Administrative Income/Administrative Costs Costs Costs Mean Mean Mean Overall: Prvate .915 .529 1.44 Provincial .375 - .402 .778 Privatized .669 .736 1.41 1994: Private .764 .543 1.307 Provincial .524 .386 .911 Privatized .413 .798 1.211 1995: Private .883 .494 1.377 Provincial .245 .339 .584 Privatized .633 .751 1.385 1996: Private 1.088 .506 1.595 Provincial -.026 .425 .399 Privatized .758 .742 1.501 Data Source: B.C.R.A. Notes: The privatized sample includes only those banks that had completed their privatization as of March, 1996. Only these cases provided at least six months of post-privatization data on income and costs. The cases include Chaco, Corrientes, Entre Rios, Formosa, La Rioja, Misiones, Rio Negro, and Salta. The private sample includes the ten largest private banks in Argentina as of 1996. All public provincial banks are included in the provincial sample, including pre-privatization observations for the eight banks in the privatized sample. In those years where data were available for only a sub-sample of months for a given bank, the data were annualized. For example, for Salta in 1996, costs and income data were totaled over the six months for which data were available, and then multiplied by two. Fiscal Pressure and Poor Bank Performance Increased the Likelihood that the Provincial Government Would Decide to Privatize the Public Provincial Bank 98. The previous discussion makes it clear that, at every point in the privatization process, policy decisions were guided as much by political reality as by economic theory. In an effort to better understand the political economy of privatization the decision to privatize is modeled as a function of the quality of the provincial bank, the fiscal situation of the province, the political incentives of decision makers, and exogenous shocks to the banking sector.3' Table 16 describes the variables used in the econometric analysis and Table 17 presents the coefficient estimates. 31 The econometric methodology is described in detail in a background paper included as an annex to the main report. -42 - 99. Bank quality, fiscal pressures, and political incentives all played a role in privatization decisions. Poor bank performance greatly increases the probability that the provincial government will decide to privatize its bank. The estimation indicates that a 1% decrease in net worth over liabilities increases the rate at which banks were privatized by 3%.32 These results are consistent with the hypothesis that provincial governments are less willing, or able, to support poorly performing banks. Likewise, provincial fiscal difficulties increase the probability that the government will decide to privatize the bank. A 1% increase in the provincial government's deficit (as share of total provincial revenues) increases the rate at which banks were privatized by 0.7%. This result is consistent with the hypothesis that political crises can force governments to reform. Finally, when the governor belonged to the Partido Justicialista, the rate at which the provincial government privatized public provincial banks was more than six times the rate when the governor was a member of the Uni6n Civica Radical or an independent party. 100. The estimation method used in the analysis controls for the possibility that exogenous shocks might affect the probability of privatization. It appears that one exogenous shock, the Tequila Crisis, might have had a pronounced effect on privatization decisions. Facing a severe liquidity crunch in the wake of substantial deposit flight, many public provincial banks were privatized after the crisis. Public provincial banks appear, in general, more likely to be privatized after the Tequila crisis, even after controlling for bank quality, fiscal performance and the political party controlling the provincial government. 101. These results should help inform advocates of bank privatization, such as the World Bank, as to when political pressure to privatize is greatest. They should have implications not only for Argentina's remaining provincial banks, but also for other countries that have state-owned banks frequently in need of re-capitalization. 32 Elasticities are calculated at the sample means of all variables. - 43 - Table 16 Independent Variables in the Estimation. Variable Comments Does the PJ control A dummy variable indicating that the Partido Justicialista (President the assembly and the Menem's party) controls both the Governorship and provincial executive? (1=yes, Congress. If the provincial congress is bicameral, this indicates that O=no) both the Senate and the Chamber of Deputies are controlled by the PJ. Can either the UCR or ................ A dummy variable indicating that the Union Civica Radical (UCR) or a independents block? single independent party controls at least one chamber of the provincial (1=yes, O=no) Congress or the Governorship. In practice, since when either the UCR or a single independent party controls either chamber the governor also belongs to that party, this indicates that the governor is either a member of the UCR or an independent party. ~~~~~~~~... ...... ....... .. ,- .... ............................................................................................................................................... Can UCR block in the A dummy variable indicating that the Uni6n Civica Radical controls (at assembly or the least one chamber of) the provincial Congress or the Governorship? In executive? (1=yes, practice, when the UCR controls either the Chamber of Deputies or the O=no) Senate, the governor also belongs to that party. Can independents A dummy variable indicating that a single independent party control block in the assembly either (at least one chamber of ) the provincial Congress or the or the executive? Governorship? In practice, when either a single independent party (1=yes, O=no) controls either the Chamber of Deputies or the Senate, the governor also belongs to that party. .............................. ..................... .............................................I................................................................................................................. ...... Bank Net Worth over The bank's net worth over its liabilities (lagged six months). Liabilities ..ii...................................................... ......................................................................................................................................................... % of Bank Loans to Percent of the bank's loans made to the public sector (lagged six Public Sector months). ............................................................ ................................................................................................... ............................................................ % of Bank Loans Percent of the bank's loans that are not overdue. (lagged six months). considered "normal" The definition of bad loans and reserve requirement against lrans (either definition) changed in 1994. Therefore, we include an extra term allowing the coefficients on this term to change when the definition changed. Government Deficit as Government deficit as percent of total revenues (lagged one year). Share of Revenues Government deficits are positive and surpluses are negative. This is used, rather than deficit as share of provincial GDP because good measures of provincial GDP were not available. Change in Net Worth The change in the banks net worth over its liabilities (lagged six over Liabilities months). An increase in net worth is positive. .................................... ............................ ...................................................................................................... ........................................ Can the Governor Run A dummy variable coded one if the sitting governor is (constitutionally) for Re-Election? able to run for re-election. (1=yes, O=no) - 44 - Table 17 Proportional Hazard Model for Bank Privatization (Buenos Aires Included) I (1) 1 (2) (3) 1 (4) Proportional Hazard Model (Monthly Data) Prov. de Buenos Aires Included Yes Yes Yes Yes j g ~ i ....................................... ..............................2 2 .......................... ................................ ~~~~~~Banks ~~~~~~~~~23 23 23 17 . .................................................................................................... ............................... ............................... ............................... .......................... .of Privatizations 16 16 16 12 ...................................................................................................... ............................... ............................... ....................I........... .......................... # of Months at Risk of PrivatizatiQn 1010 1010 1010 825 Does the PJ control the assembly and -0.732 -0.625 -0.598 -1.156 the executive? (1=yes,O=no) (-0.91) (-0.78) (-0.75) (-1.02) .....................~~~~~~~~~~.................... ..............................I............................... ............................... ................. ......................................... Can either UCR or independents -1.865** -1.822** -3.070** block? (1=yes,O=no) (-2.19) (-2.19) (-2.02) ...........................................................- i i iw o ... h ........ ..........i..... ......... ....... ......................... ...... ......................... .......................... ... Can UCR block in the assembly or the -1.466 executive? (1=yes,Ono) (-1.53) ........................... ..... ... .......................................... ................. ............................... ............................... ............................I... ................. .an independents block in the -2.553** assembly or the executive? (-1.92) (1=yes,O=no) Bank net worth over liabilities -16.989** 16.622** 17.09** j 3.631 ** (-2.63) (-2.64) (-2.83) (-1.99) ./0of bank loans to public sector 5.368** 5.645** 5.552** 2.134 (2.09) (2.22) (2.20) (0.47) .................................................................I................................... ............................... ............................... ............................... ......................... %of bank loans considered "normal" -7.321 ** -7.289** -7.423** -7.441 ** (either definition) (-3.34) (-3.22) (-3.32) (-2.45) ...........~ ~ ~~~............................................................ ............................... .....................I.......... ............ ........ .. ................ ........... .................... % ofbank loans considered - "normal" 2.807** 2.499* 2.514* 2.975* (old derinition) (2.04) (1.90) (1.90) (1.92) overnmen def cit as share of 8.01 6** 8.038** 8.01 8** 8.842** revenues (2.44) (2.46) (2.46) (2.40) Change in net worth over liability -2.841 -2.080 (-0.35) (-0.27) ..d i ii ... ................................................................................ ................. .............. ................(-35 0.2 .... .......... ................. ....... ............. Can the govemor run for re-election? -1.605 (1-yes, O=no) _ (-1.36) Log Likelihood -25.56 -25.84 -25.88 -16.11 - 45 - CHAPTER III: CREATING A MORE FLEXIBLE FISCAL SYSTEM A. REFORMING THE REVENUE SHARING SYSTEM WITH THE PROVINCES 102. Over the past decade, the Bank has conducted extensive analysis of provincial finances and the incentive problems posed by the system of intergovernnental fiscal relations.33 The major conclusion is that the present system is not conducive to fiscal efficiency at the provincial level. Provinces tend to rely too much on transfers and neglect the potential of their own tax revenues. On the other hand, the allocation of intergovernmental transfers across the provinces has lost its rationale leading to questions of equity. The country authorities have well understood the importance of the issue whose complexity is a major obstacle to its own solution. By Constitution, the overall system was due for reform in 1996, but given the apparent lack of agreement between provinces and the national government on a strategy and direction of the reform, the deadline for the new regime has been postponed to 1998. 103. Before this deadline, the central government and the provinces would have agreed on the basic pillars of the new revenue sharing system: (i) the composition of the tax pie to be shared between the federal government and the provinces; (ii) which taxes will be appropriated by the federal government or the provinces and which would be included in the revenue pie; (iii) how to distribute the overall provincial share among individual provinces taking into account tax potential and capacity, population needs and poverty, economies of scale and cost differentials in providing public services. 104. The present chapter contributes to this wide theme. Based on a thorough analysis of existing data on the financial flows between the federal government and the provinces as well as the potential for tax collection at the provincial level -- presented in detail in Annex III the report simulates policy options in greater detail than in the past. It addresses three questions, in particular: a) What are the totalfinancial revenueflows from the federal to the provincial governments? Due to the evolution of the inter-governmental transfer system, the actual flows to the provinces are much larger than the basic coparticipation system. New estimates were made of the total federal fiscal flows to the provinces, including direct federal expenditures, which shed some light in an otherwise complex transfer system. b) How independent could the provinces be in financing their own spending? What is the provincialpotentialfor tax collection? On the revenue side, estimates were made of (i) the federally collected tax revenues collected in each of the provincial jurisdictions; (ii) the potential revenues that provinces 33 Argentina: Provincial Government Finances (1990), Argentina: Towards a New Federalism (1992), Argentina: Provincial Finances Study (1996), and provincial sections of the economic report, Argentina: the Convertibility Plan (1996). -46 - could collect through their own tax instruments; and (iii) the financing gap that would exist for each province under a regime in which provinces were to collect provincial taxes at their full "potential" level and federal transfers were allocated according to the estimated share of federal taxes collected in each jurisdiction. c) How to finance the unavoidable financing gap? While some provinces would be able to finance their own expenditures out of their own taxes, for the majority of the provinces there are evident needs of redistributrion. The report simulates the impact of implementing a Canadian-style equalization formula for compensating provinces where financing gaps were due to a below average per capita tax base. 105. The results of the analysis, although preliminary, are important. First, the structure of total flows to the provinces changes significantly when all the federal- financed programs are included. The direction of the change does not reflect such factors as population or poverty; some provinces end up loosing part of the cake, while others gain. Second, there is a substantial potential for increasing tax revenues -- provinces' own tax revenues could increase by 45% by emulating the collection record of the most efficient province -- making provincial governments more responsive and fiscally responsible to their constituencies. Third, per capita expenditures vary widely among provinces and such variation has no apparent justification on poverty accounts or population density. Per capita expenditure seem to follow transfers availability which indicates large margins for savings. Fourth, devolution of tax revenues to the provinces on an origin base has a dramatic effect in the way that largest provinces could finance tneir needs but leaves a clear role for equalization programs due to the concentration of tax bases. Fifth, using piggy backing system on the income tax has a relatively small impact which reflects the limited role of this tax in the present Argentine tax system. Sixth, using a Canadian equalization formula brings very sound results on the way equalization could work taking into account differences in provincial tax bases. The formula could be enriched in the future including factors such as the cost of providing public services. While further analysis is needed to guide the reform agenda, these preliminary results reveal the potential that can be realized due to the new data collected, organized and estimated for this report. 106. The chapter is organized into five sections. The first section outlines general principles to guide policy reform in the area. The second section describes the present system of inter-governmental transfers as it has evolved since 1988 when the Law of Coparticipaci6n de Impuestos was approved. The third section estimates the potential for additional tax revenue at the provincial level and the alternatives to financing the fiscal gap. The last two sections summarize the main conclusions of this chapter and the agenda for future work. - 47 - Principles 107. Literature and previous work have identified three overlapping objectives to guide policy reforms in intergovernmental fiscal relations: increase the degree of correspondence, enforce hard budget constraints and improve the equity of the current system. 108. First, the new system should increase the degree of correspondence so that, to a large extent, the local population pays for the services that it receives. The degree of correspondence does not need to be 100 percent, since there are externalities which require some form of federal intervention -- for example, basic levels of education for a mobile population or disease control. For these reasons, there may be a need for federal transfers to compensate for regional differences in tax bases, allowing poorer provinces to have the resources to meet minimum service standards. The general rule remains, however, that a province's public services should be financed primarily by that province's population, under the basic principle that efficiency in public expenditures will be enhanced when voters directly confront the cost of local policy makers' decisions. 109. Secondly, the reforms should provide incentives for provinces to achieve fiscal sustainability. The rules of taxes and transfers should be fixed and predictable, so that provinces will face a hard budget constraint. Also, the rules for provincial contracting of debt should be fixed to avoid the perception of implicit federal guarantees. 110. Thirdly, the reforms should reduce clear distortions in the distribution of resources across provinces, improving the degree of equity of the current system. By "distortions," one means differences in per capita resources so large that they cannot be justified by economies of scale or differences in the degree of economic development or natural resource base. Background. 111. Provinces are responsible for spending about 40 percent of all public sector resources in Argentina (including the municipal level). The quality and efficiency of public goods and services in Argentina are determined to a large extent by provincial government performance. Most importantly, provinces are responsible for primary and secondary education and basic health services. 112. In the context of Bank studies and lending activities34, the Bank and the Argentine federal government have identified a number of inefficiencies in public service delivery at the provincial level: over staffing and overly generous retirement benefits, to name just two. The results in many provinces have been poor quality of education and health services; poor access to basic water supply and sanitation services; and, deteriorating infrastructure. These service delivery problems were combined with persistent and large 34 For example, C6rdoba: Public Sector Assessment, the Provincial Development Loans, and the Provincial Reforn Loans. - 48 - budget deficits, on the order of 5 to 10 percent of total provincial revenues in recent years. 113. The system of intergovernmental fiscal relations can provide critical incentives or disincentives for provincial populations to demand better services and for provincial government officials to maintain fiscal equilibrium. The Current System of Intergovernmental Transfers. 114. Despite the fairly high degree of decentralization of expenditure responsibilities, Argentina has traditionally maintained a highly centralized system of taxation. Intergovernmental transfers fill the huge financing gap. There are two broad categories of federal to provincial transfers: (1) untied, general revenue-sharing, known as "coparticipation," and (2) sector specific or specific purpose transfers. Table 18 Share of Provinces and Federal Government in Total Expenditures and Revenues Year: 1996 Share of Expenditures Share of Revenues Collected Provinces 45% 20% Federal Government 55% 80% Total 100% 100% Source: Accounts on a cash basis, chapter I of this report. 115. Coparticipation. Federal coparticipation of taxes (Federal Coparticipaci6n de Impuestos) consists of untied, automatic transfers to the provinces. The current law gcverning this system was established in 1988. The revenue-sharing pool is made up of all the major federally collected taxes, with the exception of the fuels and energy taxes (which are shared according to different rules to be discussed below) and payroll taxes, which remain at the federal level to finance the national social security system. The "primary" distribution of this revenue-sharing pool, that is the distribution between the federal level and the provincial level is: 41.95 percent for the federal government, 57.05 percent for the provinces, and the remaining 1 percent for a pool of discretionary transfer 35 fnids.3 The "secondary" distribution, or the distribution of the provincial share across provinces, is based on fixed coefficients that were agreed in the 1988 basic law.36 Table 19 depicts those coefficients which corresponded to the weight each province had in total provincial spending at the time. 116. Modifications in the 1988 basic law. A number of modifications have been introduced into the system since 1998. The most important are the following: ATNs, or national treasury grants. Although the source of funding is automatic, the distribution across provinces is purely discretionary. 36 Prior to 1988, the formula included several parameters, including population, unsatisfied basic needs. - 49 - (a) The final phase of decentralization of primary-secondary education and basic health services to the provinces occurred in 1993; along with this, a fixed amount of the regular revenue-sharing would be earmarked for those sectors and distributed among the provinces according to the cost of providing the services transferred, rather than the fixed coefficients; (b) The Fiscal Pacts of 1992 and 1993, during a period of rapidly growing revenues, established (i) the earmarking of 15 percent of the gross pool of coparticipation for financing the federal social security system, (ii) a fund of $550 million per year deducted from the coparticipation pool which would be distributed according to agreed fixed amounts for each province, and (iii) the federal government made a commitment to guarantee a minimum monthly floor of $740 million to be distributed among the province in cases where monthly revenues fell to unusually low levels; c) Changes in the distribution of the income tax before it would feed into the revenue-sharing pool were introduced, with 10 percent of these revenues going to the province of Buenos Aires (up to a maximum of $650 million), another 4 percent to the rest of the provinces (plus the excess over $650 million) according to indices of "Unsatisfied Basic Needs;" d) Deduction of 11 percent off of the VAT revenues before those revenues enter the coparticipation pool--9.9 percent for the federal social security system and 1. 1 percent for the provincial pension systems; and e) Personal assets tax revenues, which had been earmarked mostly for social security, would now enter the general coparticipation pool. 117. Special purpose transfers. There are also a variety of special purpose automatic transfers, the most important of which are associated with the sharing of the fuels tax revenues. Of federal fuels tax revenues, 21 percent are earmarked for the federal social security program, 29 percent are general revenues for the federal government and the rest go to the provinces, but earmarked for several special public works programs in areas like housing (FO.NA. VI-- see chapter III.B), rural access to electricity services (FEDEI) and road construction (Fondo Vial). The chart on the next page presents a description of the various revenue-sharing programs and special programs. 118. Combining both the regular revenue-sharing law, its modifications and special transfers, results in a system which is difficult to understand and almost impossible to justify on economic or social grounds. The general revenue-sharing coefficients do not follow clear economic, social or expenditure need criteria, and the various partial attempts to remedy the system have resulted in a tremendous degree of complexity and lack of transparency. The large differences between provincial shares of transfers and provincial population shares result in transfers per capita that vary widely across jurisdictions. Not surprisingly, provincial government expenditures per capita also vary widely as seen below. ~The proine.CanO GU'AT YEAR 1997 M ASSIGNMENT OF FEDERAL TAX REVENUES Fot - ntaxandthefrst tFORMATION OF THE REVENUE-SHARING POOLS INCOME TAX 60% Elec. Energy Tax pre-deduction ~~~~~~~~~~~~~~#q~~~ inioal40 [ Miion per Yeai 0lion ( g) 79 % 21EL 66% aoh22 ohl"ime Tax t over 650,mill. to otherprovinces, -e Fl to .w .f m Ead 9)-9
Groupe de la Banque mondiale · Pre-2003 Economic or Sector Report
Argentina - The fiscal dimension of the convertibility plan : a background report (Vol. 1 of 2) : Main report
Voir le document original
Le texte intégral est hébergé par l’organisation qui le publie. lawenc.com indexe les métadonnées et renvoie vers la source officielle.
Texte intégral
Informations clés
Organisation
Groupe de la Banque mondiale
Type de document
Pre-2003 Economic or Sector Report
Pays
Argentine
Source
Banque mondiale