Report N.A. 8371 -BR Brazil EconomiC Stabilization with Structural Reforms January 31, 1991 Country Operations Division Country Department I Latin America and the Caribbean Region FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. FORWARD The document presented here had been completed in early March 1990, with only a few minor changes introduced into the present version, to reflect discussions with the Brazilian Government. It thus represents the situation prior to the inauguration of the new Administration. President Collor declared on March 16, 1990. immediately upon assuming office, a comprehensive program ("Brasil Novo") of stabilization and structural reforms. In both its general tenor and in many of its components this program, as announced and implemented since, is similar to the set of policy measures proposed here. Naturally, following the basic policy change, some of the structural attributes of the economy discussed here have undergone a fundamental transformation. But the major issues analyzed in this report are still crucial for the economy's performance; and the set of recommendations for a medium-term policy course is still valid. This docuument has a restricted distribution and may be used by recipients oniv in the performance of their officirl duties. Its contents rmay not otherw!se be disclosed without World Bank authorization. GLOSSARY OF ACROt'YMS AGF Adquisides do Governo Federal (Federal Government Acquisitions (Min. Price Prog.) BEFIEX ComissAo para Concessao de Beneficios Fiscais a Programas Especiais de Exportaiao (Special Program of Fiscal Incentives for Exporters) BNDES Banco Nacional de Desenvolvimento Econ6mico e Social (National Economic and Social Development Bank) CACEX Carteira de Comercio Exterior (Foreign Trade Office) CDI Conselho de Desenvolvimento Industrial (Councii of Industrial Development) CONSIDER Conselho de NAo-Ferrosos e de SiderCrgia (Steel and Non-Ferrous Metals Council) CVRD Companhia Vale de Rio Doce (Valley of Rio Doce Company) FAE Fundo de Assistencia Escolar (Student Assistance Fund) FCVS Fundo de Compensa,co de Variac§o Salarial (Fund for Compensation of Salary Variation) FIESP Federacao Industrial do Estado de Sao Paulo (Industrial Federation of the State of Sao Paulo) FINSOCIAL Fundo de Investimento Social (Fund for Social Investment) FISET Fundo de Investimento Setorial a (Fund for Sectoral Investment) FUNDABEM Fundacao Nacional do Bem-Estar do Menor (National Foundation for Child Welfare) IBC Instituto Brasileiro do Cafe (Brazilian Coffee Institute) IBGE Instituto Brasileiro de Geografia e Estatistica (Brazilian Institute of Geography and Statistics) ICM Imposto sobre Circulacao de Mercadorias (Value-Added Tax) GLOSSARY OF ACRONYMS (Continued) INAMPS Instituto Nacional de Assitgncia Medica da PrevidAncia Social (National Institute for Social Insurance Medical Care) INAN Instituto Nacional de AlimentasAo e Nutri,co (National Institute for Food and Nutrition) INPS Instituto Nacional de Previdencia Social (National Institute for Social Insurance) IPEA Instituto de Pesquisas Econ6micas Aplicadas (Institute of Applied Economic Research) IPI Imposto sobre Produtos Induqtrializados (Industrial Products Tax) LBA Legi§o Brasileira de Assistencia (Brazilian Legion of Assistance) LCT Leis Consolidadas do Trabalho (Consolidated Labor Laws) LFT Letra Financeira do Tesouro (Treasury Bill) MEC Ministerio da Educa,Ao e Cultura (Ministry of Education and Culture) ORTN Obriga,6es Reajustaveis do Tesouro Nacional (Indexed National Treasury Bond) OTN Obriga,oes do lesouro Nacional (National Treasury 3ond) PASEP Programa de Assistencia ao Servi,o Pflblico (Asset Accumulation Program for Public Servants) PCA Programa de Complementa,co Alimentar (Program of Complimentary Feeding) PIS Programa de Integra,ao Social (Program of Social Integration) PNAE Programa Nacional de Alimenta9&o Escolar (National Program for School Feeding) PSA Programa de Suplementagao Alimentar (Food Supplementation Program) SEST Secretaria Especial para o Controle das Empresas Estatais (Special Secretariat for the Control of State Enterprises) GLOSSARY OF ACRONYMS (Continued) SENAC Servic,o Nacional de Aprendizagem Comercial (National Service for Commercial Apprenticeship) SENAI Servico Nacicnal de Aprendizagem Industrial (National Service for Industrial Apprenticeship) SFH Sistema Financeiro da Habita,ao (Housing Finance System) SINPAS Sistema Nacional de Previdencia e Assist4ncia Social (National System of Social Security and Assistance) 1] -i- BRAZIL ECONOMIC STABILIZATION WITH STRUCTURAL REFORMS Table of Contents Page PREFACE ..... vii ECONOMIC AND SECTOR WORK ON WHICH THE PEPORT DRAWS .. viii PART 1: SUMMARY AND POLICY AGENDA ................................. 1 I. SUMMAr' ..................................................... 2 A. Economic Stagnation ................................... 2 B. The Inflationary Process and MacroEconomic Policy ..... 4 C. Income Distribution and Poverty ....................... 8 D. The Budget: Expenditures, Taxes and the Social- Security System ..................................... 9 E. Public Enterprises .. 13 F. External Transactions: Trade an(; Debt .. 16 G. Industrial and Agricultural Policies .. 19 H. The Financial Sector .. 23 I. Social Expenditures .. 26 II. A POLICY AGENDA .. 30 1. Short-Term Structural Reforms and Stabilization .. 30 A. The Policy Framework .. 30 B. Fiscal Reforms .. 31 C. Other Reforms .. 34 D. Structural Reforms and the Fiscal Stance .. 35 E. Other Elements of Macroeconomic Stabilization .. 39 F. The Operation and Impact of the Stabilization Policy .. 45 G. Provision of a Social Safety Net .. 49 H. The Administration of the Stabilization Policy .. 51 2. An Agenda of Medium- and Long-Term Reforms .. 52 A. The Budget and the Government's Labor Force .. 52 B. Privatization of Public Enterprises .. 54 C. The Foreign-Trade Regime .. 56 D. The Agricultural Sector .. 57 E. The Financial Sector .. 58 F. Social-Sector Policies ............................ 59 G. Overall Pattern of Policy Changes .62 -ii- Table of Contents (Continued) Pag Table 2.1: The Federal Budgetary Change .................. 37 Table 2.2: Brazil: Key Economic Indicators .............. 47 PART 2: THE BASIC ISSUES .......................................... 63 III. ECONOMIC STAGNATION ....................................... 64 A. The Rate of Growth .................................... 64 B. Saving and Investment ................................. 65 C. Labor and Land ........................................ 69 D. Capital Flows ......................................... 71 E. The Impact of Inflation ............................... 73 F. Distortionary Gove..nment Policies .75 Table 3.1: Growth Performance, 1978-88 ................ .. 64 Table 3.2: Gross Fixed Capital Formation, 1970-88. 65 Table 3.3: Saving Rates, 1970-7 .68 Table 3.4: Labor-Force Participation Rates, 1960-85 69 Table 3.5: Major Balance-of-Payments Components .71 Table 3.6: Industrial Origin of Domestic Production, 1970-87 ..................................... 73 Figure 3.1: Investment by the Public and the Government Sectors, 1970-86 ............................ 67 IV. THE INFLATIONARY PROCESS .................................. 76 A. General: The Pattern of Inflation .................. . 76 B. The Budget and the Fiscal Deficit ..................... 78 C. The Domestic Debt ................................... . 82 D. Money and Monetary Policy ............................. 83 E. Exchange-Rate Policy .................................. 86 F. Indexation ............................................ 89 G. Exogenous Shocks ...................................... 91 H. Stabilization Plans ................................... 92 I. The Inflationary Process .............................. 93 Table 4.1: Inflation Rate, 1980-89, Quarterly ............ 77 Table 4.2: The Public-Sector Deficit, 1980-88 ............ 80 Table 4.3: The Public-Sector Domestic Debt, 1980-1988 .... 82 Table 4.4: Composition of Money, 1980-88 ................. 84 Table 4.5: Ratios of Money Components to GDP ............. 85 Figure 4.1a Annual Average Exchange Rate and Price Level ....................................... 87 Figure 4.1b Quarterly Inflation and 2 Change in Cz$/US$ Exchange Rate ............................... 88 Figure 4.2: Real Wage and CPI Inflation ................... 90 I Table of Contents (Continued) PaRe V. INCOME DISTRIBUTION AND POVERTY ........................... 98 A. The Degree of Income Inequality ....................... 98 B. Regional and Functional Distribution of Income ........1 01 C. Indicators of Poverty ............ .................... 104 D. The Impact of Government Polic.es ..................... 108 Table 5.1: Income Distribution in Developing Economies... 99 Table 5.2: Gini Coefficient of Income Distribution, Developing Economies ...................... 100 Table 5.3: Relative Per Capita Incomes, by Region, 1970, 1975. and 1980 ........................ 102 Table 5.4: Relative Size of Per-Capita GDP in the Northeast, 1980-86 .......................... 102 Table 5.5: Indicators of Real Wages, 1980-87 ............. 104 Table 5.6: Incidence of Poverty, 1980-87 ................. 105 Table 5.7: Infant Mortality Rates, by Region, 1980-86 .... 107 Table 5.8: Access to Adequate Sanitation Services, 1987.. 108 PART 3: ANALYSES OF MAJOR ACTIVITIES ....... ..................... 113 VI. THE BUDGET: MECHANISM, EXPENDITURES, AND TAXES ........... 114 1. Expenditures .......................................... 114 A. An Overall View ................................... 114 B. The Budgetary Process ............................. 116 C. Subsidies ......................................... 117 D. Other Issues of Budgetary Expenditures ............ 117 2. Taxes ................................................. 124 A. An Overall View ................................... 124 B. Indirect Taxes .................................... 125 C. Direct Taxes ...................................... 128 3. The Social Security System ............................ 132 Table 6.1: Budget Composition, by "Compressibility", 1980-88 ..................................... 114 Table 6.2: Structure of Central Government Current Expenditures, by Major Functions, 1978-86 ... 115 Table 6.3: Tax Revenues by Level of Government ........... 119 Table 6.4: Frequency Distribution of Employment by Total Remuneration, 1985 .................... 121 Table 6.5i Federal Employment by Administrative Group, 1985 ........................................ 122 Table 6.6: Tax Revenues, 1970-88 ......................... 124 Table 6.7: Revenues from Major Indirect Taxes, 1970-88 ... 125 Table 6.8: Revenues from Major Direct Taxes, 1975-88 ..... 129 Table 6.9: Expenditures and Revenues of INPS, 1972-86 .... 133 Table of Contents (Continued) Pag~e VII. PUBLIC ENTERPRISES ......................................,.136 A. Origin, Size and Structure .136 B. Performance: Basic Indicators ........................ 139 C. Public Enterprises and the Government Budget: Recent Performance . 142 D. Privatization of Public Enterprises . 3.47 Table 7.1: Public versus Private Ownership of Brazil's Thirty Largest Financial Firms, 1962-79 . 137 Table 7.2: Share of Public versus Private Enterprises in the Economy, 1986 .137 Table 7.3: Indices of Public-to-Private Ratios of Output and Investment, 1975-85 .139 Table 7.4: Capital/Labor Ratios, by Branch, in Public and Private Enterprises, 1980-86 .140 Table 7.5: Profit Rates in Large Brazilian Firms, Public versus Private Sectors, 1974 and 1975 .141 Table 7.6: Profit Rates of Large Enterprises by Branch Public versus Private, 1986 .142 Table 7.7: Operational Account of Public Enterprises, 1986-89 .143 .able 7.8: Deficits and Government Transfers, Major Public Enterprises. 1988 .144 Table 7.9: Public-Sector Real Tariffs, 1985-88 .145 Table 7.10: Public-Sector Real Tariffs During 1989 .146 Table 7.11; 'Required" Adjustments of Public-Sector Tariffs 146 VIII. EXTERNAL TRANSACTIONS: TRADE AND DEBT .................... 152 1. Foreign Trade and the Trade Regime ........................ 152 A. The Size and Structure of Trade ....................... 152 B. Trade and Exchange-Rate Policies ...................... 160 C. Recommendations: Reform of the Trade Regime .......... 163 2. Capital Flows and the External Debt . ............. 166 A. Sources of the Debt ................................... 166 B. The Burden of the Debt ................................ 168 C. Attributes of the Debt ................................ 171 D. Debt-Transformation Schemes ........................... 175 E. Expected Course of the Debt ........................... 180 F. Foreign Investment: Recent Developments .............. 180 G. Foreign Investment: Regulatory Framework and Policy Recommendations .............................. 183 Table of Contents (Continued) PaLe Table 8.1: The Current-Account Balance and the Stock of External Debt, 1973-89 ...................1.67 Table 8.2: Measures of the Burden of the Debt. 1980-88,. 169 Table 8.3: Debt Ratio, Highly-Indebted Countries, 1987 170 Table 8.4: Registered and Non-Registered Debt, 1980-88 .71 Table 8.5: Public and Private Debt, 1980-88 .............. 172 Table 8.6: External Debt by Category of Lender, 1983-87 173 Table 8.7: Currency Composition of the Registered Debt, September 1989 .............................. 175 Table 8.8: Debt-Equity Conversion and Direct Foreign Investmert, 1975-88 ......................... 177 Table 8.9: IBRD RMSM Balance of Payments Projections ..... 179 Table 8.10: Registered Foreign Investment in Brazil, 1969-88 ..................................... 181 Table 8.11: Direct Foreign Investment, Capital Repatriation and Profit Remittance (Excluding Conversions) .......... 182 Figure 8.1: Exports by Commodity Group .................... 153 Figure 8.2: Imports by End-Use ............................ 155 Figure 8.3: Trade as Percent of GDP, 1965-89 .............. 157 Figure 8.4: Manufactured Export Profitability ............. 159 Figure 8.5: The Real Exchange Rate, 1965-89 ............... 161 IX. INDUSTRIAL AND AGRICULTURAL POLICIES . ............. 186 1. Industrial Policies ................................... 186 A. Policy Framework .................................. 186 B. Fiscal Incentives .................... ........... 187 C. Financial Incentives .............................. 190 D. The Impact of Industrial Policies ................. 191 2. Agricultural Policies ................................. 194 A. Policies of Direct Intervention ................... 194 B. The Impact of Agricultural Policies ............... 197 Table 9.1: Rates of Fiscal Incentives Granted by CDI, By Sector, 1986 ............................. 188 Table 9.2: Rates of Fiscal Incentives Granted by BEFIEX, 1980-85 ..................................... 189 Table 9.3: Sectoral Allocation of Lending by the BNDES System, 1986-87 ............................. 191 Table 9.4: Share of BNDES in Credit to Industry, 1980-87. 192 X. THE FINANCIAL SECTOR: STRUCTURE AND POLICIES ............. 202 A. Structure of the Sector ............................... 202 B. Competition in Financial Activities ................... 204 C. Directed and Subsidized Credit ........................ 210 D. Major Problems of the Financial Sector ................ 211 E. Recommendations ....................................... 215 -vi- Table of Contents (Continued) Pag~e Table 10.1: Financial Sector Intermediaries ........203 Table 10.2: Market Share of the Largest Commercial. Banks. 206 Table 10.3: MAarket S'iare of the Ten Largest Financial Conglome rates ................207 XI. SOCIAL EXPENDITURES* STRUCTURE aNL POLICIES... ......217 1. General.........................217 2. Education........................220 A. Structure of the Sector...............220 B. Issues and Problems.................225 C. Recommendations...................228 3. Health and Nutrition...................229 A. Structure of the Sector .... ..........229 B. Issues and Problems.................233 C. Recommendations...................234 Table 11.1: Federal, State and Loca'l Social Expenditures By Program, 1986 ...............2 Table 11.2: Public Spen.ding on Education by Level of Instruction, 1985 ..............221 Table 11.3: Primary School Attendance by Region and Type of School, 19R5.............223 APPENDIX: BASIC ECONOMIC DATA FOR BRAZIL.............236-272 - vii - PREFACE This Country Economic Memorandum of Brazil has been prepared in late 1989 and early 1990, when it became clear that a new, incoming administration will have to undertake a major effort of overhaul'iig Brazil's most fundamental economic policies. The report therefore emplhasizes the impact of government policies, and the policy reforms which will be required to lead Brazil to economic stability, growth, and equity. Michael Michaely is principal auchor of the report. Other participants are Malcolm Bale, Barbara Bruns, Rui Coutinho, Antonio Estache, Louise Fox, Maria Emilia Freire, Valeriano Garcia, Robert Kaplan, Miguel Kiguel, Geoffrey Shepherd, and Silvina Vatnick. James Stephens has provided research assistance; and Robin Gaster has edited the manuscript and prepared the Summary (Chapter I). The memorandum is to a large extent a synthesis of the economic and sector work that has been carried out in the Brazil Department during the last two years. The next page lists the reports on which the memorandum has drawn. In the specification of contributions to individual chapters, these reports will be referred to by their numbers or this list. Chapter IV is based on contributions by Rui Coutinho, Valeriano Garcia, and Miguel Kiguel and Nissan Liviatan: and draws on (1] and (2]. Chapter VI is based on the contribution of Antonio Estache; and draws on [3], (4], and (5]. Chapter VII is based on the contribution of Maria Emilia Freire; and draws on (6]. Chapter VIII is based on the contributions of Malcolm Bale, Geoffrey Shepherd, and Silvina Vr.tnick; and draws on (7] and [8]. Chapter IX draws on (9] and [12]. Chapter X has been written by Valeriano Garcia; and draws on (14] and (15]. Chapter XI has been written by Barbara Bruns and Robert Kaplan; and draws on (16], [17], [18], (19], [20], and (21]. - vill - Economic and Sector Work on Which the Report Draws 1l Brazil--An Assessment of the Current Macroeconomic Situation (in two volumes), Gray Cover, December 1988 (Report No. 7540-BR). ;' Brazil--An Update on Recent Macroeconomic Developments, Yellow Cover, = June 1989 (Report No. 7814-BR). 3] Brazil--Public Expenditure, Subsidy Policies and Budgetary Reform (in two volumes), Gray Cover, December 1989 (Report No. 7738-BR). [ 4] Brazil--An Agenda for Tax Reform (in three volumes), Gray Cover, February 1990 (Report No. 8147-BR). 5] The Macroeconomics of Social Security in Brazil, LAC Discussion Papers, June 1989 (IDP-0043). 63 Brazil: Prospects for Privatization, Gray Cover, June 1989 (Report No. 7550-BR). [ 7] Trade Policy in Brazil: The Case for Reform, Gray Cover, March 1990 (Report No. 7765-BR). ( 8] Brazil--External Debt Development and Prospects, LAC Discussion Papers, December 1988 (IDP-0034). 93 Industrial Regulatory Policy and Investment Incentives in Brazil, Gray Cover, March 1990 (Report No. 7843-BR). [10] Brazil--Agriculture Storage and Marketing, Gray Cover, March 1989 (Report No. 6928-BR). (113 Brazil--Sugar Suosector Review, Green Cover, May 1989 (Report No. 7589-BR). [123 Brazil--Agriculture Sector Review: Policies and Prospects (in two volumes), Gray Cover, July 1990 (Report No. 7798-BR). [133 Brazil--Irrigation Study, Gray Cover, September 1990 (Report No. 7797-BR). (143 Brazil: Selected Issues of the Financial Sector, Gray Cover, March 1990 (Report No. 7725-BR). [153 Brazil: The Dilemma of Brazil's State Banking System: An Analysis and Suggestions for Reform (in two volumes), Green Cover, January 1991 (Report No 8247-BR). [163 Brazil--Public Spending on Social Programs; Issues and Options, (in two volumes), Gray Cover, May 1988 (Report No. 7086-BR). [173 Brazil--Adult Health in Brazil: Adjusting to New Challenges, Gray Cover, November 1989 (Report No. 7807-BR). - ix - [18] Brazil--Women's Reproductive Health, Green Cover, December 1989 (Report No. 8215-BR). [19] Brazil--Country Assessment of Women's Role in Development, (in two volumes), Yellow Cover, September 1989 (Report No. 8043-BR). [20] Brazil--Finance of Primary Education, Green Cover, May 1986 (Report No. 6120-BR). (21] Brazil--Issues in Secondary Education, Green Cover, November 1989 (Report No. 7723-BR). PART 1: SUMMARY AND POLICY AGENDA This part of the report presents the summary and conclusions. Part 2 will then discuss Brazil's current fundamental issues; whereas Part 3 will analyze Brazil's major economic activities. The first chapter of the present part summarizes Part 2, and the analyses and policy evaluations of Part 3. The policy recommendations derived from these evaluations are then piresented in the second chapter, by way of two agendas. First, a package of structural reforms is suggested which may, and should, be undertaken in the short run. With other specified policies, it forms a comprehensive agenda for macro-economic stabilization. Second, a framework of structural reforms for the longer term is formulated, encompassing most major areas of economic activity. Taken together, the two chapters of this part thus provide an executive summary of the report. - 2 - I. SUMMARY 1.1 This chapter summarizes the report without offering policy recommendations, which are presented in Chapter II. It is designed to present the analyses and concerns that underpin those recommendations. 1.2 In general, the summary follows the structure of the report, beginning with the critical problems of growth and inflation, followed by an analysis of income distribution and poverty. A discussion of the budget, the core of all attempts to resolve Brazil's economic problems, follows. The section on the budget separately examines the budget mechanism itself, and expenditures and revenues. The budget analysis is followed by a discussion of the public enterprise sector, and in particular of the possibilities for privatization. Government trade policies are then analyzed, followed by an examination of policies aimed specifically at industry and agriculture, and a discussion of the financial sector. The summary concludes with the investigation of social expenditures. A. Economic Stagnation 1.3. Many of the sectoral problems outlined later in this chapter stem from the abrupt slowing of growth in Brazil in the 1980s. During the 1960s, growth in Brazil was respectable, at an average of 5.7 percent. In the late 1960s and 1970s, growth boomed at 12 percent or more, and remained solid between 1973 and 1982 at 7.1 percent. Since the beginning of the 1980s, however, growth has averaged only 2.1 percent. Per capita income slowed even more, from about 4.5 percent annually over 1974-80 to approximately zero in the 1980s. There were several causes for this declining gro-wth, the most important of which are outlined below. Savings and Investment 1.4 During the 1970s, investment grew very fast, more than doubling during the decade, and it grew even as a share of GDP. But that share fell sharply in the early 1980s; and despite some resurgence in the mid-1980s, it has remained at a lower level. Since the earlier increase in the capital stock implies increased levels of depreciation, net investment must have fallen even further. Overall, this implies that the capital stock has remained stable in the 1980s, doubling during the 1970s. 1.5 Government investment had increased very substantially during the 1970s, from 8 percent to 12 percent of GDP; but 1986 it had fallen to 5 percent. Most of the change reflects changes in the investment patterns of the state enterprises, whose investment rate fell from 8 percent to 2 percent of GDP. 1.6 In similar fashion, the national savings rate dropped sharply, from 27.5 percent of income in the 1970s to 20.4 percent in 1980-83. Almost all the decline occurred in government saving, which fell from 3.3 - 3 - percent in the 1970s to about zero over 1980-83 and to -7.5 percent in 1985-1987--an overall decline of approximately 10 percent of GDP over a decade. Capital Flows 1.7 Up to 1974, capital flows played little part in Brazil's economic performance. With the oil shocks and then cheap foreign loans, capital began to flow in rapidly. Soon after, Brazil's interest payments began to grcw as well. 1.8 Following the debt crisis in 1982, and the collapse of voluntary foreign lending to Brazil, interest payments grew while no net inflow of capital was forthcoming. The result was a switch from net (of interest payments) inflows of $US3-4 billion, to outflows of about US$10 billion--a net switch of US$13-15 billion, which amounts to 5 percent of GDP. And with gross investment at 15-18 percent of GDP from 1983 onward, these outflows are equivalent of between a quarter and a third of investment. The ratio of outflows to national savings is only slightly lower, and the impact on capital accumulation is therefore obvious. This outflow also partly accounts for the decline in government savings. The Impact of Inflation 1.9 While stagnation and accelerating inflation were largely the result of the same basic forces, rapidly accelerating and variable inflation rates have also caused stagnation. The structure of Brazil's economy has certainly been distorted, with heavy growth in financial services during the 1980s and a halt to the expansion of manufacturing. Clearly, the growth of financial services reflected the need for protection against inflation, and is therefore in fact a real subtraction from GDP. Capital flight has also slowed domestic growth. In addition, the nature of investment is distorted, as excessive inventories are accumulated capital goods purchases are accelerated, and changes in relative prices and demand become blurred. Distortionary Government Policies 1.10 Despite the apparently modest size of its budget, government intervention has been comprehensive. It will be discussed in more detail below, but two points should be made here. First, the distortionary effects of government policy must have increased during the 1980s, especially in relation to the foreign-trade regime, Second, even with unchanged policies, the likelihocd of impeded growth increases over time. Newly-encouraged activities expand immediately, but once thr immediate gains from these activi:ies have been reaped, the benefits of the policy decline, while the cumulative cost grows. This is especially reflected in policies which after a time come to favor incumbents, create powerful barriers to entry, and thus severely reduce efficiency and distort resource allocation, while allowing the harvesting of economic rents by incumbents. The 1980s were the decade in which much of the costs of pervasive government intervention during earlier decades finally surfaced. -4 Labor and Land 1.11 Before the early 1980s, the participation rate of the labor force increased steadily; since then, it has been stable. This is entirely due to changing female participation rates, and while not completely unexpected, the effect is to slow GDP growth. Hard data are not available, but it is also estimated that additions to the stock of usable land have also slowed in recent years. B. The Inflationary Process and Macro-Economic Policy 1.12 Inflation has been rising, as a trend, for a full decade in Brazil, but 1986 marks a turning point: since then inflation has greatly intensified, to over 1700 percent iP 1989. The nature of the cycles has also changed, from a progression of semi-stable levels before 1986 to a series of sharp breaks, with short periods of decelerating inflation followed by longer periods of acceleration. The key elements in the recent pattern are the budget, public debt, monetary policy, exchange-rate policy, indexation, exogenous shocks, and packages of stabilization policies. The Budget Deficit 1.13 Three major measures of the budget deficit are relevant in Brazil: the 'nominal deficit" is the government's net borrowing requirement, the 'operational deficit" is the nominal deficit less the monetary correction, and the 'primary deficit" is the operational deficit less real interest payments. 1.14 The nominal deficit has increased continuously, as should be expected. The operational deficit is rather substantial, at around 5 percent of GDP, but it has been relatively stable. The primary balance is mostly negligible, or even in a slight surplus. Hence the entire operational defic t reflects reai interest payments by the government; and as the real interest rate has been rising during 1989, the operational deficit for the year is expected to be 7-10 percent of GDP. 1.15 The central government, local governments, and the state enterprises all contribute significantly to the operational deficit. Part of the operational deficit--an average of 2 percent of G?P in 1980-85, increasing to 3 percent over 1986-88--has been financed by the inflation tax. The rems'ning operational deficit adds to the public debt. Government Borrowing 1.16 Net government borrowing from abroad has been rather small in recent years--only US$10 billion over 1985-88, including the assumption of considerable private debt. Hence most of the deficit has been financed by domestic borrowing, and domestic debt has indeed grown relative to GDP, as has the share of credit to the government in overall credit in the economy. 1.17 The real interest rate largely determines the rate at which public debt accumulates; this constrains the use of monetary policy in Brazil, as the increased real interest rates required for a contractionary monetary policy would increase the government's real interest bill and its deficit. - 5 - This link is heavily rei:.forced by the composition of the government's debt, as maturities have fallen sharply and continuously, and now average 1-3 months. With a daily repurchase agreement, most of the debt is in any event held at variable rates tied to overnight fund rates. 1.18 The size znd composition of the debt critically affect monetary policy. The high liquidity oi mos, debt makes it a close substitute for money proper, and these instruments are thus included in Brazil's M4. Within M4, M1 (unindexed and non-interest-bearing money) has declined continuously and substantially, as would be expected in periods of high inflation. M1 has also declined in relation to GDP. As indexed (in effect) money, government bonds have become increasingly popular, amounting to more than half of M4 by 1988. That is why the nominal budgetary deficit has monetary significance, because it is the government's primary way of injecting money into the system without using the Central Bank. Clearly, U money which is mostly indexed would accommodate any rate of inflation. 1.19 Monetary policy thus becomes ineffective. Increased interest rates increase the operational deficit, thus the public's perceived wealth and, most important, the quantity of money (M4). Hence the classical measure of tightening the money market via increased interest rates generates a perverse outcome: it tends to increase money and liquidity. 1.20 Open-market operations are similarly constrained. Open-market sales by the Central Bank would indeed lower the monetary base and reduce the share of M1 in the money stock; but it would also place additional securities in the hands of the public, which would increase the stock of money. 1.21 Nor can the Central Bank easily control the money supply by changing the banks' reserve ratios. In Brazil, there are more than a hundred such ratios, amongst which asset holders can easily shift. The result is quite large fluctuations in the money multiplier. Thus for M4, variations in the monetary base (not surprisingly) explain next to nothing. The Exchange Rate 1.22 In principle, the government has aimed to establish, mostly by using the crawling peg, a stable and constant real exchange rate, except for a few periods of fixed nominal rates under stabilization policies. Hence the foreign exchange rate was, de facto, indexed to domestic inflation; so, like the quantity of money, the exchange rate has not served Brazil as a nominal anchor. Indexation 1.23 Brazil's economy is highly issdexed. This has numerous representations. First, a pervasive informal indexation has developed after years of inflation, in which individual prices are adjusted in accord with expected general price increases. More formally, the price of labor is indexed to the general price level, with lags which have declined in recent years under successive stabilization plans. In fact, actual movements of nominal wages and prices do not correspond closely; fluctuations in real wages are substantial, and there is no clear - 6 - relation?fhip between changes in wages and inflation. So indexation is not a primary determinant of wages, except perhaps for government sector wages. 1.24 The indexation of financial assets, on the other hand, is crucially important. It is certainly pervasive, and plays a major role in perpetuating inflation. Discrete Events: Shocks and Stabilization Plans 1.25 Over the last decade, Brazil has encountered two exogenous shocks. The second oil shock of 1980 tripled oil prices at a time when Brazil imported almost all its oil, and that certainly had powerful price effects. The debt crisis of 1982 had important, through less direct effects. Notably, however, there have been no similar shocks since then, and these have been the years of rapidly accelerating inflation. If anything, exogenous developments have been favorable. 1.26 There have been three recent stabilization plans. The Cruzado Plan (February 1986) froze prices and the nominal exchange rate (the latter without a devaluation). Wages were raised slightly and then became subject to a trigger indexation mechanism, the nominal overnight rate was fixed at 1.5 percent per month, and fiscal policy also became somewhat more expansionary. The Bresser Plan (June 1987) also imposed a wage and price freeze (the latter was to be gradually relaxed after three months), and the wage indexation rule was changed. After an initial devaluation, the exchange rate was to be subject to a series of mini-devaluations. The main short-term financial instrument was indexed and the real interest rate was raised. But the government deficit was not, in the event, reduced. The Z'iumer Plan (January 1989) introduced an indefinite price freeze, and an increase in real wages of 26.5 percent along with ongoing indexation. While the main short-term indexed instrument was eliminated, real interest rates were kept positive on short-term treasury bills. Public sector prices were initially realigned and then frozen--which increased the government deficit. And no real expenditure cuts iccurred. 1.27 As has become painfully clear, none of the stabilization plans addressed the root causes of Brazil's stagnation and inflation. The Process of Inflation 1.28 The process of inflation in Brazil is complex, and cannot be entirely explained by a single model. The data from 1980-85 show inflation doubling in the early 1980s, stabilizing at around 100 percent, and then doubling again in 1983 to 200 percent. This pattern is largely explained by the exogenous shocks of the second oil price increase and the debt crisis, exacerbated by a loose monetary policy. The experience of this period has colored later views, establishing the perception of "inertial' inflation, which can be stopped simply--and solely--by breaking the inertia. 1.29 Since the end of 1985 the pattern has been different. Inflation has accelerated persistently, and deep cyclical movements occurred that were obviously related to the stabilization plans. While the persistent acceleration is not linked to similar increases in the size of the - 7 - budgetary deficit, the cumulative deficit has indeed increased, and this has contributed to the rise in inflation. Also, as debt grows the public becomes less willing to hold government short-term assets at a given interest rate. Hence real and nominal interest rates must be increased, with the effects noted above. Increases in nominal interest rates have in addition come to reinforce inflationary expectations in Brazil, acting as a signal that the government itself expects inflation to increase. 1.30 Expectations have now become dominant. This in turn causes preemptive price increases. Similarly, wage rates now respond faster to past price increases. The lag is now a month or less. 1.31 All the inflation cycles are related to the stabilization plans. In essence, the cycle begins with a period of accelerated inflation, either because this acceleration leads directly to the introduction of the stabilizat..on policy, or because firms anticipated expected price freezes with preemptive price increases. Both likely play a part. 1.32 The drastic reduction in inflation that accompanies announcement of a stabilization plan may result from earlier beat-the-freeze price increases, which allows prices to be held stable for a while; or even sub- optimum prices may be maintained if the price freeze is seen as temporary; or prices could in fact be increased despite the formal freeze in a variety of non-measurable but well-known ways. More genuine factors could include automatic improvement in the fiscal stance caused by a reversal of the Tanzi effect under rapidly decelerating inflation; and, probably more important, a change in expectations. This effect has however become progressively weaker under successive plans. 1.33 The next period of rapidly increasing inflation is partly explained by the period of stabilization: the effect of beat-the-freeze price increases eventually runs out and further increases become necessary, and for the same reason price controls are increasingly flouted. 1.34 More fundamentally, essential policy elements need to be changed, and those changes must be seen as permanent. Under the three plans, the budgetary deficit did not decline, the public debt continued to increase as before, and the money supply remained accommodating. The absence of credible pclicies must have hurt the Cruzado plan, but it must have made succeeding plans unbelievable almost from the start. 1.35 Changes in the black-market exchange rate also have played an important role as signals in the resumption of rapid inflation. After the currency appreciated with the progress of the plans, the balance of payments began to deteriorate. As real devaluation loomed, demand for foreign assets increased, and with it the black-market premium. This signalled the imminent collapse of the plan. 1.36 When rapid inflation resumed, it did so in all three cases at higher levels than before the plan. Aside from the general trend for the decade, the higher level is tied to expectations. Failure reinforces expectations of persistent high inflation, and the pre-plan, exceptionally high level may be seen as a level that the government will tolerate in the future. These expectations are self-fulfilling. - 8 - C. Income Distribution and Poverty Incidence of Poverty 1.37 Poverty is an enormous problem in Brazil. It holds back economic development in many ways, and it has yet to be addressed by effective government policies. In fact, government policies have often only made matters worse. 1.38 Brazil's income distribution is highly tnequal. During the 1970s, it was among the most unequal in the developing world, although by 1982 the ratio of incomes of the top and bottom quintiles in Brazil had moved closer to the norm. Other measures tell the same story. 1.39 Regional effects dominate income distribution patterns in Brazil: per capita incomes in the Southeast are 3.5 to 4 times those in the Northeast. This disparity declined somewhat during the 1"70s, and relative incomes in the Northeast grew rather rapidly up to 1986. 1.40 Poverty-level is defined as a per-capita incomes of 25 percent of the real minimum wage in 1980. Households falling below the poverty line are heavily concentrated in the Northeast, where about half the population is in poverty by this measure. In the country as a whole, between one third and one half of the rural population are poor, compared with 10-20 percent of urban dwellers. These patterns have not changed substantially during the 1980s. 1.41 Infant mortality rates are also high in the Northeast and lowest in the Southeast; but they did decline substantially in all regions during the 1980s. Rural habitation is sometimes even more important than income for determining some aspects of poverty: thus, for instance, a majority of even upper income households in rural areas do not have adequate sanitation. 1.42 Neither pove,ty nor unequal income distribution were affected by a shift in the income of shares of labor and capital, as these shares remained relatively stable throughout the period. Real wages of workers in the formal sector stagnated during the last decade. However, the share of informal-sector workers was increasing rapidly, and per capita incomes in this sector are very low. This effect was partly offset by the shift of workers from agriculture--where incomes were even lower--into the informal sector. Government Policy Failures 1.43 Income distribution is mostly the result of factors other than government policies, but the latter have played an important, and undistinguished, role in the perpetuation of poverty in Brazil. In most countries the tax system is a primary tool for redistributing income to reduce poverty. In Brazil, the personal income tax and the corporate income tax provided little revenue, and are themselves not especially progressive. Most revenues come from one of the least progressive forms of taxation, payroll taxes. 1.44 Fiscal subsidies have proved largely ineffective for addressing poverty, even though many are in principle specifically designed to aid the poor. In practice, they have largely been collected directly or indirectly by the rich. At the same time, the economic rents that these policies create are paid, not received, by the poor. Credit subsidies have been even less effective. Received predominantly by large-scale producers, they have encouraged the concentration of land-holdings. In industry, both sets of subsidies lead to monopolistic organizations and monopoly rents, via the same mechanism: the creation of powerful barriers to entry. 1.45 Trade policies have been equally biased against the poor. They have supported incumbents, and promoted capital-intensive industries. Ir Brazil, these effects are felt in discrimnination against the agricultural sector in particular. 1.46 Social expenditures are the most direct means of addressing poverty. They are normally a powerful mechanism for income redistribution, but do not play this role in Brazil. In education, resources are poured into university education. Each undergraduate student receives a direct transfer of 5,500 dollars annually. This subsidy almost entirely benefits the middle- and upper-income classes. Education spending is also biased against rural children, who must largely depend on the poorer municipal schools while state funding is directed to urban schools. 1.47 Similarly, public health spending strongly favors formal sector employees, despite recent reforms. Public health spending is also heavily focused on the richer regions. 1.48 Finally, the poor have suffered from the general failure of government economic policy. The burden of inflation does not affect all individuals equally: those holding the highest percentage of assets in cash or unindexed assets are hit hardest--and those people are, inevitably, in the low-income category. D. The Budget: Expenditures, Taxes and the Social-Security System 1.49 The budget is the fundamental representation of the government's relation to the rest of the economy. It shows priorities and provides some measure of how well objectives are achieved. In Brazil, absence of budget transparency is an obstacle, as recent reform attempts have acknowledged. Even after the reforms, the budgetary process is complex, and large amounts of government spending somehow find themselves placed off-budget. 1.50 The substance of the budget is equally problematic. The tax system is failing in several ways, while expenditures are high, inflexible, and distributed among sectors in distortive and non-equitable ways. The social security system represents some specific areas of difficulty, and illustrates some of these wider tendencies. - 10 - Expenditures 1.51 Brazil's budget represents a share of GDP that is about average for comparable countries. It is however highly--and increasingly- -inflexible; and there is a large gap between the figures in the budget and the reality of government expenditures. 1.52 By 1988, the flexible element in the budget had declined to 4 percent of GDP, or 18 percent of tax revenues. And as government investment has already been squeezed hard, there is little room for maneuver. Pressure has also been intensified by the one major recent change in government expenditures: the dramatic growth in real interest payments, which has been matched by a corresponding increase in the overall size of the budget. 1.53 So far the budget process is concerned, many expenditures are not at all transparent, despite recent improvements. Several major schemes were estimated at 20 percent of GDP in 1988--yet only 0.3 percent could be found in the budget. 1.54 Three categories of budget expenditures are especially significant. The state enterprise sector is very largc and complex. Financial regulation of the sector has improved since SEST was established in 1979, and planned and actual budgets match quite closely. But there are still some significant gaps due to various effects of inflation; and the sector's deficits have usually been 20-25 percent larger than planned, and have averaged around 2 percent of GDP. 1.55 Intergoveramental transfers were 27 percent of Federal revenues, or 3 percent of GDP in 1987, and the new Constitution has further increased transfers to state and local governments. These transfers have allowed the two lower levels of government to increase their real spending substantially. 1.56 Salaries are of course a very large part of public sector spending, and they are higher than salaries in the private sector, This is especially so in the state enterprises, but that partly reflects higher skills. The real problems with the compensation system are its complexity and remarkable heterogeneity. Equal skills are rewarded very unequally, sometimes within the same branch of government. This is partly the result of decentralization and the continuing development of "supplements" (which circumvent general pay guidelines during inflationary periods). In consequence, the system provides the wrong incentives, and costs too much for the size of workforce concerned. 1.57 The size of that workforce--public sector employment--is on paper tightly regulated. But despite a permanent formal freeze, central government employment increased by almost one third during the 1980s. The hiring of "temporary" workers is loosely controlled, a loophole that has been heavily abused; and workers "seconded" from public enterprises to the central government have often become permanent fixtures. - -~~~~~~~~~~~~~~~~~~~~~~ - 11 - 1713 Ttie structure of Brazil's tax system has, as noted earlier, exat i rl,it ed the problem of poverty. But the system is also failing to gvr,trat. .ffit ient revenue to finance needed expenditures, and its actual operationns provoke increased inefficiency in many sectors of the economy. Thus the Btazilian tax system falls to perform well on any of the usual criteria: revenue-raising, efficiency, and equity. 1.59 In fact, the revenue position is deteriorating. W-;.le the declin,' in indirect tax revenues was offset by increasing direct ta -venues between 1970 and 1982, both have declined since then--by ab;.L 5 percent t f GDP overall, or 25 percent of the 1982 levels. This decline is about equal to the operational budget deficit in 1989, or the decline in public investment levels during the 1980s. It is difficult to exaggerate the importance of this problem. 1.60 Close to two thirds of indirect tax revenues in Brazil come from the two VAT-type taxes, and these have become even more important under the new Constitution, as several excise and specific taxes have been folded into one of them (the ICM). 1.61 One of these taxes, (the IPI) is levied by the Federal government on industrial production; the other (the ICM) is levied by the states mostly on the circulation of industrial and agricultural goods. While both taxes include the other in their bases, which makes the posted rates inaccurate, the most serious distortions stem from the numerous exemptions.1 1.62 IPI rates vary from zero to 300 percent. They also fall very heavily on only four industries, as about 70 percent of revenues come from the tobacco, beverages, weapons, and perfume industries. The ICM has been the most important indirect tax in Brazil, and in principle cnly services are now not covered by it. Posted rates under the new Constitution range from 12 to 25 percent, with effective rates being somewhat higher. 1.63 The VAT tax bases suffer from substantial exclusionq, as the IPI base covers only 30 perceat of GDP and the ICM base 70 percent. And within the base, a wide array o' exemptions are granted, mostly through fiscal incentives. In addition the taxes act more as cascading sales taxes than as uniform levies on the value added at a given stage of production. They are enormously complex, far more so than in most comparable countries; and until recently, they suffered heavily from erosion by inflation. 1.64 The upshot is that these taxes impinge on economic activity in varied and complex ways. Their effects are unpredictable, and outcomes are often not those anticipated by lawmakers. Their complexity also makes these taxes relatively easy to avoid. 1/ There are two kinds of favorable tax treatment. An activity which is exempted from paying a tax--and almost all industries treated favorably in the tax code are so exempted--pays no tax on its own production, but does pay the tax on inputs. Activities which are zero-rated pay the tax on neither their own product nor inputs. Zero-rating applies mainly to exports and to production in the Zona Franca de Manaus (ZFM). - 12 - 1.65 Direct tax revenues have also declined substantially in recent yelrs, though not by as much as indirect tax revenues. Pure personal income taxes, paid mostly as withholding taxes, are not very significa:nt, producing revenues of less than 4 percent of GDP. 1.66 Reforms in 1989 simplified the personal income tax considerably, widened the base, and reduced deductions while making them more uniform- -although agriculture remains privileged. However, the reforms will reduce revenues by 18 percent, or 0.1 percent of GDP. 1.67 The corporate income tax generates revenues that are only about one fifth those in comparable developing countries. It too was reformed in 1988, when the standard rate was being reduced from 35 percent to 30 percent (plus an extra 5 percent tax on high incomes), while a "social contribution" of 8 percent was added. However, numerous fiscal incentives have radically reduced the incidence of the tax. Agriculture is especially favored, and agricultural incentives have permitted large-scale tax avoidance in other sectors. 1.68 Payroll taxes are the main form of direct taxation, and they have been the largest source of revenue in the Brazilian tax system. In principle, they are imposed on all formal sector workers, but exemptions and avoidance mean that only about 60 percent of formal sector workers are in practice subject to the tax. Payroll taxes are a collection of several taxes, with both flat rate and progressive components. Aggregate rates range from 38 to 45 percent, with marginal tax rates declining above 10 minimum salaries. 1.69 Payroll taxes are largely earmarked for specific expenditures (mostly for social security), which lowers the scope for changes in the structure of public expenditures. 1.70 As this brief description shows, the direct tax system has some key weaknesses. Income taxes play a relatively small role in the tax system, aside from some specific problems. The system is inefficient largely because fiscal incentixes are so large and numerous. Payroll tax -.-es are relatively high but their incidence is constrained, making these taxes a source of distortion. 1.71 The system is also inequitable. Capital incomes are treated more favorable than labor incomes. Large firms are better able to exploit the complex loopholes in the corporate income tax. And the high rates and narrow coverage of the payroll taxes imply highly unequal treatment of incomes from different sources. The Social Security System 1.72 The social security system displays a potentially wider growing imbalance between revenues and expenditures in the economy. Within the system, the imbalance in funding can be resolved only by increasing the payroll tax or reducing expenditures. As payroll taxes are already very high, the latter solution appears inevitable. Yet there are structural reasns to expect expenditures to increase. - 13 - 1.73 The system's current structure was set in 1977. Under th overall SINPAS framework, there are three operating institutes: INPS, which provides insurance for retirement, disability, and survivorship; NAMPS (direct medical care); and LBA (social assistance for the non-ins red poor). FUNDABEM (social programs for poor minors) was under SINP S for many years, but was recently moved to the Ministry of the Interio . In 1984, INPS expenditure accounted for 70.5 percent of total SINPAS spending, and INAMPS for another 25.3 percent. 1.74 INPS expenditures are financed by the payroll taxes. Bo h insurance coverage and the income of contributors expanded rapidl before 1976, leading to surpluses. Since then, problems have begun to nerge. Growing expenditures are now inevitable as the population ages; ad expenditures will grow faster than revenues as the ratio of contributors to beneficiaries declines sharply, from 4-5 during 1955-77 to 2.7 by 1981 (the latter being quite low by international standards). 1.75 Retirement options are generous, although poor indexaticr does make them less attractive. Currently, urban workers are eligibl for pensions at age 65 for men and 60 for women, at levels equal to 0-95 percent of their wages. Early retirement is possible, and benefi s at 70 percent start after only 5 years of contributions. Rural retire s receive 75 percent of the minimum wage, with no early retirement option. 1.76 In principle, benefits are indexed to the cost of living. In practice, inflation has eroded benefits to less than half their egislated values. This may change under the new Constitution, which stipu ates full indexation. E. Public Enterprises 1.77 The weaknesses of the existing structure of public enteiprises are now widely accepted in Brazil. These enterprises vary significartly from each other; and some of them are well run and efficient. The se tor as a whole, though, is inefficient by several criteria, notably profitability and capital intensity. It has also generated significant deficit.s in recent years, partly because of government pricing policies. 1.78 The core of Brazil's public er.terprise sector was large ly in place by the mid-1950s, but Lhe sector grew still further as private s ctor firms were bailed out and taken over, and other enterprises were estab ished directly by BNDES. By the late 1970s, virtually all of Brazil's largest non-financial firms were government-owned. Government enterprise; represent almost half of all capital in the economy, almost a quarter of n t operational revenues, and employ just short of a fifth of all wo kers. Public sector enterprises took more than 30 percent of all inves ment by the mid-180s. - 14 - I/9 While the sector contains numerous firms and is highly diveLsified, it is dominated by 9 large companies, which account for about half of total activity int the sector. 1.80 Over the years, many of these companies have ge..erated enough profits to give them substantial independence from the government, a development that was further supported by diversification and the creation of numerous subsidiaries. Independence has in general persisted despite nominal control by SEST. 1.81 Although enterprises are owned by all three levels of government, the Federal part of the sector is by far the most important. 141 of the 175 Federal enterprises are defined as "productive state firms,"3 and are in principle supposed to cover their own operational expenses, turnirng to the Treasury only for investment. 1.82 Several measures indicate that the public sector is inefficient. The ratio of public to private-sector investment was seven times greater in 1985 than in 1975, yet the ratio of outputs less than doubled. This implies a large increase in the capital/output ratio in the public sector, but labor productivity grew much less rapidly in the public sector than in the apparently more labor-intensive private sector. Public sector entPrprises are generally more capital intensive than private sector firms even within the same branch of industry. Clearly, the availability of cheap capital for public sector industries encouraged capital investment, which is necessarily inefficient in the allocative sense. 1.83 Public sector profits were both lower and declined faster during 1975-85 than those of private sector firms. This difference is not explained by other variables such as size, and once again is a strong evidence of inefficiency. 1.84 Since the early 1980s, the sector's persistent current surplus- -which used to cover about half of its investment--has declined substantially (while borrowing in the market has also been cut). The resulting operational deficit, averaging around 3 percent of GDP in 1986-88, accounts for 70 percent or more of the government's overall operational deficit in 1986-87 (declining to just over half in 1988, and to perhaps one third in 1989). Significantly, the major enterprises in the sector (except for CVRD and PETROBRAS) produce most of the deficit. 1.85 Recently, the deficits have been heavily affected by government pricing policies under stabilization strategies. Although each plan has had somewhat different effects, and these effects have also varied by sector, 2/ Three of the nine companies are in electricity (ELETROBRAS, ITAIPU, and INB). The others are PETROBRAS (petroleum), SIDEBRAS (steel), CVDR (iron mining), TELEBRAS (telecommunications), PORTOBRAS (ports), and RFFSA (railways). 3/ Of the others, 18 handle government services and 16 are Federally-owned banks. A -~~~~~~~~~~~~~~~~~~~~~~~~ - 15 - public sector tariffs often did not increase enough to match increases in costs or in the general price level. The decline in real prices was in many cases very substantial--reaching on average 30 percent over 1989. Privatization L 1.86 The weakness of the public enterprise sector has led to the widespread acceptance of privatization in Brazil. A legal framework for privatization was introduced in 1985, and modified in 19S8. 1.87 There has been little actual privatization since then. Eleven firms have been privatized, but their assets total only US$160 million out of the sector's total book value of US$60 billion. More firms are in the pipeline, however, and--including share tenders without the transfer of ownership--they amount to more than 6 percent of the sector's assets (three-quarters of that coming from USIKAS, a subsidiary of the Federal steel company). 1.88 A number of industries--about 25 percent of the sector by value- -are protected from privatization by the new Constitution.4 However, subsidiaries of these companies may still be privatized, so that the Constitutional protections apply to perhaps 20 percent of the sector. 1.89 The market value of the public enterprise sector is estimated at around US$25 billion, so that assets worth about US$20 billion could potentially be sold off. That is not an overwhelming amount for the Brazilian capital markets to absorb. There are several important sources of potential demand for privatized assets, domestically and outside Brazil. The stock market is certainly large enough, with traded companies' equities valued at US$30 billion in 1987, and new equity issues at US$1.2 billion in 1986. Institutional investozs alone managed more than US$22 billion in assets in 1987. These figures suggest that the domestic stock market could absorb about US$1 billion worth of shares annually. 1.90 Foreign possibilities are even more significant. Debt/equity swaps could be substantial as the two privatizations linked to such swaps so far have indicated. Foreign commercial banks currently hold US$70 billion of Brazilian debt; and participation in privatized firms would not, in principle, be limited to current debt-holders. Debt/equity swaps amounted to US$8 billion in 1988. 1.91 Workers in pri,atized firms may also provide demand for shares, and experience in other countries (Chile, the U.K.) shows that worker responses to opportunities such as discounted shares are quantitatively significant. 1.92 Finally, domestic government debt can be swapped for equity. As short-term government debt is five times the aggregate potential market value of privatized enterprises, such swaps could provide an important avenue for converting this debt into long-term instruments--which would, moreover, not be government instruments. 4/ These industries include the exploration for, extraction, refining, and transportation of oil, as well as mining. - 16 - F. External Transactions: Trade and Debt 1.93 Brazil's external economic relations have, since 1982, been dominated by the need to deal with its large external debt. This has led to two major sets of developments. First, the character of Brazil's trade flows has changed dramatically: a deficit of the balance of trade and non- factor services has become a surplus (needed to finance interest payments), and the nature of Brazil's foreign trade regime has changed as well. Simultaneously, Brazil has pursued policies to reduce the burden of the debt. These two developments are discussed in the following sections. Foreign Trade and the Trade Balance 1.94 Brazil's trade policies must have affected the allocation of resources within Brazil, especially since the turning of the foreign exchange controls and licensing requirements into a highly-restrictive system with the onset of the debt crisis. The misdirection of investments has increased, and producers have been given what amounted to monopoly rights which permitted both substantial economic rents and inefficiencies. The working of the price system has tended increasingly to be replaced by administrative controls. 1.95 Brazil's pattern of foreign trade has changed rapidly over the past 20 years. Notably, exports are now predominantly manufactured goods, completely eradicating the historically strong dependence on coffee. Within manufactures, traditional labor-intensive exports (like textiles) grew faster than capital-intensive exports until the late 1970s, when this trend was reversed. Within agriculture, soybeans have become an important export crop. 1.96 The debt crisis of 1982 also led to radical changes in government policy. Since then, import-regulation has been aimed at the twin goals of servicing the debt via the creation of export surpluses, and reducing inflation at home. 1.97 During the 1980s a persistent current account deficit turned into a substantial and increasing surplus--largely in response to changes in capital flows mandated by the debt crisis. After years of expansion, the ratio of imports to GDP began to decline, while that of exports continued to increase, approximately doubling between the late 1970s and late 1980s. 1.98 Exports responded both to changes in domestic-market activity--like the slack domestic demand in 1988 and early 1989--and to price changes, that is, to movements of the real exchange rate. The strong export growth of most of the 1980s is closely related to the increase in real exchange rates (until 1987), which even with the reduction in export subsidies has increased the profitability of production for exports. 1.99 Despite these changes, Brazil's economy remains one of the most closed in the world, especially for imports. Even at their early 1970s level (6-7 percent of GDP), imports were extremely low; and the ratios are now lower still. - 1 7 - 1.100 Foreign-exchange controls have been a permanent fixture in Brazil since 1947, and they have been a major instrument of government policy. By the early 1960s, import-control policies were being supplemented by a variety of export promotion schemes, thus lowering the bias against exports and in favor of import substitutes. A substantial discrimination exists within each of these two categories--particularly among import substituting activities. 1.101 The bureaucracy imposed by the regime that controls foreign exchange is heavy, especially that surrounding the key document, the import license. In licensing, CACEX has widely extended the "Law of Similars' so that all of Brazil's industrial production is protected. There is also a de facto import-retention scheme which links import licensing to export performance. 1.102 Exports are less heavily controlled than imports. The major fiscal incentives go to firms with BEFIEX agreements for the export of specific goods (especially autos, textiles, and capital goods). And while _ export subsidies remain opaque and complex, they have been much reduced h since 1984. Capital Flows and the External Debt 1.103 While Brazil's foreign interest payments are substantial, they are now not overwhelming by international standards. Among highly-indebted countries, Brazil's debt/GDP ratio is among the lowest; and even though Brazil's debt/export ratio is high, the intensity of the economy's response to changing circumstances suggests that debt-servicing capacity is constrained by domestic saving-creating capacity rather than by the potential availability of foreign exchange. 1.104 Brazil's debt increased rapidly following the first oil shock in 1973, and the ample availability of cheap foreign credit in the late 1970s and early 1980s. The virtual disappearance of voluntary foreign lending since 1982, and the adjustment policies followed in response, have sharply reduced the current account deficit, to about zero since 1984. The nominal size of the debt has not increased since 1986. 1.105 After the current account stabilized, changes amo.g exchange rates of third currencies became increasingly important in affecting the size of the foreign debt. In fact, the increase in the dollar value of the debt since 1985 is entirely explained by the devaluation of the dollar relative to other currencies in which Brazil's debt is denominated. Capital flight has contributed to the size of the debt; rough estimates place this at about US$15 billion cumulatively over 1980-87, or 30 percent of the addition to the debt. 1.106 No single measure accurately captures the burden imposed by the debt. As a ratio of GDP, the stock of debt increased usstil 1984, with some decline since then; in term of exports, debt has declined steadily since 1982. Flow criteria show a burden that increased substantially in the early 1980s, but has declined since 1985. This largely reflects the decline in interest rates on Brazil's debt, in turn mainly the result of declining world interest rates. - 18 - 1.107 The composition ot the debt varies in several dimensions. To begin with, it was registered debt--largely to foreign official sources- -that graw until 1984; non-registered debt (usually commercial debt) remained roughly flat. Registered debt now accounts for about 90 percent of all debt. Second, the public sector's share of the debt increased during the 1980s, to almost 90 percent. This was largely caused by the public assumption of private debt. Third, about two thirds of the debt is owed to foreign commercial banks, but the share owed to foreign official sources has been increasing; within that share, Paris Club debt has increased sharply to about 7 percent of total external debt by 1987. Fourth, between 7) and 80 percent of the debt contracted during the past decade carries variable interest rates, so that in general the interest bill on Brazil's debt moves in line with world interest rates. Finally, about 30 percent of the debt is denominated in non-dollar currencies, primarily Japanese yen and German marks. And as that currency composition differs from that of Brazil's trade flows, appreciation of the German mark or Japanese yen tends to increase the real burden of its debt. 1.108 Among the many mechanisms available for manipulating the debt and reducing the debt burden, only a few schemes can he introduced unilaterally by the Brazilian government. Debt/equity swaps have been by far the most important of these. 1.109 Debt-equity swaps can be done formally via the Central Bank, and have been implemented since the 1960s. Fiscal incentives have applied especially since the end of 1982. This program was suspended in February 1988, but should it be reopened, about US$60 billion worth of debt would be eligible for it (all debt which has matured, and all private sector debt before it reaches maturity). These formal swaps have been important relative to "normal' direct foreign investments since 1983, and predominated in 1988 (at 3.6 billion dollars). Debt conversion thus partly substitutes for other forms of foreign investment. 1.110 Informal swaps started in 1985 and have increased sharply since fall 1987. Through various procedures, the result is an investment by a foreign multinational company in its Brazilian subsidiary. Estimates place informal cor.versions in 1988 at US$4.6 billion. Formal and informal conversions thus totalled US$8 billion, which if maintained in the future could substantially and quickly affect the size of the foreign debt. 1.111 The World Bank has used an RMSM model to project the course of the external debt over 1990-94 leaving aside debt conversion. The current account deficit is estimated at approximately US$3 billion, direct foreign investment at US$1-1.5 billion annually, and foreign lending by Brazilian banks at about the same level. Thus, over the 5 year period, net medium- and long-term debt is expected to increase by US$20 billion. The debt/GDP ratio is expected to decline somewhat, and interest payments will fall even further assuming a decline in interest rates)--from 3 percent of GDP to less than 2 percent by 1994. 1.112 G.ven these estimates, debt-transformation schemes could turn an increase in the stock of debt into a significant reduction, both relative to GDP and absolutely. While that would not change the transfer of - 19 - resources to foreigners via interest or dividends, it would have important positive effects on the budgetary burden that debt-servicing payments impose and, presumably, on demand for foreign exchange. G. Industrial and Agricultural Policies 1.113 The Brazilian industrial sector has recently accounted for 30 percent of GDP; and agriculture, despite a gradual process of relative decline, still accounts for 10 percent. Many of the policies summarized more directly elsewhere in this chapter have powerful effects in these sectors. The government has persistently sought to influence economic activity in industry and agriculture by specific policies. These are briefly described, and the problems that face are analyzed. Industrial Policies 1.114 There are three primary components of industrial policy in Brazil: import protection and export promotion policies; fi3cal incentives; and financial subsidization. The Council for Industrial Development (CDI) has been the key agency in defining industrial policies, while BNDES has served as the major instrument in providing the subsidies. The CDI licenses investment. More important, it certifies industrial activities for government financial assistance, as meeting domestic content requirements,5 and for eligibility for various fiscal incentives. Failure to gain CDI approval implies substantially higher costs for the project. U 1.115 CDI's criteria are not entirely definable, but they have been aimed at maintaining a balance between demand and supply in a market, while ensuring that a minimum efficien: scale exists for the production of most goods. 1.116 Industrial policy chaaged in 1988. The reforms were aimed at increasing the transparency of the process, enhancing competitiveness, reducing government interve,.'ion in investment decisions, and emphasizing both technological development and higher quality production. However, they also allowed CDI to differentiate between firms within the same sub- sector, enhancing its discretionary powers. 1.117 Before the 1988 reforms, fiscal incentives were very large. Exemptions from, or reductions in, tariffs and in the two VAT-type taxes (the IPI and the ICM) were granted by a wide array of bodies. In particular, BEFIEX runs an export promotion program which offers incentives to firms meeting export targets, while CDI offers incentives to particular projects or products. CONSIDER provides incentives for various metal industries. 5/ Domestic content certification is necessary for BNDES funding and for sales to the public sector, and for permitting purchasers to avoid both the payment of import duties and the using up of part of their annual import quotas. - 20 - 1.118 CDI offered relatively large incentives--an average of 15 percent of project value--which varied a great deal by sector, with consumer goods attracting the largest incentives. BEFIEX granted very major concessions, reducing tariffs that reached 70 percent to nearly zero. This support constituted a substantial percentage of overall investment, and an even larger percentage of investments made by recipient firms. Overall, fiscal incentives amounted to around 0.5 percent of GDP in the early and mid- 1980s. Regional schemes and the application of accelerated depreciation made the real value higher. The 1988 reforms widely broadened eligibility for incentives, but they also stipulated an annual maximum size. How these possibly conflicting mandates will be resolved is unclear. 1.119 Financial incentives are largely administered through the BNDES system, which provides medium- and long-term financing for capacity expansion in targeted industries and regions, subsidized credit for sinal. enterprises, and--particularly during the 1980s--assistance to industries in financial distress. BNDES financing has, along with retained earnings, been the major source of industrial financing, and almost the sole source of long-term financing. While BNDES lends throughout the economy, close to half of its lending went to industry during the 1980s, and heavy industry was the main beneficiary. 1.120 After growing fast during the 1970s, BNDES financing slowed, declining absolutely at times. Administratively, BNDES followed much the same guidelines as CDI. although new guidelines in 1989 stressed the economic soundness of the investment as the sole relevant criterion. 1.121 Several significant economic problems flow from the structure and practice of industrial policy in Brazil. The most important are: 1 1. The establishment of barriers to competition. During the 2980s, industrial policy switched from encouraging entry into markets to protecting incumbents. CDI policies now support large firms; so, to a lesser extent, do those of BEF.EX and BNDES, especially as large firms are likely to be incumbents, and attract a disproportionate share of BNDES support. BNDES credit is also highly concentrated among a very small number of firms. As that credit has increasingly gone to firms in distress, it must also have discouraged exit from an industry. 1. Distortion of factor proportions. Almost all fiscal incentives have been designed to lower the cost of both domestic and imported capital goods to selected industries. These policies have therefore encouraged capital-intensive production, and consequently discouraged the use of other factors. This resulted in exceedingly high levels of capital investment per job created in CDI projects (US$230,000 in petrochemicals, chemicals, and phprmaceuticals, for example). BNDES credit has had the same impact. 1. Distortion of sectoral allocation. By design, the incentives have been focused on specific sectors, especially intermediate-goods sectors, notably chemicals and steel. -1 - 21 - 1.122 Rents. Most incentives have gone to large incumbent firms. Where entry is constrained, these incentives have probably merely provided rents for these firms, at considerable budgetary costs for the government. !gr.cultural Policies 1.123 Overall, government policies probably discourageA the agricultural sector, even though a number of policies did work in the opposite direction. Different sub-sectors have certainly fared very differently, as export crops have been taxed more heavily than domestically-consumed crops. This effect has to some extent been offset by subsidies to larger producers. 1.124 The government has operated numerous policies in the agricultural sector. The Minimum Price Program (MPP) is the government's most comprehensive direct price intervention program in agricultural product markets. It covers most major crops (except coffee, whose price is supported through a different program), which are purchased by the Commodity Financing Corporation and are then sold in commodities markets. The program aims to provide market services where required and to stabilize prices, and provides related loans covering crop storage for up 180 days. In recent years, MPP has been mainly funded by the Bank of Brazil and the state banks. Through ownership or rental, the government also operates about half of the country's total storage capacity, and subsidizes its operation. 1.125 Self-sufficiency is the goal of the wheat subsidy program, via direct subsidies and price controls. Recent attempts to shift the cost of the subsidy tc consumers have not been very successful. Similarly, the sugar-industry complex is highly regulated, with price controls, production quotas, and regulated trade. Prices are largely determined by the goal of substituting alcohol for imported fuels; and they are also variable across regions. 1.126 So far as trade in agricultural products is concerned, government intervention is generally more important for exports than for imports. Exports are constrained for several reasons; in recent years, a reduction in the price of domestic goods with a large share in the consumer price basket has been seen as an important direct way of reducing inflation rates. The government intervenes directly in trade mainly through government monopolies, intermittent trade bans on certain products (e.g. rice, maize), monitoring of exports which are in principle free of controls (e.g.cocoa, orange juice), and the regulation and monitoring of coffee I marketing.| 1.127 The government has also provided subsidized rural credit. During the 1960s, direct lending by the Bank of Brazil predominated; it has, since 1986, been fully reflected in the budget. More recently, commercial banks have been forced (through "compulsory applications") to earmark a certain proportion of their demand deposits for rural loans at concessionary rates. While interest rates on subsidized loans were moderately negative from 1965-73, and strongly so from 1974-83, they turned slightly positive after ORTN-tied indexation was introduced in 1983, and were then set at 12 percent plus monetary correction under the Summer Plan of January 1989. - 22 - 1.128 Taxation in the agricultural sector is complex and far from ttansparent. The sector is essentially exempt from income taxes, making it a convfnient, tax shelter for other income, although agricultural exports do pay some ICM tax (unlike industrial exports). And while a land tax (ITR) provides some incentive for deforestation, its low rates, lower incidence, and lower still enforcement make this effect relatively minor. 1.129 Finally, there is a relatively small amount of direct govervmient investment in agriculture. At least two thirds of government spending on the sector goes on subsidized credit and subsidy programs; the remainder is mainly spent on various specialized agencies and institutes (such as the sugar or coffee institutes). 1.130 Like industrial policies, agricultural policies had several negative effects. They can usefully be discussed in terms of the programs that generated them. 1.131 The credit and fiscal subsidies programs have significantly increased the budget deficit, and they have had powerful misallocational effects. At their height (1979-80), subsidies to the sector were running at around 2.3 percent of GDP, but they have declined significantly since then. Allocationally, subsidized inputs (like fertilizers) are used excessively, while the artificially high capital/output ratio leads to equipment being under-utilized. In addition, the complexity of the system has generated significant compliance costs for ag;icultural producers. The distributional effects have also been negative, as land holding has become more concentrated, and credit has gone disproportionately to highly- commercialized regions and to larger producers. Subsidies also generated incentives for the cultivation of virgin forest land in some ecologically fragile areas. 1.132 The MPP was designed to stabilize prices and increase the predictability of markets. It has clearly failed to do either. It seems to have had very little effect on prices; if anything, intervention probably increased uncertainty. Government control of stockholding has also been a failure, as the financial and--at times--technical performance of public storage companies has been poor. 1.133 The wheat program has proved extremely costly, losing some $US1.8 billion in 1986, or 0.7 percent of GDP. It has created important allocational and distributional distortions, as the subsidy's negligible effect on the consumer price index went along with benefits for the rich and middle-class, mainly in the richer South and Southeast. The sugar- ethanol program has fared much the same, with annual losses running at around US$1-1.3 billion, or 0.3-0.4 percent of GDP, and similar allocational distortions. 1.134 While expenditure policies in agriculture are of minor significance, those in other sectors may have substantial effects on agriculture. Investments in road transport have generally been very effective, but investments in human capital were much less useful due to their concentration in higher education. -1, - 23 - 1.135 A series of other governmenit policies also affect the agricultural sector. The exchan e rate policy of maintaining sub-equilibrium exchange rates acts as a gerneral tax on agriculture, which is a net provider of foreign exchange to the economy. Agriculture has also beern hurt by tariffs and quantitative restrictions imposed on important inputs into the sector. H. The Financial Sector 1.136 he financial sector includes commercial banking, development banking, housing finance, credit cooperatives, securities exchanges, insurance, and the commodities and futures markets. Of these, the highly diversified banking system is the most important. The broadest problems in the financial sector lie in the use of directed and subsidized credit, which now affects a majority of all assets in the banking system, and the low levels of competition caused by the development of conglomerates together with barriers to entry and exit. This section will outline the structure of the sector, followed by en examination of directed and subsidized credit and of intra-sector competition. It concludes with a discussion of the major difficulties facing the sector. Structure 1.137 Between 1964 and 1988, the banking structure was dominated by laws aimed at separating the institutions that would serve each sector of the economy, and at ensuring that each received a separate but adequate flow of funds. Despite the segmentation, many enterprises formed themselves into financial conglomerates which provided the full array of services, often with shared staff and offices. Institutional segmentation was eliminated in 1988 with the "multiple service banking resolution," which also liberalized the licensing process for entry. 1.138 Of the 103 commercial banks, 24 are owned by state governments, and these hold 23 percent of demand deposits and 28 percent of all commercial bank loans. These state banks are largely financed by Federal funds, as deposits account for less than 25 percent of assets. The Brazilian system for house financing (HFS) was created in 1964. It is segmented, and funded by savings deposits and an 8 percent payroll tax on salaries and wages. 1.139 The primary stock exchanges in Sao Paulo and Rio de Janeiro have suffered from thin markets and weak supervision. This allowed stock manipulation and hence lack of confidence. Capital is very heavily concentrated in a few stocks. Competition in Financial Activities 1.140 Changes in the shares of total assets, total deposits, and equity held by the largest commercial banks (excluding the Bank of Brazil) show that concentration increased sharply from 1965 to 1975. By 1987, the twenty largest banks held 90 percent of all assets held by commercial banks. And the Bank of Brazil alone held about 44 percent of total assets and total equity. - 24 - 1.141 Thi,; * -iitI ration sesms to result from deliiberate poliicy (thi)ices. not f rom the drivii.g totc(e of economies of scale. If this hypothesis is corret(t, a turrn to rmatket forces via liberalization would %i)t give conglomteralt- ol' ig.r ly power. 1.142 Wlile (oncentration is high in Brazil, that is not necessarily anti-competitive. The (riti(_al element governing competition under a deregulated system is ease of entry; even a monopoly can be regulated by the market it entry is unhindered. 1.143 l Brazil's case, regulations have controlled entry quite directly, permitting only an increase in branches, and that in specified regions. Following important financial failures, a points system was created in 1985 which completely closed entry, except for foreign banks under reciprocity provisions. The points system was abolished in 1988, and new institutions will be allowed to enter the system, subject to approval from the Central Bank. About 90 applications are now pending. 1.144 Exit from the market via bankruptcy or liquidation is not common in Brazil. The economy has grown consistently, allowing non-performing loans to be covered by profits; the government has helped by limiting entry and capping interest rates on demand deposits, thus giving the banks part of the inflation tax. Mechanisms are required (like deposit insurance) for dealing with the negative externalities that stem from exit, and these have been anproved in principle, although not yet in detail. Directed and Subsidized Credit 1.145 The government intervenes very substantially to direct credit and to subsidize it. That intervention in financial markets has increased steadily, and by 1987 about 61 percent of total liabilities were under the direct or indirect control of the government (41 percent in the portfolios of government-owned institutions, and 20 percent through mandatory government applications of demand or savings deposits to required reserves or directed credit.) 1.146 Official credit programs, in tandem with reserve requirements, severely constrained the system's portfolio. In 1986-7, these programs accounted for 80 percent of the tctal average stock of credit--high even in relation to levels in other developing countries with extensive directed credit programs. Thus, for example, savings banks could lend only about 20 percent of passbook savings accounts (the predominant source of funds) freely; pension funds, insurance companies, and short-term mutual funds had to hold 25-50 percent of assets in government securities. 1.147 The subsidy implicit in a sample estimate of the largest directed credit programs (including the HFS) amounted to 7-8 percent of GDP. More than one third came from direct Central Bank advances to the Bank of Brazil or to state commercial banks, eitlier through discount window operations or -t through the so-called Monetary Fund. Inflation lags sometimes contributed, especially after the collapse of the Cruzado Plan. =I - 25 - Mlj_y Problems in the Financial Sector 1.148 Official subsidized directed credit has an income and a price effect. Though neither are directly measurable, both are perverse and distortionary. Forced portfolio allocations have certainly increased the price and scarcity of non-forced loans, and reduced the rates received by depositors. 1.149 These subsidies are poorly corrected for inflation, which undermines budget calculations as well. Thus directed credit programs have complicated the task of controlling monetary aggregates, and they have also substituted for the primary issue of securities, thus inhibiting the deveelopment of capital markets. 1.150 The capital markets themselves face both supply and demand problems. On the supply side, state interventions and regulations have seriously distorted signals to investors. On the demand side, government intervention has strongly pushed capital seekers to look primarily to bank credit and internal sources of capital, rather than to the capital markets. Clearly, the higher the quantity of subsidized credit available, the lower the incentive to enter non-subsidized capital markets. Reluctance to risk family control of businesses is also sometimes cited as a reason for avoiding capital markets. 1.151 In the housing sector, the deep recession and high inflation of the early 1980s created a major crisis at HFS. Borrowers failed to service their debts, and foreclosures became both ineffective and politically unfeasible. The result was a bailout financed by FCVS. At the same time, as interest rates on savings accounts were fixed at negative real levels, massive withdrawals occurred, causing a liquidity crisis. 1.152 These problems were only exacerbated by the Cruzado Plan, and most borrowers now pay only a fraction of the amount required to amortize the principal or even to cover interest payments. FCVS covers the losses, and it is now expected to require financial assistance soon. By the end of 1989, heavy portfolio switching out of savings deposits and into government bonds had put great additional stress on the system. 1.153 Among the state banks, the state commercial banks are of special important because they alone operate in a relatively unfettered manner as the true financial agents of their states. From experiences in the mid- 1980s and again in 1987-9, we can conclude that in periods of strain state banks have been able to borrow what they need from the Central Bank, and that the latter has been unable to insist that agreed reforms are actually implemented. 1.154 This borrowing had very significant monetary effects. In 1985, the amount of debt to the Central Bank already approached half the monetary base, and after the reversal of the Cruzado Plan it had reached that level again by mid-1987. That borrowing allows the states to increase their share of the inflation tax, but at high cost to the economy; and this mechanism must be eliminated for any true stabilization policy to work. In - 26 - addition, the states seem to have used their banks to finance real increases in their expenditures, in the (accurate) expectation that the banks will be eventually bailed out by the Central Bartk. I. Social Expenditures 1.155 By virtually all indicators, the social welfare of most of Brazil's population lags well behind that of countries with similar income levels. The average child completes less than 5 years of schooling--the lowest level in South America except for Bolivia. In 1987, infant mortality rates were more than twice those of Korea and three times those of Chile. And large regional disparities mean that these averages conceal even more disastrous problems in the Northeast. 1.156 Yet Brazil spends more per capita on social programs than many similar developing countries. Clearly, inefficiency and mistargeting are major problems: in 1986 the poor (41 percent of the population) received only about 20 percent of social expenditures. Brazil achieves less because spending has been partly targeted at pensions and urban housing, which have not helped the poor. Mistargeting is reflected in both the education and public health systems. Education 1.157 Brazil's education system is very diverse, with all three levels of government providing services, alongside a strong private sector. In 1985, overall government spending amounted to about 4 percent of GDP on education, and the private sector spent another 1-1.5 percent. While much education spending has been financed by earmarked taxes, these taxes ha e probably not provided additional resources for education. The diversity of the system leads to coordination problems between the efforts of different levels of government and with the private sector. These problems are reflected in complex and inefficient patterns of financing. These in turn show up in sharp regional disparities in per-student spending. 1.158 The new Constitution decentralized many administrative arrangements and responsibilities, and also mandated considerable increases in state and municipal spending on education, from sources to be generated over time by the 1988 tax reform. 1.159 The educational system has serious problems with its management of human and physical resources, reflected most directly in the very heavy - bias in funding toward higher education, which cannot be justified on any rational economic calculus. This bias is partly caused by the Federal emphasis on higher education spending, while the relatively poorer states and municipal authorities pay for primary and secondary education. 1.160 Pre-primarl education has grown rapidly in Brazil, and today about . 25 percent of the age group is enrolled. Enrollees come primarily in urban areas in the Southeast, and few are from lower-income households--the children who most desperately need pre-primary help. Pre-school education needs expanding, and efforts to insist on public sector expansion alone do not seem promising. -~~~~~~~~~~~~~~~~~~~~~~~ g - 27 - *e t.:v-,- (School education has also expanded during the past 30 y t*ar, c i;now nearly all eligible children are enrolled. Once again, toWeel |, :t g'na1 disparities are important. Although state schools a.,r,- {fo mote than half of enrollmer.t and municipal schools only 30 ppei. e:t , :.r than 70 percent of rural students are taught in municipal sh. :,K. PIrvate schools cater to 12 percent of the students. Now that a-kk.. 1 ,prrmary education is almost universal, the key problems are qcjalitv ir-' student attainment. Less than 40 percent of entrants ever ( orviet e;t ,iary edu'ation, and most who do require substantial and costly re,pet it Ion. 1.162 Most se;ondary students are taught in state and private schools (nearly 60 percent and about 33 percent respectively). The large role of the private sector is striking, and so is the very large number of night classes (covering 56 percent of all students). The share of secondary schooling in total public spending on education is 8 percent, compared with an average of 26 percent elsewhere in Latin America. This is only partly explained Ly the large private sector. 1.i63 Secondary education Faces low enrollment and high repetition and drop-out rates, at the same time that demands for higher skills in the workplace are increasing. Currently, there are huge disparities in spending by school type and source of funding. The distinct problem of declining quality in night classes is also growing. 1.164 Brazil does spend heavily on vocational training, which is an important alternative to secondary schooling. The vocational programs are Federally funded and run by industry. They train about 1.7 million workers annually, and cost about US$555 million, financed by earmarked payroll taxes. 1.165 Undergraduate university education in Brazil expanded rapidly in the 1960s and 1970s. In 1985, the 1.4 million students were split equally between the 80 universities and the 800 isolated faculties. The Federal government pays for 45 percent of university enrollments, and the states another 15 percent. The private sector accounts for the remaining 40 percent and for about 80 percent of students in isolated faculties. These faculties have met the growing demand for higher education from those denied entrance to the (generally elite) universities. 1.166 In 1985, almost 60 percent of Vederal expenditures on education went to public higher education. Public system students pay no tuition at the university level. The university system also shows signs of significant inefficiency: it has one of the lowest student-teacher ratios in the world, and salaries absorbs more than 90 percent of the higher education budge'. Health and Nutrition 1.167 Over the next few decades, Brazil's health system needs to change its approaches to traditional problems (e.g. malaria, diarrhea), improve its care for women and children, and deal with new challenges (AIDS, and - 28 - the growing case load of cancer and cardiovascular problems as the population ages). Current weaknesses limit the system's capacity to respond. 1.168 Since 1968, participation in social security has been mandatory for all formal sector workers. In the 1980s between one-quacter and one- third of the social security budget went to INAMPS to finance health services for contributors and their families. This funding accounted for two-thirds of all Federal spending on health and over one-half of all public health spending. Most of INAMPS-financed services were provided through contracts with private physicians and hospitals. 1.169 As INAMPS' role in the health sector became dominant, the health sector became preoccupied with the provision of personal curative services. The share of total public health spending allocated to curative personal services rose from 36 percent in 1965 to 85 percent in the 1980s. Brazil allocates about 70 percent of its public spending on health to hospitals--a level higher than that of industrialized countries and much higher than in similar developing countries. 1.170 In the 1980s there have been strong pressures to reform the health system. The reform process, now enshrined in the Constitution, involves two major elements: opening up the social security to all citizens (not just those who work in the formal sector and thus pay social security taxes); and decenteralizing responsibility for operating health services from INAMPS to state and municipal governments. 1.171 An autonomous private sector (known as 'group medicine" in Brazil) also emerged, principally in the major metropolitan areas of the South and Southeast. This sector is financed primarily by contracts with enterprises, in which both the employers and workers contribute. The group medicine organizations receive no direct funding from INAMPS or any other public source. Competition amongst group medicine organizations is vigorous. This sector has grown extremely rapidly, and currently covers about 30 million people. 1.172 Brazil is in the midst of an epidemiologic and demographic transition. Over 70 percent of the population lives in urban areas. Fertility rates have fallen dramatically in recent decades: the average woman in 1965 gave birth to 6 children during the course of her life; in 1985 the corresponding figure is below 3.5. Between 1960 and 1980 infant mortality rates dropped by almost one half and death rates due to infectious and parasitic diseases declined by about 70 percent. These changes pose new challenges for the health sector. In 30 years time, 85 percent of the population will live in urban areas, the proportion of elderly will have doubled, and cardiovascular diseases, cancers and injuries will account for 3 of every 4 deaths. It is estimated that the joint effect of these epidemiologic and demographic changes will be to double the per capita cost of providing the current quality of medical treatment to the currently served proportion of the population. 1.173 Along with these 'new' challenges, Brazil faces formidable "old' health challenges. Child health remains poor. About 6 percent of children born die in the first year of life, a level much higher than in most - - 29 - countries at a comparable level of development. Endemic diseases (including malaria, yellow fever, Chagas' disease and schistosomiasis) are also still important in the less developed parts of Brazii. For example, the reported cases of malaria more than doubled between 1979 and 1986. 1.174 The reproductive health of women is especially poorly served in Brazil, which has high rates cf maternal mortality and morbidity, unsafe abortion and unnecessary caesarian rates. Information about, and choice of, contraceptives is also insufficient, especially among the poor. 1.175 An important dimension to the health challenge in Brazil is the association of poverty with high levels of mortality. This is widely understood for the "old" causes of death. For example, the infant mortality rates in the Northeast (double the national average) are worse than those in many poor African countries. Vaccination levels in the Northeast are low, and many children continue to die of diarerheal diseases. The poor are also at much greater risk of death from the "new" causes of death--cardiovascular diseases, cancer, and injuries. 1.176 Finally, Brazil faces an AIDS epidemic of major proportions. Already Brazil has the second largest number of AIDS cases in the world. All indications are that the disease will not be confined to limited, "high-risk" groups, but will continue to spread to the population at large. -~~~~~~~~~~~~~~~~~~~~~~~~ - 30 - II. A POLICY AGENDA 2.1 This thapter draws policy conclusions from the analyses that wiJ! be detailed in the next two parts of the report, and which have just been summarized. The policy recommendations will be divided into two SeCti.ns. In the first section, short-term reforms will be suggested, conmi~,ting of policies which should be carried out during the next few mionths. In the following section, an agenda for a longer-term structural reform is presented. It consists of policies which may either need time to plan oi whose implementation is lengthy, or which should best be undertaken gradually rather than in a single step. In each section, the policies are to be regarded as related to each other, forming components of a policy package, rather than separate and independent steps. This applies with particular strength to the short-term policy agenda where macro-economic stabilization is the overriding concern which must direct and bind the use of individual policy instruments. 1. Short-Term Structural Reforms and Stabilization A. The Policy Framework 2.2 Under the present extreme degree of macro-economic instability, stabilization must be the first order of business. But such a policy cannot, and should not, be considered in isolation; it must be closely tied to an agenda of structural reform policies. This is the case for two * important reasons. First, the sustainability of stabilization depends on it being founded on solid structural changes. A stabilization policy, particularly in its fiscal aspects, that does not adopt the required structural steps is bound to be seen as reversible, not durable, and hence not credible: several structural reforms are essential to reduce the fiscal deficit on a sustained level. At the same time with the possible exception of reforms in soc'Rl programs, structural reforr.is cannot be - introduced and presented in a macro-economic environment as volatile as in Brazil today. Such reforms must therefore be contemplated only in the context of a stabilization policy. 2.3 The second reason derives from political-economy considerations. The macro-economic crisis has provided a sense of urgency to policy changes. It should therefore be easier now to introduce reforms which would normally face substantial--often overwhelming--political resistance. Experiences in many other countries suggest that at times of crisis, major structural reforms may sail through; if that opportunity is missed and fundamental reforms are delayed, they are likely to be shelved for many years or avoided altogether. 2.4 This report therefore specifies a short-term agenda of structural * reforms which should be carried out, as a first step, in conjnction with a - stabilization scheme; the discussion will be extended further to other essential components of a stabilization policy. The policy steps are lI - 31 - selected for their economic sigr,ificance, and for being rational beginnings of the process of structural reform in their respective spheres. They must also be consistent with the requirements of stabilization. It is also worth mentioning that very few steps which would be justified on structural grounds need to be rejected on stabilization grounds: the need to make trade-offs is, fortunately, very limited. 2.5 The short-term stabilization scheme presented below should be viewed as a package which, to be successful, must be adopted in toto. This is a structure which would collapse if one of its building blocks is removed altogether, or if its size is not calibrated correctly in view of the other blocks. This does not mean that no substitutions at all are possible. Thus, for instance, more of one tax may substitute for less of another, or one instrument of monetary policy may--with greater difficulty- -replace another. But the major instruments must all be used jointly, in the direction and rough orders of magnitude that are indicated. The experience of Argentina's stabilization policies (which have had to be devised with great haste) offers a clear lesson: a policy would fail if its essential structural components (of fiscal, monetary, and exchange rate policies) are not implemented fully and in conjunction with each other. If the reform measures are not seen as substantial and serious attempts, they will also not be seen to be credible, and the reforms will fail. 2.6 The policy reforms specified are strong and the orders of magnitude are large. This is due to two inter-related considerations. First, given the history of rapid macro-economic deterioration and repeated failures of stabilization attempts, the policy package must be both strong and have an immediate impact to be seen as credible--and policy credibility is of paramount importance; second, errors in evaluating the feasibility of implementing a policy change, in estimating the extent of the realized change in the policy variable (e.g., the increase in revenues resulting from a given change of a tax regulation), or in predicting the economic impact of this change, are bound to happen; and, to prevent a collapse of a well-meant policy, an over-shooting should be aimed to counter the risks of over-optimistic estimations. 2.7 In suggesting these recommendations, we realize that there may be some political constraints. In some cases there may be certain vested interests that would oppose the changes. These are not discussed here. In other cases, particularly in the components of fiscal reform, certain changes may require congressional approval whereas others may be implemented by executive decisions. This is discussed later in this chapter. B. Fiscal Reforms 2.8 In the sphere of fiscal reforms, three areas of policy will be discussed: government expenditures; taxes; and subsidies. The first - 32 - concerns the real activity of the government, whereas the other two, in combination, determine the net taxation imposed by the government.6 Excep)t for the impact of one of the measures on the budgets of local governments. the discussion will be confined to the Federal government. Estim,ates of the impact of these policy reforms on the fiscal deficit will be indicated later in Table 2.1. Government Expenditures 2.9 First, an immediate reduction in the central-government labor force should be undertaken. In the long run, it is generally estimated that the same amount of government services can be provided with a substantial reduction of the labor force; but this would involve time- consuming procedures and studies, and should therefore be left to the longer-term agenda. However, a smaller reduction is believed to be both legally feasible and technically manageable without much disruption. The fiscal saving of such a step is of only secondary importance. The main significance of its introduction is that, due to its high visibility, it would lend credence to the government's announced plan of fiscal contraction. In addition, real wages of government employees at the higher part of the wage scale should be reduced, in a manner which will be specified later in this discussion. 2.10 We also recommend that virtually all discretionary transfers from the Federal government to state and local governments be eliminated, leaving in operation only the transfers using transparent criteria. Reduction in inflation will simplify this process, since high inflation helps provoke "discretion' by administrative manipulation of supplementary allocations. To strengthen the Federal government's redistributive role, transfers to those states and muncipalities that are benefitting relatively more from the 1988 changes in revenue sharing could be reduced. Any reduction in transfers (combined with an effective ban on additional borrowing) would inevitably lead those state and local governments to some combination of the following steps: contract their real activity, by reducing the size of their labor forces; reduce real wages; increase taxation; and impose (or increase) payments for services--that is, again, raise their net taxation. Transfers to the poorer regions, and for activities in sectors such as education and health that are heavily financed by centralized Federal 'social contributions,' cannot be reduced _ in the short run. 4~~~~~~~~~~~~~~~~~~~~~~~~~ 6/ The definition of a measure as a change in subsidy or (in the opposite direction) a change in tax is very often arbitrary (and of little economic consequence). Thus, the elimination of a subsidy granted through a tax exemption--a quite common feature of Brazil's both direct and indirect taxes--may be labeled a tax increase or a subsidy reduction. The change in net taxation would obviously be the same whatever the classification. I - 33 Taxat irn 2.11 Severa1 significant reforms of the taxation systtem ale proposed. In fact, most of the important reforms are candidates ior short-run adoption. They aie required for stabilization and can be 3mplemented in a relatively short period (several months). 2.12 Regarding indirect taxes, first, the coverage of the two VAT-type taxes should be extended. The base of the ICK should be widened. The coverage of the IPI, the other tax, should also be broadened, eliminating its many exemptions: and a flat rate should be applied to most activities. This would both increase revenues and remove distortions (see the discussion in Chapter VI). An increase in the average tax on oil products is also recommended, by removing the substantial subsidy elements which some of these products currently enjoy (at the same time that other oil products are subjected to taxation). This would, again, lower distortions by bringing relative prices of varieties of oil products closer to the structure of international prices. 2.13 Under direct taxes, several major reforms should be implemented. An increase in the highest marginal rate of the personal income tax is justified, though it would have relatively little quantitative impact. This rate has been lowered recently to 25 percent. Part of this reduction may best be rescinded (to bring the rate to something like 35 percent). Recommended changes in the corporate income tax are more important. The many fiscal incentives now granted through exemptions from this tax should be eliminated. A whole important sector, agriculture, is currently exempt altogether, and it should become subject to the tax at the same rate as other sectors. This would primarily increase tax revenues from activities in other sectors, which now use agriculture as a tax shelter. In addition, consideration should be given to the introduction of an alternative minimum tax, which would ensure tax payments by activities that will continue to be exempt for one reason or another. Subsidies 2.14 There are several fiscal and credit subsidies in the system that cause a major drain on the Federal budget. These subsidies should be eliminated. The most important ones include subsidies to wheat and to the sugar-ethanol industry, and directed, subsidized credit in the agriculture sector, 2.15 In addition, the substantial subsidy element involved, at present, in the maintenance of low prices for public-sector goods and services (such as electricity, steel products, and telecommunications) should be removed. - 34 - At a minimum, these prices should be restored to their real level at the end of 1988--prior to their reduction under the "Summer Plan."7 C. Other Reforms 2.16 Beyond the fiscal area, short-term reforms should also be undertaken primarily in the foreign-trade regime. These reforms would also have a substantial fiscal impact. They should include: (a) The removal of all quantitative restrictions of imports. They should be replaced by several categories of surcharges, which would lead to rough equivalence of the protection granted through QRs, except for the goods whose imports are now completely prohibited (and taking into account the added protection through accompanying real devaluation, as we shall note shortly). A time range for the elimination of such su:charges should be announced upon their introduction. The removal of all QRs should also include the repeal of the "Law of Similars' and the "Law of Informatics.' Only a very narrow prohibited list should remain (based on safety hazards etc.). (b) The elimination of all duty preferences, with the exception of the drawback scheme for exports. Any imports of inputs or of capital goods even formally free of import duty should becore subject to a tariff. The longer run objective of this proposal is the Lmplementation of a uniform tariff on inputs and capital goods, at some 0-15 percent. This is a correct measure, of course, only because a present most outputs are subject to higher tariffs (or, usually, to QRs which grant higher levels of protection); and because eventually a uniform low tariff on outputs should be established. (c) The elimination of all export subsidies (the drawback scheme, it should be noted, is not a subsidy). This recommendation, however, strongly depends on the implementation of the recommendations below concerning the foreign-exchange regime. 7/ The nature of this recommendation (and consequent quantitative estimate) should be clear: it merely advocates, as an immediate step, the restoration of public-sector prices to their (real) levels at a given starting position, i.e., December 1988. This by no means implies that these are, by one definition or another, optimal prices; most likely, they are not. A World Bank study of prices in the energy sector, which is nearing completion, will indicate optimum levels of prices for electricity and oil products. - 35 - (d) A substantial increase in the real exchange rate (i.e.. a real currency devaluation). It is generally believed that the present average rate of exchange8 is too low (the cruzado is over-valued). In addition, two of the policy changes suggested above call for further devaluation (that is, a formal devaluation would be required to maintain a given effective rate of exchange). The first is the abolition of export subsidies, which would by itself lower the effective exchange rate for exports, and the second is the abolition of QRs. Although the latter would be accompanied by the imposition of tariffs, or surcharges, to offset the removal of protection through regulation, compensation would not always be full. Thus, a substantial real devaluation is required. In addition, a recommendation made later (Chapter VIII) to abolish the tourist rate and establish a uniform exchange rate, would call for a still greater devaluation of the existing formal rate. Later in this chapter we will discuss the implications of the above for a nominal devaluation. 2.17 Beyond trade policy, the fundamental reforms that will be suggested in this report are almost all longer term. There are two exceptions. One is in higher education. On both equity and efficiency grounds, the government should introduce changes that will increase the financial accountability of the Federal universities. We recommend the immediate introduction of charges for such non-educational services as meals, and the payment of laboratory fees; the introduction of prospective budgeting; and a crash program to inform the students and the public of the privileged status of public-university students. 2.18 The other reform is suggested in the area of nutrition, where an immediate, though modest, increase of Federal spending is called for to enhance the social safety net during the transition period following stabilization; this is elaborated upon later in this chapter. The increased expenditure could be partly offset by the elimination of non- targeted nutrition programs. D. Structural Reforms and the Fiscal Stance 2.19 We shall now relate the fiscal implications of the short-term structural reforms proposed above to the extent of the fiscal change that a macro-economic stabilization policy would indicate. 2.20 In the long run, an operational balance (i.e., a zero deficit) should be both required and sufficient (along, of course, with other appropriate policies) to sustain price-level stability. A zero balance 8/ The average rate of exchange is the weighted average of the official and the tourist exchange rate. - 36 - would mean that the government neither creates directly excess demand, nor insufficient demand; and also that it makes no addition to (or subtraction from) the size of the monetary base and the money supply.9 For the immediate stabilization period, however, this would not be sufficient: within the stabilization scheme, a substantial operational surplus has to be generated, for at least three reasons. First, to change expectations and lend credibility to the program a large fiscal shock must be introduced. Second, planning can never be precise, and a scheme may not be fully implemented; erring on the side of large surplus is advisable if unexpected shortfalls are not to abort the program. Third, a substaiutial change in the structure of expenditures is required in the medium term: public-sector investment, which has been faltering in recent years, must increase (in infrastructure--not in public enterprises), at the expense of consumption. Overshooting the required budgetary surplus for a year, or perhaps two years, would p'rmit a future increase in public-sector investment without the reintroduction of a budgetary deficit.10 2.21 Achieveme..t of an operational surplus will require a large reduction of the real interest payments on the government's debt, and the creation of a substantial primary surplus. In a later section, we will examine ways for lowering the real interest bill. Here we will discuss the primary balance. 9/ In fact, in a growing economy some positive government contribution to the monetary base would still be consistent with price stability. 10/ A major cause of the current hyperinflationary situation in Argentina reflects the fact that an operational surplus was not created. J4 4 - 37 - Table 2.1: THE FEDERAL BUDGETARY CHANGE Contribution to Budgetary Surolus Change (in percentage of GDP) A. Revenues Widening the base of IPI, and imposing a flat rate/a 1.5 * Elimination of corporate income-tax exemptions/a 1.0 ** Imposition of corporate income tax on the agricultural sector/a 1.2 ** Increase of highest marginal rate of personal income tax/a .2 ** Replacement of QRs by import tariffs .7 Effective imposition of tariffs on imports of intermediate and capital goods .5 * Increase of average tax on oil products .4 * Increase of regulated public-sector prices 3.5 * TOTAL 9.0 B. Expenditures Removal of export subsidies .6 * Removal of rural-credit subsidies/b .5 * Removal of fiscal subsidies in agriculture (wheat and sugar- 1.0 * ethanal) Reduction of discretionary transfers to local governments .5 * Reduction of public-sector labor force .3 * Reduction of public-sector real wages .1 *** TOTAL 3.0 Aggregate improvement of the primary balance 12.0 * Can be implemented by executive decision. ** Requires congressional approval and budget revision. *** Requires agreement and possible legal changes. /a Close to one-half of the increased revenues from these taxes would be the share of the states. Strictly speaking, this part would then not be counted as an improvement of the Federal budget. -b Other credit subsidies are estimated to amount to as much as 5 percent or more of GDP; and their recommended removal would in substance improve the primary balance to that extent. These implied subsidies are not recorded, however, ir. the budget; hence, in the accounting sense their removal would not lead to a budgetary change. (Put differently: they would first have to be recorded, increasing the deficit; then removed). 38 - z 2.22 Table 2.1 sums up the size of changes ir, the budget tor govet:wnent revenues and expenditutes, respectively, whi.h fkAlow from the recommended short-term structural reforms. If all the suggested rt-forms are implemented, in aggregate. the primary balance would improve by some 12 percent of GDP.11 While in most years the primary budget was balanced, or even tesulted in a slight surplus, it is estimated that in 1989 a small deficit, of about 3 prcent of GDP, is likely. Hence the changes proposed. of 12 percent, would create a primary surplus of about 9 percent of GDP. 2.23 Implementation of the fiscal changes--even if all measures are defined at the inception of the plan--is a time-consuming process. The process of translating a decision on principle into concrete action is rather involved. Some important steps require Congressional approval, which could take some time. Also, once approved, changes in taxes would legally become effective only in the forthcoming fiscal year (that is, starting in January 1991). Hence, for the first few months, the impact of the proposed measures on the budget will be considerably less than that estimated here. Table 2.1 indicates which of the proposed steps may be implemented immediately and which may depend on legislation, and would have to wait for implementation in January 1991. Thus, an impression may be gained of a feasible time path for the realization of the fiscal changes. The full impact of the fiscal changes should be realized in the calendar (fiscal) year 1991 in which the operational surplus should approach the target of .4 percent of GDP. In calendar/fiscal year 1990, on the other hand, an approximately balanced budget should be expected with the proposed fiscal reforms. The target of a 4 percent surplus is illustrative, rather than definitive. It is based on the assumption that slippages in implementation would still allow some operational surplus: the sign of the budgetary balance should be of considerable importance. 11/ This aggregate could presumably be achieved by substituting other fiscal changes for those proposed here; but in their impacts on the economy's allocation of resources and income distribution--as analyzed further in this report--the proposed measures appear to be superior. In this context, a warning should be sounded against the use of one source ot government revenue, namely, the sale of assets to the public. Such sales should often be undertaken on their own merit--as will be discussed extensively later in this report in reference to the privatization of public enterprises. Even, however, were it possible to conduct such transactions within the relevant time for stabilization, it should be realized that the sale of a government asset to the public is not a tax; hence, it should not be perceived as affecting the fiscal deficit. It changes the form, but not the size, of the private-sector's wealth. In this, it could function like an open-market operation; but given the other forms, considered here, of government operation in the money market, no useful monetary role may be assigned at present to government sales of real assets. - 39 - 2.24 An added factor--unaccounted tor here--will most probably contribute to some further increase of the surplus. This is the reversal of the Tanzi effect,12 which should increase substantially the real level of tax revenue with the expected major deceleration of inflation following a successful stabilization program (despite the fact that lags in the collection of Faxes have been shortened considerably during 1989). This effect on real tax collections may, however, be offset by an increase in the level of public-sector wages which would result from indexation of wages to past price changes when inflation is decelerating.13 E. Other Elements of Macro-Economic Stabilization 2.25 Three other policy components must form part of a comprehensive and sustainable macro-economic stabilization: monetary policy; exchange- rate policy; and incomes policy. Monetary Policy 2.26 The creation of a substantial budgetary primary surplus is a necessary condition for che success of a stabilization policy; but it is not sufficient: it is li"zely to fail in the absence of a sharply contractionary monetary policy. First, as noted, while basic decisions about the course of fiscal policy and their announcement must be made as part of ary first step, with no delay, the realization of fiscal measures is a time-consuming process; possibly close to a year would pass before they could become fully operative. Even once all the laws and rules are adopted, I. would take time to implement various cilrponents fully. Second, even if all changes in tax revenues, subsidies, and other government expenditures were fully effected on the introduction of a stabilization program, the public would only recognize these changes after a time lag. A public which has continuously experiencea announcements of fiscal changes which do not materialize has become highly skeptical about such announcements and will need time to be persuaded that the changes have proceeded beyond mere intentions. Third, the creation of an operational surplus would, of course, lower the size of the government's debt. Howeve:, given the large internal-debt stock, this change in flow would, in 12/ The Tanzi (or Oliveira-Tanzi) effect describes the impact of inflation on the fall of real tax revenues due to tinme lags in tax collections. An acceleration of the rate of inflation would lead, ceteris paribus, to a further reduction of the real value of the tax; whereas a deceleration of infiation (even when some inflation is still maintained) should lead to an increase in real tax revenues. 13/ The accounting also abstracts from two offsetting (in direction) effects the real devaluation (which will be recommended shortly): an increase of local-currency value of (government) external-debt payments; and an increase in the size of revenues from import duties. A real devaluation of 25 percent is estimated to lead, in these ways, to a net increase of government expenditures by some 0.5 percent of GDP; hence, the surplus is reduced by that amount. . - 4 0 - thIIe Sh e. :t - If, 1-ive ot, ly a mtargin al ttect on the siz- of thLe interiaal deb)t st o.k, ci x :.' t y anl 1 i quidit v it th.Et system. As out analysis shows (in Chapter IV), the t:m.aior tor(t-* dr-ivinig the inflationary pro(ess today is the size atnd nadtuioe of thti internal debt. A montary shock raust therefore be a fundanmental elemont of a sustainabie program. History tells us, in fact, that no economy wit h hyperinf lat ion has been stabilized without a thorough monetary ref orm. 14 2.27 In Brazil, money is mostly the government's short-term internal debt. The size and nature ot this internal debt present a major obstacle to the implemertadtion of a contractionary monetary policy, as well as of the required fiscal change. Given this size and structure, any attempt to contract the monetary base and raise real interest rates leads (aq it has most clearly in 1989) to an increase of the short-term debt and, thus, an increase of the quantity of money (M4). 2.28 The government must therefore make a major effort to change the attributes of its domestic debt: to convert it from a highly-liquid asset, essentially part of the public's money, to a component of the public's long-term investment. This may be achieved through the introduction of various long-term financial assets, with yields that would make them attractive as investments. Such yields would be considerably lower than the real interest rate during most of 1989 on the government's short-term debt (the LFT's). Hence, the government's real interest bill on its domestic debt would be reduced. If, for example, the real level of such a yield were to drop to only slighly more than the level of interest on (indexed) saving deposits, the interest bill on the government's domestic debt could drop to around 2 percent of GDP, rather than the current 6-8 percent. With interest payrnents on the government's ex.ernal debt at the present level of 3 percent of GDP, real interest payments by the government would then total around 5 percent of GDP. 2.29 If this approach is adopted, money and liquidit; will be tight when the stabilization policy is implemented, with a generally high level of real interest rates. The public's response will be the creation of money through external transactions. This tendency would be reinforced if, as in one of the options suggested below--that of tho fixed exchange rate --the stabilization package establishes a starting real exchange rate which is both high and expected to fall. The combination of a high domestic real interest rate and an expected decline in the real exchange rate should produce an inflow of short-term capital, mostly the repatriation of Brazilian capital held abroad. Data on the current stock I of this capital are not available for obvious reasons; even approximate estimates are not feasible. But it is believed that, while this is a minor part of aggregate Brazilian liquid assets, it may amount to USS10-30 billion. Even the low end of this range would be approximately equal to the country's monetary base, and roughly twice the size of the government's foreign-exchange reserves. Repatriation of a large part of this capital, 14/ In the two sustainable stabilization schemes of recent years--in Bolivia and Israel- -monetary reform and monetary contraction were essential ingredients. .2 - 41 - whikh may reasonably be expected with this scenario, would then have a suBstantial impact on the monetary base. 2.30 Short-term capital inflows would also be generated through short- term iborrowing from abroad. While Brazil's capacity for government 1bOrrowing, and long-term borrowing in general, is (and may be expected to romain) highly constrained, short-term, secure commercial lending is still available. Under the prescribed pattern it will probably be drawn upon substantially. This would reinforce the impact of repatriating Brazilian short-term capital. 2.31 Still another mechanism for generating mone) through external transactIons would be the trade balance. An initial real depreciation, large enough to maintain a relatively high real exchange rate even after several months of domestic price .nd cost increases, would rapidly promote exports and--to a much lesser extent at the beginning, given the prevalence of QRs in Brazil--contract imports. The reduction of domestic absorption required as a counterpart of the increase of the trade surplus would be provided by the creation of the budgetary primary surplus. 2.32 The combined outcome of these movements would boost money and the monetary base. A sterilization of these movements would not be advisable because they represent a rational response to a liquidity squeeze, and to entirely offset them would probably mean the onset of a severe recession. On the other hand, since these movements may be highly volatile and unpredictable, it is also not suggested that an unconstrained market response to such an increase in the monetary base should be allowed; that is, we suggest that the multiplier effect of such an increase in the base money supply should be blocked. This would be achieved by making the credit supply the monetary target. Real credit supply should be expanded beyond the expected real growth rate of the economy: with the proposed creation of the operational balance, a reversal of the 'crowding out' will now take place, enabling the expansion of credit to the private sector. The nominal target for credit supply should then, of course, incorporate the expected changes in prices.1 2.33 Given the volatility and speed of potential changes--particularly of repatriation of short-term capital--we suggest that credit, during an 15/ A model used by the World Bank indicates that, given the other proposed changes (in the budget and the exchange-rate level and regime), real credit supply could be increased during the first year at the annual .-ate of some 25 percent. The inflation target should not be zero--an unrealistic expectation unless a severe recession is allowed. Recent experiences in other countries following sustainable stabilization- -Bolivia, Chile, Israel, or Mexico--show the persistence of annual inflation in the range of 15 percent for several years following polizy implementation. Although no a priori reasoning explains this level of inflation, this consistent experience leads to expectations of similar inflation levels in Brazil for some time after stabilization is implemented. - 42 - initial period of 4-6 months, be determined through direct regulation (by decree) of its size. After that initial period, changes in bank reserve ratios and, primarily, open-market operations can be used. The Exchange-Rate Policy 2.34 Basically, there are three options for an exchange-rate policy as a component of a stabilization policy. All offer opportunities and risks. 2.35 One alternative is to unify the exchange rate a.ad let the foreign- exchange rate float freely, with no attempt at government intervention. This approach has several virtues. One lies in the absence of intervention itself, which would be consistent with the general agenda of short-term structural reforms and stabilization aimed at removing distortions created by the Brazilian Government's pervasive intervention in the economy. This approach would also obviate any balance-of-payments problem, whose potential existence would endanger the credibility and sustainabilfty of the stabilization policy. 2.36 Against these benefits must be counted several drawbacks. A floating exchange rate would probably exhibit, at least during the early stages, a strong element of instability, which may undermine the credibility of the policy as a whole. In their direct impact, movements of the foreign-exchange rate are presumably, in Brazil, only a minor source of the country's aggregate price changes. This is because the country's trade is very small in relative terms (imports amounted, in 1988, to only somewhat more than 5 percent of GDP), and because imports are largely regulated (at the moment) through QRs. But indirectly, as a signal, the exchange rate in Brazil--as in other countries experiencing strong and volatile inflation --does play a substantial role in the inflationary process: exchange rate movements have become a major determinant of inflationary expectations. For an initial period, in which inflationary expectations must be changed, it is therefore important to change the signal. 2.37 The other risk involved in a floating-rate policy is that while it avoids a balance-of-payments problem, a balance-of-trade issue may well be expected. Specifically, the anticipated large-scale capital inflow would lead to a low real rate of exchange, which would in turn inhibit exports and promote imports. Although no foreign-exchange scarcity would result, the contraction (or lack of sufficient expansion) of the tradeable activities would aggravate any recessionary impacts of the stabilization policy, and may lead to the perception that the trade-liberalization policy--undertaken in conjunction with stabilization--has failed. International experience suggests that such circumstances greatly increase the prospect that trade-regime reforms will be aborted. 2.38 The deficiencies of one approach are the virtues of the other. The main virtues of the alternative--fixed--(nominal) foreign-exchange rate--is that it offers a signal of stability. Such an exchange rate would remain fixed for a short period (say a few months) and the regime would I - 43 - the; ,. :,.i te oither a free or, a managed float. Provided that the new fi.x..! eX . ate was determined appropriately, it would also grant the fle e .-.i: v su~; ;t tor the traue-policy reform. The major risk, on the o)ther- h; , 1s that a fixed nominal rate will remain fixed for too long, beini,njined ever. after the real rate becomes excessively low (with the *-hargte in 1!-al prices and costs). This would result in the contraction of tiadeat'l a,tivities, a shortage of foreign-exchange reserves, a black- market ex,Lange rate whose disparity from the official rate rapidly inc-reases, and anticipation of a major devaluation which would, in turn, shake the credibility of price stability. Another risk is due to past experien(t-: the fixed-exchange-rate principle was used in most of the prior failed stabilization experiments in Brazil in recent years, so that repeatiing the policy might itself undermine the credibility of the stahilization policy as a whole. 2.39 A fixed-exchange-rate policy would make sense only if the selected level of the nominal exchange rate itself seems credible. No balance-of- payments pressure on the rate should be perceived, and no significant black market, with a substantially higher exchange rate than the official one, should develop. This implies determination of a high real exchange rate when stabilization is introduced; hence, a substantial nominal devaluation is warranted. What the level of such devaluation is, is difficul to determine precisely. During 1989, a real appreciation of some 35 percent took place. In addition, the proposed changes in trade policy (abolitior of QRs, with less than full compensation by surcharges, and abolition of export subsidies) would further lower the real effective exchange rate. Thus, s real devaluation of roughly 50 percent would be required just to restore the rate to its average level of 1982-87. Account must also be taken, in advance, of expected changes in local prices and costs. The initial rate should be high enough to let the real rate still be close to its target level after a period of several months. The increases in prices and costs may still be significant for the first 2-3 months after the introduction of stabilization. in view of these considerations, a nominal devaluation (increase in the official rate) of very substantial proportions would be required for the fixed-exchange rate policy to be sustained, and for it to produce the expected results. If a devaluation of this size cannot be undertaken, for one reason or another, the option of a fixed exchange rate should not be entertained. 2.40 An alternative which combines elements of both exchange-rate regimes--incorporating probably more of each one's virtues than of its deficiencies--would be a system of a managed floac. This regime would also require a substantial nominal de-aluation as already discussed in paragraph 2.40. The guideline for management of this rate would be the prevention of violent fluctuations of the exchange rate, particularly in the first few days and weeks following the introduction of the stabilization plans, while changes that are more obviously due to an existing rate being out of equilibrium should be allowed, Like a fixed-rate regime, this system would remove the risk of violent exchange rate fluctuations introducing an element of instability into expectations and lowering the plan's credibility, though the fixed-rate regime can be more efficient in this - 44 - direction (provided the initial devaluation is, and is seen to be, of a sufficient magnitude). Like the floating rate regime, on the other hand, the managed float would lower the risk of a sub-equilibriun exchange rate being maintained for an excessively long time. The major constraint on the use of suich a scheme is the ability of the administrating machinery to make the right distinctions and the right decisions at the right time, to lead to close adherence of the policy's guidelines. This should be considered in the general context of the administration of a stabilization policy, which will be discussed shortly. Incomes PolicY * 2.41 Under some circumstances, not relevent for today, a good case may be made for a 'heterodox" element of an incomes policy which would be, basically, a temporary freeze of wages and prices. This would accompany the 'orthodox' fiscal and m-nonetary policies. Essentially, the argument for a heterodox element is that given the change in the "fundamentals'--money and the budget--inflation is no longer inherer.t. But economic agents do - not realize that right away. Their push to increase input and output prices would lead to one of two alternatives: maintenance of the new fiscal and monetary rules at the expense of heavy unemployment and loss of output; or abandonment of the fundamental principles of the stabilizatiorn Ai policy, leading to its failure. 2.42 This argument is important. Had Brazil been contemplating a stabilization policy for the first time, that could have been convincing enough to support the incorporation of a wage and price freeze as part of the stabilization package. But given the history of the different stabilization attempts over the last four years, such a policy would row be counter-productive. 2.43 The three stabilization plans of 1986-89--the Cruzado, Bresser, and Summer Plans--were all primarily based on the heterodox element: a price freeze and (less universally) a wage freeze. These plans have all failed--not because they involved this element, but because other essential ingredients--particularly, a strong fiscal element--were missing. Nevertheless, the public by now associates a price freeze with an ill- conceived plan, doomed to fail, an attitude that was already present when the Summer Plan was introduced. In terms of the program's credibility, and of changing inflationary expectations, a wage-price freeze would most likely be a strong negative element. It is, therefore, suggested that such freeze should not form a component of the stabilization policy. The cost will be some added recession which, while being very undesirable, has to be anticipated and acknowledged. 2.44 However, some wage policy should be practiced at the outset in the public sector. In view of the relatively high level--and large increase in recent years--of public-sector salaries at the higher range, the real level of salaries at this range should be lowered with stabilization. This could best be achieved, if legally feasible, by the graduated removal of indexation, for a limited period (without compensation for the loss when -I - 45 - full indexation is reintroduced). Thus, the higher the salary, the lower should be the rate of indexation--with, at the margin, no indexation at all for very high salaries. 2.45 Finally, it is definitely not recommended that public-sector prices he frozen (or be adjusted by less than the rise of the price level); the loss that this would cause, by increasing the budgetary deficit, would more than outweigh its moderating effect on increases in the price index. F. The Operation and Impact of the Stabilization Policy 2.46 When the fundamental elements are corrected--as they would, if the policy agenda suggested here is followed--inflation will be reduced substantially. The speed and timing pattern of this event cannot, however, be predicted with any certainty. Experiences in other countries, in which a sustained stabilization policy was introduced following high or hyperinflation, indicate that inflation is reduced dramatically, within a short time, not to return. Even in unsustained stabilization programs, such as the Cruzado Plan in Brazil or the mid-1989 program in Argentina, price increases were halted immediately when the plans were introduced. 2.47 This will most likely be the case in Brazil, when a stabilization policy like the one recommended here is introduced. The major determinant of a probability of such an event, and of the speed of deceleration of inflation, will be the extent and magnitude of the fiscal and, even more importantly, the monetary shock. With a major contraction of money, as is proposed here, large-scale draw-down of inventories would be necessary to secure cash. This is bound to push prices down immediately. Also, the expected large-scale inflow of short-term capital and rapid build-up of foreign-exchange reserves would greatly reinforce an immediate perception that the program is sustainable and credible. That would immediately change inflationary expectations, and consequently actual patterns of price setting. 2.48 If by the time stabilization is introduced the economy has already been undergoing hyper-inflation, in which normal economic activity cannot be carried out, any short-term yardstick must show an improvement following stabilization. But in comparison with some "normal" state of affairs, the short-term economic performance in Brazil--say, in the first year- -following a stabilization policy along the lines suggested here, is likely to be as follows: (a) Some unemployment is bound to develop. The reduction in the government's labor force--even if modest--will be combined with layoffs by firms facing entirely new circumstances. The draw-down of inventories, following the liquidity squeeze and the high real interest rate, would lower aggregate demand as would the stabilization program's impact on other forms of investment and on consumer spending. Increased export activity would cffset these tendencies, but only partly. Business failures might be common, despite the reputed high liquidity of Brazilian firms. Also, currently many firms must be losing money on their real activities, and the - 46 - losses are being covered (or, probably, have even been inadverI-it)1y concealed) by profits on financial assets. With the disappearzon.& ot the latter, the losing real activities are bound to be contracted. An: im( rvase of real wages, following the drastic deceleration of inflation (d.t,d when wages are adjusted to past inflation) may also contrihute to urnu1loymrint (b) Production may also fall, but not as much as empl,yment. The short-term reduction of employment will be partly offset by imrmt=diate increases in efficiency and productivity, which universal experienm e shows to follow the implementation of stabilization and, in particular, of trade liberalization. At the same time, it must be realized that the expected longer-term acceleration of g7owth, following stabilization and struutura] reforms, would not materialize within the first year; perhaps not even within the second. (c) Both exports and imports should increase: the former due to an increased real exchange rate and slack domestic demand, and the la.ter due to the lowering of imports barriers, which should (by design) more than offset the effect of low domestic aggregate demand on the demand for imports. The effect on exports should be large and immediate: Brazil's exports have in the past responded fast and strongly to changes in both p rices and aggregate domestic demand.16 Short-term capital inflows are also expected to be substantial--perhaps massive--with the introduction of the monetary squeeze. With a regime of a fixed nominal exchange rate thib would be reinforced by expectations of a real currency appreciation. Under such regime, or a managed float, a large-scale accumulation of foreign- exchange reserves may hence be expected. 16/ The recorded increase in exports may appear to be dramatic, since much of the short-term capital inflow may be counted as a change of exports: an under-invoicing of exports, which by all accounts is prevalent at present (as it was on other occasions in Brazil, when the official exchange rate was low and disparity between it and the parallel rate was high), will turn into over-invoicing following the assumed large devaluation and liquidity squeeze. J Table 2.2: BAI - KFY E-CONOMI4C INDICAIORS I----------------Act.al- --------------I Prale- Eat. I.PrOj.t.a--- 1980 1961 1982 198.3 1984 1985 1986 1987 1988 1969 1990 1991 1992 1993 1Q94 CMP Cro.th Rate (1) 9 3 -4 4 0.6 -3 5 5 I 8 3 7 6 3 6 -0 3 3 0 0 0 2 0 6 0 a 0 8 0 CDP /cap.ta CroetS Rate (1) 6.9 -6 6 -1.7 -S 7 2 a 6 0 5 4 0 a -2 4 0 9 2? 1 -o 1 3 8 5 5 6 con".uption/capita CroetS Rate (6) 3.1 -6.4 1.5 -4.3 -1 5 4.6 7 4 0.9 2 0 0 6 -3 5 -' 1 6 4 1 5 6 Tot.aI DOD (in U1.8. mulliso) 2/ 64.244 73.963 85,364 93.566 102.039 105,126 111,045 121,264 114.041 115.302 118.998 122.341 126.776 131.435 135 486 (of oh ico, IWF) - - - 560 2,644 4.186 4.619 4,490 3,938 3.3-57 2.519 1.764 1.764 1.764 1.764 1.764 D00DXCS 2/ 2 76 3.49 3.64 3.84 3.38 3 58 4 42 4.22 3 13 3 12 3 00 2 62 2 32 2 05 180c DOD/CDP 2/ 0 268 0.356 0.314 0 453 0 483 0 460 0.396 0 380 0 315 0 310 0 305 0293 0 273 0 260 0221 Debt sar-ics (;n US58miIc. 2/ 14.146 17.801 20,766 18.409 19,995 22.319 24.243 25,274 16.421 17.226 19.610 18,293 18.426 16.452 20.994 Debt S*rn.ce/XOS 0.61 0 66 0.88 0 76 0.66 0 76 0.96 088s 0.45 0 47 0 49 0 39 0 34 o 2. 0 28 Debt Sor-sce/COP 0 0.59 0 068 0 076 0.069 0 095 0 098 0 067 0 079 0 045 0 046 0 050 0 044 0 040 0 035 0 035 Inter.mt/X0S 2/ 0 320 0.383 0.538 0 422 0.379 0.384 0 414 0335 0284 0300 0295 0244 0212 0287 0159 Intorest/CWP 2/ 0 031 0.039 0 047 0.050 0.055 0 060 0 038 0 030 0 029 0 030 0 030 0 027 0C025 0 023 0 020 C,o.. I-n.etm*nt/O0P 3/ 0.233 0 230 0.210 0.165 0.162 0.170 0 192 0 222 0 170 02180 0 199 0 232 0 253 0 268 0 271 Domost.c San.ng./a0P 3/ 0.211 0.227 0.204 0.189 0 203 0.221 0.217 0 254 0 221 0 223 0 235 0 265 0 281 0 292 0 293 Hational Sac.;ng/(fDP 3/ 0 179 0 186 0.153 0 132 0.148 0 169 0.173 0 218 0 179 0 190 0 194 0 230 0 249 0 26? C 267 lhrginal National Sa.ingo Rat* 0 159 0 23 0.620 -3 593 0 647 0.917 0 231 1 078 4.754 0 242 6 116 1 658 0 552 0 433 0 316 F-eod In,eatusnt/Wl 0 229 0 228 0.214 0.179 0 162 0.170 0.192 0 222 0 170 0.180 0 19. 0 232 0 253 0 268 0 271 ICfR 2.460 -7.168 18.567 -6.592 2.812 1.877 2.015 6 231 -74 159 5 600 0 000 9950 3 851 3 176 3 357 Export Volume Growthk Rato (1) 4/ 22.6 21.3 -9.2 14 3 22.1 -2.0 -23 6 14 5 17.6 2 3 5 4 14 6 13 0 12 6 12 9 Export Value Growth Rate (1) 3/ 4/ 30 8 16 7 -13 9 7.5 22.6 -4.2 -12.6 17.7 26 9 1 1 6 9 18 1 17 5 17 6 17 4 EpaortefQlP 3/ 4/ 0 089 0 094 0 079 0 112 0 135 0 120 0.085 0 095 0.109 0106 0114 0 128 0 137 01243 0 149 Igmport Volume Croeths Rato. (6) 5/ 0.6 -12.3 -6 0 -17 4 -3.0 "1.1 Is50 -3.6 -0 a 1s a 20 6 23 3 21 4 16 6 16 5 Import Valu* Cromtls Rat. (5) 3/ 5/ 27.9 -2 1 -8 9 -21.1 -9.9 -3.8 4 5 7.1 -1.4 21 5 22 2 28 6 26 2 22 2 19 9 Import*/GDP 3/ S/ 0111I 0098 0 066 0.089 0 079 0.069 0.061 0055s 0.058 0065s 0 079 0 095 0 109 0 119 0 127 Iaport Elasticity 5/ 0.07 3.73 -6.67 6 96 -0 53 0.37 1.61 -1 32 2 57 5 21 11 66 3 55 2 34 2 07 Trad* Salanco (in 11S8 Millions) (,.'.82) 1.202 780 6.470 13.089 12.486 8.305 211.167 19,064 16,002 14.422 14.747 14.925 15.659 16.977 Current Account (in U1.6 million.) (12.807) (11.735) (16.310) (6,837) 45 (241) (5,304) (1.428) 4,448 374 (1.777) (1.342) (2.219) (2.624) (1.970) Resource Balance/CDP 3/ -0 022 -0 004 -0.007 0.024 0 056 0.050 0.024 0.029 0 050 0 040 0 032 0 029 0 024 0 022 0 020 Current Account/GDP 3/ 6/ -0 053 -0 044 -0 061 -0 033 0.000 -0.001 -0.017 -0 004 0 012 0.001 -0 005 -0 003 -0 005 -0 005 -0 003 Term ofTrade Index (1987s100) 103 3 91.3 88 2 81.5 87 8 86.2 109 0 1000 110 5 102 4 103 0 101 9 102 0 102 8 103 3 (1) Hiatorical figures based on 180E nat,onal ac-out, data. April 1989. (2) Includes IW (3) OroetS rate, and ratio. calculated using current U116 nalues. (4) Faporte of goods and non-factor ser-ices (5) lmporta of goods and non-factor sor-,ces (6) Projected values se.mum au;ntenance of the 1987 real oxcltange rate - 48 - 2.49 Table 2.2 represents projections of short- and medium-term (1990-94) performance of Brazil's economy, following the suggested implementation of a stabilization plan, along the lines proposed here, by the end of the first quarter of 1990. The projected development reflects the course that has just been described. Its most important feature is the resumption, within several years, of Brazil's traditional rate of growth--a tradition that has been grossly disrupted by the turbulence of the 1980s. This should be contrasted with the alternative of not just an absence of growth but of a major collapse of the economic mechanism which would follow the continuation of present policies and the avoidance of structural reforms. In the short run, some sacrifices will have to be borne; but even these are "sacrifices" only in reference to a normal period, and not to the chaotic outcome of present policies. The projected values of the major real variables are: (a) The economy's growth would be nil in the first year (1990), and very modest in the second (1991); the growth rate picks up a year later, and then assumes a longer-term annual level as high as 8 percent- -reflecting the effects of the projected structural reforms. (b) The economy's saving rate, and with it domestic investment, increase substantially (each by about 7 percent of GDP). This is a reflection of the projected substantial increase of the government's saving (i.e., the change in the operational deficit). (c) In the longer term, per capita consumption should also increase at a much higher rate than during most of the 1980s (in which it barely changed); but at the beginning (1990 and 1991), when the economy does not yet grow, per capita consumption must fall: disposable income would decline following the large increase in net taxation without much of an increase of gross income. (d) Finally, both exports and imports are expected to grow rapidly. Imports should increase faster than exports; but the absolute size of the trade surplus should not differ much from its level during 1988 and 1989. - 49 - C. Provision of a Social Safety Net 2.50 The fiscal adjustments and other stabilizatior elem.,enits proposed in this program can be expected to have beneficial eftecis on Brazilian income distribution.17 The major subsidies reconumiended for elimination predominantly benefit upper income groups. The share (f real wages in income is not expected to fall, except for wages of highly-paid public sector employees, whose income levels are far above the economy's average. Unemployment will probably be mostly confined to labor in the urban formal sector, where average wages are relatively high. Finally, the inflation tax on the holding of money, which is imposed disproportionately on low- income households, will disappear, and a higher share of fiscal revenues will be generated by taxes which are progressive, such as the corporate income tax. 2.51 In short, these changes can be expected to improve the aggregate welfare of low-income groups in Brazil. However, it is also likely that among the poor certain groups will suffer in the short run. Reduced unemployment in the short run could affect low-wage workers in the formal sector disproportionately; in 1985 almost 20 percent of all poor worked in the formal sector. An overall reduction in demand is likely to reduce incomes of informal sector workers as well. The effects of stabilization on the poor will also depend upon the speed and comprehensiveness with which the program is implemented, and the rate at which the economy responds and growth resumes. 2.52 Given these uncertainties, an important element of the stabilization program should be an explicit "safety net" of social assistance programs to protect particularly vulnerable groups (such as infants, children, female-headed households, and the elderly) as well as individuals experiencing short-term dislocatiLns under the program. An effective social safety net can also help ensure the political sustainability of the stabilization program over the first few difficult months. 2.53 We do not recommend immediate institution of new employment or retraining programs; these tend to be ccstly, do not generally reach the 17, A study of the recent experience in the Philippines indicates the following: "'Good' policies also have a distributional payoff. Reducing inflationary pressures, avoiding real exchange rate over- evaluation, and attaining positive real interest rates all have a desirable incidence, while undiscriminating expansionary fiscal policies, with no attention to public expenditure composition, will probably result in a higher skewness of the distributional curve." Mario I. Blejer and Isabel Guerrero, "The Impact of Macroeconomic Policies on Income Distribution: An Empirical Study of the Philippines," Review of Economics and Statistics, 1990 (forthcoming). The Israeli stabilization of 1985 is also judged to have contributed to equalization of income distribution, primarily through its impact on tax collections. - 50 - truly poor, and are difficult to terminate once instituted. We recoumend instead car2ful evaluatior, of existing unemployment programs to ensure they reach low-wage workers that are laid off, and er-phasis in the short run on income transfers to the very poor through expansior. of targeted nutrition programs. 2.54 Basic principles for the design of a social safety net are: (a) to make sure that assistance is available quickly, existing programs should be adjusted rather new ones designed; (b) spending on those existing programs that are most effectively targeted to needy groups should expand; (c) spending on non-targeted programs, such as the wheat subsidy or the PAT (Programa Alimentar do Trabalhador), should contract; and (d) the impact of the economic stabilization program on the poor in general and on target groups in particular should be monitored. 2.55 The basic elements, in turn, of a social safety net in Brazil would be: (a) Expansion of nutrition programs targeted to young children and pregnant and nursing mothers, such as the PSA (Programa de Suplementacao Alimentar) and LBA programs of this type. (b) Protection of expenditures on basic health services, such as vaccination programs and Maternal and Child Health services, and actions to ensure that budget cuts do not jeopardize spending on non-salary recurrent - expenditures, such as basic drugs and other medical supplies. (c) Protection of expenditures on day care and pre-school programs serving low-income neighborhoods, to ensure that mothers who are the sole source of family support can continue to work. (d) Establishment of an emergency system to monitor the access of target groups to the "safety net" programs and to track changes in their nutritional and health status. Results of the 1989 National Health and Nutrition Survey, which are expected to be available shortly, should be helpful in this regard. 2.56 A social safety net of this type can and should be implemented rapidly at the outset of the stabilization program. The overall structure of the safety net need not change over the medium term. Spending on certain purposes, however, is likely to be counter-cyclical. In particular, the demand for spending on nutrition is greater during recession. Depending on the rate at which economic gtowth resumes, there may be a need to supplement the safety net with an emergency social fund, similar to funds established in other Latin American countries implementing economic adjustment programs. Such funds are generally designed to support income-generating activities and small-scale social projects developed by local governments, community groups, and other non-governmental organizations, in low-income neighborhoods that may not benefit immediately from improved aggregate economic gains during the stabilization/adjustment -~~~~~~~~~~~~~~~~~~~~~~~ - 5 1 - period. The need for implementation of such additional measures should be evaluated by close monitoring. This would be part of the function of the administration organ to which we now turn. H. The Administration of the Stabilization Policy 2.57 In designing and implementing a stabilization policy, many uncertainties are involved. Both directions and magnitudes of changes and of policy impacts must be guessed--and may easily be guessed wrong. Many problems and issues are certainly not foreseen, and would not appear until the program is actually implemented. Some such issues may not be of crucial importance in any fundamental sense; but, given the overwhelming weight which must be assigned to the credibility of the program, they may nevertheless undermine even the best-conceived plan if not handled appropriately. Similarly, for the same reasons that the collection of measures which form a stabilization policy must be considered as a package, changes in such measures during the implementation must be viewed in toto and be well coordinated, 2.58 All of these lead to a clear-cut recommendation, namely, that an administrative machinery should be prepared beforehand to monitor the implementation of the policy, and implement fast and well-coordinated changes and adjustments as the need arises. This machinery should consist of a body of policymakers representing the major governmental units which must share in the conduct of policy. It should also embody the knowledge (legal, organizational, etc.) which must be involved when decisions of policy principles are translated into concrete actions. For a specified period of time, this policy-coordinating body should be granted the authority to make binding policy decisions, insofar as such decisions are within the domain of the administration (i.e., do not require congressional approval), without need to refer to any further governmental deliberations. 52 - 2. An Agenda of Medium- and Long-Term Reforms 2.59 This section specifies the set of longer-term, tundamental reforms which are essential for achieving greater efficiency, the rekindling of growth, and a more equitable distribution of the country's income and wealth. It will be assumed, in this discussion, that the short-term agenda of policies has been fulfilled: that the appropriate first-step, quick- implementation measures are in place, and that macro-economic stability has been achieved. 2.60 These policy reforms will be classified into the following areas: the budget and central government activities; public enterprises; the foreign-trade regime; agriculture; the financial sector; and social expenditures. But this classification should not be viewed as rigid; many policies obviously belong to more than one category. Likewise, a substantial degree of interdependence exists between policies in different areas. Perhaps in no case is this interdependence so strong as to make the implementation of one set of policies a prior condition for the introduction of another, but on various occasions, policies will be mutually supporting. To illustrate: privatization of public enterprises depends, to a large extent, on reforms in the financial sector; and recommended policies in agriculture, or privatization, would make better sense if the foreign-trade regime was reformed. A. The Budget and the Government's Labor Force The Government's Real Expenditures 2.61 Broadly categorized, the government's expenditures consist of spending on its labor force, purchases from suppliers, subsidies, and other transfers--in Brazil's case, primarily to local governments. The two latter categories are presumably handled as part of the short-term policy package, and purchases from suppliers have drastically dwindled with the decline in government investment. Long-term policy on government expenditures will thus primarily address the part of real expenditures which involves the government's labor force. This policy will have to handle the (inter-related) issues of size, structure, and the remuneration of government employees. 2.62 it is generally accepted that the present level of government services may be provided with a substantially reduced public-cector labor force. That reduction would have to be carried out gradually, and would be achieved over a number of years. Time is required both for a thorough evaluation of organizational structures, functions, and labor requirements, and for allowing natural attrition, rather than large-scale firing, to be the primary mechanism in the reduction process. 2.63 This labor-force reduction should be complemented by a thorough reform of the compensation sys.em in the public sector to improve the mix of employees, increase transparency, and raise the sector's productivity. A set of principles should be formulated to simplify the compensation structure. These principles should include the definition of criteria for -4 - 53 - jo!) grading, t he re lat ioroshi p of job grading to c omlen"sat ion antd t aboI- mctrket t end.l , incent ives for performance, and traitning and cart-er velkPIme lt .Given the principles, new salary scales should be designed, itn which the number of scales is drastically reduced and most supplements are eliminated. Once the simplified structure has been established, a program of p)ha,ed salary adjustments can be designed and implemented. TaKes 2.64 In indirect taxes, the VATS should be thoroughly reformed. At present, ht,th the IPI and the ICM ex lude services, which constitute about halt of the national product. The t4o separate taxes also overla), and lead to double taxation of the industrial sector. The two taxes should be incorporated into a single VAT, covering all sectors of Brazil's economy. By extending the coverage, the single tax rate could be lower than the average VAT rates imposed today on the sectors which are at present subject to both the IPI and the ICM, yet it would lead to larger total revenues. These revenues could then be allocated among the three levels of government (federal, state, and municipal) in a manner which would prevent a tax loss to any of the three. This proposed change would require negotiations with the states, and Constitutional revisions. 2.65 Among direct taxes, the principal sources of revenue today are the payroll taxes (or "social contributions"). These taxes should be redefined and greatly simplified. The tax base should be broadened, to cover all public-sector employees, including the estutarios (the tenured employees). While this coverage may not increase the public sector's net revenue, it would reallocate net revenue from local to the Federal government. Since this is, in effect, the major income tax, a higher degree of progressivity should be introduced into it. Tax rates should also be uniform over the various activities. In the long run these rates, very high at present, should be lowered. Local Governments 2.66 This branch of the public sector has grown most over the decade. The increase of its real expenditures, and in particular its labor force, has exceeded any increase in the services that the sector provides. A thoroughi evaluation and reform, of both the expenditures of state and municipal governments and their manner of finance, must have a high yield.13 The Social Security System 2.67 Entitlement to benefits from this system should be reduced. Thus, legislature should lower "replacement rates," and in particular eliminate the exceptionally generous schemes for specific sectors or professions. Also, a closer relationship should be established between entitlement and coverage period, through measures such as the reduction of disproportionate early retirement benefits. 2.68 On the other hand, the effective benefits should be those that have been legislated. In particular, indexation of benefits should be as 13/ The World Bank has recently undertaken a study of state finances, which should be concluded with a detailed agenda of reforms in this sector. - 54 - complete as is techniically possible: There is no justification for imposing a capricious inflation tax, specifically on the recipients of social security benefits. rhis issue should become less severe following stabilization, but it will remain of son.e concern even in the long run. 2.69 The system should also shift to a scheme in which benefits are clearly related to contributions, so that payroll deductions contain less of a tax element and become more payments for deferred benefits--a change of far-reaching consequences in its efficiency implications. Even beyond that, a shift should be contemplated in the pension provision function from the government's social security system to mandatory participatior. in government-monitored private pension funds--along the lines recently introduced in several Latin American countries, such as Chile, Costa Rica, or Uruguay. An interim measure might be a reduction of the scope of the program. This would be done by lowering the replacement rate, and setting ceilings on benefits and contributions. This would encourage higher-income workers to supplement their pensions by using private schemes; while at the same time provide an adequate safety net for low-income workers. B. Privatization of Public Enterprises 2.70 The mechanism for privatization, established in 1988, is adequate. The technical organization which handles the process is of a high quality, and the guidelines and procedures under which it operates are by and large appropriate. With this mechanism, a comprehensive privatization policy should be feasible. The recommendations are: (a) The process of privatization should greatly accelerate. The main existing constraint is not the absorptive capacity of the financial * market, particularly if the recommendations below concerning sources of A finance are followed. The market would not be disrupted by annual offerings of equities valued at several billion dollars. The capacity of the privatization agency to prepare the individual er.terprises for privatization falls, however, short of such amounts, and thus forms today the effective constraints. Such preparations may include legal and other requirements surrounding the offer of equities, deregulation of prices, or restructuring of enterprises before it is offered to the market. While bringing many enterprises to the market may strain the administrative capacity of the agency, and the absorptive capacity of the market, privatizing companies too slowly would question the government's .:ommitment to the program, and would postpone the gains from privatization. The optimum speed cannot be decided beforehand; but it is, undoubtedly, much faster than the practice so far. (b) A "negative list, should be prepared, of enterprises that are not subject to privatization. Presumably, this should include only instances in which the Constitution prescribes public ownership. All enterprises not on the list would be candidates for privatization, and some work to begin the process should start quite early in each. (c) Few broad rules may be offered about the sequencing of privatization. Logical first candidates are (and have actually been) enterprises which are former private companies. Next would come activities which typically belong, in industrial countries, to the private sector, - 55 - such as steel, fertilizers, pulp and paper, or warehousing. Within this classification, it is important to select at the very early stages several large-scale enterprises, which draw attention and whose privatization would signal the government's serious intentions. These should be mostly enterprises which have been doing well, rather than obvious failures. (d) Privatization should be 'true' and full fledged. In several recent instances, the govermnent has still intended to maintain majority control of the enterprises. This would be inappropriate for three reasons. First, it misses the main objects of privatization; second, it sends the wrong signal. indicating governmental equivocation about liberalization; and third, it leads to lower prices for the shares sold in the market. (e) In addition to financing privatization through sales of equities in the stock market as it exists today, three additional sources should be developed and promoted. One is foreign financing, in particular through swaps of external debt and shares of the privatized enterprises. Another is basically a similar swap of short-term domestic government debt, by partially converting it into equity. Third, equity acquisition by employees of the privatized enterprises should be encouraged, through discounts and the establishment of credit facilities. (f) An essential part of the privatization process must, on many occasions, be the removal of regulations and price controls: without this, privatization may make little sense, and may not be feasible. At the time of privatization, the potential purchasers should be reasonably confident that the privatized firm faces a stable policy environment, and output and input prices which reflect basic market factors rather than government interference. Industrial deregulation, retormulation of industrial policy, reform of trade policy, and removal of price controls should thus be viewed as necessary accompaniments to privatization. When existing monopolies are privatized and price controls lifted, the government should ensure beforehand contestability in the market--first and most, by allowing free trade in the product. (g) When the privatization process starts, it would be important to install in public enterprises managers who are committed to privatization in general, and of their own enterprises in particular. In the absence of such commitments, the process would be stalled and be carried out in a less-than-optimal fashion. (h) Close attention should be paid to the interests of employees. Those who remain employed should, as suggested above, participate in the fate of the enterprise by acquiring shares on concessionary terms. Some workers would be displaced, as part of the efficiency-enhancing process of privatization. Natural attrition would lead in the short-term to some reduction in the labor force, and early retirement schemes would also help. For those displaced involuntarily, safety-net schemes should be prepared. These may include severance packages which would consist, inter alia, of allotments of company shares, job search allowances, and job placement policies. Losses for some of the existing labor force are inevitable, but the privatization process would be greatly facilitated if, as experience proves, all efforts are made to minimize them. - 5b - (i) Finrdl v, tare should ie taken to ensure that gover.ment receipt s f ron. the sale ci puhliic enterprises ate nut perceived as const itut ing partt of *j,,vt!rnment revenue. For the pur(haser of erquit y in the pt ivat izd fir>, this is a substitution of assets, and not a reduct ion in his wealth (as a tax would be). Specitically, government receipts from this source shoald rct ic used to increase its current expenditures or its investment, arti u±arly in other public enterprises. Instead, all such receipts sh-:ould Le used to retire government debt (with debt-equity swaps, of both ext ernal and riternal debt., this would of course ha ppen automatically). It- this way, a reversal of the 'crowding out process will take place--leading, indeed, to thie growth of private investmenit and the private sector. C. The Foreign-Trade Regime 2.71 The opening of Brazil's market--at present one of the world's most insulated--to trade with the world market would be the single most important means of increasing both allocative and technical efficiency, ending monopolistic market structures, and realizing Brazil's huge potential for growth. 2.72 The first stage of the trade-regime reform, recommended as part of the immediate-action package, is the removal of QRs and their replacement by tariffs or tariff-like instruments. This change in the form of protection is of the utmost importance. But beyond it, the discrimination among activities should be removed. A substantial body of international evidence ana experience demonstrates that wholesale discrimination is costly--and nothing in Brazil's experience would tend to contradict this finding. Once non-tariff barriers have been removed, the Brazilian government should proceed to gradually lower the level and dispersion of the tariff system--a measure begun, in fact, in 1988. The progress of implementation should be steady and predictable, and the process should follow the "concertina method;" i.e. lowering, at each stage, the highest tariff rates in the system. 2.73 Several general guidelines for protection, or intervention, should be adopted for the end of this process. First, any market failure which generates a call for infant-industry support should be tackled at the source. This may mean reform of domestic capital markets, improved training opportunities or incentives, or better protection of the intellectual property rights of Brazilian firms. Second, the mechanism of granting infant-industry support should be as transparent as possible, so that the benefits to the economy of such protection should be clear and that this argument is not used for a wholesale provision of protection. Third, infant-industry protection should be limited to very few activities rather than being applied to a broad range of industries; obviously, the more widespread is such protection, the less effective it is in shifting resources to the targeted activities. Brazil's large and diversified industrial structure would suggest that an infant-industry argument might be valid only on very rare occasions. Finally, whenever it is granted, infant-industry support should have a finite (and not too distant) horizon. The more it turns into a permanent privilege, the more likely it is for an unbridgeable gap to open between local and world prices. NJ - 5 7 - 2.74 Beyond such exceptional cases of infant-industry protection, a uniform, low tariff (within the range of 15-20 percent) should in the end be applied to all imports (of final goods and intermediate inputs). At this level, a substantial amount of tatiff revenue can be generated, but the protective, distortionary impact of the tariff would only be moderate. 2.75 The process of tariff reduction would need to he offset by a real currency devaluation (in other words, the tariff component of the effective exchange rate of imports would be replaced by the component of the formal, official exchange rate). This should start from the level to which the real exchange rate is raised by the devaluation envisioned in the short-run policy package. International experience suggests that without such an accompanying exchange-rate policy, trade liberalization is not likely to be sustained. D. The Agricultural Sector 2.76 The most important steps in reforming the government's agricultural policy would be undertaken as part of the short-term policy agenda. They would include changes in taxation, fiscal subsidies, and credit subsidies. Several significant reform measures should, however, be undertaken as part of the longer-term package. 2.77 One is the elimination of price-stabilization programs. While one could conceive a perfectly-managed program, experience in Brazil (as well as elsewhere) shows that these schemes have been costly and wasteful. Consequently, the minimum-price program (MPP), as well as the stock- purchase program (AGF), should be eliminated. The reform should start by reducing the number of crops covered and by formulating prescribed rules, which directly link the established prices to international prices. Complete abolition of the programs should then follow. It should be noted, though, that the longer a full reform is delayed and the existing system continues, the more difficult it will become to make the necessary adjustments in the future. 2.78 Public storage operations should also be gradually discontinued. Instead of providing storage, the Federal and state governments should undertake an oversight function, improving licensing and inspection practices and procedures. 2.79 In wheat and sugar, marketing should be privatized. In sugar this has started, by removing the monopoly status of IAA in the export market. The IAA should, however, be abolished altogether, and the administered price and quota system which now regulates the sugar industry should be dismantled. 2.80 In trade policy, agriculture should be liberalized in the same way as other activities. However, a concern specific to this sector is the intermittent controls of agricultural exports, through quota or outright bans. These are designed to ensure that demand at home (at some low price levels) is fully matched by domestic supply. It is recommended that this policy be discontinued. - 58 - 2.81 In its real expenditures on agriculture, the government should withdraw from operations which compete with, or replace, the private sector. It should instead focus on providing support services and investments in infrastructure which complement and facilitate the private sector activity. Particular attention should be paid to areas of research and extension services, inspection, grading, market information, job training, or general education; and public infrastructure, such as conservation works, off-farm irrigation facilities, telecommunications, power, or road transport. E. The Financial Sector 2.82 A variety of steps should be undertaken to introduce a larger degree of freedom and competition ir the financial market, and to lower the government's dominant role in it.14 2.83 A major step would be the abolition of regulations requiring mandatory investments of pension funds, insurance companies, and short-term mutual funds in government securities, except when there is clear evidence that some such requirement is needed under reasonable and prudential diversification guidelines. 2.84 State banks should be closed, or privatized. There is no relevant banking function that private banks cannot provide at least as efficiently as the state banks. While state banks still operate, they should be closely audited, and prudential regulations should be applied to them effectively. 2.85 Exit from the banking system, when warranted, should be facilitated as a means of increasing competition and efficiency. Adequate mechanisms have to be introduced to facilitate exit, minimizing its negative side effects. One possible step is the establishment of a limited deposit-insurance scheme, which would repay small depositors when a financial institution liquidates or restructures. 2.86 Conversely, the banks also need to be better protected from their clients' failures. Bankruptcy protection procedures (the "concordata") should be adapted to recognize the effect of inflation on borrowers' * liabilities. At present, the liabilities of a business filing for bankruptcy protection are frozen in nominal terms. This is, obviously, an easy way for borrowers to reduce indebtedness at the expense of financial intermediaries. At current rates of inflation in Brazil, such indebtedness is almost wiped out within a few months, but even with mild inflation, this procedure provides an incentive for bankruptcy, weakens payments habits, and can potentially affect the solvency of the creditor institution. 2.87 Fees on banking services should be deregulated (especially if and when banks are authorized to pay interest on demand deposits). At present, the Central Bank most often sets very low fees, leading to increased demand for various banking services--especially current accounts. This results, 14/ A study of Brazil's capital market is now being undertaken by -he World Bank. It should conclude with a detailed agenda of reforms in this market. -A d - 59 - in the eind, in higher operating costs, which are passed on to borrowers in the form, i h gher lending interest rates. 2.88 All temi reulated credit programs (primarily out of budget) should be gradually phased out, on a prepared and announced timetable (in various instances, this process has already started). Earmarking of taxes for lk,ng-term investment through Federal financial institutions should be eliminated, in phases. As a first step, the transparency of these programs should be enhanced, and interest rates should be so determined as to move the price of long-term credit toward what would be a market price. Beyond its allocative effect, this would also reduce the prospects of future structural government deficits. The compulsory applications of demand deposits held by commercial banks for directed, subsidized credit, as well as the compulsory application of required reserves, should be phased out. F. Social-Sector Policies 2.89 The two major social sectors which we address are education and health. Several principal themes are common to the specific recommendations for policy reforms in the two sectors: 2.90 (a) Targeting of public expenditures should concentrate spending on high-priority programs. The targeting could be improved by reducing the growth--actually, cutting down the present size--of programs that subsidize middle- and higher-income classes; and expanding public expenditures on those programs which are most likely to benefit the poor. Generating up- to-date information on liv'ng standards of groups across the country would contribute to the proper design and monitoring of the programs. 2.91 (b) Private-sector activit- should be encouraged. The effectiveness of social spending would increase if recipients of social services could choose among competing providers of services, with the government restricting itself to the provision of evaluative information. The private sector could then fully charge the middle- and upper-classes for services, with the government subsidizing access for the poor. 2.92 (c) Equity of financing should be enhanced. Reliance on payroll taxes should be reduced, and the earmarking of tax revenues for particular types of expenditures or individual agencies eliminated. Introducing transparent processes for the allocation of resources, and increasing the - accountability (hence efficiency) of the governmental units would also be needed. 2.93 The specific reform measures in the two sectors concerned should be as follows: Education 2.94 Public education systems throughout Brazil today are large, centralized, bloated bureaucracies. They lack incentives for teachers to perform well within the classroom, and sanctions for those who do not. Both bottom-up and top-down approaches will improve school accountability at the local level (for primary and secondary schooling). The steps required ate decentralization, increasing parental participation, and - 60 - providing parents with objective feedback on school quality. At the national and state levels, the steps are to strengthen education policy analysis, evaluation, and provision of technical assistance to support innovative approaches to improving school quality and efficiency. Specific recommendations include: (a) Improving the allocation of resources across levels of education. The highest public spending priority is primary education, which has the highest social returns and currently suffers from the most intractable problems. Universities could reduce their burden on the Federal treasury by increasing their internal efficiency, (I'smissing non- productive staff, and introducing and gradually increasing student changes to cover some costs. Introduction of changes should be accompanied by strengthened scholarship and loan programs to assure access regardless of income. At the primary level, repetition and drop-out must be reduced; dramatic administrative reforms are a first step in redirecting more real resources to primary classrooms. New approaches to teaching literacy and numeracy, and crash programs of in-source teacher training are needed if poor children are to succeed in school. (b) Improving the efficiency of physical and human resources in public school systems by strengthening management and increasing accountability. At the university level, a priority reform is to grant schools greater budgetary autonomy in exchange for more efficiency and accountability in resources use. At the primary and secondary levels, devolution of responsibility to the municipal level, supported by automatic transfers of federal and state resources to the poorer municipalities, would increase accountability to parents and communities who could then put pressure on local officials to improve quality. Federal and state authorities could reduce their role in direct service provision, and build capacity in planning and evaluation, oversight provision of technical assistance to local school systems seeking to improve performance, and dissemination of information. Local school systems are trying out a diversity of approaches to improving quality and efficiency of primary education; all could benefit from the Federal government systematically collecting and disseminating information about these approaches. By publishing the results of standardized student achievement testing in public and private schools, government would help parents to base school selection decisions on objective information on school quality, and create an incentive for public and private schools to improve quality. (c) Developing coherent and complementary roles for the public and private sectors in the delivery of education services. Brazil can build upon the strength of the existing private sector by focusing public efforts on the enhancement of private school's performance, rather than their displacement by publicly-delivered services. At the pre-school level, the public sector cannot afford to take on full responsibility for providing educational services to all pre-school aged children, and must therefore find innovative ways to develop a cooperative relationship with not-for- profit community groups. At the primary and secondary levels, the presence of a strong private sector can reduce the overall costs of schooling for the public sector, by providing high quality education to families who can afford to pay for it, as well as by providing middle- and lower-income students with a competitive alternative to public schools. The presence of -i - 61 - a competitive private sector in education improves efficiency by giving public school administrators strong incentive to improve quality. The Federal and state governments could maximize these positive effects by playing a stronger role in monitoring and evaluating private school quality and by avoiding counterproductive regulation of private school tuition levels. Health 2.95 The public health systems in Brazil face formidable challenges in coming decades. As the population ages and medical technology continues to develop, there will be large increases in per capita demands for medical services. These demands will place an additional burden on government during a period when public resources are likely to shrink or, at best, not to increase. In order to respond to the challenges, government will need to focus its efforts on high priority actions. Specific recommendations include: (a) Strengthening of controls and implementation of incentives to contain costs and improve quality of publicly-financed personal health services, in public as well as private hospitals, and by private doctors. Brazil finances most hospital care through its public health insurance system. A major challenge in the 1990s will be to contain costs in this system, through greater reliance on both supply-side controls (prospective budgets for hospitals, use of diagnostic-related pricing, overseeing the pricing of high-cost, high-technology equipment) and implementation of as yet underused demand-side incentives (selective user charges for those able to pay). Public policy can also encourage the growth of the non- contractual autonomous private sector that does not rely on public insurance payments, partly by providing to consumers with information on the nature and quality of services, so that they can make informed choices. (b) Priority for public goods (e.g. malaria control and immunizations, from which all individuals benefit) and for cost-effective, largely preventive programs; these generally overlap. In most instances, actions which prevent disease or detect it early (such as smoking prevention campaigns and cancer screening) are more cost effective than actions designed to treat disease. Larger amounts of public resources need to be spent on cost-effective prevention, both to combat the classic t1hreats to public health (such as diarrhea, schistosomiasis, Chagas' disease, dengue, and other endemic communicable diseases) as well as to address the "new" problems posed by AIDS and the increasing incidence of chronic diseases (including cardiovascular diseases, cancers ar.d injuries). (c) Targeting of a higher share of public expenditure to the poor. There is a social imperative (as well as efficiency justification) for greatly expanding the coverage of cost-effective basic health services and nutrition programs to the underserved rural and urban poor. Preliminary results from the Second National Nutrition Survey, carried out in mid-1989, will provide information to map malnutrition patterns in Brazil, identifying needy population groups not currently served by nutrition programs. - 62 - (d) Development of complementary roles across levels of government. The new Constitution and implementing legislation currently before Congress provide a firm basis for progress in the decentralization of authority to the state and local levels, and the establishment of automatic and transparent criteria for the transfer of funds to these levels of government. Implicit in these changes is a major change in the role of the Federal government toward a concentration on setting standards, controlling the quality of pharmaceuticals and foodstuffs, providing consumers with information, and investing in cost-effective public health and disease prevention programs. G. Overall Pattern of Policy Changes 2.96 Common threads to all the policy changes recommended in the agenda of medium- and long-term structural reforms are: removal of distortions introducrd by government policies; increasing the role of the private sector, by letting the government concentrate on what it can do only or what it can do best; and targeting government taxes and, in particular, spending to improve income distribution, reducing the incidence and intensity of poverty, and increasing economic opportunities for all. 2.97 Some minimum modicum of macroeconomic stability is a necessary condition for these changes to be completed, and to have their desired impact. Implementation of the short-term agenda of reforms and stabilization would achieve that target. The introduction, following it, of the proposed longer-term reforms would then ensure the realization of Brazil's enormous potential for growth; and the more universal distribution of the country's richness among its population. 4 -63- PART 2: THE BASIC ISSUES Three issues--two of relatively recent origin--form the core of Brazil's economic problem. One is economic stagnation, in sharp contrast to Brazil's long-term experience of rapid economic expansion. The other new issue is macroeconomic performance: an economy with moderate, roughly stable inflation has become one with a highly volatile, unstable macro- economic environment. The third, long-standing issue is the extent of poverty caused by the lopsided distribution of income. These issues will be discussed, respectively, in the three chapters in this part of the report. -64- III. ECONOMIC STAGNATION A. The Rate of Growth 3.1 Aggregate economic performance is summed up best by the changes in aggregate GDP and in per-capita GDP. With no radical fluctuations in the rate of expansion of population in Brazil, these two representations give roughly the same indication. This may be seen from Table 3.1. Table 3.1: GROWTH PERFORMANCE, 1978-88 Aggregate GDP Per-Capita GDP Rate of Charnge Rate of Change (in b.Cz of (in Cz of of (1) of (2) 1980) 1980) (in percent) (in percent) Year (1) (2) (3 (4) 1970 5.42 56.5 71 6.04 61.5 11.4 8.8 72 6.76 67.2 11.9 9.3 73 7.70 74.8 13.9 11.3 74 8.36 79.0 8.3 5.6 75 8.76 81.1 5.1 2.7 76 9.65 87.3 10.2 7.6 77 10.13 89.5 4.9 2.5 78 10.63 91.7 4.9 2.5 79 11.35 95.7 6.3 4.4 1980 12.40 102.2 9.3 6.8 81 11.85 95.6 -4.4 -6.5 82 11.93 94.1 0.6 -1.6 83 11.52 83.8 -3.5 -5.6 84 12.10 91.3 5.1 2.8 85 13.11 96.7 8.3 5.9 86 14.11 101.6 7.6 5.1 87 14.62 103.1 3.6 1.5 88 14.58 100.6 -0.3 -2.4 89 n.a. n.a. n.a. n.a. n.a. = not available Source: IBGE national accounts data. April 1989. 3.2 During the 1960s, Brazil's GDP grew at an average annual rate of 5.7 percent--not spectacular, but a respectable growth performancq. In the late 1960s and early 1970s growth was phenomenal: between 1970 and 1973 -65- GEP gtew by 12.6 percent per year. From then on, growth declined; but it still remained at 7.1 percent from 1973 to 1980. During the 1980s, on the other hand, the average GDP growth was only 2.1 percent. Put differently: GDP increased by a mere 17.6 percent from 1980-88; whereas over the preceding 8 year period, it grew by 83.5 percent.20 Per-capita income shows even a more radical change: an average increase of about 4.5 percent per year over the period 1974-80 (and, obviously, a co:nsiderably higher rate in earlier years) has become roughly zero growth over the 1980s. 3.3 A full analysis of the sources of growth, and cf its absence, is beyond the scope of this report. But the major faciors which have played important roles in Brazil's transformation frGm rapid growth to near stagnation may nevertheless be indicated. Economic growth is a function of three components: changes in the stock of factors of production; changes in technology; and changes in the extent to which factors of production are used efficiently. The drastic turn for the worse in Brazil's performance would be explained by a combination of such changes. B. Saving and Investment Table 3.2: GROSS FIXED CAPITAL FORMATION, 1970-88 (in billions Cruzados of 1980) Year Investment 1970 1.11 1971 1.29' 1972 1.50 1973 1.82 1974 2.06 1975 2.26 1976 2.41 1977 2.39 1978 2.50 1979 2.60 1980 2.84 1981 2.49 1982 2.33 1983 1.95 1984 1.95 1985 2.19 1986 2.68 1987 2.67 Source: IBGE, Ibid. 20/ Preliminary data for 1989 indicate, again, GDP growth within the range of 3-4 percent. -66- 3.4 Table 3.2 presents the size of gross domestic investment in fixed capital (i.e., excluding the accumulation of inventories): and its share in GDP. The trend clearly changed in the mid-1970s, and has worsened since 1980. During the first half of the 1970s, investment grew very fast. In 1980, annual investment was over two and one half times its level in 1970 (although investment was already growing much more slowly between 1976 and 1980). Income, of course, was also rising rapidly, but even as a share of income, investment increased substantially during the first half of the 1970s, and fell only moderately in the second half of the decade. In the early 1980s, on the other hand, investment's share of the national product fell dramatically, and despite some resurgence in 1986 and 1987, it remained much lower in the 1980s than in earlier decades. Even in the two exceptionally good years, 1986-87, the absolute level of investment only matched that of the late 1970s. Since the higher stock of capital implies a higher share of depreciation, net investment must have fallen considerably more, as a share of GDP. Its absolute value in the 1980s is therefore significantly smaller, on average, than in the second half of the 1970s. The addition to the nation's stock of capital has consequently slowed down in the 1980s. 3.5 Estimates of the capital stock do, indeed, attest to radical changes in the 1980s. From 1970 to 1980, the aggregate capital stock more than doubled, increasing by 116 percent--an average annual increase of 8 percent. Between 1980 and 1987, on the other hand, the stock increased by 16 percent--an annual increase of just over 2 percent. The stock of capital per capita, which nearly doubled over the 1970s, remained roughly constant from 1980 onwards.21 21/ The estimates in this section are taken from Donald V. Coes and Marcelo Bianconi, Macroeconomic Policies, Crisis and Growth in the Long Run: Brazil Country Study. Draft, World Bank Research Project 673-99. - 67 - Fig ure3.1: INVE:STMFNT BY TIlE PUBLIC AND ITH: (WVERNMENT SLCTORS, 1970-86 r--,A '4_ 0 {a -. - ~ 19701971 1972 197319741975 1976 19771978197919801981 1982 1983 198419851986 kE/ Privote rn Goverment Stute Enterises b LiVhSTK6AT S//AS OF GROSS DORiMIC PRODUCT 26 to 0.8 a 1 197019711972197319741975197619771978197919801981198219831 98419851986 CM N Private Investment/GoP State Enterprisesnt/GP S S Government investment/GPD Source: Donald V. Coes and Marcelo Bianconi, Brazil: Ilacroeconomic Policies, Crisis and Grow-th in the Long Run. Draft, mimeo., Mtarch 1989, Figures 3-13. -68- 3.6 The government's investment behavior changed substantially from one decade to the other. Figure 3.1 shows investments, both as absolute value (panel a) and as a ratio of GDP (panel b), by three (ategories of investors: the private sector, state enterprises, and the government proper. Private-sector investment did not change much from the 1970s to the 1980s. Its rate of growth was only slightly lower in the latter period, and its share of GDP did not decline by much. Government investment, on the other hand (including both the government proper and the state enterprises), changed drastically. During the 1970s it increased substantially--in absolute size, as a share of GDP, and as a proportion of total investment. During the 1980s, on the contrary (omitting 1980), government investment declined in all these three dimensions. From 1970 to 1979, the share of government investment in GDP increased from about 8 percent to 12 percent; by 1986 it had fallen to roughly 5 percent. Within the government sector, investment by state enterprises changed most, increasing from about 3 to over 8 percent of GDP, and then declining sharply to just 2 percent by 1986. Over the 1980s as a whole, government investment was only about one-half its level in the 1970s. 3.7 The drastic decline of government investment thus led to the substantial change in the size (and composition) of the country's stock of capital. And similar conclusions emerge from analysis of the economy's saving pattern. Table 3.3: SAVING RATES, 1970-87 (in percentage of GNP) Aggregate Gross Private Government Period Saving Saving Saving 1970-79 27.5 24.2 3.3 1980 29.6 28.5 1.1 81 28.5 27.4 1.1 82 28.2 28.6 -0.4 83 21.1 22.5 -1.4 84 17.1 20.0 -2.9 85 18.0 26.4 -8.4 86 22.2 29.5 -7.3 87 23.6 30.4 -6.8 1982-87 22.7 26.4 -3.7 Source: IBGE , Ibid. I -69- 3.8 Table 3.3 compares the savings rates in the 1980s, and the 1970s. The change is quite striking. The rate of aggregate savings in the economy has declined substantially, although this is evident only from 1983 onwards. From a rate of 27.5 percent in the 1970s, and 28.8 percent in 1980-82, savings fell to 20.4 percent in 1983-87. But practically the entire decline is explained by the behavior of governrment saving. The private savings rate has remained rather stable, at a level of some 25 percent of GNP,22 Government saving, on the other hand, declined from an average of 3.3 percent of GNP during the 1970s to about zero in 1980-83, and to a negative rate of 7.5 percent of GNP in 1985-87--a decline of about 10 percent of GNP. This decline is, in turn, the outcome of several factors to be indicated later in this chapter and in other parts of this report: it is essentially due to a drastic decline in tax collection, and an increase in government interest payments, without an offsetting change in the government's other current expenditures. Table 3.4: LABOR-FORCE PARTICIPATION RATES, 1960-85 (in percent of working-age population) Year Male Female Total 1960 76.2 16.6 46.5 1970 '0.1 19.8 44,9 1979 73.2 29.2 53.6 1981 73.4 33.4 53.4 1982 74.6 35.4 54.9 1983 73.6 36.2 54.9 1984 73.2 36.2 54.7 1985 74.6 37.6 56.1 Source: IBGE, Estatisticas Hist6ricas do Brasil. C. Labor and Land 3.9 Table 3.4 presents the participation rates in the labor force. Prior to the late 1970s and early 1980s, the aggregate participation rate had increased significantly. Since then, it has been roughly stable. The change is entirely due to participation of women ,n the labor force, which approximately doubled between 1960 and 1981, after which the increase slowed markedly. This slowdown might be expected as more women are drawn 22/ But a comment is called for: estimates of private savings, made as a residual, may have a large margin of error (although no bias is indicated). Similarly, in the absence of data of disposable income, inferences about saving behavior of the private sector must be tentative. -70- out of the pool of potential labor force participants. Nevertheless, its effect has beEn to lower GDP growth, at least in the way it is estimated in practice (in which housework is not counted). 3.10 Data for land cannot be established in a similar manner. But Bank studies indicate23 that land coming under cultivation in recent years has been of substantially lower productivity than the (large) addition to the stock of usable land in earlier years. This would then be ar.other reason why the rate of accumulation of productive resources in the economy has contracted. -~~~~~~~~~~~~~~~~~~~~~~~~ 23/ See Brazil--Agricultural Sector Review: Policies end Prospects, June 1989. -71- D. Capital Flows Table 3.5: MAJOR BALANCE OF PAYMENTS COMPONENTS (in billions of dollars) Balance of goods and Balance of Factor Current-Account Balance Non-Factor Services Service Income ( = (1) + (2) 1970 -0.23 -0.35 -0.58 71 -0.89 -0.42 -1.31 72 -0.97 -0.52 -1.49 73 -1.02 -0.71 -1.73 74 -6.24 -0.90 -7.14 75 -4.71 -1.77 -6.48 76 -3.98 -2.19 -6.17 77 -1.18 -2.86 -4.04 78 -2.40 -4,66 -7.04 79 -4.66 -6.10 -10.77 80 -5.29 -7.69 -12.98 81 -1.00 -10.93 -11.93 82 -1.87 -14.43 -16.30 83 +4.90 -11.85 -6.95 84 +11.84 -11.97 -0.13 85 +11.50 -11.89 -0.39 86 + 6.64 -12.03 -5.39 87 + 9.59 -11.09 -1.50 88 +18.23 -13.79 +4.44 Average 1970-73 -0.73 -0.50 -1.28 1974-78 -3.70 -2.48 -6.17 1979-82 -3.21 -9.79 -13.00 1983-88 +10.45 -12.10 -1.65 + = an export surplus. Source: Banco do Brasil, and IMF Balance of Payments Statistics, various issues. 3.11 Of major significance to the growth process in Brazil (on par, perhaps, with the change in the government's saving) is the change that has taken place in capital flows and in their derivative--external interest payments. The basic relevant data are presented in Table 3.5 (where the net capital inflow--which is equal by definition, in opposite sign, to the current account balance--may be read from column (3)). -72-- 3.12 Untril l 71, capital flows did not play a substantial role in Brazil's etznom.i. ptifotnar. e. During the early 1970s the net annual inflow of some US$1.3 billion, on average anmounted to) less than 1 percent of GDP. In the ahsence of a large accumulated debt, interest payments abroad (eolwur 2) \ere also minor. Starting in 1974--with the oil crisis and the availability of cheap ±runds in the international capital market- -capital inflows grew rapidly. Arid as debt started to accumulate, Brazil's interest pavmenrts to the outside world grew as uell, so that in later years a large further increase of capital inflows was matched by the higher interest payments, to leave about the same amount of net resources flowing from the outside world. 3.13 A radical change took place in the second half of 1982, with the onset of the 'debt crisis' and the unwillingness of external capital markets to provide further capital to Brazil. Since then, the not capital inflow from abroad has been approximately zero, or very small. Essentially, the existing debt was rolled over, but not increased. At the same time, interest payments abroad became very large indeed, due to the size of the cumulative debt. The net outcome has been a change from net inflows on the annual order of US$3-US$4 billion, to payments of over US$10 billion, on average, from 1983 onward24 US$14-US515 billion.25 3.14 This is a change of considerable proportions, amounting to about 5 percent of Brazil's GDP. With gross investment of 15-18 percent of GDP from 1983 onward, the change amounts to some 30 percent of investment--or close to 25 percent of investment levels in the 1970s. The change also amounts to about one quarter of national savings. The impact on the economy's capital accumulation, and growth, is therefore obvious.26 3.15 To a considerable extent, the change in interest payments (though not of capital flows) helps to explain the reduction in government savings noted above. As explained in Chapter X, the government of Brazil has assumed most of the country's external debt, and interest payments abroad 24/ This statement, and the accounting involved, should not be understood as implying that the capital inflow was harmful. Whether or not the large capital inflow of the 1970s and early 1980s was beneficial depends on factors such as the marginal productivity of investment or the impact of the inflow on the economy's saving behavior and on the relative prices of tradeables and non-tradeables. 251 An estimate using constant prices--which is more appropriate--would have shown a still larger relative change. 26/ Capital inflow is not in full an addition to investable resources: on the contrary, both theory and international experience show that the existence of capital inflows tends to increase consumption and lower domestic saving; depending on individual circumstances, this would apply to both enterprises (which would lower the rate at which they reinvested profits), the government, and households. In different ways, this also applies to interest payments abroad. It is most likely that this has been the case in Brazil. Nevertheless this offsetting effect is only partial: investment would normally increase by an amount equivalent to the major part of the added capital inflow. -73- becime a government commitment. With no reduction in the government's other current expenditures, th.ese interest payments thus became a source of government dissaving and hence of the reduction in Brazil's aggregate saving. E. The Impact of Inflation 3.16 To a large extent, Brazil's economic stagnation and the process of inflation are due to the same original forces (such as, for instance, the change in external capital flows). But the rapiG acceleration in inflation and its increased variability (which will be analyzed in the next chapter) has itself contributed significantly to the onset of economic stagnation. Table 3.6: INDUSTRIAL ORIGIN CF DOMESTIC PRODUCT-ON, 1970-87 (in percentage of GDP) Manufacturing Services Mining and Public Con- Finan- Public Agriculture Utilities struction cial Admin. Other GDP Year (1) (2) (3) (4) (5) (6) (7) 1970 11.6 30.4 5.4 6.0 9.2 37.4 100.0 1975 10.7 34.2 6.2 6.5 7.5 34.9 100.0 1980 10.2 33.9 6.7 7.9 6.5 34.8 100.0 1981 9.5 32.2 6.9 10.0 8.5 32.9 100.0 1982 7.7 33.6 6.7 Q.8 7.0 35.1 100.0 1983 9.0 32.2 5.6 11.3 6.4 35.5 100.0 1984 9.3 34.3 5.1 10.5 5.4 35.4 100.0 1985 9.1 33.3 5.4 11.0 6.6 34.6 100.0 1986 9.3 33.1 6.8 7.6 7.5 35.7 100.0 1987 7.7 31.3 7.3 13.2 7.2 33.3 100.0 1985-87 8.7 32.6 6.5 10.6 7.1 34.5 100.0 Source: IBGE, op. cit. 3.17 One impact is reflected in the structure of the economy, which is outlined in Table 3.6. Some of the trends revealed by these data are quite "normal," such as the decline in agriculture's share of GNP, which has continued in the 1980s and which may be expected at this stage of economic development. But the GNP share of manufacturing, which had increased significantly during the 1970s (and in earlier decades), stopped increasing in the early 1980s (after the increase slowed by the mid-1970s). However, -74- services increased substantially. Some of that increase is a normal part of development, but the growth was unusually large in Brazil. Moreover- -and this is the crucial point--the most substantial increase was in financial services, whose GDP share more or less doubled from the early 1970s to 1985-87. In absolute terms, financial services contributed about 5-6 percent of GDP more in 1985-87 than in the early 1970s, and about 3 percent more than in the late 1970s. This "contribution" is, however, better viewed as a subtraction. The increase reflects the extra increased need for financial services originating from the increased need for portfolio management to protect financial assets from erosion by inflation. While the provision of such services is in a formal sense part of the process of production, and part of the national product, and is so counted, it does not in fact contribute to what production would have been without inflation.27 Moreover, this does not take into account the assignment of resources by households and, primarily, tirms for portfolio management induced by the process of inflation. These resources must include the ablest managers, who are assigned to this function because decisions about physical invebtment, or about the production process must be of secondary importance in creating profits or preventing losses. While this effect cannot be estimated, it could easily be larger than the increase in financial activity by the financial institutions. The aggregate effect is possibly much higher than 5 percent of GDP.28 3.18 Another effect of inflation, related to portfolio management, is the increased share of foreign assets held by Brazilians as part of their financial assets--the so-called "flight capital." This activity represents an investment abroad rather than in the domestic market. It almost certainly reduces the national product, since the yield on that investment (the real interest on the capital held abroad) must be lower than the productivity of a real investment in the economy; it obviously reduces the domestic product. The size of this investment abroad is probably lower in Brazil than in other countries with roughly similar inflationary trends, due to the widespread availability in Brazil of domestic financial assets which are either formally indexed or which otherwise provide effective protection against inflation. Nevertheless, this factor too must have been quantitatively significant.29 3.19 The qualita, or nature, of investment has also probably been damaged by inflation and inflationary expectations, especially when these 271 This is equivalent to an increase of the country's military expenditures following a deterioration of its defense posture. Given the new situation, these services obviously fulfill a need, and should be counted as income; but in compariaon with the situation in which the need did not exist, this is a subtraction from useful activity. 281 This method of estimating an important component of the cost of inflation has been suggested, and applied to the case of inflation in Israel, in Ephraim Kleiman, "The Costs of Inflation," Working Paper No. 211, Department of Economics, Hebrew University of Jerusalem, May 1989. 29/ A World Bank estimate, mentioned in Chapter VIII, puts the cumulative size of capital flight from Brazil from 1980 to 1987 at roughly 5 percent of GDP--almost all in recent years. -75- are <:)i:.ed with periods of price controls and rationing. Inventories of goods at all stages--final goods, intermediate products, and inputs--will often be excessive. In times of high and rising inflation, this would also be true ttr capital goods, where purchase or installation would be advanced due to inflationary anticipations. In other words, capital is likely to be underut.ilized by firms in various sectors. Similarly, anticipations of relativt, clBanges of prices of outputs and inputs, and demand for those goods are likely to be blurred or grossly distorted in the face of very high, a. elerating, and volatile inflation--thus misdirecting investment. 3.20 Altogether, therefore, although little of the effect can be quantified, the process of high and accelerating inflation has most likely been an important cause of economic stagnation in Brazil.30 F. Distortionary Government Policies 3.21 The rElatively moderate size of the government's direct activity in Brazil obscures the more comprehensive role of the government in the economy. Through a widespread network of regulations, restrictions, fiscal measures, or similar means, the government's intervention in economic activity in Brazil is pervasive and highly distortionary. 3.22 In other chapters in this report, some of the more obvious seccoral policies, and their effects will be surveyed and analyzed. Here, only two observations will be made. First, the distortionary nature of government policies has become even worse in the 1980s. This was especially true of the foreign-trade regime, following the oil crisis and, in particular, the external-debt crisis. Second, even with unchanged policies, the likelihood that growth will be impaired increases over time. In the early stages, the instruments of intervention may be doubled-sided, at least in circumstances of less than full employment of the economy's physic,l and human resources. While the distortionary effect of a discriminatory policy exists at any time, it may for a while be offset by the expansion of activities which, at first, do not come fully or immediately at the expense of other (more economically profitable) activities. Thus, the benefits are realized more immediately than the high costs of the policy. As time passes, these costs are realized. Moreover- -as we will see in later chapters--a policy which first encourages some new enterprises and entrepreneurs turns into a system which primarily protects existing activities and enterprises, stifling any threat of competition. It is likely that the 1980s--in contrast with the two earlier decades--has indeed been the period in which the cost of the government's pervasive intervention in Brazil's economy has surfaced. While this factor cannot be quantified, it has most probably contributed substantially to Brazil's economic stagnation. 30/ On the other hand, one should avoid the pitfall of identifying--as is sometimes done in Brazil--the accumulation of domestic financial assets (primarily the government's short-term debt) with a "lost" investment in physical capital. On the contrary, it is likely that in the absence of these domestic financial instruments, domestic investment would have been smaller, both because saving would have diminished and because more would have been invested abroad. -76- IV. THE INFLATIONARY PROCESS A. General: The Pattern of Inflation 4.1 A change in the course of inflation in the 1980s, and particularly in the most recent years, is a very distinctive feature of Brazil's economic performance. Inflation, represented by the movement of the consumer-price index, is shown in Table 4.1 and in Figure 4.1 (pp. 87-88). -77- Table 4.1: INFLATION RATE, 1980-89, QUARTERLY (percentage change of IGP/DI) YEAR QTR Quarterly Annualized Inflation Rate 1980 1 18.9 100.2 2 18.8 99.3 3 22.3 123.9 4 21.9 120.9 1981 1 22.7 126.6 2 19.9 106.4 3 17.2 88.3 4 15.5 77.8 1982 1 18.7 98.5 2 20.4 110.3 3 19.5 104.2 4 15.3 76.7 1983 1 24.4 139.6 2 29.1 178.0 3 39.0 273.3 4 36.8 250.0 1984 1 32.3 206.0 2 31.3 197.2 3 33.2 215.1 4 37.0 252.1 1985 1 38.0 263.0 2 29.0 177.1 3 32.2 205.6 4 38.3 266.2 1986 1 46.1 355.8 2 3.7 15.9 3 2.4 9.8 4 6.9 30.6 1987 1 36.4 245.9 2 77.1 884.1 3 43.6 324.7 4. 37.2 251 7 1988 1 62.4 595.0 2 70.7 748.7 3 82.0 998.0 4 105.3 1,675.0 1989 1 88.2 1,154.0 2 32.8 210.7 3 161.4 4,571.1 4/a 186.1 6,603.2 la Fourth-quarter 1989 based on average inflation for October and November. Source: Conjuntura Econbmica (various issues), and IBGE. -78- 4.2 The increasing trend in the level of inflation in Brazil hi 1ŽHn apparent for a full decade. But the course of inflation was diftt-rer.t prior to 1986. First, and most important, the rate of increase o! inflatior. has intensified greatly since then. From 1980 to 198S, arnu;.l inflation approximately doubled; in the four years since, it incieased (eightfold), to an annual level of over 1700 percent during 1989 (a:id a still much higher rate in early 1990). Second, the nature of inflationary cycles has changed. From 1980 to late 1985. inflation increased from one semi-stable lev-1 to another: from about 100 percent per year between early 1980 and early 1983 to somewhat over 200 percent from then to the fall of 1985. Since late 1985, on the other hand, the cyclical pattern is one of a sharp deceleration of inflation, once in a while, followed hy an even sharper and longer period of accelerating inflation. 4.3 The pre-1986 inflation will not be discussed at any great length here; but will be referred to in a summary way later, since the process of inflation during that period is a necessary background for understanding more recent inflationary patterns. 4.4 The discussion will begin by describing and analyzing the components which seem to be involved in the recent inflationary process; then move to see how their interaction explains the process. These components are the budget, the public debt, monetary policy, exchange-rate policy, indexation, exogenous shocks, and stabilization policy packages. B. The Budget and the Fiscal Deficit 4.5 The fiscal deficit cannot be universally and uniformly defined; the definition would vary according to the intended use of the concept and to economic circumstances. In the context of Brazil's economy in the 1980s, three major variants of the "deficit' must be distinguished. 4.6 The "nominal deficit" is the excess of government expenditures, of all sorts, over its total revenues from any source other than borrowing (that is, primarily, from taxes and other compulsory payments). This deficit represents the net borrowing requirement of the government. As such, it represents the public sector's contribution to the increase in the nominal financial claims of others, i.e., those of the domestic private sector and of the outside world. 4.7 However, in the inflationary circumstances of Brazil, this deficit would not be an addition to the real net wealth--financial or total--of the non-government sectors. A considerable part of government expenditures consists of interest payments to the local private sector, which to a major extent simply reflect compensation for the reduction of the real value of the principal of the debt by the prevailing inflation. The size of interest payments needed to maintain the real value of the principal is the "monetary correction,' and it has to be deducted from the nominal deficit -79- in order to calculate the size of the government's contribution to the real increase in financial (and total) wealth of the lenders.31 This deduction yields the "operational deficit.' 4.8 The latter concept may still be separated into two components. One is the size of the real interest payments made by the government. Deducting this element from the operational deficit gives the "piimary deficit.' The latter defines the size of the deficit derived from government operations other than interest payments; and, due to the distinctive nature of interest payments, merits separate consideration. It reflects the real fiscal effect under the control of the government. 4.9 The "operational" deficit is thus the relevant variable when analyzing the fiscal impact of increased private-sector wealth (disregarding 'Ricardian equivalence") on inflation; whereas the 'nominal' deficit is the appropriate concept for analyzing the impact of the budget on money creation. This will be elaborated upon shortly. 31/ One way of formulating the matter is to think of the governn. obligation held by the public as an indexed bond. Part of tl.. nominal interest payment is not genuinely interest but the component required, by the indexation, to maintain the original real value of the bond. Only the excess over it represents true interest payments. Only if money illusion in present--a very unlikely assumption in an economy with continuous inflation, like Brazil's--would the increase in nominal value due to indexation be perceived by the bond holder as an increase in wealth. -80- Table 4.2: THE PUBLIC-SECTOR DEFICIT, 1980-88 (percentage of GDP) 1985 1986 1987 1988 1981-84 1985-88 Central Government Nominal Deficit 10.2 4.9 13.6 10.8 5.5 10.0 Monetary Correction/a 9.1 3.7 10.5 7.1 3.8 7.7 Operational Deficit 1.1 1.3 3.1 3.7 1.7 2.3 Real Interest Payments/a 2.7 1.6 1.1 2.4 n.a. 1.9 Primary Deficit -1.6 -.3 2.0 1.3 n.a. 0.4 Local Governments Nominal Deficit 6.5 2.7 7.9 13.0 4.9 7.5 Monetary Correction 5.6 1.8 6.3 12.8 4.2 6.6 Operational Deficit 0.9 0.9 1.6 0.2 0.7 0.9 Real Interest Payments 0.7 0.7 0.8 0.9 n.a. 0.8 Primary Deficit 0.2 0.2 0.8 -.7 n.a. 0.1 Public Enterprises Nominal Deficit 11.3 3.5 9.8 24.8 8.0 12.e Monetary Correction 9.0 2.2 9.0 24.4 5.4 11.1 Operational Deficit 2.3 1.4 0.9 0.4 2.6 1.2 Real Interest Payments 2.0 2.2 2.9 2.3 n.a. 2.6 Primary Deficit -0.7 -0.8 -2.0 -1.9 n.a. -1.4 Total Public Sector Nominal Deficit 28.0 11.1 31.4 48.5 18.4 29.8 Monetary Correction 23.7 7.5 25.8 44.3 13.4 25.4 Operational Deficit 4.3 3.6 5.5 4.3 5.0 4.4 Real Interest Payments 6.4 4.5 4.8 5.6 n.a. 5.3 Primary Deficit -2.1 -0.9 0.8 -1.3 n.a. -0.9 n.a. = not available /a Discrepancies between interest-rate data here and in Table 6.2 of Chapter VI are due to difference in sources; specifically, the assignment of payments to "real interest" vs. "monetary correction" is not uniform. "Nominal deficit" minus "monetary correction" = "operational deficit". "Operational deficit" minus "real interest payments = "primary deficit". "Monetary correction" plus "real interest payments" = nominal interest payments. (minus sign indicates a surplus) Source: Brazil Economic Program, various issues, and Central Bank of Brazil. -81- 4.10 Table 4.2 presents the three concepts of the budgetary deficit tor the public sector as a whole, as well as for its malor organizations: the central government (including the social-security system and independent agencies); local governments (states and municipalities); and the public enterprises. Data are for the years 1985-88, along with annual averages for 1981-84. 4.11 The nominal public sector deficit has in fact increased substantially: it was about twice as high during 1985-88 than during 1981-84. This is not surprising with rising nominal interest rates accompanying the accelerated inflation. But the operational deficit, reveals a different pattern. The deficit is rather substantial--around 5 percent of GDP; but it is, at the same time, rather stable, whether comparing year to year or 1985-88 and 1981-84. The primary deficit is mostly negligible; on average, there was even a very slight surplus. The whole of the operational deficit may thus be "explained," in the accounting sense, by the public sector's interest payments on its debt. In the current year, 1989, this relationship is actually specified by law: the government cannot borrow to finance a primary deficit--its borrowing is limited to the finance of its debt service payments. Since the real interest rate has been rising substantially during most of the year, the operational deficit in 1989 is expected to be of the order of 7-10 percent of GDP. To some extent, this increase is also due to the deterioration in the primary budget (which had a surplus of 1.3 percent of GDP in 1988), due to increased expenditures and reduced public enterprise revenues (the latter paid higher real wages while it suffered reduced real tariffs); and the deteriorating position of the social security system. 4.12 All major components of the public sector--the central government (including independent agencies), local governments, and public enterprises--usually contribute significant parts of the operational deficit. So far as the primary balance is concerned, the central government shows deficits in 1987 and 1988, which are roughly offset by surpluses of government enterprises in these two years. 4.13 A large part of the operational deficit has been financed through the inflation tax, that is, the reduction in the value of the monetary base due to inflation.32 Other financing has come from the government's seigniorage gain. But the latter could not be of much sigrificance in an economy like Brazil's in recent years, in which as a consequence of 32/ To be precise, the variable subject to the inflation tax is the monetary base (high-powered money): the difference between this and the size of money (M1) is the value of credit, which reaps inflationary gains. That is, the inflation tax is imposed on holders of money. It is therefore partly received by the government, and partly goes to banks or to recipients of bank credit. The inflation tax on money as a whole has distributional implications, discussed elsewhere in this report. -82- inflation demand for money (Ml) has fallen drastically and the ratio of the monetary base to income is very small.33 In combination, these two elements amounted to about 2 percent of GDP during the period 1980 to 1985, increasing over the last three years (1986-88) to 3 percent. The excess of the operational deficit over this tax must be represented by an increase in the public debt, to which we shall now turn. C. The Domestic Debt 4.14 Government (as well as Brazil's overall) borrowing from abroad has, in net terms, been small in recent years: from end-1985 to end-1988, the public sector's external debt increased by about US$10 billion--of which a substantial part was the assumption of private debt. Hence, the overwhelming part of the government's budget deficit has been financed through local borrowing, which increased the stock of the government's domestic debt. A nominal budgetary deficit would increase the nominal stock of the debt, whereas an operational deficit increases the real size of this stock. The size of the domestic debt is shown in Table 4.3. Table 4.3 THE PUBLIC-SECTOR DOMESTIC DEBT, 198G-88 (in percentage of GDP)/a Aggregate Net Debt Excluding Treasury Securities Domestic Debt the Monetary Base Held by Private Sector Year (1) (2) (3) 1980 n.a. n.a. 3.4 1981 11.9 8.2 6.4 1982 14.3 11.1 6.6 1983 17.7 15.9 4.4 1984 20.4 18.4 7.9 1985 21.6 19.7 10.2 1986 20.6 16.3 7.4 1987 20.3 17.9 9.9 1988 21.1 19.6 12.5 1989/b 48.0 n.a. 23.0 n.a. not available /a End-year real stock of the debt divided by annual real GDP 1/5 Preliminary Source: World Bank calculations, based on Central Bank data. 33/ In a sense, though, as we shall see shortly, an increase in tne public debt may also be viewed, in the context of Brazil's economy in rerent years, as the realization of seigniorage gains. -83- 4.15 The stock of the domestic debt increased substantially, in relation to GDP. In tite earlier years of the 1980s, this resulted partly from the government's assumption of the external debt--a process t. be discussed in Chapter VIII. But in more recent years, the predc,ninant factor is the budgetary deficit, which has also led to a continuous increase in the share of credit to the government in total credit in the economy (i.e., to the government and the private sector): in mid-1980 this share was 19 percent, by 1984 it was 34 percent, and by 1988 it had reached 45 percent. 4.16 The level of the real interest rate is a major determinant of the rate of accumulation of the public debt: it determines real .nterest payments which, with the general absence of a primary deficit, are the major contributors to the existence of (and variations ir., the operational deficit. While the level of the real interest rate fluctuated violently- -due primarily to the episodes of 'stabilization plans"--no trend is apparent until very recently. It will be argued shortly that the tight relationship between the real interest rate and the size of the short-term public debt severely restrains the use of monetary policy in Brazil: an increased real interest rate, which would be prescribed for a contractionary policy, works forcefully to increase the government's real payments and its deficit. 4.17 This problem is heavily reinforced by the composition of the debt. Brazil's domestic public debt consists of federal bonds, local-government debts, and federal-government borrowing from the financial system. The average maturity period of the domestic federal debt has declined continuously and drastically. By the end of 1988, it had fallen to about 2-4 months, and since the Summer Plan of early 1989 it has fallen further to 1-3 months. Most of it is held in the form of government securities (LFT's--Letra Financeira do Tesouro) which, through a daily repurchase agreement, are issued at an interest rate which varies daily, corresponding to the average rate for overnight funds. At the end of 1988, these securities constituted about two-thirds of the federal bonds in the hands of the public. The accumulation of debt in this form has far-reaching implications for the conduct of monetary policy, just as it does for the fiscal stance; and the discussion will now turn to them. D. Money and Monetary Policy 4.18 The size and co-position of the domestic debt are the most important factors in the discussion of money, and monetary policy, in Brazil in recent years. 4.19 The high liquidity of most of the domestic debt makes it a very close substitute for money proper: although debt instruments cannot be used directly as means of payment in a transaction, they can be converted to money daily (sometimes momentarily) at low transaction costs. This part of the debt is hence commonly classified as part of money in the broader sense, designated in Brazil as M4. The components of M4 are presented in Table 4.4. Table 4.4: COMPOSITION OF MONEY, 1980-88 (percentage of M4) Currency plus Demand Time & Saving Government Total Deposits (M1) Deposits Bonds Year (end) (1) (2) (3) (4) 1980 40.6 42.4 17.0 100.0 1981 31.6 43.7 24.6 99.9 1982 25.0 48.7 26.2 99.9 1983 19.8 59.8 20.5 100.1 1984 15.2 55.7 29.1 100.0 1985 15.2 49.7 35.1 100.0 1986 31.6 43.3 25.1 100.0 1987 16.3 47.8 35.8 99.9 1988 9.6 47.2 43.2 100.0 1989(August) 6.6 43.5 49.8 99.9 Source: Central Bank of Brazil. 4.20 The pattern revealed by Table 4.4 is obviously that expected during a process of high (and accelerating) inflation. The share in M4 of Ml--the component of money which is neither indexed nor bears interest--has declined continuously and substantially. As Table 4.5 indicates, this has also been manifested in a substantial decline in the ratio of M1 to GDP. Savings and time deposits, on the other hand, have increased their share in the aggregate amount of money. They have also maint'tined a roughly stable ratio to GDP. These accounts are mostly indexed (and interest-bearing), to a large extent (but not quite fully) protected against inflation. Government bonds have increased beyond that; their share of aggregate money has risen continuously, to about 50 percent by the end of 1988. Their ratio to GDP has also increased substantially. A large shift has thus occurred from other forms of money to the holding of government bonds, which are now the largest component of money. The other side of this coin is the manner in which the government's fiscal activity injects money into the economy. The nominal budgetary deficit, i.e., the government's borrowing requirement, indicates the size of such injection. But the channel it takes is the issuance of the government's short-term bonds, rather than its borrowing from the central bank. 1 -85- Table 4.5: RATIOS OF MONEY COMPONENTS TO GDP (end-year stocks, in percent) Year M1 M2 M3 M4 1979 14.1 22.8 29.5 36.1 1980 11.5 19.5 24.1 29.0 1981 10.2 20.2 25.8 35.0 1982 8.7 19.9 26.4 35.9 1983 7.3 22.5 30.6 38.7 1984 6.6 22.5 32.5 46.0 1985 7.9 23.3 33.9 52.1 1986 12.3 21.2 29.0 38.7 1987 8.9 26.3 34.4 53.5 1988 7.5 34.3 44.5 78.4 1989(August) 3.6 22.2 27.4 54.6 Source: World Bank calculations, using monetary data from Central Bank of Brazil. GDP figures are from IBGE. 4.21 The composition of money has important implications for the process of inflation. In substance, if not always formally, the important component of government bonds is indexed, since the nominal interest payment on the bonds rarely falls in any measurable degree below the rate of inflation.34 In this manner, '-his part of the quantity of money behaves in a way which leads it, automatically, to accommodate inflation of any origin. Monetary policy, under such circumstances, must be much less effective in containing inflation. 4.22 The fact that a high proportion of money is held in the form of government bonds also constrains monetary policy in a different way. An increase in the real interest rate tends to increase the operational deficit, and thus the public's perceived wealth. Moreover, such an increase is manifested in the most liquid form of wealth, namely, in the quantity of money (since the government bonds yield an interest rate which is variable and equal to the market rate of the day). The classical restrictive measure of tightening the money market--via an increase in the (real) interest rate--thus yields the opposite impact in Brazil: it tends to increase money and liquidity. 341 As noted earlier, such government securities may best be viewed as indexed bonds, yielding normally some real interest rate. This real yield is rarely negative to any significant extent. -86- 4.23 By the same token, open-market operations are highly ineffective in Brazil. An open-market sale by the Central Bank would, on the one ha..d, lower the monetary base and the M1 component ot the money stock; but it would offset the latter impact by placing additional government securities in the hands of the public, thus increasing the stock of money. Put differently: given the nature of the government's borrowing from the public, the difference between its impact and that of borrowing from the Central Barik is not as radical as it is in most other economies. What determines money supply is the size of the government's nominal budgetary deficit, regardless of the way it is financed. 4.24 Nor can the Central Bank easily control the size of money by changing the banks' reserve ratio. A 'reservc ratio" as such does not really exist in Brazil's financial system. In most other countries, too, there is no single reserve ratio, but the number of separate ratios is small, and one or two predominate. In Brazil, reserve ratios change by type of liability, type of bank, regional location. and various other criteria--resulting in over one hundred reserve ratios. Moreover, asset holders can easily shift assets among classes subject to varied reserve ratios. The result is quite large fluctuations in the money multiplier, that is, in the size of money (narrowly defined--Ml) with any given monetary base--fluctuations which are not primarily the outcome of wholesale changes of the set of reserve ratios. Regression analysis has shown that even in explaining changes in Ml, changes in the size of the money multiplier are significant in comparison with changes in the size of the monetary base. The money multiplier appears already to contribute more to changes in M3 (a definition which excludes the short-term debt) than the monetary base. When it comes to M4, the broadly-defined stock of money, variations in the monetary base seem (not surprisingly) to explain next to nothing. E. Exchange-Rate Policy 4.25 Brazil's exchange-rate policy will be surveyed at some length in Chapter VIII, in the context of foreign-trade policies. For the present purpose, it suffices to state that as a general principle, the government has traditionally, for at least a generation, aimed at maintaining a stable real exchange rate. This may be seen from Figure 4.1, which shows the movement over time of the price level and of the nominal exchange rate. The mechanism for achieving approximately parallel movements in these two was, for most of the time, the use of a crawling peg policy. This rule has been maintained in recent years as well, except for few occasions--to be noted shortly--in which the government fixed the nominal exchange rate in the context of a stabilization policy. These attempts were unsuccessful, and the fixed nominal rate was always abandoned within a very short time- -not more than a few months (unlike, for instance, the longer-term 'tablita" experiments in Argentina or Chile). Thus, as a rule, the foreign-exchange rate was in fact--though not in any formal way--indexed to domestic inflation.35 Much like the quantity of money the foreign-exchange rate has not served in Brazil as a nominal anchor (except on rare and short occasions). 35/ But the fall of the real exchange rate during attempts of stabilization has not always been fully compensated later. One recent, and important, instance is the experience of 1989. - 87 - Figure 4. la Brazil - Annual Average Exchange Rate and Price Level (1977X100) Index Level 500.0 400.0 - 300.0 - 4- 200.0- 100.0- 1966 1968 1970 1972 1974 1976 1978 1980 Year Exchange rate - General Price Index Source Data from International Financial Statistics 11 I r . I ! Brazil - Quarterly Inflation and % change in Cz$/US$ Exchange Rate percent 120.0 100.0- T 80.0- 60.0 - 40.0 - II 4.~~~~~~~~4 0.0 1234123412341234123412341234 1234 123412 80 81 82 83 84 85 86 87 88 89 Period o Inflation - Exchange Rate 1/ Period average inflation and exchange rate. 2/ Inflation measured by the IGP/DI, Sourk. u 1indtc ial Stat ist ics 1nd 6azcLta .rcaritil. -89- F. Indexat ion 4.26 Brazil's economy is highly indexed--a factor of large potential importantce in the inflationary mechanism. T'iis has various manifestations, which will be mentioned here in a summary way. 4.27 First. and arguably most important, a pervasive informal indexation will normally be found in an economy like Brazil's, in which inflation has been a 'ong-term phenomenon. Even in the 1960s and 1970s, price increases were substantial--by international yardstick--and quite permanent. The accelerated inflation of the 1980s must have reinforced the perceptions about the constant need to adjust to inflation. It is thereiore likely that, at least so far as shorter-tetm price movements are concerned, individual prices of outputs and inputs are changed in concordance with general price increases--either those just realized or those expected. 4.28 In a more formal way, the price of the most important input- -labor--is indexed to the general price level. The mechanism and rules of this indexation have changed with time. Before the Cruzado Plan (February 1986), wage adjustments to changes in price levels were made every 6 months (having been shortened from one year towards the end of 1919).36 With the Cruzado Plan, this was changed into a "trigger" adjustment of wages whenever the cumulative price increase (since the last adjustment) exceeded 20 percent. With the "Bresser Plan" (of June 1987), a time interval was again introduced: adjustment was made monthly, but using the average price increase of the last 3 months over the average of the 3 preceding months--a procedure which stiLl leaves a substantial time lag. This lag has been reduced, following a temporary freeze, with the "Summer Plan" of January 1989; in recent months it has been, thus, a simple monthly adjustment. 36/ The index used for this adjustment roughly reflects the index of consumer prices, but discrepancies betweer the two are frequent. Similarly, the degree of adjustment has sometimes varied according to the wage level--with full adjustment of low wage brackets, and incomplete adjustment of higher brackets. Figure 4.2 REAL WAGE AND CPI INFLATION 1986 JAN 100 % IN MONTHLY TERMS I20 4 40 Ito 30~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 100 ,,' t " ', ,' 20 09 v_S*0 1p 0 80I , 70 .. -1100 I 1985 1986 1987 1988 1989 Source: Data from Coniunctura lCconomica. -91- 4.214 V3e ,Ppite the prevalence of formal wage indexation, actual movements of nomnial] wages and of prices do not correspond very closely, as real wages~ luc tuirite substantially. This is seen in Figure 4.2, which shows mcrnthly Pzovem,ients of real wages and of irflation for 1985-89. Two phenomena appear very clearly from this diagram. First, fluctuations of real wages ate substantial, whether shorter or longer periods are concerned. Thus, for instance, the index of real wages increased from al,.,:t 80 in early 1985 to about 110 in the fall of 1986, fell to about 90 by mid-1987, and increased again to 120 toward the end of 1988. Second, and prui:ably even more significant, is the absence of clear relationships between changes in real wages and changes in inflation. Indexation of wages to past. performance of prices--even the most recent past performance- would, if strictly adhered to, mean a decline in real wages when inflation accelerates and an increase when inflation decelerates. This, in face, appears to have been clearly the case only during the 'Cruzado Plan' period; to a lesser degree, it was true also in part of 1985. Otherwise, no such regularity appears. Moreover, from mid-1987 to the end of 1988 real wages increased substantially while inflation sharply accelerated. The formal indexation of wages in Brazil should thus not be interpreted as meaning a tight link between wages and prices--at least not in recent years. The limited coverage of formal indexation of wages, the large imperfections in the adjustment mechanism, and the fact that 'indexed' wages move often beyond the change that would be automatically called for, all imply that wage indexation is probably not a crucial element in wage determination. An important exception would be, for short- term changes, wages in the government. 4.30 The indexation of financial assets, on the other hand, is of crucial significance. Such indexation is, of course, pervasive in Brazil's economy. As we have just seen, this is true in effect for the overwhelming part of the stock of money (all of it except for M1, which is now less than 10 percent of the total). Since private long-term bonds are of minor significance, this means that most fixed obligations in Brazil are indexed (and the latter, in turn constitute the majority of financial paper, since the aggregate value of equity shares is relatively small). We have also noted that the nominal rate of foreign-exchange is, in fact, mostly indexed to the movement of domestic prices. Financial indexation in Brazil is, therefore, widespread indeed--more so, probably, than in practically any other economy in recent times.37 In the perpetuation of inflation, it is this indexation--particularly of money--which plays a major role. G. ExoRenous Shocks 4.31 Over the last decade, two major exogenous shocks have occurred, which might reasonably be related to the process of inflation. One is the second "oil shock," of late 1979, in which the price of oil approximately tripled. For Brazil, which at that time imported almost all its oil, this price increase (and income loss, due to terms-of-trade deterioration) contributed directly and indirectly to price increases. The second event 37/ The economy of Israel may be the only similar case. -92- was the "debt crisis" of mid-1982, whose impact on the price level was not as direct as that of the oil-price increase but was nevertheless substantial. 4.32 But no similar exogenous shocks, not even on a smallcr scale, have taken place after 1982. Developmeats in external markets were, if anything, favorable overall to Brazil: the decline in oil prices (relative PA and absolute) and world interest rates being probably the most significant such developments. Since 1985, the important "shocks" have all been policy shocks, introduced by the government of Brazil. H. Stabilization Plans 4.33 Three stabilization policy-packages have been implemented in recent years, at about 18 month intervals: the "Cruzado Plan" (in February 1986); the "Bresser Plan" (June 1987); and the 'Summer Plan" (January 1989). Despite some differences, all three plans shared several important components. 4.34 The Cruzado Plan (launched officialiy on February 28, 1986) amounted mostly to the freezing of prices and the nominal rate of exchange. Prices were frozen, initially, for a period of up to a year; in fact price controls were removed by November of that year. The exchange rate, on the other hand, was supposed to be subject to periodic devaluations, but was in fact kept at a constant level for 8 months--through October. No devaluation was undertaken, at the time that the plan was implemented. Wages were raised 8 percent above their average level over the preceding six months, and were, after plan implementation, subject to the trigger adjustment when cumulative inflation reached 20 percent (with the further assumption that any gap in the adjustment of real wages would be closed during annual contract negotiations). The main indexed financial instrument issued by the government--the OTN--was eliminated. The nominal overnight interest rate was fixed at 1.5 percent per month--less than 20 percent per year--with the declared aim of preventing recession. Fiscal policy too became, if anything, more expansionary, as public-sector prices were frozen at an initially low level, thus contributing to an increase in the government deficit. 4.35 The Bresser Plan (June 12, 1987) shared with its predecessor the most important element--a price freeze. In this instance, the freeze was imposed for three months, specifying a gradual release consequently. The wage indexation rule was changed--from a trigger of 20 percent to monthly adjustment to prices in the last three months; in addition, nominal wages were frozen for the period of the price freeze. Unlike the earlier plan, a devaluation (of 9.5 percent) was introduced along with the plan, and future m'ni-devaluations, rather than a fixed nominal rate, were specified. Indexation of the main short-term financial asset--the OTN--was retained. The real interest rate was raised--again, unlike the Cruzado Plan. But the government's deficit, despite an initial pledge, was not significantly lowered. 4.36 The Summer Plan (January 15, 1989) is the most recent. Its major element has again been a price freeze, for an indefinite period. The real wage was actually raised, by 26.5 percent (over the average of the previous -93- 12 months), and there was no interference with further wage indexation. The indexed short-term financial asset, the OTN, was eliminated, but the interest rate on short-term Treasury bills (LFT's) was to be kept high enough to yield a positive real interest rate. Public-sector prices were initially realigned, and were to be kept frozen during the period of the general price freeze, which contributed to an increase in the budgetary deficit. There was no real cut in expenditures, except for the symbolic gesture of eliminating five ministries. A plan to fire some 60,000 government employees was scrapped within days of its announcement. The government pledged--and has adhered to its commitment--not to borrow in the market, except for servicing the debt. However, as the primary balance had in the past too been mostly in surplus--or only marginally in deficit--this step has not actually meant a deficit reduction. I. The Inflationary Process 4.37 The process of inflation in Brazil is complex, and probably nct s:Abject to an interpretation which is either unique, conclusive, or definitive. It is doubtful whether any single model adequately explains the process. The following outline seems, however, to be a reasonable analysis of the pattern of inflation. 4.38 First, two periods must be distinguished: the years from 1980 to late 1985; and the years since then. The 1980-85 inflation will be discussed briefly, mainly because it provides a background for developments and policies in recent years, which are the major concern here. 4.39 In the period 1980-85, a distinct pattern is apparent in the development of inflation. The rate of inflation about doubled, and then remained stable on the new plateau for several years. This pattern was repeated twice: in early 1980, when inflation about doubled, to a new annual rate of roughly 100 percent; and in early 1983, when it doubled again to a level of 200 percent or somewhat above--a level sustained until the second half of 1985. The trends from 1980 onwards may, again, be seen from Figure l.b (but note that inflation rates there are quarterly). 4.40 The most probable source of this pattern are the two exogenous shocks of this period: the second oil shock of end 1979, and the debt- crisis shock of 1982. The first obviously gave a direct impetus to inflation--in the same way that both oil shocks pushed inflation upwards in almost all oil-importing countries. The second event--the debt crisis--did not lead directly to inflation, but it has led to relevant policy steps; namely a substantial devaluation and a set of fiscal measures which affect prices directly, such as the contraction of subsidies or the increase in regulated prices. In each event, the high degree of indexation in the economy maintained inflation at the higher level reached following the shock. While this describes the original push of the process of inflation, the sustainability and persistence of inflation was the result of monetary policy: with the same shocks, and a different monetary policy, inflation might have been short-lived. Given the choice between sustained inflation and a substantial increase in unemployment (which would follow from a contractionary monetary policy), the Central Bank opted fo- expansion of the money supply and persistent inflation. -94- 4.41 The process during the first half of the 1980s colorted the views of policymakers in later years. Self-sustained inflation that persisted for a long time, without an apparent fundamerital source, has established the percepticn of "inertial' inflation, which can be stopped by initerrupting tbe inertia witiout any need for fundamental change. This perception is reflected in all three stabilization plans, and was a nmajor so.irce of their failure. , .42 The inflationary process in the latter half of the 1980s has, however, been radically different. From the end of 1985 onwards, the pattern of inflation is marked by persistent acceleration, accompanied by deep cyclical movements. The cyclical movements are obviously related to the stabilization plans, and will be analyzed shortly. We shall start, however, by analyzing the forces of acceleration. 4.43 First, a negative assertion must be made. Although the persistent inflation must be related to the budgetary deficit, its acceleration cannot be assigned to a rising deficit. The budgetary (operational) deficit has been rather stable. at the level of close to 5 percent per year, and no trend is apparent in the second half of the 1980s. The accelerated inflation thus cannot be explained by a rising inflation tax: the same tax could evidently be collected only at continuously increasing rates of infl :ion. 4.44 However, this leads to the observation of the cumulative effect of the budgetary deficit, namely, of the public sector's debt. We have seen that the size of this debt has increased substantially, thus constantly increasing the private sector's perceived financial wealth. Moreover, its practical indexation has led thJs debt to be substituted for other forms of money. The size of norninally-fixed money--Ml--has shrunk continuously (from 8.6 percent of .DP at end-1980 to 2.8 percent at end-1988, and 1.8 percent in August 1989). Along with that, the monetary base has shrunk. Since this is the base subject to the inflation tax, a constant _mount of tax must require increasing rates of inf'lation. The budgetary deficit, although roughly constant, has in this way contributed to the acceleration of inflation. 4.45 Other, less clear-cut contributions to the process can be noted. The growing size of the debt has, presimably, reduced the public's willingness to hold this asset--in contrast to real assets or foreign financial assets, for example--at given interest rates. This is due both to the increased relative size of the debt and to the increased risk of default that must accompany its expansion. As a result, the government has found it necessary to keep raising the nominal and real interest rates. The increased real rate leads to increased pressure on the budget, the increased stock of short-term debt, and, consequently, the increased money supply (in the broad sense). In addition, an increase in the nominal interest rate has apparently come to fulfill an important role in creating inflationary expectations in Brazil. The government's policy is, by and large, to fix a yield on its bonds which will imply a positive real rate--or at least not a significanitly negative one. Hence, an increase in the nominal interest rate has come to be regarded as an expectation by the government of a higher level of inflation. This, in turn, leads economic agents to accelerate their price increases. I I -95- 4.46 Expectations, in general, have come to play a dominant role in this process. With the accelerating and fluctuating rate of inflation, economic agents are less likely to assume that the past level of inflation will be sustained, and to react to past price increases in their inputs (by producers) or in final goods (in the case of labor). Instead, signals of future developments become more important. In that context, it appears that the public's faith in the goverrment's abiiLity to contain inflation has greatly deteriorated. This is due to the extended stretch of inflation and its less predicable nature, to the failure of past stabilization attempts, and to the realization by the public that no nominal anchor exists in Brazil's system, so that any rate of inflation becomes possible. In the second half of 1989, even expectations of hyperinflation became common. Given these expectations, preemptive price increases became predominant. In the same vein, this learning process must have effected the demand for money, contributing to an acceleration in the declining demand for real balances, and hence a more rapid increase in inflation. 4.47 Another response to this pattern is the shortening of reaction time to past developments. Reactions of wage demands to prices are of prime importance in this context. As we have seen, the time lag involved has kept getting shorter--becoming, approximately one month during the laet year (&nd sometimes even shorter in the private sector). Presumably, also, increases in individual prices in response to either past or expected inflation have become quicker. While this is a reaction to accelerated inflation, it obviously also contributes to further acceleration. 4.48 We come now to the pronounced cycles apparent in the inflationary process since the end of 1985. The three cycles are all obviously related to the stabilization plans (the Cruzado, Bresser, and Summer Plans). Typically, a cycle starts with accelerated inflation; then, with an introduction of a stabilization program, inflation goes down dramatically. After a period of roughly half a year, the rate of inflation picks up again, to a level which exceeds its norm prior to the beginning of the cycle. 4.49 The acceleration of inflation at the s'art of the cycle could ba explained in two alternative ways. One is the policy respense: it is, presumably, when inflation accelerates beyond what comes to be regarded as a "norm" that the government feels compelled to introduced a stabilization plan. The other is a response to policy: anticipating a plan which would be based on a price freeze, firms implement particularly large price increases to reduce the impact of the imposed freeze. Both of these e>-ments share in the process, with the second playing a major role just prior to the introduction of the plan, which is usually widely expected. 4.50 The drastic reduction in inflation that follows the announcement of a stabilization plan may again have more than one explanation. It could be related directly to the point noted above: if price on the eve of stabilization plan is above optimum, it would be rational for the firm not to raise it further for a while. During the Bresser and the Summer Plans (but not the Cruzado Plan), this effect was supported by an initial decline in real wages. -96- 4.51 Alternatively, even a suboptimal price may be maintained by the firm, without rationing, when price controls exist and the firm regards this as a temporary situation. Another possibility, prevalent under price controls, is that prices (and the rate of inflation) are actually higher than those shown by measurement--pcices being increased, in fact, in well- known ways such as qutality changes, changes in the designation of products, chang;es in -edit terms, etc. Another element in an explanation, relevant oi.l2 to the Cruzado Plan, is the fixing of the nominal exchange rate, which contributes to stable prices for tradeables.38 Beyond these factors, there are some more genuine factors working toward price stability during these episodes. There is some automatic improvement in the fiscal stance due to a reversal of the Tanzi effect on real taxation when inflation greatly decelerates. Probably of much greater importance is a change of expectations--indeed, the very target of the stabilization plan. For a while, expectations do change, and such changes must be self-fulfilling. This was apparently true primarily during the Cruzado Plan, which must have been credible for a while. The credibility of the second attempt (the Bresser Plan) and of the third (the Summer Plan) must have become progressively smaller. The fact that the periods of greatly decelerated inflation got shorter from one stabilization plan to the next is most probably related to this gradual decrease in the ability of the plan to change expectations. Another indication of declining credibility is the fact that follcwing the introduction of the Cruzado Plan, real money balances increased substantially, and this phenomenon was not repeated with the two later plans. 4.52 Why did the inflation rate start rising again--quite dramatically, and within a short time--after each period of relative stability following the introduction of stabilization programs? Part of the answer lies in the earlier explanation of stability. The cushion created by preemptive price increases, on the eve of a program, must disappear within a short time. So must the willingness to obey price- control regulations, and to provide goods and services at suboptimal prices. 4.53 But this is relevant only within the general context of failing stabilization, which is explained by the wrong policy package. The inflationary momentum could be more permanently interrupted only if the fundamental policy elements which lead to inflation in the first place are changed; and when these changes ara regarded as durable and credible. This aspect was missing from all three stabilization episodes in Brazil. The budgetary deficit, and the increase in the public debt, continued during the implementation of the plans as before. So did existing constraints on monetary policy and its accommodating natare. While the reappearance of inflation took some time during the Cruzado Plan, which at first seemed credible, the repetitions of basically the same plan, with the same fundamental weaknesses, rendered the two other experiments highly incredible, and thus ineffectual, almost from the beginning. 38/ Under Brazil's circumstances, with the prevalence of quantitative restrictions on imports, this factor only forms a "contribution" to price stabilization, since the exchange rate--given the foreign price- -would not actually determine the local price. -97- 4.54 The black-market foreign-exchange rate served as a catalyst to the reappearance of rapid increases in inflation. In the normal pattern, the real exchange rate started falling (i.e., the currency started appreciating) after the introduction of th3 plan, and the balance of payments started deteriorating. In anticipation of real devaluation, demand for the holding of foreign assets increased, and with it the black- market premium. As in other experiences of high or hyperinflation, this increase in the black- (or free) market rate served in Brazil dS a signal that the policy was about to collapse. It led the government to devaluation and relaxation of price controls, and the public to reaction to the expected rekindling of inflation. 4.55 When rapid inflation resumed, following the end of the low- inflation stage, why did it (in all three cycles) reach higher levels than the "norm' prior to the start of the cycle? Part of the answer must be found in the generally-accelerating nature of inflation in Brazil--the reasons for which have been discussed earlier But specific reasons also help to explain why failed stabilization plans~ should lead to accelerated inflation,39 again via expectations. Failure of a government's stabilization policy reinforces expectations for persistent high inflation. Specifically, economic agents probably come to regard the particularly high rate of inflation on the eve of the policy implementation as a level which the government, following the failure (and without a renewed attack), will tolerate in the future. When the level of inflation is driven to a large extent by inflationary anticipations, these expectations would again become self-fulfilling. 39/ Brazil shares this phenomenon with the 1980s experiences of inflation in Israel and, to some extent, in Argentina. -98- V. !NCOME DISTRIBUTION AND POVERTY A. The Degree of Income Inequality 5.1 The well-being of the *average' citizen, as well as of the lowest- income recipients, is primarily a function of the income level of the country as a whole--of its productive capacity and the efficiency with which this capacity is used. But the pattern of income distribution, which represents the manner in which the economy's aggregate product is distributed among individuals, is also an important determinant of the well-being of individuals and groups. And the pattern of income distribution is a matter of particular concern to lowest-income recipients, as it implies a milder or a sharper degree of poverty. 5.2 Compared to other developing countries, available data suggest that Brazil's income is distributed in a highly unequal fashion. Admittedly, international comparisons in this sphere are somewhat shaky; measures of income distribution provide ambiguous indications, are not universally available, and differ in definitions, coverage, and timing among countries. Nevertheless, some general impressions may be gained from comparative data. These are preseiLLed in Tables 5.1 and 5.2, which cover only developing countries.40 40/ The data suggest that among the highly-developed economies (the roughly 20 countries classified in the World Bank publications as'high-income economies,* excluding the oil exporters), income is distributed much more equally than in almost all developing economies. -99- Table 651: INCOME DISTRIBUTION IN DEVELOPING ECONOMIES (shores of household incomes) Ratio of Income of Percentage of Aggregate Income of Top to Top Bottom Bottom Country Year Top Decile Quintile Quintile Quintiles Argentina 1970 36.2 50.3 4.4 11.4 Bangladesh 1981-82 29.6 46.3 8.8 6.9 Brazil 1972 5068 88.8 2.0 33.3 Brazil 1980 46.8 61.4 3.9 16.7 Costa Rico 1971 39.6 64.8 3.3 18.6 C6te d'Ivoire 1986-86 43.7 81.4 2.4 26.8 Egypt 1974 33.2 48.0 5.8 8.8 El Salvador 1978/77 29.6 47.3 6.6 8.6 Hungary 1982 20.6 36.8 8.9 6.2 India 1976-78 33.8 49.4 7.0 7.1 Indonesia 1978 34.0 49.4 6.8 7.5 Kenya 1978 46.8 80.4 2.6 28.2 Korea (Rep.) 1976 27.5 46.8 6.7 7.9 Malaysia 1973 39.8 65.1 3.6 18.0 Mauritius 1980-81 46.7 60.6 4.0 16.1 Mexico 1977 40.6 67.7 2.9 19.9 Panama 1978 44.2 81.8 2.0 30.9 Peru 1972 41.9 61.0 1.9 32.1 Philippines 1986 37.0 62.6 5.2 10.1 Portugal 1973-74 33.4 49.1 6.2 9.4 Sri Lanka 1980-81 34.7 49.8 6.8 8.8 Thailand 1976/76 34.1 49.8 6.8 8.9 Trinidad A Tobago 1976-76 31.8 60.0 4.2 11.9 Turkey 1973 40.7 65.6 3.6 16.1 Venezuela 1970 36.7 64.0 3.0 18.0 Yugoslavia 1978 22.9 38.7 8.6 6.9 Zambia 1978 46.4 81.1 3.4 18.0 Source: Data from World Bank (1989) World Development Report 1989, Table 30. For Brszil in 1980, data from CEPAL (1986), Antecedentes de Is diatribucidn del Ingreso: Brasil, 1960-83. Santiago, U.N. 5.3 Table 5.1 shows that income distribution in Brazil is indeed more skewed than in the large majority of countries. If 1972 indicators are used, both the share of the top decile (10 percent) and top quintile (20 percent) in aggregate income, and the ratio of the income of the top to that of the bottom quintile are higher in Brazil than in any other developing country. During the high-growth period of the 1970s, Brazil's income distribution seems to have become slightly more equitable. By the - 100- 1980 estimate Brazil is less of an outlier, although the shares of the top decile and top quintile in aggregate income, or of the lowest quintile, still place it among the countrie6 with the most unequal income distributions.41 The ratio of top to bottom quintiles falls in 1980 closer to the "norm," reflecting some relative increase in the income share held by the lowest- income recipients. Table 5.2: GINI COEFFICIENT OF INCOME DISTRIBUTION, DEVELOPING ECONOMIES Country Year Coefficient Bangladesh 1981-82 .39 Brazil 1983 .57 Chile 1971 .46 Colombia 1971 .57 Costa Rica 1982 .42 Cote d'Ivoire 1985 .55 Egypt 1974/75 .38 El Salvador 1976/77 .40 Honduras 1967/68 .62 Hong Kong 1981 .45 India 1975/76 .42 Iran 1973/74 .46 Korea (Rep.) 1982 .36 Malaysia 1984 .48 Mexico 1977 .50 Nepal 1976/77 .53 Pakistan 1979 .36 Panama 1970 .57 Philippines 1985 .45 Sierra Leone 1967/69 .59 Sri Lanka 1981/82 .45 Thailand 1981 .47 Trinidad & Tobago 1975/76 .47 Turkey 1973 .51 Source: Gary S. Fields (1989), "Changes in Poverty and Inequality in Developing Countries," mimeo. 5.4 These impressions are strengthened by the observations of Table 5.2, which presents the income-distribution pattern through a single synthetic index--the Gini coefficient (the higher this coefficient, the 41/ After 1972, a new series of income distributions is published in Brazil, reflecting a new survey. -101- less equal the distribution). Again, Brazil is among the countries with the highest degree of inequality (topped only by Honduras and Sierra Leone, and equalled by Colombia and Panama). 5.5 Although as we noted the data are imperfect--including, again, large differences in the dates for estimates from the various countries-- it does indeed appear that Brazil's income distribution is highly skewed, among the least equal in the developing world, and, a fortiori, in the world as a whole. 5.6 The absence of recent data makes it impossible to judge whether this aspect of Brazil's economy has imvproved or deteribrated in the 1980s. Some indicators will be discussed below, but they refer to shares in income of certain categories, rather than to individual household incomes. Available estimates for three years--1977, 1980 and 198142 show, by all the yardsticks that we have employed, some reduction in income inequality over these years. But the period covered is short, and does not necessarily indicate developments over the Last decade. B. Regional and Functional Distribution of Income 5.7 The inequality of income distribution in Brazil takes, to a large extent, the form of differences among the country's regions--although within each region substantial inequality remains with regard to the incomes of the urban vs. the rural sector, or the income distribution of individual households. 42/ In CEPAL, source cited in Table 5.1 -102- Table 5.3: RELATIVE PER CAPITA INCOMES, BY REGION, 1970, 1975, AND 1980 (index, average for Brazil 100) Region 1970 1975 1980 Brazil - Average 100 100 100 Northeast 39 38 41 North 56 47 63 Center-West 71 72 87 South 94 108 106 Southeast 153 149 144 Source: Thompson A. Andrade, 'Desigualdades Regionais: Tendincias de Longo Prazo," mimeo., IPEA, Rio de Janeiro, 19C9. 5.8 Table 5.3 presents the relatit,_ levels of (average) per-capita incomes in Brazil's five major regions, in ascending order of the income level. The South represents roughly the country's norm, the Northeast, the North, and the Center-West are the regions with below-average levels, and income in the Southeast far exceeds the average. The income level in the Southeast, the country's richest region, is about three and one-half to four times that in the Northeast, the country's poorest. Table 5.4: RELATIVE SIZE OF PER-CAPITA GDPIa IN THE NORTHEAST, 1980-86 1980 1981 1982 1983 1984 1985 1986 Ratio of Northeast to Brazil's Aggregate (Z): 43 45 50 52 55 55 58 /8 GDP at factor cost, at constant 1980 prices. Source: Thompson A. Andrade, ibid. 1 -103- 5.9 However, Table 5.3 data also show that the inter-regional differences narrowed somewhat during the 1970s. Equivalent data are not readily available for the 1980s. Table 5.4 presents partial evidence for this period, up to 1986, related only to income levels in the Northeast, the poorest region. The data show a steady and substantial, considering the short period surveyed, improvement. 5.10 The functional distribution of income is also relevant--in particular, the share of wage recipients vs. that of others. Data are, again, both scarce, and tentative due to definitional issues and methods of estimation. But some inferences concerning developments over the last decade, up to 1987, may be drawn from the indications provided in Table 5.5. It appears from the first three columns that, as a trend, real wages in the formal sector remained about constant, although annual fluctuations were often large.43 In view of the approximate constancy of per-capita income over this period, and witn only a minor increase in labor force participation, this would also indicate that the share of the aggregate wage bill in the economy's aggregate income remained approximately stable. Columns (4) and (5), showing respectively, changes in the shares of labor and capital, do indeed conform with this observation: excluding the starting year 1980, no trend changes in the sizes of the two shares are apparent. 43/ Large f'_ctuations of real wages may indeed by expected during a period of high, accelerating, and volatile inflation; and the annual data presumably mask much larger fluctuations over shorter periods. It may be noted, in particular, that all three indicators of real wages (columns (1) to (3)) show a large increase in wages in 1986--the period of the 'Cruzado Plan." -104- Table 5.5: INDICATORS OF REAL WAGES, 1980-87 (indices, 1980 = 100) Labor Share Profit Share Real Wages, in Income, in Income, Private Real Wages, Real Average Private Private Industry Total Formal Earnings, Formal Formal Sao Paulo Sector Formal Sector Se;tor Sector Year (1) (2) (3) (4) (5) 1980 100 100 100 100 1CG 1981 107 101 115 116 85 1982 114 107 132 125 62 1983 106 92 95 119 84 1984 99 87 87 109 93 1985 105 98 105 116 91 1986 117 105 129 123 98 1987 108 n.a. 105 120 92 n.a. = not available Source: M. Louise Fox and Samuel A. Morley, 'Who Paid the Bill? The Impact of Brazil's Macroeconomic Policies on Poverty in the 1980s," mimeo., December 1989, Table 4; based on IBGE/PNAD survey tabulations. 5.11 So far as these data go, then, no significant change has taken place in this important aspect of income distribution. But this evidence may well be biased, since it refers only to the formal sector. Data on "wages' in the informal sector ("wages' here includes incomes of self- employed labor), also indicate no trend change (but large fluctuations) over the 1980s. But these wages are only around 40-45 percent, on average, of wages in the formal sector. And the evidence suggests a significant increase of the relative size of the informal sector. This should then indicate increased inequality in the distribution of income within the urban sector, and a reduction in the share of national income going to labor. It is still possible, though, that this has been accompanied by a (net) migration of labor from the rural sector, in which labor income is particularly low, to the informal urban sector. In this case--and this would seem quite possible--the (presumed) substantial increase in the size of the urban informal sector should not necessarily indicate a worsening of income distribution in general, or a reduction in the share taken by labor income in particular. C. Indicators of Poverty 5.12 "Poverty" is a loose term. In this report, we will refer to several magnitudes which indicate a particularly low standard of living. -105- Table 5.6: INCIDENCE OF POVERTY/a 1981-87 (in percent of population in c ategory) 1981 1983 1985 1987 'razil Aggregate 24.8 30.9 25.4 23.3 Urban vs. Rural Urban 14.9 21.6 17.1 14.8 Rural 46.8 54.2 47.1 46.3 By Region North 18.0 24.8 18.0 16.8 Northeast 44.9 52.5 46.3 44.2 Southeast 13.5 19.4 15.5 13.0 South 16.6 25.1 17.4 17.3 Center-West 23.1 28.1 20.9 18.5 /a Defined as per-capita income below one-fourth of (real) 1980 minimum salary. Source: M. Louise Fox, 'Poverty Alleviation in Brazil 1970-87,n LAC Internal Discussion Paper, July 1990, Table 12. 5.13 Table 5.6 shows the share of the population living in households whose income is classified as sub-minimal.44 In this case, the 'poverty' level has been defined as a per-capita income of one quarter of the 1980 (real) level of the minimum wage.45 5.14 These data indicate a clear and persistent phenomenon: the regional co.icentration of poverty. Poverty, in the sense defined here, is common in the Northeast, where the term applies to about one-half of the = population, and it is infrequent in the Southeast. Not surprisingly, the incidence of poverty closely follows the pattern of regional income distribution noted earlier. 44/ Since this income is still above the level of Malthusian subsistence, below which the income recipient would perish, this must be a relative term, hence being in effect another representation of income distribution. 45/ If a family of five is assumed to be an average, this would mean a family income slightly exceeding the minimum salary. The per-capita indicator ignores potential economies of scale in the family (adult- equivalent data are not available). Hence changes in this indicator are more meaningful the lower the changes in family sizes of the population. -106- 5.15 The difference between urban and rural populations is just as persistent. Once more, one-third to one-half of the rural population is "poor," whereas the incidence of poverty in the urban sector is within the range of 10-20 percent.46 5,16 No clear inference emerges from changes in the pattern of incidence of poverty ove. the 1980s. Fluctuations abound, both for the aggregate and for the specific regional or sectoral categories. The year- to-year changes are sometimes violent--which probably reinforces the limited uses of these indicators. So far as the evidence goes, the 1980s have seen no reduction in the incidence of poverty, but no increase either. Since per-capita income changed little for the population as a whole, this means that incomes of the poor (as a group, rather than individually) stagnated along with other incomes in Brazil during the 1980s. 5.17 An important attribute of poverty (not shown in Taole 5.6) is its close association with family structure; specifically, a variety of indicators suggest that female-headed households tend to be poor. The share of such households in the total has continuously grown over the years. Thus, it was 5.2 percent in 1960; 11 percent in 1965; and 20 percent in 1980. For the metropolitan areas of Brazil, this ratio was 21 percent in 1987. The evidence shows that almost two-thirds of the children in female-headed households live in poverty (as defined before). 5.18 As might be expected, poverty is strongly associated with a low level of education. In 1980, close to 60 percent of heads of poor households had no schooling at all. This clearly suggests implications for education policy, which will be discussed later. Unlike the situation in many other countries, poverty in Brazil is not related to any significant extent to unemployment. 46/ To a large extent, these classification criteria overlap: the Northeast tends to be rural, whereas the Southeast tends to be urban, with the other regions falling accordingly inbetween. But the differences among regions persist when only urban, or only rutal, populations are compared. Similarly, the differences between urban and rural populat'ons persist within each of the five regions. Hence, the intersections of classifications yield extreme results. The incidence of poverty is extremely high (60-70 percent) among the rural population of the Northeast; and particularly low (around 10 percent) in the urban sector of the Southeast. -107- Table 5.7: INFANT MORTALITY RATES, BY REGION, 1980-86 (deaths per 1,000 births) Region 1980 1981 1982 1983 1984 1985 1936 Brazil 76.2 69.7 62.7 63.8 68.2 55.0 53.^ No.th 60.1 58.8 52.6 63.5 63.8 56.2 58.4 Northeast 120.1 109.3 91.1 101.1 113.7 81.6 74.7 Southeast 53.5 50.1 57.8 44.9 46.5 39.4 38.4 South 54.7 47.5 42.6 39.0 44.4 38.6 37.0 Center-West 59.0 56.5 39.4 62.9 47.1 45.5 40.7 Source: IBGE. Perfil Estatistico de Criangas e MAes no Brasil: Mortalidade Infantil e Saude na decada de 80. Rio de Janeiro, 1989. Corrected Statistics from the Civil Register of Brazil. 5.19 Other, partial representations of a particularly low standard of living do show some improvement over the 1980s. Thus, Table 5.7 presents infant mortality data over the decade (up to 1986), for the country as a whole and by regions. A substantial difference exists among regions, with a far higher mortality in the Northeast than in the Southeast. At the same time, mortality rates declined substantially over this period, for the country as a whole as well as for each of its major regions. -108- Table 5.8: ACCESS TO ADEQUATE SANITATTON SERVICES, 1987 (shares in Brazil's total population, in p)ercent) Per-Capital Income (in minimum salaries) kv Sector Total Below 1/4 1/4-1/2 1/2-1 1-2 Above Brazil Total 100.0 10.8 17.5 23.6 22.6 25.5 Adequate 50.4 0.9 3.5 9.9 14.7 21.4 Inadequate 49.6 9.9 14.0 13.7 7.9 4.1 Urban Total 75.7 4.2 10.1 17.8 19.6 24.0 Adequate 48.1 0.8 3.2 9.2 14.0 20.9 Inadequate 27.6 3.4 6.9 8.6 5.6 3.1 Rural Total 24.3 6.6 7.4 5.8 3.0 1.5 Adequate 2.3 0.1 0.3 0.7 0.7 0.5 Inadequate 27.0 6.5 7.1 5.1 2.3 1.0 Source: Calculated from data in Perfil Estatistico de Criangas e MAes no Brasil, 1987, Table 5. 5.20 Table 5.8 indicates another basic component of standard of living- -access to what has been defined as "adequate" sanitation services (either water or sewerage treatment). This reflects a survey conducted in 1987 for almost the whole population (over 32 million households). Access appears, once again, to depend on income and on employment sector: it is common among higher-income families, and almost non-existent in the lowest-income group. Within the same income group, it clearly depends on location-- being much more common in the urban than in the rural areas (in the latter, in fact, adequate sanitation is found in a minority of even higher-income households). D. The Impact of Government Policies 5.21 Much of the pattern of income distribution must be due to 'objective," given factors such as availabilities of production factors in the economy, available technology, or world prices. Much also depends on factors that may have been influenced by past policies, but are given at present--such as the distribution of human capital (skills), or the patterns of ownership of various other components of the country's wealth. In a large degree, however, current economic policies help to determine the shape of income distribution. To some extent, our analysis will discuss q1 - 109 - distributional pattern beyond that revealed by the data surveyed--partly because we shall refer to disposable income (actual or potential), as well as to wealth effects, whereas the data refer to gross incomes. 5.22 Many economic policies in Brazil in important spheres obstruct,.-by omission or commission--improvements in income distribution. ' occasion, government policies simply fail to address the issue of inconr stribution in a context in which they might have done so, and in which the international experience is rather different; on other occasions, policies that actively contribute to inequality may be observed. We shall discuss policies in the following major areas: taxation, fiscal subsidies, credit subsidies, trade policies, social-sector expenditures, and the impact of inflation. The inferences are based primarily on the analyses of major activities to be found in subsequent chapters. Taxation 5.23 The system of direct taxes would normally be a major government instrument used towards equalizing real incomes (or real expenditures). This would be done through a progressive personal income tax schedule; and through the imposition of a corporate income tax which would also most often be progressive (due, in this case, not to progressivity of tax rates but to the source of income). 5.24 In Brazil, both income taxes yield only relatively small revenues. Over the 1980s, this amounted to roughly 4 percent of GDP, and about half of this was collected through 'withholding taxes," in which progressiveness is limited. Instead, payroll taxes (as well as several minor o;her taxes) constituted the majority of direct taxes (6-7 percent of GDP). These taxes are, essentially, a form of prorortional taxation (although the poorest are exempt from them). Although as such it does not lead to a deterioration of income distribution, this form of taxation abrogates the important equalizing effect which incrome taxes have in most countries. Fiscal Subsidies 5.25 Several subsidies have been designed to improve income distribution. They are granted to activities in which either the producers are assumed to be poor or where typical consumers are assumed to be low- income. Sometimes, subsidies are given on a regional basis, favoring specifically the poorest region in the Northwest. 5.26 As the discussion in Chapter IX shows, however, subsidies have by and large gone ta the rich rather than the poor. Large firms (which are relatively capital-intensive), rather than small-scale producers, were able to avail themselves of the myriad of subsidies--especially those which reduced tax liabilities. In conjunction with trade and credit policies, fiscal subsidies have also imposed entry barriers on many industrial (and some agricultural) activities, leading to the widespread existence of transfers and economic rents--a fact which no doubt contributes to income inequality. -110- Credit Subsidies 5.27 The same comments apply here as well--perhaps even more strongly. Credit subsidies--particularly for agriculture--so predominantly to large- scale producers. Moreover, in combination with fiscal subsidies these subsidies support land acquisition by these enterprises, which leads to large land concentrations rather than the diversification of land ownership among many smaller-scale producers. In industry, subsidized credit for investment tends to be enjoyed primarily by large-scale, incumbent producers. By creating in effect an entry barrier, subsidized credit-- like fiscal subsidies--tends to lead to monopolistic organizations and monopoly rents, benefiting specific, high-income groups, and hurting the relatively poorer local consumers of the goods produced by these enterprises. Trade Policies 5.28 The current restrictive trade policies tend to hurt income equality in two different ways. Much like fiscal and credit policies, these trade policies encourage incumbents, the stifling of competition, and the creation of large-scale economic rents--enjoyed by the relatively richer part of the population (including, it might be noted, the more prosperous part of the labor force, such as organized urban labor). Beyond that, the allocational impact of the trade regime must be detrimental. Studies of internationa'l experience clearly suggest that a restrictive trade regime, in countries like Brazil, usually leads to the expansion of capital-intensive activities and the contraction of labor-intensive activities (in comparison to activ ties under a free trade regime). These policies therefore increase the share of capital, and reduce that of labor, in the country's income. In Brazil, specifically, freer trade would most likely give relative encouragement to (most) agricultural activities, and increase relative as well as absolute incomes in the rural sector. Sccial Expenditures 5.29 Social expenditures are normally a major vehicle for improving the real incom.e distribution. Beyond their immediate impact on the redistribution of current income, these expenditures are of great significance in determining the future, long-term pattern of poverty and income distribution. In Brazil, this is probably the area in which government policies have failed most profoundly. 5.30 Nowhere is this more apparent than in education. The poor receive relatively little benefit from Brazil's spending on education, both within and across education levels. Thus, to cite an important instance, public spending on the federal university system, wi'ich is tuition-free, results in a direct transfer of about US$ 5,500 to each undergraduate student (the total amounts to nearly US$1 billion per year). This is appropriated, in essence, by the well-off part of the population. one-half of all public- uni-ersity students (but only 6 percent of the general population) come from families with incomes which exceed 10 minimum salaries. Only 1 percent of this student body come from the bottom quintile of income distribution. This subsidy is, by its nature, granted to individuals whose prospective incomes should be particularly high. -111- 5.31 Rather than contributing significantly to social mobility for the poor and equalizing opportunities, the Brazilian education system appears to act as a screen, filtering out children from low-income families and allowing children from higher-income origins into higher education, where the private returns are highest. At the primary and secondary education levels, public expenditures are low, resulting in a poor quality of schooling, high drop-out rates, and the very low educational level attained by much of the population. Public spending on education is, moreover, particularly low in the Northeast. State spending there is about 60 percent. per primary-school student, of that in the Southeast. The disparity is even much more pronounced for students in municinal schools: within the Northeast, spending per primary-school student is three times higher in state-run schools, virtually all of which are in urban centers, than in schools run by municipalities, most of which are rural. 5.32 In health, much the same inferences can be drawn. Public spending benefits the poorest least, although it is they who tend tc suffer most from infectious and non-communicable diseases. Until recently, only formal-sector employces had access to health care funded by social security. Currently, these employees are bearing an increasing share of the costs of their health care, but they continue to receive a substantial health-care subsidy from the government. Public spending on health is heavily concentrated in the more-developed regions of the country. For instance, the canacity of hospital beds and outpatient facilities in the South and Southeast is twice that in the country's other regions. In rural areas, and in the Northeast in general, the poor are still heavily under- served by the public-health network. The Burden of Inflation 5.33 In general, empirical as well as a priori inferences about the impact of inflation on the distribution of current income are not easy to come by. But in one respect, inflation clearly imposes a relative burden on the poor, via a tax on their wealth in particular. 5.34 The inflation tax is a wealth tax imposed on the holders of the most liquid wealth--money. Although the part collected by the government is only the fraction of the tax imposed on the monetary base, all money holdings are subject tc- the tax (the difference being the gains of commercial banks and recipients of credit). In Brazil, most money is indexed (formally or effectively), and is therefore not subject to the inflation tax: the tax is paid only by holders of M1, the non-indexed form of money. It may safely be assumed that the ratio of M1 to income is far higher among the poor than among the rich: in general, the ratio must increase as the income level declines. Thus, a disproportionate part of inflation tax is imposed on low-income recipients. This tax must, indeed, be much more regressive than any of the explicit taxes in the system. 5.35 Strikingly, this broad survey of the impact of economic policies on income distribution and poverty indicates that the changes in policies called for on equity grounds do not conflict with the policy implications mandated by efficiency concerns: by and large, efficiency and equity point to the same policy requirements. This congruence of policies was also -112- noted when the need for macro-economic stabilization was contrasted with the requirements of structural reforms called for on efficiency grounds. Overall, therefore, not many tradeoffs have to be faced in designing policy changes in present-day Brazil. -I 4 ~~~~~~~~~~~~~~~~~~~~I -113- Part 3: ANALYSES OF MAJOR ACTIVITlES In this part, analyses of major sectors or major sphetes of economi( activity will be presented. It is not a comprehensive discussion--sectors like housing and construction, or energy and power, are obvious omissions; nevertheless, a predominant part ot the economy is covered. The six chapters of this part will be devoted to, respectively: the budget; public enterprises; the external transactions; industrial and agricultural policies; the financial sector; and social expenditures. -114- VI. THE BUDGET: MECHANISM. EXPENDITURES, AND TAXES 1. Expenditures A. An Overall View 6.1 The budget is the most fundamental and crucial instrument of economic policy. It represents the government's intervention in economic activity, expressing the system of priorities set through the process of collective decisions. 6.2 The overall size of Brazil's budget reflects a degree of intervention which is not unusual; as a share of GDP, Brazil's government expenditures, or revenues, would fall close to the median of countries at its level of development. Brazil's budget deviates from that of others in two main respects. First, it is highly inflexible--a phenomenon which has become progressively more severe in recent years. Second, there is a very large gap between government intervention as reflected in the budget and in reality. Table 6.1: BUDGET COMPOSl-^Zi!, BY "COMPRESSIBILITY", 1980-88 (in percentage of GDP) 1980 1984 1985 1986 1987* 1988* A. Gross Tax Receipts 24.2 21.7 22.2 25.5 22.6 22.1 B. Current Expenditures 21.8 21.0 22.9 24.3 23.9 25.2 B.1 'Non-compressible"/a 15.4 16.8 18.5 19.9 19.5 21.2 B.2 "Compressible" 6.4 4.2 4.4 4.4 4.4 4.0 * Preliminary /a "Non-compressible": Mandated transfers, (real) interest payments, and social remittances. Source: National Income Accounts. 6.3 Table 6.1 indicates the major components of the budget. A comparison of lines A and B in the table shows that the public sector's saving pattern has radically changed over the 1980s. An excess of gross tax receipts over current expenditures of 2.4 percent of GDP in 1980 has turned into a deficit of 3.1 percent in 1988. This change has already been discussed at some length in earlier chapters. Here we want to note, in particular, another aspect: given that fundamental legislation is binding, the maneuverable part of current expenditures had, by 1988, declined to 4 percent of GDP, or some 18 percent of gross tax revenues. As government -115- investment expenditures have already declined drastically--by some 80 percent in the last two annual budgets alone (from 1988 to the proposed budget for l990)--the room for change under existing constraints appears to be almost nil. A significant reduction in the size of expenditures, or a change in their pattern, would inevitably require fundamental legislative changes--including the amendment of constitutional stipulations. Table 6.2: STRUCTURE OF CENTRAL GOVERNMENT CURRENT EXPENDITURES, BY MAJOR FUNCTIONS, 1978-86 (in percentage of GDP) General Expenditures Real Public Social on Production Interest Services Defense Expenditures/a Activities/b Payments Total Year (1) (2) (3) (4) (5) (6) 1978 3.7 .9 9.5 3.8 1.5 19.4 1979 2.4 .7 9.4 4.2 1.6 18.3 1980 3.9 .7 8.8 5.2 1.0 19.6 1981 3.4 .7 9.3 4.9 1.4 19.7 19A2 2.5 .9 10.0 4.5 2.3 20.2 1983 2.4 .8 9.5 4.9 2.9 20.5 1984 2.2 .8 8.6 2.8 2.4 17.8 1985 3.0 .8 8.2 2.8 4.7 19.5 1986 3.1 .9 9.1 4.2 5.4 22.7 /a Primarily education, health, and social and security welfare. /5 Primarily fuel and energy, agriculture and forestry, mining, manufacturing and construction, and transportation and communication. Note: Due to differences in sources and coverage, total expenditures recorded in this table vary somewhat ftom those of Table 6.1. Source: I.M.F., Government Financial Statistics. 6.4 Table 6.2 shows the composition of the central government's current expenditures, by major function. It appears that from the late 1970s to 1986--the latest year for which these data are available--little change has taken place either in the total size of expenditures (as a ratio to GDP) or in its composition. There is one exception, as real interest payments have grown dramatically, and total expenditures grew by approximately the same amount. This increase was due, of course, to the increase in the external debt (and in the interest rate due on it) until the early 1980s, and to an increased interest bill for the domestic debt thereafter. Social expenditures have, throughout remained the major -116- component of central government expenditures, amounting to approximately one-half of tne total. Of these, in turn, the majority--about two-thirds-- have usually been payments of the social-security system. B. The Bu4getary Process 6.5 The process of planning and implementing the budget has been undergoing fundamental transformation--particularly during the last two years, following he adoption of the new Constitution and the establishment of Congress as a ma,or participant in the decision-making miechanism. By and large, a substartial improvement has been in progress, with the inclusion of the quasi-fiscal deficit in the budget as perhaps the most important element. The process is still not fully articulated, and it keeps evolving. (a) Congress. The new Constitution establishes Congress as the primary source of formulating the budget. Congress has to adopt, sequentially, three pieces of legislation: i. A multi-year budgetary plan, which establishes a general budgetary strategy--objectives, targets, and basic principles. ii. A budget-guideline law. This legislation defines the guidelines for the preparation of the annual budget: goals, priorities, constraints (such as the size of the deficit), major components, changes in the laws, or lending policies. iii. The annual budget law. This is the full budgetary specification, which has three components: (1) the fiscal budget; (2) the investment budget for state enterprise; and (3) the social-security budget. The annual budget must adhere to the guidelines law, as Congress itself may shift allocations among current expenditures, or among investment activities, but it is not permitted to change either total expenditures or major aggregates (such as total current expenditures or total investment). Thus, while the detailed system of preferences is represented by the concrete budget, fiscal policy is determined by the guidelines law. 6.6 Partially integrated with the fiscal budget document is a 'credit budget", which represents some of the government's major subsidization schemes--particularly in the agricultural sector. As will be noted later, thoigh, despite recent progress, this representation is grossly biased downward, as major resource transfers through credit activities do not yet find an appropriate and explicit expression. 6.7 As of early 1989 (with the Summer Plan), the budgetary process has increased in complexity, and become a more continuous process. By law, monthly expenditures are bound by two constraints. First, they are determined by the annual budget (with adjustments for seasonal factors), with a correction being made for inflation (the latter is predetermined, being assun.ed at the start of the year). Second, expenditures (excluding debt service) should be equal to revenues. The assumed inflation has been below the level actually expected at the time, and certainly below the realized level (for calendar 1989, it was assumed at 10 percent per month), whereas nominal revenues have increased at approximately the actual level of inflation. As a result, monthly expenditures have been below revenues. These "surpluses' are, in turn, allocated by Congress through a series of -117- quart e l'y up, 1lt.::t ary budgets. Thus, beyond the annual budget, three sepa Iat e !-:get s ale prep ared durirng the year. It is likely that this procedure will be tollowed again in 1990. 6.8 The budgetary pLocedures and mechanisms have been undergoing various transformations in recent years. By and large, these have represented improvements. The monetary (credit) budget has been partially incorporated into the fiscal budget, while government's capacity to bypass the budget has de(lined drastically, either through borrowing from the Central Bank (now prohibited) or through the administration of subsidies through the banking system. Nevertheless, very large elements of government fiscal intervention are non-transparent. In the following discussion of subsidies, this will be indicated more specifically. C. Subsidies 6.9 A large amount of subsidization is involved in government activity in Brazil, and only a fraction is explicitly expressed in the budget: various subsidization schemes find no representation in it. 6.10 Subsidization is provided primarily--but not exclusively--in the agricultural sector.47 Subsidization is also involved, explicitly or implicitly, in the foreign-trade regime.48 D. Other Issues of Budgetary Expenditures 6.11 Beyond subsidization, three major categories of budgetary expenditures deserve discussion: the state enterprises; intergovernmental transfers; and employment compen;ation in the public sector. State Enterprises49 6.12 The state enterprise sector in Brazil is of major importance. It has 178 federally-owned companies, 90 percent of which (by sales) are in the electricity, petroleum, steel, telecommunications, iron mining, and transportation (railways and ports) sectors. The nine largest holding companies in the country are all state enterprises, and they generated about one-fifth of total fixed investment in the economy during the last decade. 6.13 The process of observing, regulating, and providing budgetary control of the sector has improved significantly over time. A major change took place in 1979, when the Secretariat of the Control of Public Enterprises (SEST) was established. This agency provides the government's supervisory mechanism, through: the approval of each enterprise's budget 47/ But this statement should not be read as an ind.cation that the agricultural sector as a whole is, on average, subsidized; see the discussion of agricultural policies. 48/ For a detailed discussion of subsidies in agriculture and in trade, see, respectively, Chapters IX and VIII, in which policy recommendations are also presented. 49/ Chapter VII is devoted fully to this subject. It includes policy recommendations. -118- (including ceilings on expenditures and on domestic and external borrowing); the allocation of foreign exchange; the approval of specific investment projects; and the review and approval of changes in each company's legal status. 6.14 The improved supervisory mechanism has led to a cl(ose match between actual budgetary performance in this sector and planned budgets. Substantial lags remain, however, mainly due to inflation, which influences these gaps in a variety of ways. Budgets tend to assume an inflation rate below the realized level (and probably even below the level actualiy anticipated at budgeting time). Price freezing, or substantially lagging, tends to be used on products and services provided by this sector, as a means of combatting inflation. By and large, the sector's deficits have been larger than budgeted--although not dramatically so (by some 20-25 percent). Over the last five years for which data are available, 1983-1987, the aggregate deficit of the sector amounted to some 2 percent of GDP. Intergovernmental Transfers 6.15 Transfers from the Federal government to the states and H municipalities have been large. In 1987 they amounted to some 27 percent of total Federal treasury receipts, or 3 percent of GDP. The new Constitution has led to a further increase in the share taken by local governments; local governments' share in receipts from income tax and from the tax on industrial production has increased from 33 percent to 44 percent, while various other taxes (such as the electricity tax--IUEE), which had been shared by the Federal and local governments, have been incorporated into the value-added tax (ICM) whose receipts go entirely to the states. In addition, revenues from taxes imposed only by local governments have increased substantially. The share of the latter in total public sector expenditures has thus, in recent years--and particularly since the 1988 new Constitution--assumed a substantially higher level. The changes in tax shares of the government at the three levels, prior to the new Constitution, are summed up in Table 6.3. -119- Table 6.3: TAX REVENUES BY LEVEL OF GOVERNMENT Federal Period Government State Municipalities Total a. In Percentage of Total Revenues 1979-81 50 35 15 100 1982-84 50 35 15 100 1985-87 44 39 17 100 b. Index, 1983 = 100 1979-81 102 109 109 105 1982-84 98 106 105 102 1985-87 102 142 151 122 c. In Percentage of GDP 1979-81 7.5 5.2 2.2 14.9 1982-84 7.6 5.3 2.2 _5.1 1985-87 6.6 6.0 2.7 15.3 6.16 The new Constitution also assigns new functions, particularly in education and health, to states and municipalities. But it is the Federal government which will continue to collect revenues in the form of earmarked taxes, thus making necessary a continuous transfer to local governments. 6.17 The recent changes in the revenue-sharing formula arising from the Constitution have exacerbated the macroeconomic problem, by not having addressed the expenditure decentralization issue fully. The problem is now that the richer states and municipalities are receiving much higher levels of revenues while having unchanged expenditure responsibilities. The pressure on state resources has been increased by the decree that froze state debts at their December 1988 levels. This pressure is being offset, to some extent, by the distribution of discretionary transfers, and the negotiation that accompanies the determination of debt rollover ratios. This continuing ambiguity of expenditure responsibilities, and the related transfer and debt rollover negotiations, are de facto structural impediments to fiscal adjustment. Transfers will necessarily be a part of Brazil's attempts (driven by political and not necessarily by efficiency considerations, as is true in all Federal systems, notably Canada, Australia, etc.), to address inter-regional equity considerations, and the critical issue is that these transfers should be based on explicit and -120- transpar ent , i itt r :a. To a large extenlt, present legi.l1ation already p)ovide~. for it. Thus, the States Participat ion Fun-d assigns, by law, 85 percent of its tiansters to the North, Northeast, ar_d Cvnter-Weqt, and 15 percent to the re: ai ii g t wo regiions. Transters from the V-uni(ipalities Part ic-i)at ion Futnd are allc(ated through a formula the com,onen t s of whicih c are, )r irat ily. po;ulation, area, ana per capita in o me. 6.18 Recommendations. The expenditure de-rntralization and related issues of transfers and debt roll-overs are too important to be treated through ad hoc measures from year to year, and indeed are major structural impediments to medium-term fiscal adjustment. While the details of a recommended set of measures are not yet available,50 it is clear that the following principles will have to be implicit in any final measures: (a) Transfers should be based soley on explicit and transparent criteria and the role of discretionary transfers should be eliminated or reduced to an insignificant share, for emergencies only; thus, all inter- regional equity issues should be directly addressed through the determination of the criteria for transfer distributior. (b) A clear definition of the decentralization of expenditures must be agreed upon urgently, with a mechanism for reviewing these criteria over a one-to-two year period, to fine-tune them if necessary. (c) A uniform policy for all states should be established that requires tl.em to pay all interest and a uniform proportion of principal on their outstanding debts; any offsetting resources for poorer states must be addressed entirely through the non-discretionary transfer criteria. (d) Clear guidelines for the determination of state and municipal creditworthiness must be defined and access to new borrowing by all states and municipalities must be based entirely on these criteria. (e) Mechanisms to encourage fiscal efforts on the part of states and municipalities should be put in place, and non-discretionary transfers should include an element of matching funds. Compensation and Employment Policy 6.19 Table 6.4 shows that salaries of Federal workers tend to be high in relation to salaries in the private sector--partly reflecting higher skill levels. Thus, for instance, employees receiving three minimum salaries or below amounted to about 64 percent of the total in the urban formal sector and only 33 percent in the federal public sector, while employees receiving 7 minimum salaries accounted for about 32 percent in the public sector and only 10 percent in the other. The difference between the public and the private sectors would actually appear even more substantial if the latter is separated out from the ~urban formal sector,n which is itself a combination of the two sectors. Within the public sector, salaries are particularly high in public enterprises--extremely so among those in the financial sector. Salaries in general administration are lowest, not much higher in general than in the private sector. 50/ A forthcoming World Bank study of State finances will address these issues. - 121 - TaMt' *.4: FREQUENCY DISTRIBUTION OF EMPLOYtMENT BY TOTAL REMUNERATION, 1985 (in minitnum salaries) Sector < 1 1-3 3-5 5-7 7-10 10-15 15-20 20+ Unknown Total Federal Public 2.5 30.8 19.8 11.8 12.4 10.3 4.6 4.6 2.9 100 Urban Formal 8.0 56.0 15.2 6.4 4.4 3.0 1.3 1.4 2.9 100 Source: ATb/SES, "Emprego e Salario no Sector PCiblico." 6.20 The compensation system in the Federal government is not uniform. The compensation system distinguishes among eight groups: (a) productive enterprises; (b) departmental entities (ETG); (c) financial enterprises (public banks); (d) regional utilities; (e) the social security and health system (SINPAS); (f) foundations (research institutes and universities); (g) autonomous institutes and (h) direct administration--the core of the civil service. Each group's share of total public-sector employment is shown in Table 6.5. -122- Table 6.5: FEDERAL EMPLOYMENT BY ADMINISTRATIVE GROUP, 1985 Administrative Number of Employees Share of Total Group Thousands (percent) Enterprises Productive Enterprises 582 32 Departmental Entities 30 2 Financial Enterprises 179 10 State Utilities 132 7 Total Enterprises 916 51 Central Government Autonomous Agencies (including SINPAS) 438 25 Foundations 100 6 Direct Administration 329 18 Total Central Government 867 49 Total Federal Sector 1,783 100 Source: MTb/SES, Ibid. 6.21 The compensation system in the Federal government is not uniform. First, it differs among the administrative groups. Part of the labor force is subject to the Consolidated Labor Code (CLT), which governs the public enterprises, the autonomous agencies, and most employees in foundations. This is a legal framework which sets standard work rules and benefits, and provides for participation in the national health insurance, social security, and unemployment insurance. Under the new Constitution, most employees in direct administration are, on the other hand, estatutarios- -tenured employees, not covered by the CLT and benefitting from a separate federal pension scheme. 6.22 Within the central government, compensation is now divided more or less equally between a "cash base" and "salary supplements." The latter have developed with the process of inflation--in effect, as means of circumventing general guidelines. They are not subject to any policy regulation, and vary widelv among the different ministries. Salaries have, in recent years, tended toward a higher degree of dispersion, as a result of the 'supplements." The decentralized bargaining structure and the consequent expansion of supplements, have led salary scales for employees with roughly equivalent levels of education and experience to vary substantially. Some of this variance may be explained by specialized knowledge and market forces, but much cannot be attributed to "objective" -123- fao ti. 'IUh, tor instance, the dispersion is particularly high among y :h i It t, ondary-school education--a category in which spetA l. l,'.i it and expert ise are not common. To illustrate: start ing SdaaL1t, l% ste.onlary-school. giaduates are 4.7 times the average of this grou,, in In Auditing Department, and 8 times the average in the Federal polii e. A:t-olet illustration: average remuneration in the Ministry of Fit.ant- i,, i . times the level in the Ministry of Education, although the averag* 1'.lt oi er,ployee education is roughly the same in the two 6.23 Tn sum, it appears that the compensation system in the central governm.ont is now chaotic. This causes the very high dispersion noted above; it also leads to the wrong incentives so far as attracting a desirable labor force, or deflecting others, is concerned; a.ad to an overs-endir.g on compensation, given the size of the labor force. 6.24 Employment in the central government is, on the surface, highly regulated and centrally controlled. Each department is authorized to have a total number of positions at each grade, and hiring for most positions is strictly regulated. Permanent positions in the civil service can be exnanded only by authority of Congress (through the budgetary process). In fact, since January 1982, no new permanent positions have been authorized. However, the system is broadly circumvented through two major avenues. "Temporary' labor hiring is not regulated nor closely monitored, and many of the 'temporaries" have stayed in their positions for years. And there is a system of 'secondment" from a ministry's public enterprise to the central government, wh.ch may again become, in effect, permanent rather than transitory (and may, in addition, contribute to the distortion of salary scales). By using these loopholes, employment in the central government increased during the 1980s--a period of presumed freeze--by about 150,000 or roughly one third of its level at the beginning of the decade. Twice during the decade, the government in effect created formal positions for the long-term "temporaries", to improve administration and budgeting. And in 1988, the new Constitution granted tenure (making them lestatutAriosn) to all employees employed for more than five years. 6.25 Recommendations. In the short run, the fiscal component of a stabilization policy should call for reduced compensation in the public sector. This would be achieved through a combination of some lowering of real remuneration--concentrated primarily among categories of workers which have enjoyed a substantial increase of real wages in recent years; and by a reduction of the size of the labor force. The latter could be implemented through a freeze of all hiring, allowing natural attrition to work effectively, combined with a concerted program of voluntary departure built on the lines of the program announced by the goverrment in April 1988; ard through a modest amount of firing, which the Constitution authorizes the President to undertake. In the local governments, this reduction would be induced by the aforementioned contraction of transfers from the federal government. 6.26 In the medium term, a thorough reform of the compensation system should be undertaken, with the aim of improving the mix of public-sector employees, increasing transparency, increasing productivity in the sector and, eventually, performing more and better services with lower total spending. The government should elaborate and promulgate a set of principles to help simplify the structure of compensation policy and to -124- meet the requirements of the new Constitutioui. These principles should cover the following topics: the definition of criteria for job grading; the relationship of job grading to compensation and to labor-market trends; procedures for the revision of job grading; incentives for performance; advancement within grade and prom'otion; training and career developmzent. New salary scales should be designed in accordance with the uiew principles. The number of scales should be drastically reduced, and most supplements should be eliminated. 2. Taxes A. An Overall View 6.27 Brazil's taxation system has failed in two fundamental respects. First, it raises total revenue which, given the limited size of potential reduction of the government's expenditures, is substantially below the amount required for macroeconomic stability--and which is, incidentally, quite low from an international perspective. Second, the system's allocative and distributional impacts are often undesirable, in comparison with other feasible taxation systems. These basic deficiencies have characterized the system for some time, but the system has deteriorated in recent years--at least so far as the amount of revenies is concerned (see Table 6.6). Table 6.6: TAX REVENUES, 1970-88 (in percent of GDP) 1970 ;975 1980 1982 1984 1988 Direct Taxes 9.2 11.3 11.2 12.6 11.2 9.7 Indirect Taxes 16.8 13.9 13.5 12.5 10.2 10.2 Total Tax Revenues 26.0 25.2 24.7 25.1 21.4 19.9 Source: National Accounts, IBGE. 6.28 Between 1970 and 1982, indirect tax receipts fell substantially, but this was roughly matched by an increase in direct taxes. From 1982 onwards, both categories declined, leading to a decline in aggregate revenues on the order of 5 percent of GDP--about one-fifth of overall revenues in 1982. Thie is about equal to the size of the budgetary operational deficit in 1989, or to the decline in public investment during the 1980s. -125- 6.29 Brazil's tax structure suffers from three basic flaws: (a) It is not simple. Increased complexity makes tax administration cumbersome and expensive. The system is opaque, both to taxpayers and to lawmakers. And it has increased both the incentives for and tL. likelihood of successful tax evasion. (b) It is not efficient. Some allocative distortions must be created by practically any tax system, but Brazil's current system contains numerous distortions wnich should be easily avoidable. (c) It is not equitable. There are two widely accepted criteria of equity, or fairness for a tax system--that taxes be applied uniformly to all economic agents under similar circumstances, and that the system be progressive. Both are violated by the Brazilian system. B. Indirect Taxes 6.30 Close to two thirds of total revenues from indirect taxes in Brazil comes from two types of value-added tax (VAT) (see Table 6.7 for indirect tax revenues). The reiative importance of these taxes must increase further with the new Constitution, which has incorporated into one of them (the ICM) several excise taxes on specific industries. We shall confine the present discussion to thes.. two taxes, with only a scant reference to other indirect taxes. Table 6.7: REVENUES FROM MAJOR INDIRECT TAXES, 1970-1908 (in percentage of GDP) 1970 1975 1980 1988 IPI (VAT on industrial production) 4.4 3.5 2.1 1.8 ICM (VAT on circulation of production) 6.9 5.5 5.0 4.6 Specific Taxes 1.7 1.2 .7 .7 Trade Taxes .7 1.0 1.2 .4 Other Indirect Taxes 3.1 2.7 4.5 2.7 Total Indirect Taxes 16.8 13.9 13.5 10.2 Source: Secretaria da Receita Federal, and IBGE, National Accounts (April 1989). 6.31 One of the VATs, the IPI, is a tax on industrial products levied by the Federal government; the other, the ICM, is a state tax imposed on the circulation of (mostly) industrial and agricultural goods. As we can see from Table 6.7, yields from both taxes have declined substantially (as percentage of GDP) throughout the years. -126- b.32 The two VATs are not only separate, but for each the tax !a'5e includes the other. Thus, the effective tax rate of each is higher th:an the posted rate. On the other hand--and of much more importance- -exclusions from the tax (either or both) are so frequent as to, make the posted rates resemble only little the actual average tax rate:'. Ex lusion from taxation may take two forms. One is the exemption of an activity from payment of the VAT. This applies to almost all industries whi(h are treated favorably in the tax code. In this case, though, the activity concerned is not reimbursed for the taxes imposed on its inputs. Thus, in effect, it remains subject to taxes on inputs, although exempt from taxes on its value added. The other form of exclusion, which applies to industrial exports in general and to production within the Zona Franca de Manaus (ZFM), grants a zero rate under the VAT. In this case, the activity is not subject to the VAT, and it is also entitled to a refund of the VAT paid on its inputs. The IPI 6.33 This tax is imposed only on the manufacturing sector: agriculture, minerals, and services are exempted. Within manufacturing, the tax rate as a rule differs among branches more than in VAT schemes in must other countries. Rates vary substantially, with the highest rates being paid by the tobacco and alcoholic beverages industries. Exemptions from the tax, or partial refunds, have increased during the years--which partly explains the radical decline in IPI revenues. In fact, about 70 percent of total revenues from this tax come from just four sectors: tobacco, beverages, weapons, and perfumes. The ICM 6.34 This form of VAT has been the most important single tax in Brazil; - and revenues from it have declined by less than those of most other forms of taxation. It is levied on sales of goods at all stages of production, and excludes in principle only products subject to specific taxes--mainly, since the reform under the new Constitutic-, services. In agriculture, the tax paid on inputs is not generally reimbursable, but numerous exceptions have been introduced, such as those for specific intermediate imports, fertilizers, pesticides, and similar inputs. 6.35 The tax rate was, until recently, 17 percent for intra-state transactions, and 12 percent for most inter-state transactions. In 1989, with the new Constitution, the uniform 17 percent rate on intra-state transactions has been transformed into a multi-rate scheme, in which goods heavily consumed by the poor are taxed at a rate of only 12 percent, whereas "luxury" goods may be subject to a rate of 25 percent. Since the tax is imposed on a base which includes the tax itself, the actual level of a rate posted as 17 percent becomes 20.7 percent. 6.36 The VATs, and the indirect tax system in general, thus have the following deficiencies: (a) Exclusion. Indirect taxes fall differentially on the agricultural, industrial, and services sectors. In sum, the IPI base covers ahout 30 percent of GDP (just manufacturing), and the ICM roughly 40 percent (manufacturing and agriculture). -127- (b) Exemptions. A very wide array of exemptions is granted, mostly through fiscal incentives. (c) Multiplicity of tax rates. Even within the covered activities, and disregarding the exemptions, both VATs, but mostly the IPI, impose multiple tax rates. Most revenues are derived from just a small number of heavily-taxed industries. (d) Cascading Taxes. The two taxes are referred to as VATs. But instead of a uniform levy on value added, with minimum distortions, both taxes (in particular the ICM) operate more as cascading sales taxes than as VATs. (e) Complexity. Noted above, this point deserves to be stressed. For example, the principles of the taxes, which with normal VATs occupy just a few pages. occupy a large volume of regulations in Brazil. This increases the cost of collection for both the government and the taxpayer. It also increases the incentives and opportunities for tax evasion. (f) Erosion by inflation. The acceleration of inflation has led to serious revenue losses (the "Tanzi effect"). This factor has become less important recently, however, with new mechanisms and regulations which have substantially reduced the collection lags. 6.37 Of other indirect taxes, we shall note only the taxation (or subsidization) of oil and its products. Overall, oil products are subject to a tax at a weighted average rate of over 10 percent. This consists, however, of quite high taxes on some products (particularly gasoline), and the subsidization of others. Specifically, two important oil derivatives- -fuel oil and Liquified Petroleum Gas (LPG)--are sold at prices substantially below the international level (or the cost to Brazil). The latter's subsidization is based on the assumption that the subsidy is targeted to the lower-income groups--an assumption not justified by the facts. 6.38 Recommendations. Tax reforms following the new Constitution have mostly been in the right direction, especially the integration of almost all excise and special taxes into the ICM and the reduction in the complexity of the tax system. But the fundamental problems just enumerated basically renain. To respond to them, we suggest the following proposals: (a) widen the base of the ICM, to include goods and services currently excluded, with at most very few exceptions; (b) widen the base of the IPI, to include all industrial activities currently excluded, and grant true zero-rating to agricultural and services inputs into industrial-sector production; (c) eliminate all current exemptions--mostly granted through fiscal incentives and the ZFM exemptions--within the broad tax base; (d) allow full credits for all inputs into production, to impose the taxes indeed on value added; -128- (e) simplify taxes on exports and imports, by zero-rating all exports and in.posing the full VAT rate on all imports; (f, implement a flat rate for the IPI on all industries, except the four which at present are particularly highly taxed. The actual tax rates (both the rule and the exceptions) should be determined so that total revenues from IPI increases from 1.8 percent to 3 percent of GDP; (g) intensify the measures undertaken recently to reduce collection lags; and (h) increase the average tax on oil products by raising the prices of fuel oil and of LPG to reflect their respective costs. 6.39 In addition, include the trade regime reforms suggested in Chapter VIII beyond recommendation (e) above. 6.40 In conjunction, these measures would reduce substantially the distortions built into the major indirect taxes, and lead to the development of a more neutral indirect-tax system; and, of prime short-term significance, generate a substantial increase of the tax revenues. Assuming that the present 17 percent ICM rate is retained, and that the IPI rate is established as recommended in (f) above, revenues should increase by about 1.5 percent of GDP. The increase in the average tax on oil products would yield about 0.3 percent of GDP. The yield from tariffs imposed on imports to replace existing QRs is estimated at 0.7 percent of GDP (0.5 percent at the current levels of imports, plus the proceeds from an assumed expansion of imports). A total increase in indirect tax revenues by 2.5 percent of GDP is thus indicated. 6.41 In the long run, a further reform is required (involving a process of negotiation with the states and a consequent constitutional revision). Introduction of a single VAT covering all sectors of the economy would significantly improve the neutrality of the Brazilian indirect tax system, in particular with respect to the use of inputs of services. At present, both the IPI and ICM exclude services, which constitute about one-half of Brazil's national product. In addition, the two separate taxes overlap, and lead to double taxation of the industrial sector. The integration of the two taxes into a single consumption-based VAT taxing all sectors would also raise revenues-- primarily due to the imposition of a tax on services, at a rate considerably higher than the 5 percent currently levied on that sector. This would, in part, allow the average VAT rate levied on the sectors currently subject to both the IPI and the ICI4 to be lowered. The revenues from the single tax would be allocated among the three levels of government so as to prevent any level incurring a loss of tax revenues. C. Direct Taxes 6.42 Table 6.8 summarizes the development of revenues from major direct taxes since the mid-1970s. It shows that revenues from direct taxes have declined substantially in recent years--although not by quite as much as revenues from indirect taxes. -129- Table 6.8: REVENUES FROM MAJOR DIRECT TAXES, 1975-88 (in percentage of GDP) 1975 1980 1985 1988 Income Taxes Personal Income Taxes 0.2 0.2 0.2 0.3 Corporate Income Taxes 0.8 0.6 1.1 1.3 Withholding Taxes 1.5 1.7 3.1 2.2 Total 2.5 2.5 4.4 3.8 Payroll Taxes 7.8 7.3 5.4 5.3 Other Direct Taxes 1.0 1.4 2.0 0.6 Total 11.3 11.2 11.8 9.7 Source: Balanco Geral da Uniao. 6.43 Pure income taxes are of quite minor significance in Brazil: in 1988, revenues from them amounted to less than 4 percent of GDP. Income taxes, primarily the personal income tax (PIT), are paid monthly by withholding. Payroll taxes, on the other hand, are the major source of direct tax revenues. Payroll Taxes 6,44 The most important single source of tax revenues in Brazil payroll taxes are imposed, in principle, on all workers in the formal sector. In fact, only about 60 percent of these workers are subject to the tax--or roughly 30 percent of the total labor force. Tenured public servants at the Federal and state level are the major legal exemption, but, given the high tax rates and the weak connection betweer tax payments and benefits, small businesses tend to avoid registering their workers and paying the tax. Self-employed workers are subject to a special tax regime, but only a small portion of payroll tax revenues is derived from these workers. 6.45 Tax Rates. Payroll tax rates are very high in Brazil by any standards (developing or industrialized countries) The total tax rate is the sum of a number of different taxes earmarked for various programs. The rates vary by sector and level of salary, ranging form 38 percent for the lowest paid worker in the public sector, to 45.2 percent for the highly paid workers in the industrial, commercial and service sectors. Employee contributions to social insurance, which form about 25 percent of total tax -130- liability, are subject to a ceiling of 10 minimum salaries; all other contributions have no ceiling, and are invisible to the employee. 6.46 Payroll taxes are earmarked for certain categories of expenditures. These differ from tax to tax, but are mainly designed to finance social-security expenditures. This earmarking obviously reduces the scope of possible changes in the structure of public-sector expenditures. The Personal Income Tax 'PIT) 6.47 The 1989 tax reform has changed the PIT regime substantially, and mostly for the better. A ten-bracket schedule, ranging between zero and 45 percent, has been changed to a three-bracket schedule of zere, 10 percent, and 25 percent. The tax base has been widened, and deductions have been reduced (under the previous scheme, deductions varied by the sectoral source of income). One important exception has remained, however, as agriculture still receives privileged treatment. Implementation of these reforms will mean a large reduction in the number of taxpayers--from 8.3 to 3.2 million--and a reduction in PIT revenues by about 18 percent, or 0.1 percent of GDP. The Corporate Income Tax (CIT) 6.48 Following the 1989 reform, the standard CIT tax rate is 30 percent (down from 35 percent), with an additional tax of 5 percent on high incomes. An 8 percent "social contribution' has been added, as well as levy of 5 percent of tax revenues as a state tax. Together, these changes leave the CIT roughly unchanged. The full tax rate is, however, applied only in a minority of cases. A plethora of incentive programs, mainly to promote sectoral or regional growth, substantially reduce the incidence of the tax. These include rate reductions (particularly in agriculture), increased deductions (again, primarily in agriculture), the reduction of taxable profits, and accelerated depreciation. Many of these incentive schemes may be--and are--used cumulatively. The Withholding Taxes 6.49 These may be used either as a separate instrument or as a substitute for the PIT. In the first case, the tax is levied at source on labor or capital. The second form, applied predominantly to the tax on labor, is essentially an anticipation of the PIT, and is subject to a progressive tax schedule applied to monthly income. In the case of capital taxed at source, the tax is not progressive, and its payment at source substitutes for the income tax and frees the taxpayer from any further obligation. As we saw earlier, most personal income taxes are actually paid through withholding taxes. Deficiencies of the Direct-Taxes System 6.50 The most obvious and most glaring observation about income tax in Brazil is the minor role they play in the taxation system. In the personal income tax, this is due to the low rates--a level destined for a further sharp decline following the 1989 reform; to several important exclusions from the tax base--particularly in agriculture; and to widespread tax evasion. The corporate income tax is, in fact, nearly absent--its share of 4S -131- GDP in Brazil is only about one fifth the level in otner developing countries of similar income levels, although the posted tax rates show no differences. The main reason for this discrepancy is the widespread use of fiscal incentives, applied through tax exemptions. The loss due to these schemes is officially estimated at approximately 1 percent of GDP, but this is definitely an underestimate, since it takes into account only part of the incentives. Standing out, again, is agriculture. Indeed, many producers of non-agricultural goods lower their tax liabilities by offsetting profits in their prime activity with deductions for agricultural investment. The complexity and multiplicity of fiscal incentives also create many other loopholes in the system, contributing to large-scale tax avoidance. 6.51 Inefficiencies also abound in the system, largely generated by the plethora of fiscal incentives. An additional source of particular importance is the nature of the system of payroll taxes. In comparison with other countries, the coverage of these taxes is low in Brazil, but their rates are very high. Beside the inducement to tax evasion, this leads to significant distortions, due both to the difference in effect between the treatment of labor and of capital--the latter being more easily able to flow to non-covered activities--and to the change in relative rewards of formal versus informal sectors, high-coverage and low-coverage activities, or urban versus rural regions. 6.52 Inequity in the system of direct taxes is due to several sources. First, labor income and capital income are treated differently. Labor income--at least in the formal market--is taxed fairly comprehensively, whereas capital income is not. Also, labor income is taxed progressively, whereas taxation at source of capital is not progressive. Second, the complexity of the CIT provides a variety of loopholes, which are effectively utilized by larger firms, with ample access to legal resources, whereas smaller businesses are less likely to be able to benefit from the potential for tax avoidance. Third, the high rates and limited coverage of the payroll taxes imply highly unequal treatment of incomes from different sources. 6.53 Recommendations. To remove--at least partly--these deficiencies, we recommend the following reforms in the tax system. 6.54 In the personal income tax, the highest marginal tax rate should be raised back partly--to 35 percent. The base-widening measures built into the December 1988 Tax Law should be enforced. In conjunction, these measures should increase the yield from the tax by 0.1-0.2 percent of GDP. 6.55 In the corporate income tax, the current proposal to tax agricultural income similarly to other incomes should be implemented. This would be significant, not just for its direct effect on tax from this source but also because it would close a most important channel for tax evasion. Fiscal incentives for investment projects through the tax system should be cut drastically, and wherever investment subsidies remain, they should be administered through budgetary allocations. Even if the same overall level of subsidization were left, net taxes would increase significantly, since the loopholes created through the tax incentives would be closed. Inasmuch as tax incentives cannot be removed due to constitutional constraints, they could be offset--at least partly--by -132- introducing an Alternative Minimum Tax. These changes would increase net revenues from the corporate income tax by as much as 2 percent of GDP. 6.56 The payroll tax system should be simplified. Ideally, a singJe payroll tax should be levied. While this thorough reform would required time, an immediate measure might be the elimination (and incorporation) of several small taxes assigned for the financing of various programs (INCRA, DPC, and FUNDO AEREO). Another candidate for immediate incorporation could be the education salary tax. The tax base should be broadened, to cover all public-sector employees, including the estatutarios (the tenured employees). While this coverage may not yield net revenue to the public sector, it would reallocate net revenues from local governments to the Federal government. The tax rate should also change. Since payroll taxes are, in effect, the major income tax, more progressivity should be introduced. Tax rates should also be equalized among the various activities. In the longer run, once the macro-economic balance permits, these tax rates--very high at present--should be lowered. 3. The Social-Security System 6.57 Brazil's social-security system, evolvir.; since the 1920s, has acquired its present structure in 1977. Since t en, the system has been organized under the SINPAS framework (Sistema N cional de Previdencia e Assistencia Social--the National System of So. al Insurance and Assistance). This framework consists of three institutes: (a) INPS (Instituto Nacional de Previdencia Social--National Institute for Social insurance), which provides insurance for retirement, disability, and survivorship); (b) INAMPS (Instituto Nacional de Assistencia Medica da Previdencia Social--National Institute for Social Insurance Medical Care), which provides direct medical care; and (c) LBA (Legiao Brasileira de Assistencia--Brazilian Legion of Assistance), which provides social assistance to the non-insured poor. FUNABEM (Fundacao Nacional do Bem- Estar do Menor--National Foundation for Child Welfare), which grants social assistance to non-insured poor minors, was recently transferred to the Ministry of the Interior. Of these, only the first two institutes are quantitatively significant: in 1984, for instance, INPS expenditures constituted 70.5 percent of total SINPAS expenditures, and INAMPS--25.3 percent (most of the small remainder were administrative expenditures). We shall confine our attention thus to only these two institutes, or functions. Since the operation of INAMPS will be discussed later, in the context of the health sector, the present discussion of expenditures will thus be restricted to INPS--the pension system for workers and their survivors. -133- Table 6.9: EXPENDITURES AND REVENUES OF INPS, 1972-86 Contribution Expenditures Net Balance of Revenues Annual Change of Real Size (in percent of revenues) (percent) 1972 26.3 18.4 +6.1 1973 26.9 15.0 +9.2 1974 19.1 14.0 +8.5 1975 14.8 16.6 +5.5 1976 15.1 16.9 -1.4 1977 10.5 31.2/a - .8 1978 6.7 12.7 + .5 1979 9.1 2.3/a - .6 1980 1.8 9.6 -7.3 1981 3.0 11.8 -8.7 1982 25.5 12.0 +7.7 1983 -2.3 10.7 -1.8 1984 4.1 10.3 - .3 1985 n.a. n.a. +3.6 1986 n.a. n.a. +6.5 M /a Data of doubtful reliability Source: Francisco B. de Oliveira et al. (1988), "Current Issues in Brazilian Social Security", Processed 6.58 Expenditures are financed by the payroll taxes, and Table 6.9 demonstrates the changing balance in the system. Until 1976, a significant revenue balance is recorded, but this changes into an excess of expenditures in most subsequent years (except 1987 and 1988). The large increase in both contributions and expenditures during earlier years is the product of a substantial increase in incomes and the widening of insurance coverage--the latter leading, naturally, to a revenue surplus in the earlier years of coverage expansion. With the deceleration of both factors, the problem inherent in the system has surfaced, although only partially as yet. This problem--increased expenditures--is due partly to the inevitable 'maturation" of the system, and to increased life expectancy. To illustrate: while the number of active contribators per beneficiary was roughly stable from 1955 to 1977, at 4-5, it declined to 2.7 by 1981--a level quite low by international standards. Additionally, rising expenditures result from increased average pre-retirement wages of covered workers. Finally, rising expenditures reflect the addition of new benefits, and the relaxation of entitlement rules. 6.59 At present, urban workers are eligible for retirement payments of 70 to 95 percent of their wages, at the age of 65 for men and 60 for women. -134- Early retirement is also possible, at any age, after 30 (sometimes 25) years of work, with the benefit changing with the length of the covered period. A mininum coverage period of 5 ye-rs is required--entitling the beneficiary to the lower 'replacement" rate (70 percent)--with higher rates applied for longer periods of coverage. In certain occupations, particularly generous length-of-service provisions are applied. Rural retirees receive 75 percent of the statutory minimum wage, with no provision for early retirement. 6.60 In principle, benefits are indexed to the cost of living. But in fact, the inflationary process has led to large erosion of benefits--to half, or less, of their legi:lated values. Benefits are based on a weighted average of wages in the last 12 to 36 months of covered employment, and the levels for the last 12 of these months are not corrected for inflatton. Moreover, even where correction is made--for the earlier periods--a SINPAS index is used which is adjusted only every six months. In addition, indexation does not start until the first benefits are received, and several months may elapse before initial claims for benefits are processed. At the pace of inflation in Brazil in recent years, these flaws in adjustment process imply heavy attrition of benefits. This may change with the new Constitution, which stipulates full indexation of social-security benefits. 6.61 The increasing gap between expenditures and revenues of the social-security system may be solved by either increasing payroll taxes or lowering expenditures--given the general policy context, in which an increase in other taxes and a reduction of other government expenditures are required anyway for macroeconomic stabilization, and in which an increasp in national saving is an imperative. The level of payroll taxes, we have seen, i3 very high, and increasing it still further--at least in the present form of the tax--would be counter-productive. This leaves a reduction of benefits as the inevitable means of achieving a solvent social-security system. 6.62 Recommendations. We recommend that the entitlement to benefits from the social-security system should be reduced, and at the same time, that legislated benefits should, indeed, be the effective ones. Thus, legislation should lower "replacement rates", and in particular eliminate the especially generous schemes for specific sectors or professions. The relationship of entitlement to coverage period should also be strengthened. Today, the benefit of a worker's marginal contribution to the scheme is zero, for all but the end of his working life, beyond the minimal coverage period required for participation. This leads, inter alia, to the tendency to avoid participation in (and contributions to) the scheme until a short time before retirement age. Some of these changes would not be possible without appropriate constitutional amendments. 6.63 On the other hand, indexation of benefits should be as complete as is technically possible: no justification can be seen for imposing an inflationary tax, in a highly capricious way, specifically on the recipients of social-security benefits. 6.64 Another important recommendation, for a medium-term fundamental reform of the system, is to shift to a scheme in which benefits are clearly related to contributions. That would change perception of payroll deductions from pure taxes, to payments for deferred benefits--a change of far-reaching consequences in its efficiency and distributional -135- implications. Even beyond that, the government should contemplate a shift of the pension provision function fr m the government's social-security system to mandatory participation in government-monitored private pension funds, along the lines practiced recently in several Latin American countries, such as Chile, Costa Rica or Uruguay. -136- VII. PUBLIC ENTERPRISES A. O_igin, Size anr( St ructure 7.1 Some state ownt- ship of produ t ion ernterrises began quite early in Brazil. A major impetus in this direction, however, anwo during Getulio Vargas' Presidency (1930-45). The fundamental approach of his admir,istration was that economic development could best be served by ex)anding the role of government. This approach was reinforced during World War II when the private sector was viewed as having failed to meet the country's wartime needs, such as steel; Brazilian officials concluded that private enterptise could not be relied upon to develop proje(ts of .national importance." Following the war, government activities continued to expand. The establishment of the National Bank for Economic Development (BNDES) in 1952 and of the oil monopoly (PETROBRAS) in 1954 were important stPps in this direction. A host of infrastructure-related firms were established during the early postwar years. 7.2 BNDES has, since its establishment, practically been the sole source of long-term project financing. As such, it has largely determined, through its allocation policy, patterns of development. Mainly following defaults of p;ivate firm debts to BNDES, the government gradually took possession of existing private-sector enterprises, besides establishing its own enterprises. 7.3 The role of government enterprises continued to increase throughout the postwar period. The turning point came in 198., when two laws were enacted to reverse this development. One specified restrictions on the establishment of new government enterprises; the other, which mentioned for the first time the desirability of privatization, established guidelines for transferring state-owned enterprises to the private sector. In later years. the reduction -f government investment became the major constraint on further expansion of the state enterprises; and, as we shall see shortly, strongly reinforced the perception that privatization was needed. But in terms of actual scope, no significant reduction of the government enterprise's share of GDP has yet taken place. -I 4 - 137- Tl t / . ': PUBLIC versus PRIVATE OWNERSHIP OF BRAZIL'S THIRTY LARGEST FINANCIAL F1RMS, 1962-79 (niunber oi eo-terprises) Ownership 1962 1967 1971 1974 1979 Public 12 13 17 23 28 Private 18 17 13 7 2 Source: Thomas J. Trebat, Brazil's State-Owned Enterprises: A Case Study of the State as Entrepreneur, 1983, p. 59. 7,4 Some indication of the present size of the government enterprise sector, aTL of its progress over time, may be prov.'led by some elementary data. Thus, Table 7.1 shows the share of these enterprises among Brazil's 30 largest non-financial firms. Even in the early 1960s, this share amounted to 60 percent. By the late 1970s, virtually all the largest firms were government enterprises. Table 7.2: SHARE OF PUBLIC versus PRIVATE ENTERPRISES IN THE ECONOMY, 1986 (percent) Number Share Share of Net Share of of Operational of Ownership Enterprises Capital Revenues Employment Government 4.6 45.8 23.2 18.1 Private Domestic 89.0 45.2 58.3 70.1 Private Foreign 6.4 9.0 18.5 11.8 Total 100.0 100.0 100.0 100.0 Source: Visao. Quem e Qu6m na Economia Brasileira, 1987, p. 26. -138- 7.5 Another indication is found in Table 7.2, which shows the share of public enterprises in the economy by three yardsticks--aggregate capital. aggregate revenues, and aggregate employment. In terms of capital, the share is close to one-half of the total economy, and as large as the domestic private sector. It is lowest where employment is concerned--a reflection of differences in capital-labor ratios, to which we shall turn again. But even by the share in aggregate employment, public enterprises amounted in 1^86 close to 20 percent of Brazil's economy. In terms of annual investment flows, public enterprises exceeded a 30 percent share by the mid-1980s. 7.6 The public-enterprises sector includes a large number of firms--at present, there are 175 Federally-owned companies alone, and it is highly diversified. Roughly half of the total activity of the sector, however, is accounted for by nine large companies. Three are in electricity: ELETROBRAS, ITAIPU, and INB. The others are PETROBRAS (petroleum); SIDERBRAS (steel); CVDR (companhia do Vale do Rio Doce--iron mining); TELEBRAS (telecommunications); PORTOBRAS (ports); and RFFSA (railways). 7.7 Over the years, state enterprises have acquired 3ubstantial operational autonomy, ao a result of financial independence originating from internal generation of funds. In the oil and mining sectors, in particular, high profits and large cash reserves allowed large investments, and diversification through the establishment of subsidiaries and affiliates--a structure which further insulated these corporations from central government controls. In 1979, a broad mandate was given to SEST Secretariat for Budget and Control of Public Enterprises, granting it budgetary authority over all federal enterprises, and centralizing within it all the supervisory functions. Apart from rare exceptions, however- -ordered by the President--management of the enterprises has remained highly insulated from ministerial oversight. 7.8 Public enterprises are owned by the Federal, state, and municipal governments. The Federally-owned group, supervised by SEST, is by far the most important. Of its 175 companies, 141 belong to the category defined as "productive state firms' (SPE); 18 are "typical governrient entities", handling services; and the other 16 are Federally-owned banks (supervised by the National Monetary Council). In principle, the "productive" enterprises are supposed to cover operational expenditures with their own revenues or other resources, turning to Treasury financing only for investment projects. However, they also depend on the Treasury when their own resources are insufficient to cover operational expenditures. The "typical government" enterprises, on the other hand, are not supposed to self-finance their current expenditures: they provide what are considered to be public services, generating only a small amount of revenues. In 1988, 75 percent of their aggregate spending (current and capital) was financed by government transfers. As we will observe, however, the deficit of this category is in fact only a minor part of the aggregate deficit of the sector. ~~~~~~~~~~~~~~~~~~~~~~~~I -139- Table 7.3: INDICES OF PUBLIC-TO-PRIVATE RATIOS OF OUTPUT AND INVESTMENT, 1975-85 (1975 100) Year Output Investment 1975 100 100 1976 99 184 1977 105 265 1978 105 254 1949 112 336 19°0 116 201 1981 131 443 1982 135 297 1983 157 944 1984 182 495 1985 186 699 Source: Calculated from data on fixed-capital formation and indices of real output in Coniuntura Econ6mica, various issues. B. Performance: Basic Indicators 7.9 Several broad indicators strongly suggest that the public- enterprises sector as a whole is inefficient, perhaps grossly so. Table 7.3 shows the change from 1975 to 1985 in the relative growth of output and investment in the public enterprises and private sectors, respectively. By 1985, the ratio of public to private investment was seven times higher than in 1975, while the ratio of the two outputs increased by less than 100 percent during the period. This implies an increase--and probably a substantial one--in the capital/output ratio in the public sector, ccmpared to that in the private sector. Capital could conceivably substitute for labor. In that case, however, measures of "labor productivity"--output per unit of labor--should have shown cf:responding relative increase. In effect, labor productivity data (for 1980 to 1985) show an annual increase of 3.8 percent in industry as a who.e, versus only 2.6 percent in the public enterprises sector. 7.10 Beyond the increment values (over 1975-85), a difference in capital/output ratios is correlated with the radical difference in capital/labor ratios between the two sectors. The data in Table 7.2 implied that the capital/labor ratio was four times higher in the public enterprise sector than in the private sector. This might conceivably be explained away by the differences in industrial structure between the two sectors--with the public enterprises sector, despite its diversification, tending to consist more of activities which are inherently mtore capital -140- intensive (such as mining, or electricity). But the difference persists, in general, even when comparisons are made within specific branches of activity. This is shown in Table 7.4. Table 7.4: CAPITAL/LABOR RATIOS, BY BRANCH, IN PUBLIC AND PRIVATE ENTERPRISES, 1980-86 (millions of cruzados per man-year, constart 1986 prices) Public Enterprises Private Enterprises Sector 1980 1982 1986 1980 1982 1986 Mining 1.22 1.35 2.13 .55 .62 .86 Metallurgy .61 .42 1.21 .23 .31 .43 Transportation Equipment .32 .37 .28 .18 .23 .23 Cellulose and Paper n.a. 1.11 .53 n.a. ..7 .83 Chemicals and Petrochemicals 1.01 1.31 2.11 .45 .63 .83 Transportation 1.21 1.31 1.28 .14 .24 .12 Communications .51 .72 .61 .79 1.34 1.42 Electricity 1.21 l.?8 1.47 .44 .61 .70 International Trade .69 .93 2.30 .43 .65 1.37 Average .75 .99 1.32 .36 .58 .75 n.a. - not available. Source: Calculated from Vislo, op.cit., 1981, 1983, and 1987. 7.11 Table 7.4 indicates that capital/labor ratios are (usually considerably) higher for public enterprises in most sectors (communications and, for 1986, cellulose and paper being the only exceptions). Once more. the higher capital/labor ratios would not, by themselves, necessarily indicate inefficiency; and the absence of further data (such as on comparative labor productivity) makes it impossible to draw final conclusions. But at the very least, these differences would indicate the application of relatively capital-intensive techniques in public enterprises, which must be due to the availability of relatively cheap capital for this sector. Indeed, access to capital resources has been much easier and cheaper for public enterprises--at least until very recently. In an allocation sense, this must be a source of gross inefficiency in the economy. -141- Table 7.5: PROFIT RATES/a IN LARGE BRAZILIAN FIRMS, PUBLIC versus PRIVATE SECTORS, 1974 AND 1975 (percent) Sector 1974 1985 Public Enterprises 11.4 2,5/b Private Domestic Enterprises 22.9 12.0 Private Foreign Enterprises 20.6 16.0 /a Profit rates are measured as the ratio of net earnings to equity. /b The recorded drop in profitability of state enterprises from 1974 to 1985 is somewhat exaggerated by the fact that in the latter year, but not in the former, state enterprises were subject to taxes on profits. Source: Wilson Suzigan, "As Empresas do Governo e o Papel do Estado," Exame, June 1987, p. 102. 7.12 Direct evidence on technical efficiency is provided by data on profitability. First, in Table 7.5, profit rates are presented for 1974 and 1985 for the aggregates of the sectors (public enterprises, private domestic, and private foreign firms). For both years, public enterprises appear to show a substantially lower profit rate than either domestic or foreign private firms (profitability of the latter two sectors being roughly the same). Moreover, while profitability declined in all sectors over the decade shown, the fall is much more dramatic among public enterprises. 7.13 Nor, again, can this difference in performance be explained by differences in industrial structure (and even if it were, it would still be meaningful). Table 7.6 shows profitability of large enterprises in the public enterprise and private sectors, by the same industrial classification used in Table 7.4. It appears that wiTh two exceptions- -transportation equipment and cellulose and paper--profit rates are substantially higher in the private sector (10.8 percent) than among government enterprises (0.6 percent). This is rather similar to the gap recorded in Table 7.5. -142- Table 7.6: PROFIT RATES/a OF LARGE ENTERPRISES BY BRANCH PUBLIC versus PRIVATE, 1986 (percent) Public Private Sector Enterprises Enterprises Mining 7.7 12.6 Metallurgy -11.7 12.8 Transportation Equipment 11.5 7.9 Cellulose and Paper 8.3 5.2 Chemicals and Petrochemicals -2.3 11.6 Transportation 4.2 14.4 Communications 6.3 12.5 Electricity 7.6 45.8 Industrial Trade .3 26.2 Average /b .6 10.8 /a Defined as Table 7.5. /b Average applies to an aggregate which includes "others", beyond the nine sectors listed. Source: Calculated from Visao, op cit., 1987. 7.14 The aggregate data thus clearlC indicate technical inefficiency, as well as the distorted use of inputs. 1 We shall shortly make recommendations to address this issue. But first, a note will be taken about short-term issues concerning the government enterprises in the recent period of macro-economic instability. C. Public Enterprises and the Government Budget: Recent Performance 7.15 Until the early 1980s, the public enterprise sector as a whole normally produced a substantial current surplus, which financed roughly one-half of the sector's investment. The rest was financed by long-term borrowing, and by transfers from the central government. In recent years, the current surplus has declined; and, at the same time, borrowing in the market has been cut drastically (foreign lending has, of course, disappeared completely as a source for new funds). The result is the 51/ Apart from the very recent history, which will be discussed shortly, low (or negative) profits in the "productive' government enterprises cannot be assigned to price regulation. -143- existence, in recent years, of substantial operational deficits in the sector, which are roughly equal to transfers to it from the central government. The major accounting items are shown in Table 7.7. Table 7.7: OPERATIONAL ACCOUNT OF PUBLIC ENTERPRISES. 1986-89 (percent of GDP) SEST Enterprises 1986 1987 1988 1988/a Revenues 15.0 16.4 14.7 13.0 - Current Expenditures 13.0 14.8 13.2 11.4 Current Surplus 2.0 1.6 1.5 1.6 - Capital Expenditures 4.2 3.9 3.3 2.6 - Operational Deficit (deficit = -) -2.2 -2.3 -1.8 -1.0 + State and Municipal Enterprises Deficit -0.5 -1.5 -0.4 -0.2 = Total Operational Deficit -2.7 -3.8 -2.2 -1.2 Of which: Financed by Government Transfers) 1.3 2.9 1.8 1.1 /a Preliminary estimate. Source: SEST and Seplan. 7.16 The operational deficit of the public enterprises thus appears, in recent years, to be substantial--on average, around 3 percent of GDP for 1986-88. It is a major component of the overall government budgetary deficit: in 1986, the public-enterprises deficit was 77 percent of the government's aggregate (operational) deficit, and 70 percent in 1987. In 1988 the proportion declined to 53 percent--still a major share. In 1989, it was expected to fall further, to some 30 percent; but this depends on certain assumptions about prices (discussed below). 7.17 The major enterprises generate most of the sector's deficit. While the multitude of smaller firms--including the "government typical" enterprises, which are supposed by their nature to create deficits- -contribute only a minor part, as Table 7.8 shows. The major deficit producers are the electricity related firms, predominantly ELETROBRAS; SIDERBRA (steel), TELEBRAS (communications), and RFFSA (railroads), are also major contributors. CVRD (the mining enterprise) and PETROBEAS (oil), on the other hand, produced surpluses--the former quite a substantial one. All other 'productive" firms, on aggregate, produced a minor surplus, and the "government-typical' enterprises contributed an equally minor deficit. -144- Table 7.8: DEFICITS AND GOVERNMENT TRANSFERS, MAJOR PUBLIC ENTERPRISES, 1988 (billions of 1988 Cruzados) Operational Govt. Transfers Enterprises Deficit/a and Subsidies Eletrobras 1033 579 Petrobras - 154 6 Siderbras 408 679 CVRD - 844 2 Telebras 333 16 RFFSA 218 346 Portobras 41 67 Other 'Productive~/b - 120 138 'Typical Governmentw 187 218 Total 1,102 2,051 /a A minus sign is a surplus. /b A residual item. Source: SEST and Seplan. 7.18 Most recently, public enterprise deficits have been heavily affected by the government's pricing policy, part of its stabilization efforts. Under the Cruzado Plan of March 1986, and the accompanying price freeze, government enterprise prices declined, in real terms, by about 20 percent by the end of the year; the sharpest drop, some 40 percent, applied to the price of steel. In the Bresser Plan of 1987, attention was paid to the elimination of major distortions in this area, and some price recovery took place--although steel prices were excluded from this process. In 1988, the general policy was the maintenance of roughly constant real prices for the goods and services involved. The major prices of public enterprises, from 1986 to 1988, are shown in Ta1'le 7.9. .91 -145- Table 7.9: PUBLIC-SECTOR REAL TARIFFSa/, 1985-88 (end of year) (index numbers, March 1985 = 100) Sector 1985 1986 1987 1988 1989/b Electricity (Eletrobras) 104 94 134 153 127 Oil Products (Petrobras) 81 79 81 68 50 Steel Products (Siderbras) 94 65 65 58 56 Telecommunications 93 101 112 107 68 Railways 111 70 77 83 68 Port Services 111 70 81 74 41 Average 89 82 91 85 66 Average, excluding oil products 99 84 101 104 86 /a Deflated by the IGP-DI. /b November 1989. Source: SEST and SEAP. 7.19 The Summer Plan of January 1989 started with an increase in public enterprise tariffs by an average of 17 percent. The new prices were maintained, at an unchanged nominal level, for about four months--a period during which general price and cost increases substantially exceeded the initial increase in tariffs, causing a decline in real prices. In May 1989 the policy was changed to, in principle, the indexing of tariffs. But, since the previous month's inflation was used for adjustment, and inflation was accelerating, the real tariff level was not even maintained at the post-freeze level. By June of 1989, real price reductions (from the end of 1988) were estimated to be 25 percent, on average for the sector as a whole. The heaviest reductions occurred in petroleum and electricity (32 and 29 percent, respectively). In July, some upward adjustment of real prices started. However, in October, an agreement was reached with private sector enterprises that nominal prices would, each month, increase by 90 percent of the general price increase of the preceding month. This agreement has also been applied to public sector prices--leading once again to real price reductions. -146- Table 7.10: PUBLIC-SECTOR REAL TARIFFS DURING 1989 (index numbers, December 1988 - 100) Sector June 1989 November 1989 Electricity 68.5 83.1 Oil Products 71.3 73.3 Steel Products 99.7 98.2 Telecommunications 79.9 63.3 Railways 84.7 81.3 Port Services 65.2 55.6 Average 75.3 78.4 Source: Calculated from SEST data. 7.20 The levels of real prices are shown in Table 7.10 for June 1989 and for November 1989--the latest date for which these data are available. It is obvious that real public sector prices have declined, on average, 22 percent during 1989. The decline would be even larger if the comparison was made with the end of 1987, rather than of 1988, since there was also a price fall (about 7 percent, on average) in 1988. Table 7.11: 'REQUIRED* ADJUSTMENTS OF PUBLIC-SECTOR TARIFFS (percentage of increase over November 1989 level) Adjustment to the Level of Sector End 1987 End 1988 Electricity 5.2 20.4 Oil Products 62.0 36.3 Steel Products 14.3 1.9 Telecommunications 65.1 57.9 Railways 13.9 23.0 Port Services 96.3 79.8 Average 36.3 27.5 Increase of Revenues, in Percentage of GDP 5.6 3.9 Source: Calculated from SEST data. --4~~~~~~~~~~~~~~~~~~~~~~~~ -147- 7.21 The restoration of public sector prices to the levels before their (unsuccessful) manipulation for stabilization purposes would go a long way toward eliminating the enterprises' operational deficits. And that wou'ld be a change of very substantial proportions in relation to the aggregate size of the budgetary deficit. This is illustrated in Table 7.11 where, on the bottom line, we see that a restoration of public sector prices to their end-1988 level would (in comparison with the November 1989 prices) yield revenues of close to 4 percent of GDP. Restoring prices to end-1987 levels would yield an additional 1 percent of GDP. D. Privatization of Public Enterprises 7.22 Recognition of the economic losses involved in the government ownership of enterprises have led to increased acceptance and, by now, a widespread conviction in Brazil of the need for, and importance of, privatization of these enterprises. As mentioned earlier, the first reference to privatization as government policy appeared in 1981. Little had happened, however, until the close of 1985, when a new legal framework was established for the privatization process. Under this system, some privatization had actually started--although on quite limited scale. 7.23 The intense interest in privatization has led to the recent introduction af a new framework, through new legislation (in March 1988) and subsequer.t regulations (August 1988). This established a Federal Destatization Council (Conselho Federal de Destat!zaq&o--CFD), chaired by the Minister of Finance and consisting of other economic Ministers and representatives of the labor unions and of private enterprises. This Council makes decisions on the eligibility of enterprises for privatization, and its technical organ handles the necessary ground work. The Council follows several major guidelines: (a) Use publicity to ensure transparency. (b) Use an external audit to establish transparency, fairness and legality of all phases of the procedures. (c) Use the stock exchange as an intermediary whenever control of a corporation is transferred to the private sector. (c) Use bidding procedures, when appropriate, in accordance with stock-market principles. (d) Use, when feasible, mechanisms which promote dispersion of capital ownership. (e) Establish ways of facilitating the acquisition of shares by employees of the enterprise which is being privatized. (f) Establish practices which comply with relevant legislation, including antitrust laws. -148- 7.24 As of now, there has been little actual privatization. Sixteen enterprises have been removed from the public sector--eleven through privatization, and the rest through liquidation. With a single exception. all sixteen are 'reprivatized' companies, acquired in the past by the government (BNDES) following insolvency. The total assets of these firms amount to US$160 million--a negligible percentage of the public enterprise sector's estimated total assets of US$60 billion. Still, the legislative changes of 1988 are likely to accelerate the process. Sixty-one firms are now being processed for privatization, many of them subsidiaries of the major state-owned holding companies. The largest in terms of assets are USIMINAS (held by SIDERBRAS), TELEBRAS (held by t'ie Federal Union), and two subsidiaries of PETROBRAS. TELEBRAS and the PETROBRAS subsidiaries are targeted, though, for the public offering of shares without transfer of control. The aggregate market value of the 61 firms is approximately US$4.5 billion, of which USIMINAS accounts for US$3 billion. 7.25 Under existing legislation or accepted practice, a substantial fraction of government enterprises are not candidates for privatization. The Constitution reserves the exploraticn, extraction, refining and transportation of oil, as well as mining, for the Federal Union as a monopoly. So, by law (although not by Constitution), are telecommunications and power services. In the past, an accepted doctrine- -without legislation--has also required the government to maintain a majority position in what it regards as "strategic enterprises" (such as flat steel). The "constitutional" monopolies amount to close to one- quarter of the total sector (by assets). The constitutional constraint is weaker, however, than this fraction would suggest, since it does allow the privatization of many activities (or subsidiaries) of the enterprises involved, which do not perform constitutionally-reserved functions. 7.26 Although the aggregate book value of public enterprises in Brazil is some US$60 billion, the market value of non-financial enterprises (and after the value of private holding of shares has been deducted) is estimated at only about US$25 billion. Of this, some 20 percent would at present be barred by the constitutional constraints. This leaves approximately US$20 billion as the total market value involved in an exhaustive privatization of the sector. Given the relevant magnitudes in Brazil, this is not an overwhelming amount even under present capital market circumstances, in which almost no long-term credit financing of investment is available outside the government channels. 7.27 One yardstick is the size of Brazil's stock market. It is estimated that in 1987, the market value of the equity of traded companies was around US$30 billion, half of which in shares that were actively traded. New equity issues amounted in 1986 to US$1.2 billion. Institutional investors--mutual funds and pension funds--managed assets of over US$22 billion in 1987; 20 percent of which were equity shares. These magnitudes suggest that the stock market could acconmodate an annual -149- trarnsfer to private ownership of some US$1 billion worth of shares (one- fourth to one-third of I percent CDP), without much disruption of the market.52 7.28 The foreign acquisition of the shares, particularly in connection with debt-equity swaps, is probably more relevant. Commiercial bank holding of Brazilian debt is over US$70 billion--a large amount indeed in comp)arison with the potential of US$20 billion of privatization. We shall also see (Chapter VIII) that in 1988, debt-equity swaps amounted to US$8 billion. Even if the amount is lower in coming years, this could lead to the rapid absorption of the shares of privatized enterprises (either directly or through the purchase of shares in other businesses, which would then increase the available funds for acquisition of the privatized firms). In principle, of course, foreign participation in Brazil's market need not be restricted to owners of Brazilian debt, but the latter (whether original owners or buyers at the secondary market) are much more likely candidates for participation than completely new foreign capital. Until now, debt- equity swaps have been used in the privatization of only two enterprises, and these experiments have been considered a success. 7.29 Employees of privatized enterprises are another potential source of demand for the shares. It is highly desirable to take specific measures (as indeed present guidelines require) to ensure that these employees share in the benefits from privatizing their enterprises. A major aid would be the offer of shares at discounted prices (and, perhaps, specifically favorable credit terms) to these workers. Experience in other countries which have recently practiced widespread privatization (Chile, the U.K.) shows that workers respond well to such offers. 7.30 Finally, a large potential exists in Brazil for the swapping of government domestic debt for equity. The size of the short-term debt, and the acute problems to which it leads, have been indicated in our earlier analysis (Chapter IV). In view of the overwhelming size of the short-term debt (some four times the aggregate potential market value of privatized enterprises) the purchase of equity of privatized enterprises may play an important role within a general scheme of converting this debt into the holding of long-term instruments. 7.31 Recommendations. The discussion above has already implied several important policy prescriptions concerning the financing of (or the demand for) privatization. These included encouraging foreign participation, primarily through debt-equity swaps; encouraging the transformation of domestic short-term government debt into equities of privatized enterprises; and encouraging equity acquisition by employees of the privatized enterprises. In addition, we suggest the following considerations and recommendations: 52/ To illustrate: the British stock market absorbed, within 18 months, without any apparent significant effect on average share prices, an issue of British Telecom worth about 5.5 percent of the total of shares in circulation. -i50- (a) The mechanism for privatization, estab)lished ini 1988, is adequate: the relevant technical organization is of high quality, and the guidelines and procedures undpr which it operates are by and large appropriate. With this mechanism, a comprehensive policy of privatization should be feasible. (b) The privatization process should accelerate rapidly, and proceed at a high speed. The key constraint is not the absorptive capacity of the financial market; as we argued earlier, the market would not be disrupted (especially if the earlier recommendations concerning financing are adopted) by annual offerings of several billion dollars worth of equities. The real constraint is the capacity of the privatization agency to prepare the individual enterprises for privatization. These preparations may be of varying degrees of complexity, ranging from the legal and other requirements of the offer of equities to the necessity, on occasions, of deregulating prices, or restructuring the enterprise financially and sometimes even physically before it is offered in the market. While bringing many enterprises to the market may strain the administrative capacity and, if done massively, also the absorptive capacity of the market, placing companies into private hands too slowly would raise questions about the commitment of the government to the program. It would also postpone the gains to be made from privatization. The optimum speed cannot be decided beforehand. (c) A "negative list" should be prepared, of enterprises t!t are not subject to privatization; presumably, this should be limited to cases in which the Constitution prescribes public ownership. All enterprises not on the list would be candidates for privatization, and some work to begin the process should start quite early in each. (d) Few broad rules may be offered about the sequencing of privatization of enterprises. Logical first candidates are (and have actually been) enterprises which are former private companies. Next would - come activities which typically belong, in industrial countries, in the U private sector, such as steel, fertilizers, pulp and paper, or warehousing. Within this classification, it is important to select at a very early stage several large-scale enterprises, the privatization of which would signal the government's serious intentions. These should mostly be enterprises which have been doing well, rather than obvious failures. (e) Privatization should be "true" and full fledged. In several recent instances, the government has still intended to maintain majority control of the enterprises. This is the wrong approach. First, it misses the main objects of privatization; second, it sends the wrong signal, indicating equivocation on the part of the government in its approach toward liberalization; and third, it leads to lower prices for the shares sold in the market. (f) An essential part of the privatization process must, on many occasions, be the removal of regulations and price controls; without this, privatization may make little sense, and may not be feasible. At the time of privatization, the potential purchasers should be reasonably confident that the privatized firm will face a stable policy environment, and output and input prices which reflect basic market factors rather than government -151- interference. Industrial deregulation, reformulation of industrial policy, reform ot trade policy, and the removal of price controls should therefore be necessary accompaniment of privatization. When existing monopolies are privatized and price controls lifted, the government should ensure beforehand contestability of the market--first and most significantly, by allowing free trade in the product. Existing antitrust laws should be modified, when needed, to accommodate qituations that may arise under the new industrial environment.. (g) Finally, close attention should be paid to the interests of employees. Those who remain employed should partici4ste in the fate of the enterprise by acquiring shares on concessionary terms. Some workers must be displaced, as part of the process of enhanced efficiency expected from privatization. Natural attrition may provide part of the (short-term) reduction of the labor force; and early retirement schemes would help. For those displaced involuntarily, safety-net schemes should be prepared. These may include severance packages which would consist, inter alia, of the provision of company shares and job search allowances, along with job placement policies. That some of the existing labor force will lose from privatization is inevitable; but the process of liberalization will be greatly facilitated if all efforts are made to minimize this loss. -152- VIII. EXT&RNAL TRANSACTIONS: TRADE AND DEBT 1. Foreign Trade and the Trade Reg˝ie A. The Size and Structure of Trade 8.1 The size and structure of Brazil's foreign trade have been urdergoing rapid changes over the past 20 years. 8.2 The structure of exports is represented in Figure 8.1. Most obvious is the composition of exports which has shifted strongly from coffee to manufactures. Thus, for instance from 1965 to 1988, coffee exports declined from 45 percent of aggregate exports of goods to 8 percent, while manufactures (including semi-manufactures) increased from 18 percent to 67 percent. Brazil's heavy dependence on coffee exports thus vanished completely within two decades (in fact, by the late 1980s coffee had even ceased to be the most important single export item). Within exports of manufactures, "traditional' labor-intensive goods (such as textiles and footwear) at first (in the 1970s) grew faster than those of 'non-traditional" exports (mainly intermediate and capital goods). Since the late 1970s, this trend has been reversed. Within agricultural exports, the most notable development is the emergence and rapid increase in soybean exports, which by 1988 has reached roughly the same level as the coffee exports. -igure 3. 1 Exports By Commodity Group '985 '988 3 83 73 40 30 20 65 88 87 88 89 73 7' 72 73 74 75 78 77 78 79 8a 8' 82 83 84 85 86 87 88 A;, t"oe Source: Central Bank of Brazil. F;' -154- 8.3 The structure of imports, represented in Figure 8.2, has changed less dramatically--and less consistently--than the structure of exports. Both the trend, and fluctuations, in import composition are dominated by the evolution of oil imports which have, in turn, largely reflected changes in oil prices. The first oil-price shock of 1974, and the second of 1980, both led to a very large increase in oil's share of total imports: from some 12 percent in the early 1970s, this share has increased to about a half of the total in the early 1980s. Later, the downward trend of oil prices, as well as the developing production of domestic substitutes, have led to a decline in oil's share of imports to roughly one quarter in the late 1980s. A-side from oil, the composition of the remaining imports of goods did not change much among major categories--raw materials, final consumer goods, and capital goods--between the early 1970s and the late 1980s. J -155- Figure 8. 2 Imports by End-Use 100- 90 I 704 10200 1970 1975 1980 1ese FXi Raw Materials 1IAi Consumer goods Capital Goods Fuels Note: shares o' raw materials and consumer goods not available for 1966-69 Source: Central Bank of Brazil. -156- 8.4 The size of trade flows behaved very differently in the 1980s than in earlier decades. As Figure 8.3 shows, the ratios of both imports and exports to GDP grew from the mid-1960s to the early 1980s. With the fast- increasing capital inflows from the late 1960s, however--and primarily from the mid-1970s--imports increased substantially more than exports. Since the early 1980s, on the other hand, the share of imports in GDP has tended to decline (measured at constant prices, this trend has actually been apparent since the mid-1970s). Exports have increased continuously: at constant prices, their share in GDP approximately doubled from the late 1970s to the late 1980s. These divergences largely reflect, of course, the change in size--later even in direction--of capital flows since the debt crisis of 1982. A significant import svrplus in the 1970s thus turned into a substantial and increasing export surplus (of goods and non-financial services) in the 1980s. -~~~~~~~~~~~~~~~~~~~~~~~~ .1 -15 7- Flgure: 8. 3 Trade as Percent of GOP, 1965-89 (at current prices) 31 - - 2~~ u 1986 10 1S'?5 1 9 85 Irnporte Ecports (at 1980 grIcsea l8 is 1 42~ *SS *7 19rb 1980 1j86 'S EEIIMOtS E~Ot a Impon E(PortA Source: World Bank calculations, based on data from Conjuntura Economica, IBGE and CACEX. -158- 8.5 The fast increase in exports during the 1980s, in the face of a quite volatile policy environment, appears to indicate that Brazil's exports respond fast to changes in policies and economic circumstances. Part of this is the increase in exports that followed the debt crisis when domestic demand was slack; this phenomenon reappeared on later occasions, such as in 1988 and early 1989. The export response to price changes has also been impressive.53 Indeed, variations in export performance can be clearly related to fluctuations in the real exchange rate (for instance, declines of exports in 1982 and in 1985-86 lagged by a few months' declines in the real exchange rate). And the strong export growth of the 1980s is consistent with an increase in the real exchange rate from the 1970s to the 1980s. Even with the reduction of export subsidies, this increase on the exchange rate has meant an increase in the profitability of manufactu,ed exports as may be seen from Figure 8.4. Most recently, in 1989, export performance must have been better than the record shows due to resort to under-invoicing in periods when the official foreign exchange rate lagged and the black-market rate offered an increasing premium--an informal mechanism which must have met tacit acceptance by the government. 53/ Coes, using data of the period from 1955 to 1977, has found a vigorous response of exports (with a time lag of 9-10 months) to changes in the teal exchange rate, the level of subsidies to manufactured exports, and the level of domestic demand. The short-term price elasticity of exports has been estimated to be 1.67, and the long-term elasticity-- 2.37. See Donald V. Coes, Brazil, in Demetrios Papageorgiou, Michael Michaely, and Armeane M. Choksi (Editors), Liberalizing Foreign Trade. Volume 4: The Experience of Brazil, Colombia, and Peru. Oxford: Basil Zlackwell, forthcoming. -159- Figure : 8.4 Manufactured Export Profitability 100 1C 0 70 ~80 ~40 30 20- 10 0 B0- 1985 1970 1975 1080 198I E Expoft subsldies (as percentage Idex of mamnufaurod expoRt proitability (190-100) of f.o.b. value of manufactures .cpons Note: for definitom of index of peofittbuity for manufacure eaes Teabe EA) Source: World Bank calculations, drawing partly on Renato Baumann and Heloiza C. Moreira (1987), "Os Incentivos as Exportacoes Brasileiras de Produtos Manufacturados: 1968/85, "Pesquisa e Planejamento Economico 17(2). -160- 8.6 Despite the increased role of trade, Brazil's economy has remained one of the world's most closed. This is particularly true when the share of imports is considered: both the 4 percent ratio observed in the late 1980s and the 6-7 percent ratio of the early 1970s are extremely low. At its highest, in 1981 (with the large increase in oil prices), the ratio reached only 11 percent. thile precise meaningful comparisons would require some work, we can note that the import ratios in Spain and Canada- -two countries whose economic size is roughly equal to that of Brazil--were in recent years 18 and 23 percent respectively. In Japan, an economy five times the size of Brazil's but which is still relatively closed, the import ratio was 13 percent. Even the much higher export ratio in Brazil, some 10 percent in the late 1980s, would appear to be quite low in relation to the relevant international comparisons. This basic attribute of Brazil's economy--its very high degree of insulation from the world's economy--has largely been the product of Brazil's foreign trade regime. B. Trade and Exchange-Rate Policies 8.7 Foreign-exchange controls were introduced in Brazil in 1947, and have been a permanent fixture ever since. This control, at varying degrees of restrictiveness, has been a major instrument used by import administrators to protect and promote particular industrial activities. 8.8 By the early 1960s, import-control policies were supplemented by measures to promote exports, particularly of manufactures. These included: a drawback scheme for exports; a large devaluation in 1965, followed (in 1968) by a policy of maintaining a stable real exchange rate through a "crawling peg"; and direct fiscal and financial subsidies to manufactured exports, introduced in 1968 ind increasing in size particularly since 1973. Like the protection of import substitutes, export-promotion policies have been practiced in a discriminatory way, with special encouragement for specific activities (such as textiles, automobiles, or capital goods). An important accompanying role has, on occasions, been played, by policies that encourage direct foreign investment in industry and particularly in export activities. The introduction of export-promotion measures has moved Brazil's trade regime close to "neutrality' between import-substituting and export-promoting activities. But this broad neutrality is still accompanied by a high d?gree of discrimination within each category. 8.9 The debt crisis of the early 1980s led to radical adjustment measures. Through a large real devaluation, and a substantial recession in 1983, a large enough export surplus was created by 1984 to service the debt (in relation to GDP, this surplus was even higher than that of Japan or Germany). Since then, the targets of trade policy have alternated between lowering inflation and maintaining the required external balance. The main instrument was the turning on and off of the import 'tap', through foreign- exchange controls and other non-tariff import barriers. 8.10 Despite the heavy use of controls and restrictions, economic policy in Brazil has traditionally aimed to maintain over long periods a rather stable real exchange rate. This may be seen from Figure 8.5. -161- Figure 8.5: The Real Exchange Rate, 1965-1989 (Average 1980 = 1 00) (a) Annual Averages 1965-1989 o o00 - p - 90- 80 - 70- 5 60- -. - 50 -_ so-l J~~~~~~~~-1 40 - 3 0 20 - 1 0 - 0 -- I - 1965 '970 1975 1'980 985 (b) Monthly Averages, 1985-1989 1Pa 100-4 10 90 hI phug 804 i B0 - 50 - 40 30 - 20 - 10 0 t - . . . - T~ T m r 1 '-- --r' -: J-'-T T-'-' r - -- - 1985 1985 1987 I 988 1 989 %ote I'$ -ea,OW 18 Ire .rS$ .' p' e "-'ex (19&0-100 ofd-Dou J'DOi " .,S Je;a- FOeOB' R w -C of Ge'maMy. a"o 1'e UK we g'!oa by ¶"I w8arve 8a'e o ,'esv 'o, CD .t' 88 Ir 8'az a 19610 expOts (IS'ex 10' CC1101:bbe'-6Ce-be 199 'a eat -Vtoo,: trs O0"O" '810' t81' cvoa..s'- p' CS Mnx tw 9'az so.'zf woro SBo-' om8eaooervecgf.o- owoa- :Mr, l-e'ao-oa F8rva'. Sg:a¶,s -162- 8.11 In panel a of the figure, the long-term course of the rate is presented. The course was that of stability over long time periods, accompanied by several upward adjustments which increased the level of the real rate. These adjustments took place in 1973-74 following the first oil crisis; in 1979-80, following the second oil crisis; and in 1982-83, with the appearance of the debt crisis. 8.12 Recently, however, this policy of stability and adjustment of the real rate has been disrupted. As can be seen from pan.,.l b of Figure 8.5, a substantial real appreciation took place in 1980 and, primarilv, 1989. This is due to the (unsuccessful) attempt to moderate inflation through the slowing down of nominal devaluations--to the point of the use of a fixed nominal exchange rate. This policy has, naturally, led to a tighter restriction of imports, in recent years, through the foreign-exchange control mechanism. 8.13 Foreign-exchange controls are handled by the Central Bank, the Ministry of Finance, and CACEX (Carteira de Comercio Exterior--the foreign- exchange department of the Banco do Brazil and the country's trade-control agency). These organizations jointly project the coming year's foreign- exchange earnings; and, after providing for external-debt servicing requirements, allocate the remaining m oreign exchange to imports of goods and of non-financial services (the latter being about 20 percent of the import of goods). Priority is given to "essential" imports, in particular crude oil and wheat, and to inputs into exports. For most other imports, CACEX has been delegated the authority of allocating foreign exchange. It does so through "import programs", allocated to individual private and public enterprises. The agency has the responsibility for attaining monthly targets for merchandise-trade surpluses (the "superavit"), which it achieves through the size of import licences that it grants. Through these two-stage instruments--the allocation of import programs and the issue of import licences--CACEX thus maintains a dominant influence over the economy's merchandise imports. 8.14 The bureaucracy of the impoLt regime is heavy. After negotiating each firm's import program, CACEX scrutinizes applications for individual import licences for balance-of-payments, protective, price control, statistical, and other regulatory purposes. Negative lists are consulted, and the "Law of Similars" is applied. The import licence is the key document that an importer needs to obtain foreign exchange. 8.15 The bureaucracy of exports is generally less constraining: controls in this area are applied mainly to restrict certain agricultural exports. The major fiscal subsidies are granted to firms with BEFIEX (Commisao para Concessar de Beneficios Fiscais a Programas Especiais de Exportacao--Commission for Granting Fiscal Incentives for Special Export Programs) agreements (especially for exports of automobiles, textiles, and capital goods). Preferential export financing, on the other hand, is enjoyed more widely. Export subsidies, while they remain discretionary and opaque, have been reduced considerably in importance since 1984 (from an estimated 37 percent of the gross value of manufactured exports in 1981 to 12 percent in 1985, the most recent estimate available). -163- 8.16 Import licensing, is guided by two key principles. First, CACEX has widely extended the interpretation of the 'Law of Similars." The law limits application of the principle to imports which request preferential (tariff) treatment; but, except on the less frequent occasions when foreign exchange is in ample supply, CACEX appears to apply it informally to all imports (except inputs for the production of exports and inito the Manaus duty-free zone). Thus the whole range of Brazil's industrial production enjoys an absolute and total protection. Second, an informal rule is applied that links firms' import entitlements to their export performance- -a defacto import-retention scheme. The importance of this implicit export-promotion measure varies, of course, with the dearee of foreign exchange scarcity (that is, the degree to which the real exchange rate is below equilibrium). 8.17 Along with the approximate neutrality of the trade regime between exports and import-substituting activities, there has been an explicit policy of preventing a long-term decline and minimizing fluctuations of the real exchange rate. Together, these have limited the ef`ect of the bias against exports that has hurt other economies so badly. But the high degree of intrasector (iscrimination and, in particular, the almost total protection of import-substituting activities, have exerted a high toll on the effectiveness of the economy's utilization of resources. This is particularly true for the 1980s, as foreign-exchange controls and licensing measures have been reinforced since the debt crisis. This intensification of the restrictive policy must have worsened the allocation of resources, as the misdirection of investments increased. As foreign exchange became scarce at the official rate, domestic producers have often been guaranteed monopoly rights which allowed excess profits, inefficiency, or, most probably, both. The operation of the price system tended more and more to be supplanted by administrative decisions, in which higher transaction costs must be involved. C. Recommendations: Reform of the Trade Regime 8.18 We recommend several measures of trade-policy reform. They are designed to open Brazil's economy and to eliminate the substantial economic losses created by the present regime's misguided allocation of resources and gross inefficiencies. 8.19 The objectives of such reform are today widely and publicly embraced in Brazil. In fact, since late 1987 some steps in this direction have been undertaken. These included some tariff reform--a reduction in the tariff system's average level, its dispersion, and the extent of preferential concessions; and an increase in the size of the import programs, which the government perceives as the first step in the gradual process of eliminating foreign-exchange controls. But without credible reform of the foreign-exchange allocation system, these recent changes could not be very meaningful. The tariff changes may, in fact, be counter- productive as long as quantitative restrictions remain the binding constraint on imports. The very gradual process of liberalization contemplated by these recent measures is likely to be unsustainable. Liberalization would be subject to intense speculative pres-ures: in the current climate, economic agents would be convinced (in view of consistent -164- experience in the 1980s) that liberalization today would be followed by restrictions tomorrow, and a consequent surge of imports would be likely to validate such expectations, We thus recommend that even gradual liberalization--and the process must be gradual--should start with a bold step. The process should have two stages. In the first, carried out in one stroke, the bulk of non-tariff controls on trade should be removed; in the second, gradually-implemented, stage, the lon-er-term issues associated with the level of tariff protection should be tackled. 8.20 Stage 1: immediate reform. The first stage should consist of a large devaluation coupled with the dismantling of foreign-exchange controls and other non-tariff barriers, thus completely eliminating the QRs. This would obviously be a rational policy only in the context of a credible stabilization policy, which will be assumed here. That reform, likewise, could be sustained only with policies which unify the foreign-exchange markets and stabilize, over the long haul, the real exchange rate. 8.21 In a very rough way, the removed QRs would be replaced by equivalent levels of tariffs and surcharges (which would be lower with the recommended devaluation than without it). "Equivalence" should not be strict and precise. Instead, a few tariff-type categories should be determined, and imports would be classified into one of these categories. None of the tariff levels should be exceedingly high, and some 150 percent (which is likely to grant an effective protection level of several hundred percent) should be regarded as a ceiling. Thus, imports of various goods are likely to increase, somewhat lowering the level of protection while, in the main, changing the fcrm of protection. In particular, some imports of goods which are now absent altogether are likely to appear with this change. 8.22 To emphasize the transient nature of the increased protection through the price mechanism, and to facilitate its future gradual reduction, it would be desirable not to increase the existing tariff rate. Instead, extra tariff protection should be provided through a schedule of temporary surcharges. 8.23 At the same time, some minimum tariff should be specified--say, 10 to 15 percent--to which today's duty-free imports should be subject. This would raise government revenues in support of the stabilization policy. More important in this context, it would lower the degree of dispersion and discrimination involved in the tariff system and the effective protection granted to activities which use the goods subject to the minimum tariff. This minimum tariff should be effective, with no exemptions granted (except for drawbacks of import duties in exports and as required under Brazil's international agreements). 8.24 Using the occasion of a substantial devaluation, export subsidies should be eliminated (save, again, the drawback scheme). In combination with the changes in import tariffs, substantial net revenue should be forthcoming from the system of external taxes and subsidies. It should roughly offset the added fiscal cost that the devaluation would impose on the government's servicing of the external debt. -165- 8.25 Ouce foreign-exchan.~. controls .- trade have been eliminated, the import programs can have no .urther functxon and should be dispensed with immediately. But. this would still leave a second line of non-tarift defense provided by several controls exercised through the import licence system. Unless these conitrols (perhaps with a very few exceptions) are rapidly eliminated, the move to a price-based system will not have been achieved, the instruments allowing an easy reversal of the reform process could easily be reactivated, and the process would lack credibility. Removal of these restrictions will be facilitated by the substantial progress made, since 1987, in reducing the scope of subordinate non-tariff barriers; in particular, the substantial alleviation of controls exercised through 'he 'external financing requirements.' 8.26 The wholesale application of the 'Law of Similars' may be discontinued simply by the government making it clear that importers are free to avoid the Law's application if they elect to pay the full duty. In fact, if the widespread use of preferences and concessions is discontinued, as recommended here, this Law would then become irrelevant. 8.27 The other two major non-tariff import barriers are the 'list of suspended items' (which has been cut substantially since 1987), and the informatics law. Consolidating these restrictions into a simple small 'negative list", and announcing a schedule for further elimination of items from this list would add transparency and predictability to the more gradual process of removing this residual area of non-tariff import protection. 8.28 Quantitative restrictions on exports are not widespread in Brazil. They are largely limited to agricultural products, where they are often imposed in an attempt to stabilize food prices. These restrictions should be removed as part of the first stage reforms; within the context of a stabilization policy, in which the problem of inflation must be addressed, such controls in any case make little sense. A few exports of manufactures (such as shoes) which face quota restrictions in the importing countries are also subject to quantitative regulations in Brazil. To become more efficient, and secure maximum rents for Brazil, the allocation of such quotas is best done by auction. 8.29 Currently, the average exchange rate--of the 'official' and the "tourist" rates--is judged by most observers to be below its equilibrium level. Hence, the rate should be set at a level which would imply a higher ratio of the price of tradeables to that of non-tradeables. In addition, we suggest here several measures which should require a further increase in the off;cial exchange rate (that is, a further devaluation). One is a move to a uniform rate--abolishing the tourist rate, in which a significant number (although a minority) of external transactions are conducted. Another is the abolition of QRs, and their replacement in many cases with tariffs and surcharges which are less than fully equivalent to the removed restrictions. Third, there is the removal of export subsidies. The result is an obvious need, within this package of trade policies, for a substantial real devaluation. A good estimate of the appropriate size of devaluation would require thorough work which cannot be undertaken in the Wm -166- present context--including the construction of missing data. But it may be guessed that a real devaluation of some 25-30 percent would be, if anything, an underestimate. 8.30 Stage 2: Longer-Term Reform. The chaige in the form of protection--from QRs to regulation by prices--is of the utmost importance. But beyond that, the discrimination among activities should be removed. A substantial body of international evidEnce has built up to suggest that wholesale discrimination is costly--and nothing in Brazil's experience would tend to contradict this finding. Once non-tariff barriers have been removed, the Brazilian Government should proceed to gradually lower the level and dispersion of the tariff system--a process begun, in fact, in 1988. Implementation should be steady and predictabe, and the process should follow the "concertina method",--the lowering, at each stage, of the highest tariff rates in the system. 8.31 Several general guidelines for protection, or intervention, should be adopted for the end of this process. First, any market failure which generates a call for infant-industry support should be tackled at the source. This may mean reform of domestic capital markets, improved training opportunities or incentives, or better protection of the intellectual property rights of Brazilian firms. Second, infant-industry support should be as transparent as possible, so that the benefits of such protection to the economy should be clear, lcwering the risk that benefits would instead go to entrepreneurs seeking the protection as a source of extra rents. Third, infant-industry protection should be limited to very few activities rather than being applied to a broad range of industries; obviously, the more widespread the protection, the less effective it is in shifting resources to the targeted activities. Finally, whenever it is granted, infant-industry support should have a finite (and limited) horizon; the more it turns into a permanent privilege, the more likely it is for an unbridgeable gap to open between local and world prices. 2. Capital Flows and the External Debt A. Sources of the Debt 8.32 The stock of foreign debt is roughly equal to the accumulated flow of foreign capital into the economy.54 The latter is, in turn (in its net concept) the converse of the country's current-account deficit. Causality may, however, run either way--from the current account to the capital 54/ Except for changes in the valuation of assets and liabilities, the annual change in the stock of an economy's net foreign indebtedness is identical to its net capital inflow from abroad. This identity does not hold for the stock of "debt" as it is usually referred to--including in the present context--because the latter is a gross concept (which thus does not take account of changes in items such as - the country's foreign-exchange reserves) and because the stock of assets held by foreigners through a direct investment is not included as part of the country's "debt." In practice, of course, errors and omissions also add to discrepancies. -167- inflow, or the other way around. In Brazil's case, current-account developments started the process of debt accumulation. By 1973, the size of the external debt was minor. The huge increase in the bill for oil imports in 1974, together with a significant increase in the cost of capital goods, led to a substantial current-account deficit. The size of the debt about doubled by the end of this year. In the late 1970s and early 1980s, it was probably the ample availability of foreign credit, at a very low real interest rate, which must have led to large-scale borrowing from abroad. That, in turn, through the familiar channels of the transfer mechanism, led to a corresponding import surplus. The current-account deficits and the size of the external debt are recorded, for 1973 onward, in Table 8.1. Table 8.1: THE CURRENT-ACCOUNT BALANCE AND THE STOCK OF EXTERNAL DEBT, 1973-89 Current-Account Deficit, As a External Debt, Adjusted Value US$ B Percentage End of Year, of the Debt (minus sign = Surplus) of GDP USS B US$ B Year (1) (2) (3) (4) 1973 1.7 2.0 10.0 1974 7.1 6.5 19.1 1975 6.7 5.2 23.7 1976 6.0 3.9 28.8 1977 4.0 2.3 35.1 1978 7.0 3.5 46.5 1979 10.7 4.9 51.5 1980 12.8 5.4 64.2 1981 11.7 4.4 74.0 74.9 1982 16.3 5.7 85.3 87.4 1983 6.8 3.3 93.6 96.6 1984 0 0 102.0 106.8 1985 -.2 -.1 105.1 107.0 198C 5.3 2.0 111.0 108.3 1987 1.4 .5 121.2 112.5 1988 /a -4.4 -1.2 112.3 104.7 1989 /_ -1.0 -.3 111.6 106.9 /a Preliminary IE Projection Source: Columns (1) and (2)--World Debt Tables (1973-79) and Central Bank of Brazil; column (3)--Central Bank of Brazil; column (4)--see text. - 168-. 8.33 Adjustment policies carried out following the debt crisis of late 1982, and the virtual disappearance of voluntary foreign lending, have sharply reduced the current-account deficit. Since 1984 it has on average, been roughly zero. The increase in Brazil's external debt has, concurrently, slowed substantially. Since 1986, with a bump in 1987, the nominal size of the debt shows no further upward trend. 8.34 During these later years, in which the contribution of the current account to the stock c)f debt become minimal, the relative importance of the change in the value of the debt (in US dollars) due to changes in world exchange rates has grown. As will be seen later, about one-third of the debt is denominated in currencies other than the US dollar, so changes in the exchange rates of these currencies with the dol' - change the value of the stock of the debt. Column (4) in Table 8.1 shows ,he adjusted value of the debt stock for 1981 to 1989, namely, the value which would have been found had the exchange rates between the US dollar and the other currencies in which Brazil's external debt is denominated, remained constant at their t980 levels. The adjusted value still changes, from year to year, in ways which are not explained by the recorded current-accoui.t balance. But for the period 1985 to 1989 as a whole, in which the aggregate current-account deficit was about zero, the adjusted value of the stock of debt too shows no change. That is, the depreciation of the US dollar versus other major currencies, during this period, fully explains (in an accounting sense) the increase in the dollar value ot the debt. 8.35 Another contribution to the size of the external debt, during the 1980s, has been made by "capital flight"--unrecorded acquisitions of foreign short-term assets (including cash). Due to their nature, the size of these assets naturally cannot be known with any accuracy or confidence. A rough estimate based essentially on discrepancies between the recorded current-account and capital-account balances (in, of course, opposite directions), puts the accumulation of such assets during 1980-87 at somie US$ 15 billion. That is close to 15 percent of the (adjusted) size of the debt in the late 1980s, or some 30 percent of the addition to the debt during these years. B. The Burden of the Debt 8.36 Many yardsticks may be applied to evaluate the burden of the debt, and nore of these would by itself give an adequate estimate. We present, however, some rough-and-ready debt ratios, to construct a broad idea of the development of the burden. All of these measures face problems, such as the interpretation of shifts which are due to the change in real exchange rate (that is, in the relative price of tradables to --ontradables). Nevertheless, they convey some meaningful impressions. - I 69- Thtile 8.2: L4EASURES OF THE BURDEN OF THE DEBT, 1980-88 (ratios, in percent) Debt Debt interest Interest (End Year) (End Year) Interest Interest Bill to Bill to to Annual to Annual Bill to Bill to Government Domestic GDP Exports GDP Exports Expenditures Investment Year (1) (2) (3) (4) (5) (6) 1980 26.8 276 3.1 32.0 10 13 1981 35.6 275 3.9 38.3 11 17 1982 31.4 364 4.7 53.3 13 22 1983 45.3 384 5.0 42.2 13 30 1984 48.3 338 5.5 37.9 16 33 1985 46.0 359 5.0 38.4 11 29 1986 39.6 446 3.8 41.4 8 20 1987 38.0 422 3.0 33.5 n.a. 14 1988 31.5 313 2.9 28.4 n.a n.a. Source: Calculated from IBGE National-Accounts data, April 1989 8.37 The first two columns in Table 8.2 relate the stock of the debt to flow measures--the GDP and aggregate exports, respectively. By the first yardstick, the size of the debt kept increasing until 1984. Its decline since then brought it to the level of 1982, the eve of the debt crisis. The relative increase in exports during these years thus led the second measure--the ratio of the debt to exports--to decline somewhat following 1982. Altogether, therefore, the stock-of-t1.e-debt measures show approximate stability over the last few years, with some declining trend. 8.38 The flow criteria, showing the burden of interest payments on the external debt, indicate quite a different trend. In the early 1980s this variable increased substantially. Since 1985, on the other hand, all yardsticks show a substantial decline in the burden of interest payments. With the stock of debt remaining roughly stable, or falling only slightly, this decline in interest payments reflects, of course, the decline in the average interest rate charged on Brazil's debt. This, in turn, is predominantly a result of the decline in world interest rates, and to a minor extent the outcome of better borrowing terms for Brazil. The end result is that by the late 1980s. interest payments on Brazil's external 'ebt, although still substantial, are far from being overwhelming. This observation must have significant policy implications. Thus, for instance, it may be noted that a hypothetical reduction of interest payments by as much as 50 percent--through principal or interest forgiveness--would save -170- Brazil about 1.5 percen_ of GDP, or some 6 percent of government expenditures. These are not negligible magnitudes, but nor are they of crucial importance. Table 8.3: DEBT RATIOS, HIGHLY-INDEBTED COUNTRIES, 1987 (percent) Ratio of Debt Ratio of Debt Country to GNP to Exports ; Argentina 74 661 Bolivia 134 897 BRAZIL 39 432 Chile 124 327 Colombia 50 220 Costa Rica 116 315 Cote d'Ivoire 144 374 Ecuador 107 438 Jamaica 176 263 Mexico 78 363 Morocco 132 382 Nigeria 123 369 Peru 41 503 Philippines 87 318 Uruguay 59 256 Venezuela 95 285 Yugoslavia 39 147 Source: World Bank, World Debt Tables, 1988-89: External Debt of Developing Countries. 8.39 International comparisons also indicate that Brazil's external debt does not appear to be overwhelming. Table 8.3 presents data for 1987 for the seventeen countries usually defined as highly indebted. Brazil's external debt/GNP ratio appears to be at the lowest end--much smaller than is common for the group. Brazil's debt is relatively high in relation to the country's exports--a phenomenon consistent, of course, with the fact that the ratio of Brazil's exports to its national income is, as we have noted, extremely low.55 But the intensity with which Brazil's exports respond to changing circumstances would suggest that it is not the 55/ In a similar way, it is likely that in an international comparison of the share of external debt service in government expenditures, Brazil would occupy a high place. -171- availability of foreign exchange (through exports) which constrains the country's ability to service its debt; it is, rather, the problem of raising sufficient domestic resources for this purpose. C. Attributes of the Debt 8.40 The structure of Brazil's external debt can be described in several ways. To begin with, 'registered" debt can be distinguished from "non-registered' debt. The first consists of medium- and long-term debt, refinancing by foreign commercial banks and the Paris CVub, and use of IMF drawing rights. The Central Bank monitors the size and structure of this part of the debt closely. Non-registered debt, on the other hand, consists of the Central Bank's liabilities; commercial lines of credit; and Brazilian commercial banks' net liabilities. It thus includes, inter alia, arrears accumulated with commercial creditors. Table 8.4: REGISTERED AND NON-REGISTERED DEBT, 1980-88 Registered Non-Registered Total Gross Debt Year In Percentage In Percentage (end) US$ B of Total US$ B of Total US$ B 1980 53.8 83.8 10.4 16.2 64.2 1981 61.4 83.0 12.6 17.0 74.0 1982 70.2 82.3 15.1 17.8 85.3 1983 81.3 87.0 12.2 13.0 93.5 1984 91.1 89.3 10.9 10.7 102.0 1985 95.9 91.2 9.3 8.8 105.2 1986 101.8 91.6 9.3 8.4 111.1 1987 107.5 88.7 13.7 11.3 121.2 1988 102.4 91.2 9.9 8.8 112.3 Source: Central Bank of Brazil 8.41 Table 8.4 shcws the composition of the debt, by this classification, during the 1980s. When total debt grew substantially--up to 1984--it was the registered part which was growing. As a result, the registered share increased fast, and is by now overwhelming. The non- registered debt--predominantly debt created in the conduct of commercial transactions--was relativelsy dtable, despite some fluctuations (such as the increase in 1987, due to the debt moratorium of that year and the arrears that followed). The changing shares of the two categories is of some economic significance because non-registered debt--by now less than 10 percent of the total--is the part which is conducted primarily on a voluntary, commercial basis. The registered debt--the bulk of the total- -is by now involuntary, or non-market indu-ed. -172- 8.42 Debt is owed by both the public and private sectors, and this distraction is presented (for the registered debt) in Table 8.5. It may be seen that the public sector's share of the debt has been predominant throughout, and increased further during the 1980s. In contrast, private- sector debt has declined even in absolute terms. Recently, public debt has amounted to close to 90 percent of the total (of registered debt; it is obviously lower in the non-registered part). The increased public-sector share has not been the product of changing borrowing patterns but, primarily, of the process whereby private-sector debt is assumed by the government. Table 8.5: PUBLIC AND PRIVATE DEBT, 1980-88 Public Debt Private Debt Total Registered In Percentage In Percentage Gross Debt Year USS B of Total US$ B of Total US$ B (end) 1980 37.3 69.3 16.5 30.7 53.8 1981 41.8 68.1 19.6 31.9 61.4 1982 47.4 67.5 22.8 32.5 70.2 1983 60.3 74.2 21.0 25.8 81.3 1984 71.8 78.8 19.3 21.2 91.1 1985 78.7 82.1 17.2 17.9 95.9 1986 87.1 85.6 14.7 14.4 101.8 1987 93.1 86.6 14.4 13.4 107.5 1988 88.9 87.4 12.8 12.8 101.7 8.43 This transformation has occurred through implementation of a deliberate policy. tThen a private-sector loan reaches maturity, the borrowing agency does not repay the debt to the foreign lender. Instead, it rays the Brazilian Government, by depositing the equivalent amount in local currency at the Central Bank. From that moment on, the government assumes the obligation to the foreign lender. Under these "Deposit Facility Agreements", US$ 30 billion of matured debt has been transferred to the government between 1983 and fall 1988. Of this, about US$ 19 billion were owed to start with by public-sector agencies (or guaranteed by the government), but the remaining US$11 billion represent the transformation of purely private debt. In addition, legislation in effect since 1976 has established the possibility that private-sector borrowers can buy working payments on external debt before maturity. This they can do, again, by depositing the equivalent amount at the Central Bank and transferring to the government, in effect, the external liability. This procedure covered about 17 percent of the private-sector external debt by 1988. Table 8.8: EXTERNAL DEBT BY CATEGORY OF LENDER, 1983-87 Foreign Foreign Branches Comarcial of Brazilian Multilateral Banks Banks Organizations Bilateral Total Percentage Percentage Percentage Percentage Year(end) US$ 8 of Total US8 8 of Total USS B of Total US8 B of Total US8 8 1983 64.3 70.1 8.0 8.7 7.0 7.6 12.4 13.6 91.7 1984 70.7 89.2 8.3 8.1 9.7 9.6 13.4 13.1 102.1 1986 68.6 86.2 8.3 7.9 12.0 11.4 18.3 15.6 106.2 1988 70.8 63.1 8.3 7.6 n.a. n.a. n.a. n.e. 111.8 1987 76.6 83.1 7.9 6.5 14.8 11.8 22.8 18.8 121.2 Source: Central Bank of Brazil, Economic Progra -174- 8.44 Another important distinction is by foreign lender: commercial banks, foreigr. governments, or multilateral organizations. Table 8.6 provides this classification. The major part of the debt--roughly two- thirds--is owed to foreign commercial banks. But the share owed to multilateral organizations and bilateral lenders has been increasing. Together, they constituted roughly 30 percent of the total debt in 1987, versus some 20 percent four years earlier. Within the 'bilateral' sector, the Paris Club's share of the debt has increased substantially. By 1987, this part of the debt amounted to some US$8 billion or about 7 percent of the total external debt. A relative decline in the share of debt owned by Brazilian banks is a reflection of the relative reduction in the size of the "non-registered" debt: foreign branches of Brazilian banks engage mostly in shorter term financing of the country's external commercial transactions, and in inter-bank lines of credit, rather than in long-term lending. 8.45 The decline in foreign commercial-banks' share of lending may conceivably lower the base for potential debt-alleviation schemes. But this decline has been small enough--and the share itself high enough--to make this change of little relevance at the present stage. 8.46 Still another attribute of the debt is the nature of interest- rate commitment: fixed versus variable rates. Throughout the decade, the bulk of registered debt--75 to 80 percent--haa been subject to a variable rate (f3r the latest date for which this information is available, June 1988, this proportion was about 74 percent). This implies that the interest bill on Brazil's external debt moves roughly in tandem with world interest rates. As we have noted, during most of the 1980s this has worked to Brazil's advantage. -175- Table 8.7: CURRENCY COMPOSITION OF THE REGISTERED DEBT, SEPTEMBER 1989 Value Percentage Currency USS B of Total U.S. Dollar 69.5 69.8 Japanese Yen 10.6 10.7 German Mark 7.0 7.0 French Franc 3.3 3.3 Swiss Franc 2.3 2.3 British Pound 1.2 1.2 SDRs 3.1 3.1 Other 2.6 2.6 Total Non-Dollar 30.1 30.2 Total 99.6 100.0 Source' Central Bank of Brazil. 8.47 Finally, the currency composition of Brazil's debt is of interest. This is presented in Table 8.7, for the registered debt at September 1989. About 30 percent of the external debt is denominated in currencies other than the US dollar--primarily the Japanese yen and the German mark--and we have noted earlier the implications for changes in the valuation of debt, in US dollars. By itself, this is of little material consequencet the aggregate size could be arbitrarily designated in any currency, leading to alternative changes in valuation which differ in both size and direction. What is of economic significance, however, is the fact that the currency composition of the debt, and hence of debt servicing, differs from the currency composition of trade flows (in goods and non-financial services). Hence, changes in third-currencies rates of exchange do have an mpact--in one direction or the other--on the country's real debt burden.56 Specifically, appreciation of the yen or mark tends to increase this burden, D. Debt-Transformation Schemes 8.48 The size of the country's international indebtedness may be reduced--other than by a repayment--through one or another debt-alleviation schemes--which reduce either the principal of the debt or the interest charged on it. Such schemes are, however, subject to negotiations and 56/ This statement should not be interpreted as implying that Brazil's target should be the achievement of a currency composition of the debt which is equal to that of its trade: the latter is only one of the considerations in determining an optimum composition of the country's liabilities. -176- agreements between Brazil and its creditors. We shall discuss only U potential policies which may originate solely from the government of Brazil, and in which the partners to the transactions act voluntarily so as to maximize their profits. These may be schemes which transform the nature of the country's indebtedness rather than its size; or change the recorded value of the country's debt, when existing debt is replaced by another instrument of a lower face value. 8.49 Three principal forms of debt transformation may be distinguished. One is a "buyback'--the repurchase of a debt instrument with cash (normally prior to maturity of the J.nstrument, and obviously at a lower price than the face value). Another is a debt exchange--the substitution of one debt instrument for another; in this case, the new instrument would have a lower fare value, and would in one way or another enjoy greater prospects of payment of interest and principal. The third avenue is the exchange of debt for equity. In Brazil, it is predominant;y this latter form of debt transformation which has been practiced in recent years, and will likely remain so in the near future. 8.50 Debt-for-equity transformation in Brazil has been performed through "formal" and "informal" channels. In the "formal" channel, conversion is done through the Central Bank, and has in fact been conducted since the mid-1960s. Since 1978 and, in particular, since the end of 1982--following the debt-crisis--fiscal incentives have been applied to this scheme. Essentially, the swap is made by registering purchased equity at the appropriate department of the Central Bank, which leads to the release of the deposited debt payment at the Bank for the finance of this purchase of stocks. The precise regulations governing this scheme keep changing; specifically, it is on occasion opened to any owners of the debt, while on other occasions it is applied only to the original creditor foreign banks. All debt (private and public) which has matured is potentially subject to such conversion, as is private debt before it reaches maturity. This program, as such, was suspended in February 1988, but should it be reopened, about USS60 billion worth of debt would qualify for it. During 1988, this program has been supplemented by a scheme for the auctioning of stocks, through which the amounts involved became substantially higher. 8.51 "Informal" conversions started in 1985, and have increased sharply since the Fall of 1987, via the application of various schemes. Essentially, the process leads to an investment by a foreign multinational company in its Brazilira subsidiary. The existing company transfers debt certificates bought in the secondary market to a non-resident account in Brazil. The subsidiary then sells these certificates at a higher price to the original debtor in Brazil. Both public and private-sector enterprises have participated in these schemes. -177- Table 8.8: DEBT-EQUITY CONVERSION AND DIRECT FOREIGN INVESTMENT, 1975-88 (millions of US$) Formal Debt Direct Foreign Year Conversions Investment (Gross) 1975 56 1081 1976 82 1191 1977 108 1015 1978 160 1161 1979 207 1831 1980 39 1595 1981 2 1903 1982 143 1370 1983 452 567 1984 746 490 1985 581 485 1986 206 433 1987 344 651 1988 /a 2051 636 /a Preliminary 8.52 Table 8.8 shows the amounts involved in formal debt conversions, since the mid-1970s. It appears that only in 1988 was the size of these swaps substantial. No equivalent data are available, for a longer period, for the size of informal swaps. But the estimate available for 1988 puts the size of these conversions at US$4.6 billion--a substantial amount by any yardstick. Adding US$3.6 billion of formal conversion,57 that indicates a total debt conversion of over 8 billion dollars for the year. If debt conversion could be maintained at anything close to the 1988 pace, it would become a major factor in determining the size of Brazil's debt (although, of course, not of the ownership of financial assets by foreigners, which includes equity as well as debt instruments). Also as will be pointed out shortly, debt swaps of this size far outweigh the amount of other foreign investments in recent years. 8.53 In debt-for-debt conversion, Brazil's experience so far has been quite limited. Within the 1988 Financing Package, a ccnversion of up to USS 5 billion was allowed into Brazil Investment Bonds--better known as Exit Bonds. To enhance their value, three features were added to these bonds: eligibility for debt-for-equity conversion; eligibility for debt- 57/ The discrepancy between this figure and the one given for 1988 in Table 8.8 (US$2.1 billion) is due to differences of timing in recording transactions, and of methods of evaluation (face vs. market values). -178- for-export swap, if such a program is ever implemented; and a conversion option into a foreign-exchange-indexed cruzado instrument. In return, the interest rate on these bonds is fixed at the low level of 6 percent, and the maturity is 25 years, with a 10 year grace period. In effect, only about US$1 billion worth of these bonds have been subscribed and issued by August 1989. -~~~~~~~~~~~~~~~~~~~~~~~~~ -179- Table 8.9:: IBRD PI-SM BALANCE OF PAYMENTS PROJECTIONS 1987 1988 1989 1990 1991 1992 1993 1994 Merchandise Exporte 26.2 33.7 34.0 36.3 42.9 60.3 69.2 69.4 Merchandise Imports 16.1 14.6 18.0 21.9 28.1 36.4 43.6 62.6 Trade Balance 11.2 19.1 16.0 14.4 14.7 14.9 16.7 17.0 Resource Balance 10.3 18.1 14.8 12.6 12.1 11.3 '1.4 12.1 Net Factor Service Income -11.9 -13.8 -14.4 -14.4 -13.6 -13.? -14.2 -14.2 Net Current Transfere 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Current Account Balance -1.6 4.4 0.4 -1.8 -1.3 -2.2 -2.6 -2.0 Capital Account 1.6 -4.4 -0.4 1.8 1.3 2.2 2.6 2.0 Direct Foreign Investment (Net) 1.1 2.7 0.6 0.8 1.0 1.2 1.6 1.7 Short Term Capital -0.6 -1.6 0.0 0.0 0.0 0.0 0.0 0.0 Net LILT 3.2 0.6 -1.2 4.6 3.3 4.4 4.7 4.1 Net Brazilian Lending Abroad 0.1 0.0 -1.2 -1.0 -1.0 -1.1 -1.1 -1.2 Capital n.e.i. -1.6 -6.1 3.3 -0.8 0.0 0.0 0.0 0.0 Change in Reserves (- increase) -0.7 -1.0 -1.8 -1.7 -2.0 -2.3 -2.4 -2.6 -180- E. Expected Course of the Debt 8.54 The external debt has been inferred by the World Bank from balance of payments projections, using an RMSM model, for the next five years (1990-94) (see Table 2.2). These are presented in Table 8.9. These projections disregard explicitly the potential impact of debt conversicn, which has just been discussed. The current-account deficit is estimated, in this exercise, to be of the rough order of US$2 billion per year. This would therefore be the increase in the country's total net indebtedness to the outside world (abstracting from the potential impact of changes in asset valuations). Direct foreign investment is expected to be about USSl-l.5 billion, annually; but a similar value is projected for increased lending abroad of Brazilian banks. In addition, with the increased role of trade, an annual increase of foreign exchange reserves of some US$2-2.5 billion is projected. The increase of Brazil's net medium and long-term debt is thus expected to exceed somewhat the accumulation of cuLrent- account deficits; and to increase, over this p3riod as a whole, by some US$20 billion. 8.55 As a ratio to Brazil's GDP, the external debt--increasing at an annual rate of some 3-3.5 percent--is expected to fall. A more substantial fall is projected for interest payments, following an expected decline of the interest rate: from about 3 percent of GDP in 1989, the interest bill - on the external debt is expected to decline, gradually, to less than 2 percent by 1995. 8.56 In this context, debt-transformation schemes should again be H mentioned. In view of the size of expected increment of the debt, transformation into equity could potentially turn an increase of the debt to a significant reduction, not just in relation to the country's income but as an absolute level. While such transformation does not change the transfer of resources to foreigners, in the form of interest or dividends, it is of consequence for the consideration of the budgetary burden of the debt as well as, presumably, for the foreign-exchange requirements. The transformation may become particularly relevant within the context of privatization of public enterprises, as has been discussed in Chapter VII. F. Foreign Investment: Recent Developments 8.57 Until recently, Brazil has been very successful in attracting and retaining foreign investment. As Table 8.10 shows, between 1969 and 1988, the stock of registered foreign investment increased from US$1.71 to US$30.70 billion, making Brazil by far the largest repository of foreign investment in Latin America. The principal countries of origin for this investment are the United States (29.0 percent), West Germany (15.4 percent), Japan (9.6 percent), Switzerland (9.2 percent), the United Kingdom (6.2 percent), and Canada (4.6 percent). -181- Table 8.10: REGISTERED FOREIGN INVESTMENT IN BRAZIL, 1969-88 (millions of US$) Annual Direct Increase Investment Reinvestment Total (percent) 1969 1,185 525 1,710 1970 1,545 801 2,347 37.25 1971 1,789 1,121 2,911 24.0 1972. 2,080 1,323 3,404 16.9 1973 2,858 1,720 4,579 -4.5 1974 3,925 2,102 6,027 31.6 1975 4,902 2,400 7,303 21.2 1976 6,193 2,811 9,005 23.3 1977 7,539 3,688 11,228 24.7 1978 8,898 4,841 13,740 22.4 1979 10,595 5,367 15,962 16.2 1980 11,994 5,484 17,480 9.5 1981 11,994 5,485 17,480 9.5 1982 14,634 6,542 19,246 10.1 1983 15,548 6,754 22,302 5.3 1984 16,340 6,503 22,844 2.4 1985 17,925 7,739 25,664 12.3 1986 18,198 8,024 26,221 2.2 1987 18,748 9,150 27,898 6.4 1988* 20,843 9,860 30,703 10.0 *As of September 30, 1988. Source: Central Bank 8.58 Foreign investors have been attracted by Brazil's vast size (fifth largest area), large population (sixth most numerous in the world), vast market (eighth largest), plentiful natural resources, rapid economic growth, inexpensive labor, and generous tax and credit subsidies. They have also been attracted by the relatively liberal open-door legal regime, and Brazil's long history of resolving foreign investment disputes pragmatically. 8.59 The directions (and perhaps the size) of foreign investment have been heavily affected by the country's pervasive protection against imports. The 'Law of Similars", in particular, has been a critical factor in persuading many foreign companies to set up manufacturing plants in Brazil. 8.60 With the same degree of protection, however--or even bigger-- Brazil's attractiveness to foreign investors--has faded considerably in the past six years. Since the debt crisis exploded in 1982, the rate of growth of accumulated foreign investment has slowed appreciably, averaging only -182- 6.4 percent annually from 1983 to 1988 compared to almost 22 percent annually from 169 to 1982. Table 8.11 shows that new direct foreign investment (exclu6ing debt/equity conversions) has fallen steadily from a high of USSl.9 billion in 1981 to a low of US$433 million in 1986. Not only has new direct investment declined, but both capital repatriation and profit remittances have increased substantially. Capital repatriation, which had averaged only US$176 million a year between 1978 and 1985, soared to US$637 million in 1986; profit remittances, which had averaged only US$573 million a year from 1979 to 1984, climbed to US$1.237 billion in 1986, and reached a record of US$1.539 billion in 1988. Some of this incr&se may, of course, be due to the increased size of the accumulated stock; but this could not be a major explanation. Table 8.11: DIRECT FOREIGN INVESTMENT. CAPITAL REPATRIATION AND PROFIT REMITTANCE (EXCLUDING CONVERSIONS) (millions of US$) Net Direct Net New Direct Capital Investment Profit Inflow Investment Repatriation (=(1)-(2)) Remittance (=(3)-(4)) (1) (2) (3) (4) (5) 1978 1,160.6 124.1 1,036.5 560.5 476.0 1979 1,831.2 353.5 1,477.7 635.7 842.0 1980 1,595.1 147.1 1,448.0 309.8 1,138.2 1981 1,903.2 110.3 1,792.9 369.6 1,423.3 1982 1,369.9 143.0 1,226.9 585.2 641.7 1983 567.0 157.8 409.2 757.8 -348.6 1984 490.1 112.3 377.8 796.1 -418.3 1985 485.2 262.9 222.3 1,056.5 -834.2 1986 432.9 761.7 -328.8 1,236.7 -1,565.5 1987 651.4 326.0 325.4 909.4 -584.0 1988 635.9 289.7 346.2 1,540.0 -1,193.0 Source: Central Bank 8.61 Several factors account for the recent decline in Brazil's ability to attract and retain foreign investment. One is the debt crisis, with its sharp reduction in foreign lending, and the accompanying decline in overall economic growth. Another is the uncertain investment climate resulting from the unilateral debt moratorium, the anti-foreign investment measures in the 1988 Constitution, and the high levels of inflation. Still another -183- ! cause lies in changes in US tax legislation, making it less advantageous for US multinationals to earn arid to accumulate profits abroad, particularly in a high tax jurisdiction such as Brazil. 8.62 It is also quite likely that foreign investment through debt- equity swaps has at least partly substituted for direct foreign investment recorded as such. If it is assumed that such substitution has been full, and the debt-equity conversion investment recorded in Table 8.8 is added to column (1) in Table 8.11, a reduction of foreign investment since 1983 would still be apparent; but it would be less dramatic. G. Foreign Investment: Regulatory Framework and Policy Recommendations 8.63 Brazil has never adopted a comprehensive statute regulating foreign investment, nor has it set up a commission or agency to screen foreign investment. With certain important exceptions, foreign investors are free to invest directly in any kind of legitimate venture in Brazil and to set up business anywhere in the country. No special permission is required for foreign investors to buy or organize a Brazilian company. Only with respect to direct portfolio investment does Brazil presently impose serious barriers to entry. Instead, Brazil regulates the foreign investor through a flexible combination of incentives and disincentives, including exchange controls, direct and indirect taxes, tariffs, price controls, constraints on transfer of technology, subsidized financing and investment incentives. 8.64 The key to the regulatory regime is registration of foreign investment with the Central Bank. Foreign investment does not have to be registered, and substantial amounts are not registered. But unregistered foreign investment cannot be legally repatriated, nor may dividends from it be legally remitted. The holder of an unregistered investment is, however, free to enjoy the fruits of his investment in Brazil or to sell his investment to a Brazilian resident Lr other foreign investor. 8.65 Brazilian law permits a foreign investor to register any investment in the form of money or assets, such as machinery and equipment, transferred to Brazil from abroad and utilized in economic activity in Brazil. But the Central Bank takes a narrow view of what constitutes economic activity. Intangible property, such as patents, trademarks or know-how, is rarely recognized as a foreign investment. 8.66 Foreign investors are able to repatriate all registered capital. They are also able to remit all profits, provided they are willing to pay a 40-60 percent supplemental income tax on profit remittances averaging more than 12 percent per annum on their registered capital. Reinvested earnings can be added to the registered capital base for future remittance purposes. Foreign currency loans must also be registered with the Central Bank in order for interest and amortization payments to be made legally. Registration has always been regarded as optional rather than mandatory. A substantial amount of foreign investment in Brazil is not registered with the Central Bank, and when this occurs, profit remittance on, and repatriation of the unregistered investment is accomplished by resorting to a variety of legal and extra-legal mechanisms. -184- 8.67 Repatriation of registered capital is not subject to tax and may be done at any time without prior approval of t.l.e Central Bank. Nevertheless, when a foreign-owned firm seeks to liquidate and repatriate its entire Brazilian investment, Brazilian authorities treat the last fivp years of reinvested earnings as dividends rather than as return of capital and impose a 25 percent tax on those five years of previously capitalized profits. On th-e other hand, foreign investors are free to sell their investment outside of Brazil for foreign currency or stock. 8.68 Most investment incentives in Brazil are available to both domestic and foreign-controlled firms; but some are restricted to Brazilian controlled firms. A fundamental conflict exists between the New Industrial Policy, which extends a series of benefits and incentives to foreign as well as national firms, and Article 171 of the 1988 Constitution, which specifies the extension of fiscal incentives and concessions to Brazilian firms of national capital. This contradiction will need to be resolved. 8.69 The Remittance of Profits Law is also problematic. It assumes that the ratio between invested capital and net profits is similar in all forms of investment. The 12 percent net dividend trigger for the supplemental tax levy severely discriminates against firms whose registered capital base is generally quite low in relation to their net earnings such as those in the services sector; and against firms in high risk activities, such as mineral exploration or high technology, which require higher returns to compensate for the higher risks. It also discriminates against franchising operations, which normally involve outlay by the foreign franchiser. 8.70 This tax structure, combined with the foreign investment regulations, also exacerbates Brazil's external debt by offering incentives to foreign investors to register their investment as debt instead of equity. Remittable interest on foreign loans is geared to market rates, while remittable dividend income is fixed at 12 percent. 8.71 Given Brazil's size, sophistication, and present state of development, it would be preferable to repeal the Remittance of Profits Law and modify the exchange control system that provides the incentive for registration. The present system results in a thriving black market and increased transaction costs for private investors and the government, and provides incentives for foreign exchange fraud. Because Article 172 of the new Constitution mandates enactment of a statute regulating foreign investment and profit remittances, it may be infeasible to repeal the existing regime. We therefore consider piecemeal reforms that would significantly improve the existing system of regulation of foreign investment. 8.72 Priority should be given to modifying the supplemental tax on dividends for remittance to foreign investors so that it is triggered only by dividends in excess of a floating percentage of registered capital--say five percentage points above the LIBOR rate. An alternative way of improving the supplemental tax is to subject registered foreign investment -185- to monetary correction. Another desirable change is to aiter the present supplemental tax so that firms engaged if services, franchising, and high risk activities, would be taxed at a lower rate. 8.73 A less critical but nevertheless desirable reform would be to permit foreign investors to make direct portfolio investment in the Brazilian c.cock market and to liquidate or change their investment at any time. Investors contemplating investments in a thin market such as Brazil's need to be able to invest and to divest at any time. The present requirement that investments be made through mutual tunds or managed portfolios and remain in the country for a fixed period is unnecessarily rigid. Another measure that would help stimulate direct investment in the Brazilian stock market is to permit foreign investors to register earnings reinvested in the shares of Brazilian companies other than the one for which zheir foreign investment was originally registered as foreign investment, and thus allow them to avoid the 25 percent tax on foreign remittances. 8.74 FinRlly it is recommeded that Brazil should reduce the present tax differential between dividends and interest payments, in order to induce more firms to convert their own debt into equity. A foreign investor's dividend income is effectively taxed at 56.3 percent (aggregating the income and surtax taxes levied on the Brazilian payor and foreign recipient and assuming no supplemental tax liability or excess profits tax), while his interest income is taxed at 25 percent. -186- IX. INDUSTRIAL AND AGRICULTURAL POLICIES 1. Industrial Policies A. Policy Framework 9.1 This chapter will analyze the salient policies applied in the two major sectors of production of goods in the economy: industry (manufacturing and mining) and agriculture. The former accounts, in re-ent years, for some 30 percent of the GDP; and the latter, despite a gradual process of relative decline (conmon, of course, to all countries at this stage of development) still generates about 10 percent of GDP. 9.2 A substantial amount of overlap Inevitably exists between the substance of this discussion and the analyses of the budget (Chapter VI) and of the external sector (Chapter VII). To economize, we shall try to minimize repetition. In particular, we shall refer only briefly to trade policies, which have been discussed in the last chapter. But these policies are of crucial importance in both sectors, and their absence from the present discussion should not be construed as implying otherwise. To a large extent, this holds also for various tax policies which have been discussed on earlier occasions. 9.3 Industrial policies in Brazil have consisted of three primary components: import protection and export promotion policies, fiscal incentives, and financial subsidization. In addition, some direct regulation of investment has played a less significant role. Machinery for the regulation of investment has operated to apply a combination of these policies. 9.4 The key regulatory agency in Brazil is the Council for Industrial Development (CDI). Along with BNDES (the financial arm whose activity will be discussed later), CDI plays a major role in determining the application of regulatory and promotional policies. It was created in 1969; and while its importar.. diminished after 1979, a reform of industrial policy in May and Septemb. 1988 has re-established its prominence. CDI has two different, though related roles. One is regulation activity, rationing investment through administrative means. This licensing has been a major regulatory barrier only in a small number of areas: notables are petrochemicals, fertilizers, steel (rolled flats), shipbuilding, linformaticsf goods, and wheat milling. 9.5 CDI's other and more important role involves the certification of industrial activities for the receipt of financial assistance. This is done through two channels. First, CDI is the authority which certifies that enterprises meet domestic-content requirements. This has been an important instrument in the import substitution policy, specifically in connection with self sufficiency in capital and intermediate goods. The certificate is necessary for a firm to qualify for investment credit from government agencies (such as BNDES) and for it to become a supplier to the public sector. Without a certificate, the buyer of the firm's product is also required to pay import taxes on it, and the purchaser's annual import quota would be correspondingly reduced by the amount of the purchase. 9.6 The second channel of CDI's promotional activity is the evaluation of industrial projects for granting investment incentives: the agency has -187- a statutory authority for the assessment of eligibility in a broad range of subsectors. The incentives involved would be preferentiel credit treat.nent, tariff concessions on imports of capital goods and components, and other fiscal incentives. The absence of CDI approval would mean substantiallv higher investment costs, and would usually block the project altogether. 9.7 CDI deliberations are not entirely transparent, and no binding, published guidelines are available. However, two considerations seem to have been of prime importance. One is a demand-supply balance: this principle seems to have been predominant at least until the mid-1980s. Generally, project proposals were rejected whenever the supply of existing firms was sufficient to satisfy demand (presumably, at the existing prices). The notion of a minimum efficient scale is the other broadly- applied criterion. In the absence of free entry and competition, existing rents would make it profitable for plants of non-optimal level to be established; and CDI has seen as its objective the prevention of such investments. 9.8 In May and September 1988, laws were enacted to introduce a "new industrial policy." The change in procedures primarily involves the elimination of "domestic content" certification: this has been replaced by a declaration from the manufacturer that a pre-specified level of domestic content has been used which qualifies the firm's output as a 'national product." In general, the "new policy" differs from its predecessor by declaring as targets: (a) the promotion of domestic and international competition as important objectives; (b) the reduction of government intervention in investment decisions, and the simplification and increased t-ansparency of such intervention; (c) an emphasis on technological development and quality improvements; and (d) the establishment of a path of reduction over time of the incentives granted. At the same time, however, the new legislation extends the powers of CDI; specifically, it allows the granting of incentives on the basis of criteria specific to individual projects rather than entire subsectors, thus enhancing the discretionary powers of CDI. B. Fiscal Incentives 9.9 During the 1980s (until the "new industrial policy" of late 1988), three main categories of fiscal incentives may be distinguished. First, there was the reduction of, or exemptions from, tariffs and the two VAT- type taxes--the industrial-production tax (IPI), and the commodity- circulation tax (ICM)--for imports of capital goods, intermediate inputs, parts, and raw materials. Second. there was accelerated depreciation for domestically-produced capital goods. Third, there were tax credits on purchases of locally-produced capital goods and on the sale of steel by integrated steel producers. 9.10 Tariff reductions were granted by CDI, BEFIEX (Comissao para Concessao do Beneficios Fiscais a Programas Especiais de Exportacao- -Special Program of Fiscal Incentives for Exporters), CONSIDER (Conselho de Nao-Ferrosos e de Siderurgia--Steel ar.d Non-Ferrous Metals Council), and CPA (Comissao de Politica Adusaeira--TarJff Policy Council). BEFIEX has, since the late 1970s, become a major agency for the granting of fiscal -188- concessions, substituting in part for CDI. It is in charge of a selective but intensive export promotion program, and is a major source of fiscal concessions for producers able to precommit themselves to export targets. 9.11 CDI grants tariff reductions to specific products and projects. BEFIEX, on the other hand, grants these concessions to firms rather than to projects: a firm which imports a good under a BEFIEX concession may sell this good in either domestic or international markets, which made BEFIEX concessions more attractive than CDI support. Qualifying commitments for a BEFIEX support are: (a) fulfillment of specific export targets, usually within a period of ten years; (b) a positiva trade balance for the firm, either throughout the period or, sometimes, with a grace period; (c) specified investment commitments; and (d) acquisition of a pre- specified level of domestically produced capital goods. 9.12 CONSIDER. creat.d in 1968, had originally provided incentives only to steel plants. -n 1974, its scope expanded to include the industries which produce non-ferrous metals (aluminum, lead, copper, tin, nickel, silicon, and zinc). Table 9.1: RATES OF FISCAL INCENTIVES GRANTED BY CDI, BY SECTOR, 1986 (in percentage of the value of approved projects) Sector Rate Capital Goods .7 Basic Metals and Metal Products 22.9 = Chemicals, Petrochemicals, and Pharmaceuticals 13.5 Non-Metal Intermediates, Cement, and Paper 15.3 Automotives and Components 11.8 Consumer Goods 31.7 TOTAL 15.0 Source: R. Baumann (1988), "Uma Avaliacao de Estrutura dos Incentivos ao Investimento na Industria Brasileira," mimeo. 9.13 An impression of the significance of the fiscal incentives may be gained from Tables 9.1 and 9.2. The former shows the value of the incentives granted by CDI as percentage of the value of the approved investment, by ma4or industrial sectors, in 1986. Even when very broad - categories are distinguished, the dispersion of the rates of finan-ial incentives was quite high; and on average, these incentives were about 15 percent--not a negligible proportion (particularly since, as we shall - note soon, that is an underestimate). Table 9.2 shows rates of financial support through the BEFIEX, during 1980-85. Tariff concessions through * this channel were substantial, when granted--turning posted tariff rates as -I high as 60-70 percent to practically zero; these concessions were also -189- significant as a ratio of the value of exports (they would appear to be even substantially higher when calculated as a percentage of value added ir. exports). Another estimate, based on a sample of 200 projects (about two- thirds of the total), found an (unweighted) average BEFIEX support of around 60 percent of investment in fixed capital; and 35 percent of total investment. Table 9.2: RATES OF FISCAL INCENTIVES GRANTED BY BEFIEX, 1980-85 (percent) Tariffs Foregone Revenue Posted Rate with as Proportion of Year Rate Incentives BEFIEX Exports 1980 70.0 .4 23.5 1981 64.4 .5 34.3 1982 67.3 .9 23.9 1983 71.2 1.7 11.5 1984 64.2 1.6 9.9 1985 55.4 1.9 7.1 Scurce: R. Baumann, op. cit. 9.14 It is estimated that during the first half of the 1980s _Ae value of financial incentives granted through the four agencies--CDI, BEFIEX, CPA, and CONSIDER--was roughly 0.6 percent of the GDP; in 1985, the last year for which these estimates are available, they were 0.5 percent of GDP. This, however, is a considerable underestimate. First, it excludes the cost of several regional incentives schemes (as well as a national fiscal- incentives program--FISET--which is not discussed here); but, more importantly, it excludes the value of accelerated depreciation allowances, handled through CDI, which is thought to be qu'te substantial. 9.15 The change in the size of fiscal incentives following the introduction of the new industrial policy is not entirely clear. On the one hand, eligibility for incentives was widely broadened. The new regime covers a wide range of subsectors, and introduces new considerations, such as the diffusion of technology or an increase in competitiveness. On the other hand, the new regulations also stipulated an annual maximum size for these incentives. If, as may be expected, financing required by the criteria introduced exceeds the specified ceiling, it is not clear that the latter would in fact be binding; if it is, a mechanism for rationing is not specified. -190- C. Financial Incentives58 9.16 The BNDES system is the major vehicle through which finan*ial subsidization is administered. Its programs include medium- and long-term financing for capacity expansion in targeted industries and regions; subsidized credit for small-scale enterprises; and, in particula. during the 1980s, assistance to industries in financial distress. BNDES financing has, in fact, been the predominant instrument of financing investment in industry from sources external to the firm. Brazilian industrial firms have typically financed their investments either from BNDES or from retained earnings, with other sources of long-term borrowing almost non- existent and with equity capital raised in the stock market playing only a minor role. With accelerated inflation, and the absence of indexed instruments for private long-term lending, this structure is almost inevitable. However, before the inflation, it is likely that the existence of BNDES and its activities were themselves important in displacing private long-term lending. 9.17 The role of BNDES received a large boost in the early 1970s, when social security funds from the PIS and PASEP programs began to be channeled through it. After fast growth during the 1970s, iowever, the rate of increase of BNDES financing slowed considerably during the 1980s. During part of this period, this financing, in real terms, actually declined in absolute terms. Thus, while disbursements from the BNDES system financed as much as some 15 percent of Brazil's gross fixed capital formation in the mid 1970s, this ratio declined to an average of 9 percent during 1982-87. 9.18 BNDES lending has been far less expensive than credit available elsewhere in the economy. Interest rates on this lending in recent years have mostly been within the range of 9-12 percent (less than that in the 1970s) above monetary correction; and the latter has often been less than a full adjustment for the increase of price levels. The maturity of BNDES financing has been even more important--even before the accelerated inflation and the increased uncertainty associated with it have practically eliminated normal long-term lending, but particularly since such lending is available virtually only from BNDES. 9.19 Lending from BNDES is not limited to industry, but the latter has nevertheless typically accounted for close to half of the agency's total lending during the 1980s. Within industry, steel, chemicals and petrochemicals, and capital goods have been the primary beneficiaries. This may be seen from Table 9.3, where the allocation for 1986-87--the latest years for which data are available--is presented. 58/ See also the analysis of the financial sector in Chapter X. -191- Table 9.3: SECTORAL ALLOCATION OF LENDING BY THE BNDES SYSTEM, 1986-87 (percentage of total lending to industry) Approvals Disbursements Sector 1986 1987 1986 1987 Mining 3 4 6 5 Steel 26 12 33 13 Non-Ferrous Metals 4 5 6 8 Chemicals and Petrochemicals 11 15 9 13 Pulp and Paper 7 6 9 7 Capital Goods 17 25 12 19 Consumer Goods 12 16 11 15 Construction 12 11 9 13 Others 8 6 5 7 TOTAL Industry 100 100 100 100 Source: BNDES Annual ReDort, 1987. 9.20 In its past operational guidelines,59 BNDES has followed principles similar to those of CDI and of the other agencies administering the fiscal incentives: the policies in both areas have tended to reinforce each other in their impact on the economy. We shall now turn to this impact. D. The Impact of Iiidustrial Policies 9.21 Several salient distortionary impacts may be distinguished. Establishment of Barriers to Competition 9.22 The administration of both fiscal and financial incentives has led to barriers to both entry to, and exit from existing branches during the 1980s. This may be contrasted with the experience of earlier years. Until the late 1970s, the policies under consideration mostly led to the entry of firms into new activities, encouraging new ventures. During the last decade, on the other hand, with the original entry firms established and in place, the policies have been handled in a way which directs the incentives to these firms, barring in fact new entrants and prolonging the life of firms which should have disappeared. The following evidence supports this point. 9.23 Fiscal incentives administered by CDI tended to be concentrated in large firms. The size of the average firm enjoying CDI incentives has been eight times the size of the average firm in industry as a whole. Nor can most of this difference be assigned to the subsectoral speciqlization of 59/ With its reorganization in 1989, BNDES guidelines have been revised so that all its lending is based on criteria related to the financial soundness of the investment. -192- the incentives: in individual industries, the corresponding multiples ace between three and ten. Among incentives administered by BEFIEX, the bias in favor of large firms is not as pronouncel, but it is substantial nevertheless (a multiple of r:ive for industry as a whole, and two to seven in individual subsectors). This is probably evidence of support given to existing firms, rather than to new entrants which would normally be smaller firms. Note also that the large majority of CDI-approved projects during i)80-86 were designed for capacity expansion or rodernization of plants--obviously investments carried out by existing firms (the only exception, on this score, being found in the chemicals sector). In BEFIEX projects, the weight of incumbent firms among the recipients of support was even higher, Data of CONSIDER's operations indicates similar encouragement for entry at an early stage of development of a given activity, followed by an orientation towards the incumbents. As very little investment could be undertaken without the incentives (or, put in a different way, that the rate of support granted by the incentives has been very high), this bias toward incumbents must have served as a serious barrier to entry of new firms. 9.24 A similar inference can be drawn from financial incentives. Credit by the BNDES system is very highly concentrated among a small number of f'.rms. Thus, estimates for 1983, 1986, and 1987 show that over a half of BNDES credit was provided to 2 percent of its clients. The top ten firms were granted 60 percent of the total in 1983, and over one third in 1986 and 1987. Note also that BNDES clients overall constitute only a small fraction of industrial firms in Brazil. The evidence also suggests that credit from BNDES has tended to go to repeat borrowers. BNDES activity thus also seems to have been an important instrument in deterring entry of new firms into domestic industries. i Table 9.4: SHARE OF BNDES IN CREDIT TO INDUSTRY, 198-07 (percent of relevant total) Aggregate Credit Credit to Private- Credit to Public- Year to Industry Sector Industry Sector Industry 1980 21 14 63 1981 22 15 55 1982 26 18 55 1983 30 24 44 1984 30 10 63 1985 31 11 63 1986 25 9 63 1987 31 18 59 Source: Banco Central do Brasil, Relat6rio, various years. -193- 9.25 Moteovet, the increasing share of BNDES *redit ptuvided duLing the 1980s to suppoTt tirms in financial distress rnust have di;couLaged txit: it has prolonged the life of firw with losses, thus tending to irhibit growth rpther than to accelerate it. The 1989 reorganizatit,, of BNDES is also addr.essing this concern. Distortion of Factor Pr-portions 9.26 Almost all of the fiscal incentives have been designed to lower the cost of both domestic and imported capital goods to selected industries. The direct effect of these incen'ives has thus been to promote capital-intensive activities, and to discourage investment in activities and methods intensive in other factors (such as skilled labor). The result may be indicated by a few numbers. During 1980-86, the investment in tixed capital per job created in CDI-approved projects was US$230,000 in the chemicals, petrochemicals, and pharmaceuticals subsectors. In metal, metal products, and non-metallic intermediate products it was US$100,000. Together, these subsectors accounted for 90 percent of CDI-approved investment. These capital-output ratios obviously far exceed the ratios elsewhere in Brazil's economy. 9.27 Needless to say, BNDES subsidized long-term credit, predominantly allocated for investment in fixed rather than in working capital, must have had a similar impact. rhe particularly high capital/output ratios among public-sector enterprises noted earlier (in Chapter VII), must be at least partly related to the role of BNDES in financing public-sector investment. This distortion must have decreased, probably substantially, the size of output achieved with a given amount of investment and capital; it must, therefore, have played a part in constraining economic growth in Brazil. Distortion of Sectoral Allocation 9.2a by aesign, both the fiscal and the financial incentives have been heavily concentrated in a number of sectors, thus biasing resouice- allocation in the economy toward these activities. In general, intermediate-good industries have been the most hesa-:ily promoted, particularly chemicals and steel. The production cf capital goods has partly been encouraged by direct sector-specific incentives. But the main channel through which this activity is promoted has been indirect--through accelerated depreciation granted to the use of locally-produced capital goods. Indeed, the share of the latter in investment projects approved by CDI increased from 30-35 percent in the early 1970s to arounid 80 percent in the late 1980s. Rents 9.29 Most support in the 198Cs, via both fiscal and financial incentives, has been granted to large, incumbent firms. Where encry is thus constrained, it is likely that these incentives have--at least in part--often merely provided rents to the recipients rather than encouraging expanded production. This has placed a budgetary burden on the government, leading finally to the deterioration, rather than the improvement, of income distribution. It has also meart the attractior of resources to rent seeking rather than to productive act-v-ties. -194- Recomnmendat ions 9.30 In view of th? discussion above, it is clcar that the system of fiscal incentives has--at least during the 1980s--r-tarded competition, productivity, and growth. Nor do we recommend the substitution of this system by another fIllowing a different model. We simply recommend that these incen. ves--perhaps with very few exceptions--should be discontinued. 9.31 Given the absence of the normal provision of long-term debenture capital by the capital market, no similar recommendation may be made concerning BNDES arnd its financial incentives. But our set of recommendations concerning reforms of the capital market (see Chapter X) will remain relevant. 2. Aricultural Policies A. Pclicies of Direct Intervention 9.32 There are many varied government policies concerning the agricultural sector. Some are applied across the entire sector, whereas others are specific to subsectors or regions. We shall examine the following sets of policies: (a) tilC minimum-price support program; (b) government storage; (c) wheat subsidies; (d) sugar-marketing controls; (e) trade controls; (f) subsidized rural credit; (g) agriculture-specific taxes and related fiscal incentives; and (h) government expenditure policy in the agricultural sector. Minimum-Price Support 9.33 The Minimum-B?rice Program (MPP) is the gosernment's most comprehensive program of direct price intervention in agricultural product markets. It covers most major crops (coffee is managed by a minimum-price program administered by IBC-Instituto-Brasileiro do Cafe) as well as mary minor ones. The administering orgjnization, che Commodity Financing Corporation (CFP), buys each commb - W M _ D _ b n21'M OOes_s - - -D - - ,S - -g - - - - - -Df . > S 'Z - a 0 ~ ~ ~ 0 ?M. 10~~~~~~~ M M - V C S 8o Vo M CAi- i- A-Viai w3 00 --J Ci.-a- o-i o a I I O O -~2 2 81 D 0I _ I 0 - 0 0U I i _ 4 CA- D o8 w OiV O CC oA4 4- _Qi DrS n O _W AiVC i-.iV 8 o_ 8i-8oo -- -- - - - - - - - - - - - - - - - - - - - _ - _-C-a- -4- -2a - - - - C .ci D _CD O O W - O 0 O _ V MW O O ti i _ M W C iW C Vii- X s ~~~~~~~o o rIr9 O OU UO. _O a _O*o aa-2 I I , H~~CA O O V i O C- '~~~~~~~~~~~~~~~~~~~~~~~~ i't I - ,a ? . " e o >_ggr3.. ?or1i:^s^_˘- , , D, - 2 CAj CAViV'3 -J0a0Viio4_0W- * - OVi A i-50C C - - - a C 5-i Vi Oii '- a i oA !0 C r)_DG^ o W .: )S, 0 . 5-Via gVi-44 Vi iVi?8 CA.fS, _ V ~ ~ 2 ~ Table A.16: MONTHLY INDEX OF MANUFACTURING EMPLOYMENT IN G.REATER SAO PAULO, 1980-89 (1980=100) I I i I Period 1980 1981 1982 1983 1984 198S 1986 1987 1988 1969 1 January I 97.6 99.2 86.8 82.9 77.1 82.3 96.4 99.3 92.6 91.1 I February I 98.1 98.9 86.0 81.4 77.4 83.1 91.2 99.7 92.2 9W.2 1 I March I 99.4 97.4 88.8 8.56 77.6 83.6 91.7 99.3 92.2 89.8 a I I I Apri I 99.7 95.2 87.0 79.7 77.6 83.6 91.8 98.7 92.1 89.5 I May I 199. 93.1 87.5 79.3 76.1 83.8 92.5 97.8 92.1 96.8 l I I I June 1 196.7 91.7 87.7 79.9 78.4 84.1 93.6 9S.8 92.0 91.8 1 July I 101.1 90.4 87.7 78.5 78.8 85.1 96.4 93.6 92.2 92.8 1 August 191.9 88.8 87.9 78.4 79.4 85.6 96.8 92.9 92.6 ... I I Septmeber 1 191.9 87.4 87.4 77.9 60.3 86.9 98.5 93.7 92.2 ... I October 1 191.0 87.0 87.9 77.6 80.9 88.2 99.3 94.1 92.4 ... l Novemb, r 100.8 86.8 85.7 77.8 81.5 89.1 100.0 94.1 92.6 ... I December 1 99.4 86.8 94.5 78.9 81.4 88.8 98.7 92.9 91.8 ... I Source: rIESP, Industria e Desenvolvimento and Central Sank of Brazil, Golotim Mensal. s g- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - *~~ ~~~ . .0 .~_ _ _ . _._._._._._._._._._._._________.______ Rn * 11. . . . . . . . . .i 8. °O ogX 2*ee8g>@ o~~~ . . . . .J . . . . .b .n . . . . . o , ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~C sx -4 o n kY !W9 EH8OjiO °gXs}gieMwE o1 n ~ ~~~~~~~~~~~~~ ;s e ;3 W - 15 w we WJ|§9 (2 goXe8t" * cv °~~~ . . . . .Ox"~ .4 ^r°~ . .9rr . . . ' r * O2 |e°| i t g !°nPO!P co . . . . . . . ;98 3e|DVgu W>9 g _ tn LI C " m D, Ds@ W@94rrrrY> ~~~~. . . . . . . . . . . . . . . . . . . . . . . . . . 80 80 . 00U ti2e>°8fiO8eezg 3 rA Ln a, s ~ Z fs o, OC, s| ,%%* ite |eoSw|rr| °8 w~~~~~~~~~~~ . . .g . . . . . . . . 08o 8 Y w§ub & a9Qg f9!a ub~~~~~~~ 4 ti ?P (A l |e 5 Ye>9oe>o CP~~~~~~~~~ ~ ~ ~ . . . . . tr . 10 tP b o*seq~t n h 4Ob CD e* g cars r< st9 r4 r twpt !OH ee 0 4Z -tWq - D !tD Ng- D U W t2PO 4 efpD § 8 0 00 0 i O CD 4 tD 4 4 S 'b O S ^@ h: 2 s @ r tt | ED e 8 4s tP CD N r s s * 4 s z^ b w wo 4 C D. C tA ~ ~ ~ ~ ~ ~ ~~: wq e teeD_e-n ztP_W O c 1t n ;4 o w _ _ g _ _ _ _ _~~~w _ * o 'n _ O O !: tD vP b 0 0 D s P b 9 _ _ - h - tP - t - - tD - - - - - - - IP - tD - - - - tD - CD - - - CD - - - - - - - - - - - - - - - 8 &4 4w e e ° t °S Cllg o28 4gCP{ he| 3°° W Table A.18: FINANCIAI ASSETS DEFLATED BY CONSUM4ER PRICES, 1985-89 I ~ ~ ~ iI I MlarI I j . _ _ _ _ _ --- -- --- .. . .. . _. _.. __... dor tt{ Monetary Deoand T, s M2 Savings M3 P,bl M4 F1N"M AL DATE I Base Currency DePosi to TOTAL Depops to Depos ts Debt ASSI. r6 , I. - _ __ _ __ _ I____ _ _.___ _ _____ _ _.___ ____ _ -____ _ _.___ _ _____ _ _____ _ _____ - -____ _ _..-- - - _ _._ _ . ..-.- . ._.- - - . _ I a85-Dec I 607 314 1,181 1,494 ,991 3.S485 2,904 6,390 3,450A 9o19 10,'26 l W86-Dec I 1.505 707 3,125 3,8-33 2,453 6,286 2,776 0.062 3,023 12. 084 12,904 tc87-Dec 854 421 1,336 1,757 1,630 3,338 3.750 7.138 3,889 11,02 11.82? l : 188-jan 720 318 1,006 1,324 1,572 2.806 3,683 6,579 3,955 10,$3 .,31 1 l Feb I 578 269 993 1,263 1,575 2.837 3,901 6.738 3,833 IC',s2 1134? 1 'Ian I 704 300 920 1,220 1,590 2,810 4,085 6,895 3.624 10,518 11-2186 Apr l 639 275 892 1,166 1,518 2.685 4,035 6,719 3,936 10,65S 12,426 Ma,y 624 241 935 1,177 1 546 2,723 4,053 6,776 4 089 10,868 2 6;`0 Jun 1 555 239 884 1,123 1,!>49 2,672 3,955 6.62! 4. 627 11,254 I ,99 1 l Jul 1 497 231' 759 998 1,485 2,483 3,943 6,426 4. 58 12 12 I ; l2 Au 444 206 711 917 1,542 2,459 4,082 6,541 4 99S 11,536 12.74% Sep 461 231 704 935 1,496 2,432 3,966 6,398 4,692 1104 12 7?. I Oct 472 237 710 947 1,473 2,420 3,817 6,236 4,514 10,7S; 1,.4: :' No- 437 208 609 907 1,411 2.318 3,858 6,175 4.528 10,703 1 3 1 Dec l 564 316 762 1,079 1,467 2,546 3,872 6,418 4,884 11,306 1,.954 1189-Jan I 471 235 697 932 1,069 2,000 3,730 5,730 4.351 10 08(4 le841e 1 I Feb 484 22 620 842 825 1,667 3,940 5,607 45663 10,:- 1O1 02 ! Mar l 536 251 655 906 828 1,734 4,464 6,198 5,423 11,621 11.475 , Apr , 563 285 867 ,153 838 1,991 4,640 68,30 5,603 12.433 1Ui .306 I May 1 644 273 763 1,036 999 2,035 4,383 6,418 4.977 11,395 12,226 Jun 577 278 632 910 1,130 2,040 3,734 5,774 4,803 10,576 10 607 Jul 492 229 561 790 985 1,775 3,621 5.396 5,01I 10 407 10.428 Aug I 437 214 480 694 1,027 1, 7i 3,463 5,183 5,205 10,389 10 408 i I Sep l 424 235 455 690 985 1,675 3,185 4,860 5,163 10,023 1lr0431 l Oct l 396 202 479 682 1,034 1,716 3,059 4,775 5,287 10,062 10 0O82 Nov l 431 215 438 653 1,060 1,713 2,687 4.401 5,171 9,572 0,092 l 1988 1 1 667 296 973 1,269 1,579 2,848 3,890 6,738 3,804 10.542 11 '24 11 t 8606 252 904 1,155 1,538 2,693 4,014 6.707 4,217 10,925 2! 6.2 I I:l 467 225 725 950 1,508 2,458 3,997 6,455 4,756 11,211 11 s:2 l IV t 491 254 724 977 1,450 2.428 3,849 6,277 4,643 1092, 11 '4 i 1989 I 1 497 236 657 893 907 1.801 4,044 5,845 4.813 10,65R 22 44S if I 595 279 754 1,033 989 2,022 4,319 6,340 5,127 11,468l 2i 4h 451 226 498 724 999 1,723 3,423 5,146 5.114 10,.61 2.93 _ _ _ _ ________ _ ____ _ _ _ . _ _ _. . . .._ . _ _ _ a/ Balances *are deflated by the rastr-cted consuer price inde (INPC) b5 Total f-an-ial assets -clude MA plus biles of exchange, housing bills and state and m-uncpal debt Fxcludes state and unicipal debt and housing b-lle after may 1989 Qmuarterly fguras r0pestert per -d axerages 9.,' Central Park <>' 8-ni 1, :4 ,adr., I es lZ a rm d- Polit u- M-etar, .arious ,sxues _iU~~~~~~ I Table A. 19: LOANS TO THE PRIVATE SECTOR, BY FINALIEIE, 958 (millions of cruzados) MO0NEfTAkR Y S Y ST 68M I NO0N M JN IT A F Y Sfn San-.ng Banks k),.pn'tEa"sC' >4 FC I Bank of Coese-nl In-----nt. . ..4 corr.ctCon I ~~~Braz,l Bank. TOTAL Bars 6nmner C1 015 U SCI,APE State F*CCS EM~*C 11,cc 67 7, I 63,456 1508 219,524 I 49,636 60,878 104.679 24,763 82,477 :7.100 14790 -':..!24' Fc 1586 0evc 122 29 I 104,052 435,036 619,090 111,795 14.057 221,699 44,799 128,289 32,371 23q562 :73 636~8 :35 I19B7 eC. 569 82 1 807,009 1.263,295 2.070.304 J 347,918 191,503 1,047,000 214, 13.3 683,142 180,101 1:5A3 1' .40 .6 'a' 31 kia'. 34' 5u.;, 663 911 I 907.451 1,439,797 2,427,248 I 414.773 204,770 1,398,364 250,631 941,137 2051.5 :'a,373 3'.3.5 Yn ( ~ek 783 14 I 1.093,009 1,695,337 2,7880346 I 497,082 22, 437 1,473,298 295,425 963,050 250,574 154,814 F'2212 3. R4.9CC, '2724 908 52 I 1, 320. 682 1, 987,755 3,308,437 588,956 251,191 1.566.457 347.262 1,010,5,75 293.650 175.182 I C 2'4 .41,'9C i Apr 1083 68 I 1, 563,261 2,324,400 3,909,661 1 691.947 '283,455 2,273,660 414,395 1,528,617. 356,,698 207,323 :2 :7667e 393 9 8', 054 4a 276 36 I 1,954,078 2,800,357 4,754,435 I 806.60? 337,473 2,493,381 490,145 1,503,837 433,915 256 5~60 I". (.8 4183,36 11 "'C.?i7l 3.l 1525 63 I 2,342,048 3,395,305 5,737,353 96Q0,000 393,0(10 2,795,943 586,000 2,519,000 5667. DO 402.100 A,'c 232 c43~ 13 QC q44 6 lu 1892 35 1 2.789,075 4,088,000 6,877.075 J 1.114,000 502,000 3,488.000 724.000 3,190,000 686,000 46'.000 ~' ' .2:4 37' C' Is: 4 5, Aug 2283 31 1 3,544,384 4,946,480 8,490,864 1,305,000 599.000 4, 302.000 896,000 4,010.000 839,000 80' 000 34 000 1942,00 . : 382,64 nep 283 53 I 4.28,3,000 5,987,000 10,270,000 1 1,543,000 810,000 5,394,000 1,128.000 5.005,000 1.041.000 741,000 42,.000 :6"7:4.000 2'. Q44. OO I nt 3603 12 1 5,539,000 7,665,000 13,194,000 I 1,907,000 902,000 6,902,000 1,450,000 6,400,00-0 1,320,000 926,000 6' 3301.O 19. , 74. 000 3 3066 OD0, Non 4573 18 I 6,918,000 9,593,000 16,511,000 I 2. 303.000 1, 098, 000 8,948,000 1,825.000 8.0MS5,000 1,619,0(10 1, 273. 300 67.0CC' 25.19 QC, , 41 77000 IO 21.c 589 80 I 8,716,000 11, 131,000 19.847.000 I 2,615,000 1,480,000 211,520,000 2,206.000 7,500,000 2,148,000 1 616,000 8', 000 2.4 I7 ((C 4s) '~: 306 -C 1'.Jn 10029 16 1 11,0518,000 12.768,000 23,826,000 1 2.13-2,000 1,597,000 143300 27000 93000 2,503,000 2 106 000 99 3047 ISA 860 000 58. 6600 CYak 10390 20 I 14,741,000 12,171,000 26,912,000 1 2,160,000 1.541,000 17,082.000 3,129,000 10,600,000 2,513,000 2, 002.000I 104.C0 39,221.000 66IE133.000 a. 11022 96 I 13,331,000 13.271.000 26,6012,000 2,934,000 1,619,000 20, 53,U000 3,779,000 13,000,000 2,096,000 2,140,000 (F.000 46 :98,300 . 72,800 OX APr~ 11828 74 I 15,475,000 12,472,000 27,947,0001 2,326,000 1,512,000 23,239,000 4,289,000 12,950,000 2,0,47,000 2,296.000 -134 000 46 76,3 000 . 76 1IC,0cC -ay 13004 50 I 16,868,000 14,322,000 31,190,000 2,650,000 1,852,000 25,733,000 4,778,000 13,910,000 2,460,000 2, 513.00 1:i 3.000o 54 007.000 8"..19' I03*1 16233 50 1 16,856,000 15,725,000 32, 581,000 3,037,000 2,168,000 32,300,000 5.994.000 16,900,000 2,989,000 2,8434 000 1323 -00 f66, :3.0001 98 744 08.. JC, 20902 3AI1 22,380,000 18,05',000 40,417,000 1 3,850,000 2,395,000 41,500,000 8,052,000 20,870,000 4,313,000 4.130K 000 :6 000 869- 253 00C1 ::t X Aug 27035 05 !28,728,000 21,793,000 610,521,000 1 5,307,000 3,117,000 55,500,000 10,718,000 27,115,000 6:71,000 5.340 00C 243~ 136'I 1:11 i: 06I:wX :6: C(32 OX 36754 15 1 36.913.000 '26,097,000 63,010,300 I 6,848,000 4,347,000 72,600,000 14,684.000 32,800.000 8,430,000 7,128,000 3:; nOt. 14'. ,s4* k 001 h I 1- OCX 'It 5058 1 06 i 47,45,4,000 34.900,000 82,3.4, 000 1 8,771.000 5,700,000 100,400,000 20,607.000 45.000,000 11,910,000 9 8-33.006( 449 (COO -8< 0(CC, 2c1 X5 y4 7rn 15.31 74 1 59.932,000 4 6.434000CY :08,366,000 I :2,900 000 8,266,000 139,560,000 29-:54,000 63,0001,000 17,980,000 13,9'2Q (CCC- F KX0 2-1 ,C4 `C. :' '- I.7 Oar 19836 98 I 86, 000,00 8( 4 ROCOCO0 149, R00.~000 I 19,000,000 12,000,002000004,0,009,0,0 750000 2: 05 301:9 X.3 4:4 ~10 .3 4 7K'6 I 1988 ' '8 1! I 1,3i, 714 I.,707,630 2 841,344 t ScO, 2C) 226,133' 1,479,3 73 29 7. 779 9'18, 254 249,926 :5 '7 42 - , ,.r II 1295 2-2 : 19U0,462 2 84),0.02: 4,800,483I 816.16!. 3.37,976 2.6120,995 498,847 1,877 216 452.204 288S 428 24.4 s .4 111 23.35 73 i 3,538,820 5,007,160 P.,545,ss QA :320, C6' 637,000 4,394,687 916 000 4,068.333 85$ 300 6,,: ')C~ .14: :. CC . 1c IV 4688 70 1 7.057,667 9,45,9,66' 16, 717,3331 2,275,0(1 1,160,000 9.,122,667 1,827 000 7,321,667 1 69'1.86 7 I1 77 16 . s I 7 4,' . .4C, 19B' 1 10490 7' 7 :3, 04 3,333 12,736,667 25 780,000 2, 408,88f7 :.585.667 1' 329,667 3, 21 9.333 10,966.8667 2,370,667 7,:.::2 6F *5 C, . ,'-CC rl 136888 91 16,399,667 14.173,0610 10,5,72,667 2.871,000 1,844,000 2',090,867 5,019.687 14.586,667 2,498,667 2484 -ICC' 1..'1 3 PC',.6A. c ~, III 28,230850 29.,333,667 2I,982,33 51,I316,000 6,35. t,000 3 286.333 55,8~ 66,867 11,151 33.3 26.928,3.33 8.307.3.33 5' 66' 248 C, .4"'.5 1 3 IV 77316859 I 64,129,667 48. 711,333 112.840,000 13,557,000 8,621,667 146,6563.333 3:-135366 667 F,000, 000 19, 130.000 14 946,.0CC 7.1 3 3.'1 f' / 0 N.tC. Quarterly figures represent pero.d aberages aCentral Bank. 1nfornat- M-nal Snbre o Me.radc, 6nceonaus CsuOf 0 .4 01 0~~~~~~~~~~~C 1 0 u0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ C 00 0 CD10 (a o- 0 'S ZC- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 00 0 - to . --0O Z 0 09 0 01 0 0 A00 0 a WOW 1C CD m * 10 41 01 01 01~~~~~~~~~~~~~~~~O 0 o~~~~~~~~ co0o HC M~ 01 -0 C0I I'DI -W OD ~~~~~~~ ~~O - - - - - - - - - - - - - - - - - - - - - --0 - - - 4 - 4 - -9sz-~~~~~~~~~~ Table A.21: NET DEBT OF THE PUBLIC SECTOR, 1984-88 1 9 84 I 1 98a5 198a6 I 1 98 7 1CATECORYI al Iance I cls,nge ShareIB I Balace I change Share Balance I change Shar'.IBaaa I ic. change SKara CRAMC TOTAL (A.B = I.N11.11) 334120 253 6 100 0 I d42335 271 8 100 0 2064175 67 B 100 0 110964000 426 100 0 A - 0IMESTIC DEST 134207 297.0 40.2 512433 281 8 41 2 I868114 69 4 41 7 I 4353000 401 4 39 7I -Federal Co-t, and Control Bank 48423 477 4 14.5 I 77106 265 7 14 3 283522 60 1 13 6 1 760000 168 1 6 9 I I state *nd Local GCo,nernnta 31190 263 6 9 3 117969 278 2 9 5 201325 70 7 97f 1183000 487 6 108 I I -Statea Entaep,Pxee/A.tsrch.aa 54594 224 2 18.3 I217358 298 1 17 5 I383267 76.3 28 4 I 2410000 528 8 220 0 8 FOREIM OEErtI 199913 229 4 59 8 729902 265 1 5S8822116061 66 6 583 16611000 443 6 603 1 -Federal Co.t and Central Bank 81700 206.3 24.5 269801 230 2 21 7 559207 107 3 268 I 3591000 542 2 328 -Stat. and Local Co,erneenta 11053 274.6 3 3 I49371 346 7 4.0 I75930 53 8 368 f 353000 364 9 3 2 I I -Stat. Ent*rprieoo/A.t&rchkea 107160 246.0 32 1 410730 283 3 33 1 %M8024 41.4 2?79 I 2667000 359 1 24 3 I Fedoral Gavt and Central Bank 130123 271.1 38 9 446907 243 4 36.0I 842729 88 6 40 4 I 4351000 416 3 39 7 I - Domestic Debt I48423 477.4 14.5 I177106 265 7 14.3 I283522 60.1 13.6 I 760000 165 1 6 9 I - Foreign Debt 81700 206.3 24 S 269801 230.2 21 7 559207 107 3 26 8 I 3591000 542 2 32 8 I I El1. State and Local Giaorrnwonto I 42243 266.4 12 6 I 167340 296 1 13.5 1 27725,5 65.7 13.3 I 1536000 454 0 14 0 1 I I - Domest;c Debt I 31190 263 6 9.3 I 117969 278 2 9 5 1 201325 70.7 9.7 I 1183000 48? 6 10a8 I I - Foreign Debt I 11053 274.6 3 3 49371 346 7 4.0 g 75930 53.8 3 6 I 353000 364 9 3 2 1 itII State E.ntargrisee/Autarch.se I 161754 237 7 48 4I 62806 288.3 50.6 964191 53.5 46 3 I 5077000 426 6 46 3 1 1 -Dooeot.c Debt I 54594 224.2 16 3I 217338 298 1 17 5 I38367 76 3 18.4 I 2410000 528 8 22 0 1 -Fora;gn Debt I 107160 245.0 32 1 410730 283.3 33.1 SO5094 41.4 27 9 I 2667000 359 1 24 3 I IAccimulated Floe (Total Debt) I 239827 908215I 841840 I 8940000I IAccumulatad Flow/CaP(I) 1 61 4I 64.0 22.0 I 75 2 I Debt Average Ba lance I 202432 730112 I 1662449 I 5961380 IDebt A.eraga Balance/CDP(%) I 51.8 Si5 5 1 43.4 I so 2 ICross Domeatic Product I 390573 I 1418068 1 3826268 111885000 IEachange Rate (Purchase) 1 3 268 I 10.440 I 14.865 71.892 p a preliminary date source: Central Bank of Brazil, Economic Program. various mac - 260- 1 i) Ie A. 22: oF I I(.]Al ANDI) PARALLEL-'ARE.T EXCHANGE RATES, 1983-89 _I- b/ I llil P A R A L LLEL I I O F F I C I A L R A T E I P A R A L LEL I M A R K E T I l e/ lI M A R K E T I P R E MI I U I I PERIOD I End of Per;od Period Averagoe I (%) I 1983 1 0.60 0 60 I 0.9 I 60.0 1 1984 I 2.00 1.80 I 2.2 1 10.0 1 1986 I 6.50 6.20 I 8.3 1 27.7 1 1988 1 13.83 13.e6 1 22.6 1 83.6 I 1987 Jan 16.53 16.70 1 26.1 1 67.9 1 l I Feb l 19.79 18.14 | 32.0 | 81.7 I Mfar I 22.14 21.01 1 30.0 1 36.5 I pr I 26.43 23.71 1 33.6 1 31.7 1 M ay 1 33.99 30.78 1 36.8 1 8.3 I l Jun l 43.38 39.44 1 63.8 1 24.0 I l Jul l 46.02 44.93 1 57.8 i 25.6 l Aug l 48.36 47.13 i 69.0 1 22.0 l Sep l 61.28 49.87 I 65.3 I 27.3 Oct l 56.90 63.41 1 e8.6 I 22.5 l Nov 83.07 69.29 1 77.3 I 22.8 I l Doc l 72.25 67.48 1 93.6 I 29.4 1988 Jsn i 83.40 77.66 1 98.0 I 17.5 I Feb 1 98.50 90.84 1 126.0 26.9s Mar l 116.64 107.14 1 160.0 I 29.8 Apr 1 137.44 126.89 1 185.0 I 34.6 l May l 162.69 160.64 l 226.0 I 38.3 l Jun l 194.83 178.30 i 272.0 I 39.8 I Jul I 241.73 216.83 1 356.0 47.3 * l Aug l 292.49 267.41 1 480.0 I 64.1 Sep 1 382.98 326.24 1 630.0 I 46.0 I l Oct . 463.34 411.70 770.0 1 66.2 1 l Nov i 688.07 626.16 I 960.0 61.5 l Dec l 76e.66 666.56 I 1216.0 I 60.8 I I ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I 1989 Jan l 1000.0 906.7 I 1680.0 I 68.0 l Feb 1000.0 1000.0 l 1700.0 70.0 ; leMar l 1000.0 1000.0 1 1880.0 I 88.0 Apr I 1032.0 1016.2 1 2400.0 i 132.8 May i 1153.0 1098.9 I 3120.0 I 170.8 I l Jun l 1619.0 1337.3 3 3300.0 | 117.2 i l Jul I 2166.0 1914.2 I 3760.0 I 73.1 Aug I 2802.0 2474.9 I 4700.0 e 87.7 l Sep 3797.0 3268.9 I 7300.0 I 92.3 - l Oct l 6225.0 4489.7 I 11700.0 1 123.9 l Nov l 7367.8 8248.0 I 13760.0 I 86.6 Dec I 11357.4 9268.9 I 26500.0 I 133.3 I I _ _ _ _ _I I__ _ _ _ _ _ _ _ _ _ _ _ _ I__ _ _ _ a/ Annual figures. represent averages of monthly levels. b/ Parallel market premium relative to end-of-period official exchange rate. Source: Gazete Mercantil, Jornal do Brasil and Sum* Economica. . -0l'le A. 23: tqLE('TFD) PN7ES Or' RE'rt <, 1985-89 Intorfeet Ratee CZ$/US8 e STOCK MARKET INDEXES I Sevinga Ovorn;ght CDB Paralrlel I IBV 18V Boveapa Bovespa I INPC I Accounto I Market % change ! % change S change lInflat;onI ! 1986 Jan 13.16 12.66 13.97 I 3.9 I 493 1323 l Feb 10.76 12.16 11.64 4.8 17.96 I 466 -6.A8 1227 -7.28 10.92 Mar 13.26 11.94 14.07 I 6.1 10.87 I 613 10.09 1293 6.38 I 9.96 1 Apr 12.39 12.60 13.19 6.6 9.80 I 489 -4.68 1206 -6.73 I 8.69 1 May 10.6e 11.26 11.20 I 6.5 16.07 I 626 27.81 1676 3897 I 7.18 l Jun 9.75 10.22 10.53 7 T.3 12,31 I 836 33.60 2413 43.97 I 8I35 1 Jul 8.14 9.27 8.77 I 8.4 15.07 I 996 19.28 2905 20.39 | 10 07 Aug 8.72 8.74 9.35 I 9.6 13.10 1462 48.79 4046 39.28 I 11.71 Sep 9.65 9.98 10.07 I 10.1 6.32 1 1634 11.78 4521 11.74 I 9.99 Oct 9.66 9.83 9.94 I 10.7 6.94 I 2467 60.98 6561 45.12 I 10.25 Nov 11.67 9.97 12.02 I 12.6 17.76 I 2924 18.62 7286 11.03 14.18 Dec 13.93 12.70 14.34 1 15.3 21.43 I 2713 -7.22 6568 -9.84 15.76 1986 Jan 16.81 14.97 17.28 1 15.3 0.00 I 3226 18.87 7032 7.06 I 15.01 l Feb 14.93 16.15 16.43 I 18.2 18.96 j 3198 -0.84 7849 11.62 1 le 16 l 'ar 0.39 1.21 0.89 I 17.6 -3.86 I 3837 19.98 12013 53.06 I -0.11 Apr 1.28 1.24 0.94 I 19.3 10.29 I 5386 40.37 18137 50.98 I 0.41 May 1.90 1.22 1.09 I 21.0 8.81 I 5312 -1.37 17208 -6.13 1 1.10 I Jun 1.77 1.27 1.17 I 21.0 0.00 I 6598 6.38 1e321 -6.14 | 0.97 I Jul 1.89 1.95 1.77 I 22.8 8.57 I 5A74 -2.22 16821 -3.08 | 0.84 Aug 2.19 2.67 2. , ) 23.6 3.07 I 4908 -10.34 14062 -11.12 I 1.12 Sep 2.23 2.94 2.57 I 24.6 4.26 8 4109 -16.28 11639 -17.94 I 1.19 Oct 2.41 1.96 2.82 27.6 12.24 I 3631 -11.63 10271 -10.99 j 1.43 Nov 3.81 2.37 4.6 6 26.0 -5.46 j 3786 4.27 11236 9.40 i 3.29 l Dec 7.81 5.47 7.97 27.6 5.77 2936 -22.46 8922 -20.69 1 7.27 I 1987 Jon 17.40 11.00 13.36 2e.6 -3.64 I 2680 -12.13 7729 -13.37 16.82 Feb 20.21 19.81 20.98 I 31.0 18.98 I 2070 -19.77 e838 -11.53 I 13.94 Mar 16.09 11.96 13.47 I 31.6 1.61 2026 -2.17 6634 -2.98 I 14.40 Apr 21.58 15.30 20.68 I 31.6 0.00 1 2621 24.49 8677 29.29 I 20.96 May 24.06 24.63 26.76 37.0 17.46 I 2447 -2.94 7673 -10.64 I 23.21 Jun 18.61 18.02 18.98 64.0 46.96 I 3099 26.64 11126 46.00 I 21.44 I Jul 8.91 8.91 9.61 j 68.6 8.33 I 4269 37.76 13701 23.14 I 10.05 Aug 8.09 8.09 8.64 I 69.6 1.71 4117 -3.68 11477 -16.23 6.09 Sep 7.99 7.99 8.68 j 64.3 8.07 I 4624 9.89 14808 29.02 I 7.15 Oct 9.73 9.46 10.22 I 67.6 4.98 I 4460 -1.64 11954 -19.27 J 10.88 Nov 13.40 12.92 14.17 77.3 14.62 i 4479 0.65 12398 3.71 j 14.93 Dec 14.71 14.38 14.84 I 94.0 21.8C 6060 12.97 12486 0.71 13.97 1988 Jan 17.09 16.78 17.01 I 98.0 4.26 i 6704 32.49 19090 52.89 19.08 Feb 18.56 18.36 17.12 1 126.0 27.66 I 8640 27.39 21600 13.16 j 16.81 Mar 16.69 16.69 16.94 I 160.0 20.00 I 11284 32.13 39650 83.6e 18.12 1 Apr 19.88 20.26 20.06 I 186.0 23.33 I 16867 49.39 62040 31.26 18.30 May 18.37 18.65 18.37 I 226.0 21.62 I 216b9 27.89 81460 18.10 18.24 Jun 20.13 20.17 20.48 I 272.0 20.89 I 28469 22.73 70370 14.50 J 22.28 Jul 24.66 24.69 24.69 368.0 30.88 I 28921 1.76 72930 3.64 23.02 I Aug 21.26 22.63 21.98 480.0 34.83 I 32609 21.13 94200 29.16 20.83 I Sep 24.63 26.26 24.82 630.0 10.42 I 46239 38.73 136920 46.36 I 26.93 Oct 27.89 29.79 28.22 I 770.0 46.28 I 60531 33.80 180610 31.84 26.69 Nov 27.66 28.41 27.91 I 960.0 23.38 I 79100 30.68 213910 18.60 I 28.16 Dec 29.43 30.24 29.79 1 1216.0 27.89 1 121439 63.63 330820 64.6 i 28.4a 1989 Jan 22.97 22.97 7.78 I 1680.0 30.04 1 124765 2.74 324000 -2.06 36.48 Feb 18.96 18.96 19.91 1 1700.0 7.69 I 184881 48.18 497920 63.88 16.35 Mar 20.42 20.42 14.96 I 1880.0 10.69 267461 39.26 729290 46.47 6.90 I Apr 11.62 11.62 13.18 i 2400.0 27.86 I 404807 57.24 18ee0e0 69.88 8.06 May 10.49 11.43 18.16 3100.0 29.17 1 407321 0.62 1263000 7.46 16.67 Jun 26.45 26.78 29.44 I 3300.0 8.46 262672 -36.S4 727000 -41.98 I 29.4e Jul 29.40 31.60 36.37 I 3760.0 13.64 I 438266 68.91 1232300 89.50 J 27.40 1 Aug 29.99 33.31 36.88 I 4700.0 26.33 I 634610 21.98 1628600 24.04 I 33.18 I I Sep 38.83 37.44 40.88 I 7300.0 66.32 7 782130 42.68 2254400 47.48 38.3 Oct 38.31 44.10 46.87 I 10060.0 37.87 1 1142000 37.2C 3413400 61.41 I 38.78 Source: Interest rates - Central Bank of Brazil; Stock Market indices - Conjuntura Economica; Exchange rote - Gazeta Mercantil; Price index - Fundceeo Getulio Vargas. - 2 6 2- Table A.24: REA, RATES OF RERTVPU, 1985-891/ Savings Ovornight CDO Parallel Gold IHV Boveape Accounto Rate Market I 1986 Jon Feb -0.18 1.11 0.6e 6.33 .. -14.78 -16.39 Mar 3.00 1.80 3.74 0.83 ... 0.12 -4.18 1 I Apr 3.60 3.80 4.24 1.12 ... -12.22 -14.11 I May 3.18 3.81 3.75 8.30 ... 19.26 29.88 1 Jun 1.29 1.73 2.01 3.86 ... 23.30 32.88 Jul -1.76 -0.73 -1.18 4.64 ... 8.37 9.37 i Aug -2.e8 -2.86 -2.11 1.24 ... 31.40 24.68 1 Sep -0.31 -0.01 0.07 -3.34 ... 1.61 1.69 1 Oct -0.63 -0.38 -0.28 -3.91 ... 38.94 31.83 Nov -2.20 -3.69 -1.89 3.13 ... 3.80 -2.78 Dec -1.68 -2.64 -1.22 4.90 ... -19.85 -22.12 1986 Jon 1.6e -0.04 1.97 -13.05 1.62 3.36 -8.91 Feb -1.06 -0.87 -0.63 2.40 -1.98 -14.64 -3.91 1 nMar 0.60 1.32 1.00 -3.74 -8.73 20.11 53.22 | Apr 0.87 0.83 0.63 9.84 13.68 39.80 60.37 SMay 0.79 0.12 -0.01 7.62 3.00 -2.46 -6.17 Jun 0.80 0.30 0.20 -0.98 1.28 4.38 -8.05 Jul 0.84 1.10 0.92 7.8' 18.68 -3.03 -3.87 I I Aug 1.08 1.43 0.98 1.93 0.32 -11.33 -12.10 1 I Sep 1.03 1.73 1.37 3.03 14.77 -17.28 -18.90 I Oct 0.97 0.62 1.37 10.66 8.64 -12.88 -12.24 I Nov 0.60 -0.89 1.23 -8.47 -6.08 0.96 5.91 I Doc 0.61 -1.67 0.88 -1.43 -18.10 -27.70 -26.97 I 1987 Jan 0.60 -4.98 -2.96 -17.61 -9.78 -24.78 -26.85 I Feb 6.50 4.98 6.18 2.67 11.03 -29.58 -22.36 Mar 0.60 -2.14 -0.81 -11.18 -16.75 -14.49 -16.20 Apr 0.60 -4.68 -0.23 -17.33 -2.86 2.92 6.88s I May 0.69 1.16 2.07 -4.87 -3.29 -21.22 -27.39 Jun -2.33 -2.82 -2.03 20.18 11.02 4.29 19.40 Jul -1.03 -1.03 -0.49 -1.6e 0.60 26.18 11.90 Aug 2.86 2.86 3.28 -3.22 -3.13 -8.23 -20.29 I Sep 0.78 0.78 1.34 0.88 2.68 2.66 20.41 I Oct -1.04 -1.29 -0.69 -6.32 -0.16 -11.29 -27.19 I Nov -1.33 -1.76 -0.e8 -0.38 -1.65 -12.42 -9.76 I Dec 0.65 0.36 0.76 6.70 -0.26 -0.88 -11.83 I 1988 Jan -1.87 -1.93 -1.74 -12.45 -6.24 11.26 28.39 I Feb 2.37 2.19 1.13 10.14 -3.77 10.00 -2.30 i Mar -1.30 -1.30 -1.00 1.69 8.21 11.88 65.41 I Apr 1.34 1.86 1.48 4.26 1.46 26.28 10.96 I May 0.11 0.36 0.11 2.88 6.13 8.18 -0.12 Jun -1.76 -1.73 -1.47 -1.14 -6.94 0.37 -6.38 Jul 1.33 1. a 1.38 6.39 8.61 -17.29 -15.78 I Aug 0.52 1.86 1.12 11.77 6.09 0.41 7.08 Sep -1.81 -0.64 -1.66 -13.01 -18.06 9.30 14.51 l Oct 0.96 2.45 1.21 14.88 18.34 5.61 4.08 Nov -0.47 0.20 -0.19 -3.72 7.80 1.97 -7.63 l Dec 0.78 1.41 1.08 -0.42 -10.22 19.64 20.42 1989 Jan -9.23 -9.23 -20.46 -4.01 -10.90 -24.17 -27.71 l Feb 2.24 2.24 3.08 -7.62 -7.83 27.36 32.08 ,Mar 13.71 13.71 8.66 4.43 6.6e 31.49 38.31 I Apr 3.20 3.20 4.74 18.14 18.91 46.51 47.98 l May -5.30 -4.49 -0.44 10.71 6.40 -13.78 -7.89 l Jun -3.06 -2.80 0.03 -17.73 -16.92 -60.18 -66.16 l Jul 1.67 3.22 8.26 -10.80 -8.70 31.01 33.06 I Aug -2.40 0.10 2.03 -6.89 -12.84 -8.41 -6.8 I Sep 0.21 0.80 3.32 13.91 22.92 4.66 8.16 l Oct -0.32 3.86 4.98 -0.78 -6.91 -1.09 9.12 II 1/ Real rates relative to inflation as measured by tne INPC (e.g. [(1.nominal rate) / (1+INPC) - 1] * 100). Source: Interest rates - Central Bank of Brazil; Stock Market indices - Conjunturs Economics; Exchange rate - Cazeta Morcantil; Price index - Fundacao Cetulio Vargas; Cold - Macrometrice. -263- TABLE A.25: CONSOLIDATED ACCOUNTS OF THE GENERAL GOVERNMENT, 1981-86a/ (Cz$ millions) 1981 1982 1983 1984 1986 1988 1 I TOTAL RECEIPTS 7,662 16,170 38,091 114,749 460,662 1,116,404 1 I Tax revenues 6,041 12,772 29,894 84,247 311,471 911,166 1 I Direct taxes 2,969 6,416 14,376 43,990 186,304 482,880 t I Indirect taxes 3,073 6,366 16,024 40,267 140,167 428,496 1 I Other revenuee b/ 1,811 3,398 8,697 30,608 149,081 206,248 1 TOTAL EXPENDITURES c/ 8,428 17,926 42,448 136,7e6 660,641 1,746,724 1 I I Consumption expenditures 2,286 6,068 11,328 31,987 136,446 400,798 1 I Wages end salaries 1,683 8,662 7,762 21,832 96,976 273,396 1 I Other good4 and services 702 1,494 3,676 10,166 40,471 127,403 U I Subsidies 668 1,264 2,739 6,147 21,926 61,317 I I Social Socurity payments 2,016 4,336 9,808 29,977 100,109 292,849 1 Interest payments and I I monetary correction c/ 664 1,667 4,952 24,247 163,726 408,731 1 I Transfers to private sector 1,797 3,467 9,626 28,308 144,938 219,636 1 I Transfere abroad 72 220 641 2,184 16,169 74,721 1 Fixed inve-tment 637 1,167 2,143 7,330 82,081 111,403 1 Acquisition of fixod assets (not) 624 2,642 8,789 1 I Financial investments 379 708 1,098 3,836 16,118 106,314 1 I Not lending 27 32 211 1,132 26,600 76,267 1 I OVERALL BALANCE (774) (1,766) (4,366) (21,016) (190,689) (636,320)1 (X of CDP) I I TOTAL RECEIPTS 30.9 31.7 32.0 29.1 32.6 30.1 1 O of which: tax revenue 24.4 26.0 24.7 21.4 22.0 24.6 1 I TOTAL EXPENDITURES 34.1 86.1 36.6 34.6 46.0 47.1 1 Of which: consumption 9.2 9.9 9.6 8.1 9.7 10.8 1 tfixed investment 2.6 2.3 1. 1.9 2.3 3.0 1 I OVERALL BALANCE -3.2 -3.4 -3.6 -6.4 -13.4 -17.0 1 */ Includes the consolidated accounts of the Central Government, states and munic paliti.s, as well as decentralized agencies of the Central Government and of the states. b/ Revenues net of tranefers. c/ Includes only monetary correction on indexed bends (ORTNe) that was actually paid. F'i Sources: FCV, Minietry of Finance and IUF Report (1985 and 1989). TAIa E A .' "ATE'L' E''EPATI N' ''F THE ':EPTRAL ';''VERNMEtIT. 1981-85 7 ", ni i Ifl II 1981 1982 1983 1984 1986 e98e I I I. CONSOLIDATED CENTRAL GOVERNMENT a/ I Cur -nt Rovenueo 8,012 12,862 30,126 90,163 381,878 83e,782 I-- ------ - ------ - - Tax rOvon.oo 4,662 9,729 22,236 81,122 227,018 821,820 I Other revenueo 1,461 2,923 7,890 29,041 134,868 214,962 I… |Current Expend;turce b/ 4,e68 10,097 23,924 so,100 349,134 999 9^. I Wages and ae csr;a 634 1,182 2,448 8,832 29,619 73,471 I Other purcha ea of goods and serviceo 371 792 1,992 4,768 17,889 s8,s81 I Interest paymento/monotary correct;on le 1,889 6,099 27,867 163,219 443 348 I Subsidies and tranofors 3,164 8,234 14,387 40,864 148,507 424,:22 I Of which: trajifora to other I level, of government 674 1,188 2,876 8,648 42,216 125,73e I Current Account Balance 1,327 2,666 8,201 10,083 12,742 (163,'40- Capital Revenues and Grants 144 306 1,823 2,388 8,047 17,552 -- - - - - - -…-- - - - ---- -- ---- - I Capital Expenditures 2,022 4,42S 13,220 36,070 184,100 390,4e4 - Acquicition of fixed capital asseta 239 349 792 2,312 8,038 31,002 I Net lending 1,886 3,497 11,209 30,863 170,e8l 29,6es8 0 | Other 99 680 1,219 1,906 6,393 62,702 ! Overall Balance (661) (1,586) (6,397) (22,640) (186,311) (638,032)1 I II. CENTRAL ADMINISTRATION BUDGET ACCOUNTS c/ I Current Revenue, 2,887 6,402 13,771 36,778 132,196 371,889 1 - I I I Tax revenues 2,687 6,254 13,220 36,393 129,903 347,428 I I Other revenuec 100 148 661 1,386 2,293 24,441 ! I Current Expendituree b/ 2,367 6,297 13,273 48,384 241,288 702,168 - I Wages and salaries 264 620 1,113 3,013 13,645 34,162 I I Other purchases of goods and services 180 374 1,088 2,020 7,023 28,444 I Interest payments/monetary correction 631 1,747 4,933 27,041 161,686 440,e04 I Subsidise and transfers 1,394 2,867 8,169 18,290 89,134 200,958 I I Of which: transfore to other levele of government 498 1,064 2,388 7,800 38 112,822 I Current Account Balance 330 106 498 (11,686) (109,072) (330,297)1 Capital Revonuet and Grants 1 2 591 4 1,095 8,418 1 -------------------- - --- I Capital Expenditures 222 743 3,007 6,119 37,236 172,820 1 Acquisition of fixed capital assets 108 109 183 788 3,864 13,508 - Net lending 114 389 2,287 3,980 33,239 99,191 i Other -- 265 638 353 132 59,923 1 Overall Balance 108 (836) (1,918) (16,701) (146,212) (498,507)1 ... continued c.. following page TAKiA6 .A tI' ' ,ATE" FERAT: N1S F THE -_TRAL VEPIMENT, i481 ef: z 'z n: I i. 1 li it.Ju 1 1981 1982 1983 1984 1986 1986 I'' ATJONOMOUS SOCIAL SECURITY FUNDS C.rrent Revenues 2,840 6,730 16,069 48,889 219,191 425,209 1 - - ___ ________- ------ ------ _ - --- - Tax revonues 1,677 3,988 8,178 22,968 8e,849 249,588 Contribution to SINPAS 1,127 2,841 6,716 16,374 e2,671 170,393 l Contribution to FGTS 33e 713 1,488 3,821 14,317 45,425 ! Contribution to PIS 17 316 883 1,889 8,899 33,770 Contribution to PASEP 97 118 292 894 3,262 Other revenues 1,263 2,742 8,881 26,931 132,342 175,821 ICwrrent Expenditurea 1,807 3,740 8,280 26,342 81,914 230,728 Wages and salaries 78 214 396 1,346 6,862 12,9e7 l Othor purchasoo of goodc and services 84 166 284 867 3,633 9,688 ! Interest payments 63 28 60 210 402 323 ; l Subsidies and other transfer, 1,696 3,343 7,631 22,920 72,017 207,910 ICurrent Account Balance 1,033 2,990 8,799 23,646 137,277 194,481 l l Cspital Revenues and Grants 31 69 117 2 3,248 4,308 l --- --- --- --- --- --- -- l Capital Expenditures 1,649 3,180 8,822 26,412 137,073 199,001 I -- - - - -- - - - - -- - ---- ----- ------ ------- ------- l Acquisition of fixed capital assets 7 11 30 100 277 1,903 l h Not lending 1,641 3,148 8,792 28,312 138,790 197,098 I l Overall Balance (486) (111) (1,908) (2,864) 1,460 (212)1 -- - - -- - - ---- -- - - - - - - - I IV. OTHER DECENTRALIZED CENTRAL GOVERNMENT ACCOUNTS d/ Current Rovenues 339 629 886 4,496 16,092 39,704 I ---------------- ---_---_---_-----_------_ ___ ___ - l Tax revenues 170 302 268 2,382 8,883 24,804 l l Contributions to sarious workers' funds 121 214 46 1,463 6,646 17,113 l l Other 49 88 212 928 3,317 7,891 l l Other revenues 189 227 628 2,113 6,229 14,900 l l Current Expenditures 620 1,161 2,391 6,394 26,962 67,028 l -- ---__ _ _ _ ----- ----- ----- ------ ______-___--- I W ages and salaries 205 448 938 2,474 10,112 26,402 i l Other purchaues of goods and serviceo 107 263 640 1,869 7,233 22,949 I I Interest payments/monetary correction 33 116 116 408 1,261 2,421 l l Subsidies and transfers 176 328 898 1,845 7,366 16,258 l l Of which: transfers to other l levels of government a3 103 263 608 2,969 8,062 1 l Current Account Balance (181) (822) (1,606) (1,899) (10,880) (27,324)l l Capital Revenues and Cranto 113 2,244 914 2,361 4,783 e,834 l !--------------------------- --- ----- --- ----- ----- ----- I~~~~~~~~~~ i l Capital Expenditures 262 523 1,392 3,639 9,792 18,843 l I ----------------- - - --- ----- ----- ----- --- l Acquisition of fixod capital assets 124 228 679 1,427 3,896 1S,873 l Purchases of stocks 78 268 562 1,190 4,166 -- l Purchases of land and intangible assets 21 46 121 362 1,098 2,719 I Net lending 29 (20) 130 681 618 451 1 O Overall Balance (320) 1,099 (1,983) (3,078) (15,889) 239,333;i l _______ _ _----------_- _--_-- _-_----____- o/ Includes the operations of the Federal Treasury and of some 470 decentralized s9,nciae, foundations, and special funds. b/' Includoe only monetary correction on bonded debt ~ant was actually paid. c/ Excludeo transfers to social security funds and other decentralized agencies and funds. d/ Includes oporationo of a large number of docentralized agencies and special funds. Sources: M;nistry of Finance, SEPLAN, FGV, Central Bank of Brazil and IIF Report (1986 and 1987). -266- TABLE A.27: TAX RECEIPTS OF THE CENTRAL GOVERNMENT, 1982-88 (CzS millions) 1082 1908 1984 1965 lse 19e7p 19880 I Total Tex Reellpto */ 12,085 27,140 88,169 812,961 926,208 .. I l~~ ~ ~ -_ - ____ I I Federal Treosury 4,502 11,2389 6,240 130,132 849,998 1,082,984 8,099,61 1 - --- - - - --- - - - - - -- - --- ------ -- --- - I-- --_ --- - - --- - - ------ I Taxes on Income and profits 1,001 4,901 16,5406 60,so 162,210 426,284 3,610,327 1 Taxse on goode and servlice 2,467 5,116 14,612 64,917 163,409 602,462 1,740,942 1 I Tese on interentlonal trade S61 1,310 4,027 9,054 20,216 69,388 566,882 1 I Other tax revenues 284 427 1,661 6,121 18,061 z,;s66 711,702 1 Contributione to autonooue social 4,007 6,841 28,851 68,2f. ?82,456 798,918 I I security funds ----- ----- ------ ------ ------ ------- ---I I Nat'l. Sc. Security Inatit. (SINPAS) 2,041 5,716 16,874 82,671 181,646 660,280 3,716,181 I Unemployment Insurance Fund (FCTS) 713 1,490 8,621 14,817 47,389 121,768 I I Social Integration Fund (PIS) 81S 683 1,809 8,700 38,777 93,760 ... I I Public Employees' Participation Fund i (PASEP) 197 465 1,287 4,606 14,044 88,185 118,213 1 I Other autonomous federal funds 802 678 2,a82 8,603 24,604 ... ... --------_------------------- -_- - --- --- -- ----_-____ --- ---I I Varioue oorkere' funde 214 401 1,468 6,640 17,118 . *-. I I Other B8 212 929 8,817 7,891 ... ... I i State governments 2,784 6,001 19,966 77,876 246,207 ... ... ---------- --- ------ ------ -- ------ -- - ___ - -- - -- -- ----- ---- ---- I Value-added tax (ICM) 2,469 6,606 16,891 71,082 222,444 060,925 4,309,789 I I Other 265 566 1,565 6,206 23,768 ... ... f~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I Munlcipal governmente 840 826 2,254 7,457 22,7S6 ... ... -- -- ----- ---- _ -_-_-_-- --- I ((as of total) I Federal Treseury 80.2 41.4 42.4 41.7 87.7 ... .. I I Social security funds 88.6 80.7 28.1 26.8 80.6 ... *. I Autonomou ftederal funds 2.6 2.5 2.9 2.8 2.7 ... ... I i State governments 22.7 22.8 24.0 24.8 26.6 ... I I Municipal governments 2.8 3.0 2.7 2.4 2.5 ... I I~~~~~ U (as Ł of COP) I Total Tax Recepts 26.0 28.0 21.2 22.1 ... ... . Federal Treoaury 9.5 8.9 9.2 9.4 8.7 e.7 i I Social security funds o.5 7.1 6.0 6.8 7.6 6.7 I Autonomous federal tunde 0.6 0.6 0.0 0.6 0.7 I I State governmonts 6.7 5.1 5.1 5.5 6. ... ... I I Municipal governments 0.7 0.7 0.6 0.5 6 I a/ Totals differ froe those In Statistical Appendix Table 0.1 due to differencee in coverage and classification. b/ Includes operating revenues of INPS, INAMPS, LOA, FUNABEM, and IAPAS. C/ Iniudes taxes earmerked for the following special funds: Commrclal Social Services (SESC), Industrial Socail Service, (SESI), Social Services for COemGrCe (SENAC), and National Commercial Training Institute (SENAT). p preliminary date Sourcos: FGV, SEPLAN, Ministry of Finence, Central lank of Brasil and IMF Report (1986 and 1909). -267- TABLE A.28: FINANCES OF STATE GOVERNMENTS, 1981-86 (CZ$ millions) 1981 1982 1988 1984 1986 1986 l Total Receipts 2,169 4,706 9,788 32,466 138,004 414,922 l I Tax revenues 1,291 2,734 8,061 19,966 77,376 24U,207 I l ICM tax 1,197 2,469 5,686 18,391 71,082 222,444 l l Other taxes 94 265 666 1,566 6,296 23,763 l Other receipts (in.l trenforel l frrm Central Covernment) 877 1,971 3,728 12,609 60,627 188,716 1 l Total Expenditures b/ 2,408 6,382 11,406 36,113 160,670 478,748 1 Conaumption expenditureu 946 2,864 4,437 13,108 69,028 172,066 I I Wagos 732 1,637 3,689 18,263 46,862 137,483 1 l Other goods and oervices 216 427 828 2,863 12,167 34,692 1 l Subsidies 41 9W 219 616 8,878 6,925 1 l Social security payments 191 418 1,019 8,611 13,708 46,886 1 I Interest payments b/ 80 210 749 2,214 11,800 22,063 Transfers to privato sector e8 123 205 861 38566 4,944 1 l Transfers to rest of I I publ;c sector 686 1,564 3,186 10,292 42,600 149,366 1 l Transfers abroad 36 81 184 643 2,278 6,208 1 I Fixed investment 226 518 851 2,930 18,574 52,768 1 Acquisition of fixed assets (net) 62 46 (849)1 Financial investments 129 304 60S 1,948 7,696 20,143 1 l Net lending 3 20 61 39 104 292 1 l Overall balance (238) (677) (1,617) (3,647) (22,87) (83,826)1 I Net borrowing operation. 230 478 727 3,883 14,948 28,110 I Drawinge 289 682 1,813 4,069 20,790 40,476 1 I Amortization (69) (184) (208) (976) (6,842) (12,366)1 I Other 8 199 8s0 684 7,719 36,716 1 (As % of CDP) l T.tal receipts 8.0 9.8 8.3 8.4 9.8 11.2 1 l Of which: tax revenues 6.2 6.7 6.1 5.1 6.6 8.8 I I Total Expenditures b/ 9.7 11.2 9.7 0.3 11.4 12.9 1 l Of which: l consumption expenditures 3.8 4.3 3B, 3.4 4.2 4.8 l fixed investment 0.9 1.1 0.7 0.8 1.2 1.4 l Overall balance (0.9) (1.4) (1.4) (0.9) (1.6) (1.7)1 a.' Information based on official government finance statistics; includes the operations of decen- .-allized agencies attached to stat. governments but not of public enterprises under the control of states. b/ Excludes accrued monetary correction on state bonds that has not yet been pad. p = preliminary data. -268- TABLE A.29: FINANCES OF MUiNICIPAL COVERNMENTS. 1981-86 ( ,$ millions) 1981 1982 1983 1984 1985 1986 1 Total Roceipte 697 1,240 2,974 9,728 34,238 188,963 Tax roVenuoS 199 340 828 2,264 7,467 82,7s4 1 I D;roct taxos 72 148 278 779 2,308 6e,813 1 l Indiroct taxoe 127 194 660 1,476 6,161 16,141 O Other recipts (incl tranferel l from Central Government) 398 908 2,147 7,474 28,780 104,209 1 l Total Expendituroe b/ 812 1,389 3,132 10,321 42,222 137,975 1 Consumption expenditures 340 819 2,026 e,171 25,240 81,826 1 Wages 226 648 1,277 3,828 14,e26 47,994 l Other goods and servicos 116 273 749 2,543 10,418 33,831 Subsidies 9 32 74 263 1,077 3,468 l Social oocurity payments 30 80 164 646 2,221 7,074 l l Intoroest payments b/ 16 42 94 338 1,388 4,139 1 l Transfers to private sector 6 13 31 89 380 1,197 1 l Tronefors to roet of public sector 0 1 139 473 1,944 8,261 1 Transfern abroad 6 17 64 184 767 2,627 1 I Fixed investment 172 322 600 2,088 8,496 27,6e4 1 l Acquisition of fixed assets (not) -- -- -- (2) 31 641 1 I Financial investments 34 43 44 164 838 3,064 1 l Net lending 0 0 7 9 70 346i l Overall balence (16) (121) (169) (692) (7,986) 48,986 1 l Net borrowing oporations 18 124 83 342 2,077 11,876 1 l Drawings 33 1S6 169 628 2,858 9,387 i l Amortization 20 46 86 184 778 (2,488)1 I Other a (3) 78 260 6,907 4,114 1 (As X of GDP) l Total receipts 2.4 2.8 2.5 2.6 2.4 3.4 O Of which: tax revenues 0.8 0.7 0.7 0.6 0.6 0.6 l Total Expenditures b/ 2.6 2.8 2.6 2.7 3.0 3.7 1 I Of which: consumption expenditures 1.4 1.7 1.7 1.8 1.8 2.2 1 I fixed investmont 0.7 0.7 0.4 0.5 0.6 0.7 I Overall balance -0.1 -0.2 -0.1 -0.2 -0.6 -0.3 1 a/ Information baoed on official government finance statistics; includes tho operations of deeen- tralized agencice attached to state governments but not of public enterprises under tho control of statoe. b/ Excludes accrued monetary correction on otate bonds that has not yet been paid. p = preliminary data. Sources: FCV, Centrol Bank of Brazil and IdF Report (1986 and 1989). -269- TAFLE A.30: PUBLIC AND PUBLICLY-GUARANTEED DEBT. 198'-88 (US$ millions) lOEO 1981 1982 1983 1984 1985 :986 :;8s :688 '37AL EXYFNAL DEBT (eDY) _ong 'e r 0ebt 57.311 65,323 74.221 80,648 89,197 90,705 98,255 106 02; W $1 355 PQb] C G4d P.blI,cly CA... 40.706 45,531 51,097 59.136 69,893 73,528 83.614 91.S87 89 841 Pl.-sto Nol,a,.nt.od 16.605 19,792 23.124 21,5;2 19,304 17.177 14,641 14 434 ::.4 .OO of IMPF C4.6t 0 0 580 2,644 4.186 4,619 4.501 3,9?6 3 333 So,t-T. Debt 13 526 :5.321 17,451 14,204 11.500 11.017 9.286 13.868 9.9C3 PH_-IC AJD PU8LIC,y OLARANTEED _3N0-'f1E DffrT 3EB' O.S INC DNISB 5 3.53 57,902 64,970 72,697 81,430 80.043 90.59- C8 *99 :'X.852 OfiFc a. C,ed.to', 12.147 13,534 1S. 883 16,580 18.227 20,731 25.607 3C.886 3C.965 M4.t -.t*.. 6 526 7.589 e.707 10,481 10,10C 12.199 14,628 17.386 II 9c7 18RD 4.585 5,457 6.320 8.093 7,438 C ,304 1.462 :3.630 12 444 'OA 0 0 0 0 0 0 0 0 0 8 .t.r.. 5.621 5,94S 6,176 6,099 8.127 8,532 10,979 13.500 14.968 P,.-8tO Crd t4r- 39.206 44,368 50,087 56.117 63.203 Ł9,312 64.987 67,613 70,687 SW01 .@- 4 :.9 3,643 3,259 2,976 2,604 2.552 2,977 3.153 F.,.lnc.i Meo,kte 34 915 40,567 46,685 5.3.015 60.489 S6,666 62.028 64,392 DEB' OJTS DISS. (DO) 40.706 '5,531 51,097 59.136 69,893 73,528 83,614 91,587 89 841 Off.c.2 C..d..o. 7,002 7,754 8,643 10,028 12,392 15,449 20,488 25,260 24,612 Mult.l8taral 3.:26 3,467 *,109 5.122 5,621 7,358 10.027 12.311 11,413 18fCD 2.069 2,319 2.724 3,6S5 3,997 5,305 7,676 9.411 8,626 IDA 0 0 0 0 0 0 0 0 0 S.Ifter.i 3.876 4.287 4,534 4,906 6,771 8,091 10,461 12,949 13,199 PI-,at* Clod-tore 33.704 37,777 42,454 49,108 57,501 58,079 63,126 66,327 6S,229 S,s0Iio,0 1.838 1,417 1.606 2,110 1,823 2,178 2.473 2,711 F.I",c8eI Morketo 31,866 36,360 40,848 46,098 55.678 55,904 60,653 63,618 C0OMITNENTS 9,512 12,980 13,018 9,130 9,158 3.488 3.174 2,310 6,003 Cff.c.. C.od,to,o 2.070 2,640 2.808 3,457 1.699 2,527 2,485 2,084 957 M..It.Iltsol 1.289 1,564 1.662 2,514 775 1,829 1,910 1,806 871 I8RD 820 1,039 1,090 2,067 308 1,525 1,620 1,394 628 IDA 0 0 0 0 0 0 0 0 0 8, 6tral 7J81 1,085 1,146 943 924 698 575 278 86 Pr.I-to Clod-tor' 7,442 10,331 10,210 5,673 7,456 961 689 226 4,046 5,ppl..os 199 89 85 0 53 250 250 170 F.lelc.et Me,kat,o 7.243 10,242 10.12S 5.673 7,408 711 439 56 OISMRSEB.EW'S 8 23' 10,065 10,631 7,959 9,8SS 2.467 3,422 2,461 5.534 c.6.c.0I C,ordtol 1,168 1,522 2,011 2,284 2,391 1,739 2,605 1.502 1.515 M.4lt.lItalsl 535 599 974 1,441 1,653 1.124 1.890 1,168 1,297 '3RD 343 387 623 1,204 1,300 765 1,619 915 1.019 IDA 0 0 0 0 0 0 0 0 0 B. lfttta' 633 923 1.037 843 738 615 715 334 218 P".4gto Ce.. to. 7.069 8.543 8,620 5.675 7,444 728 817 959 4,":9 S.W.0P a |S342 37 549 755 59 250 267 170 F QC,9c 1 Marketo 6.727 8.506 8.071 4,920 7,385 478 550 789 PR'NCIPA; QEPA'MEWS 3,860 3.941 4,208 2.150 2,451 1.247 2.607 2 684 2.081 ff c $I Cro0tor. 514 568 966 898 1,073 899 1,620 1,603 1.495 M.ut.I,tg,ol 167 200 281 361 432 541 773 1.063 1,204 I8RD 101 137 218 274 335 A09 615 875 979 IDA 0 0 0 0 0 0 0 0 0 8. ete,*$ 347 368 685 537 641 358 847 540 291 PI..at* Cred;tor. 3.346 3,373 3.242 1,252 1.378 348 987 1.081 1.486 S.ppi'a-r 3.3 319 294 161 236 136 173 151 F.nanc,al M,katv 3,rC i 3.054 2,948 1,091 1.142 212 814 930 NUET P OVS 4.377 6,124 6,423 6,809 7,384 1.220 815 (223) 2,553 Off.c.' Clod.tor. 654 954 1,045 1,386 1.318 840 985 (101) 20 M.,It.lteral 368 399 693 1.080 1,221 583 1.117 105 93 I8R0 242 250 405 931 965 356 1.004 40 40 IDA 0 0 0 0 0 0 0 0 0 b.let...l 286 SS. 352 306 97 257 (132) (206) (73) P.l-ato Cr.d.to.. 3,723 5.170 5.378 4,423 6,066 380 (170) (122) 2,533 S.ppl;*e 5 (282) 258 594 (177) 114 94 19 F-... c.ol Mzri˘ 3.718 5 452 5,123 3,829 6,243 266 (264) (141) cont.,.ed or lo.t poss -270- TABLE A.30: PUBLIC AND PUBLICLY-GUARANTEED DEBT, 1980-88 (USS Millions) (continued) 1980 1081 1962 1963 1904 1985 1986 V.6' ;88 14TEREST PAYVE?S (INT) 4,200 5,182 5,982 5,063 S.164 5,585 4.e.40 5.263 1$.117 Off,c.al C'ed.tre 443 494 w88 573 734 066 1,020 1.318 Mult letOral 257 269 290 346 459 544 749 917 927 IBMD 170 184 199 239 339 392 SS8 690 746 IDA 0 0 0 0 0 0 0 0 0 8,18tereal 186 225 298 227 275 322 271 401 271 Preto Craedtore 3M7a7 4,68 5.,94 4,490 4,430 4,719 3.820 3.945 8.919 5.Splher 113 118 82 SO 92 57 64 59 F-anc.al Markets 3,645 4.570 5,312 4.440 4,S38 4,662 3,758 3,886 NEiT TRAGFBS 177 942 *41 746 2,220 (4,365) (4,025) (5,406) (7.564) Off,G.,. C'ed0ta'e 211 460 457 013 54 (26) (35) (1.419) (1,178) Mult.lterol :11 130 403 734 762 39 368 (812) (834) 18D 063 66 207 692 626 (37) 446 (650) (706) IDA 0 0 0 0 0 0 0 0 0 81,tIe,a1 100 330 54 79 (178) (65) (403) (607) (344) P,-ete Cred;tore (S4) 482 (16) (67) 1,636 (4, 39) (3,990) (4,067) (6.386) S.pplere (107) (400) 173 544 (269) 57 30 (40; Financiel Merkote 73 682 (109) (611) 1,905 (4.396) (4.020) (4.027) TOTAL DE8T SERVICE (TDS) 8.060 9,123 10,190 7.213 7,615 6.832 7,447 7,947 13,096 Officml Creditors 967 1,062 1.554 1,471 1,807 1.765 2.640 2,921 2.693 multileateel 424 469 571 707 891 1,085 1.522 1,980 2,131 IBMD 260 321 416 512 674 801 1.17S 1,565 1.725 IDA 0 0 0 0 0 0 0 0 0 lilete.el 533 593 983 764 O16 680 1.118 941 562 P.'.nte Creditors 7.103 8,061 8,638 5,742 5,606 5,067 4.607 5.026 10.405 Suppliers 450 437 376 211 326 193 237 210 Fi;nnc.al Mi.kets 6,653 7,624 8,260 5,531 5,480 4,874 4.570 4,816 AVERACE TER6 OF NDS COmSITh8NTS ALL CREDITORS Interet (1S) 12.6 15.1 12.6 11.0 12.5 9.1 8.5 a 2 0 6 Matw.'ty (years) 9 7 9.9 11.0 10.5 9.6 11.8 12 3 13 1 11 3 irece Per0od (yjere) 3 8 3.3 3.4 2.7 4.4 3.3 2 9 3.1 3 9 GCrent Element (S) (11 8) (21.6) (12.0) (5.4) (12.7) 3.6 5 5 7.8 1 7 -OFrCIAL REDrOARS Intrest (S) 9 1 10 7 10t 10 S 10.0 90 6 3 8 3 7 5 Met.-.ty (jose.,) 12.8 13.0 14.0 13 O 10 5 12 9 13 5 13 6 15 2 C_ece Per.od (yen,r) 3 0 3.0 3.5 3.1 2.6 3.1 3 1 3 3 4 0 Clc.'t Element (S) 6 9 1.1 (1.7) (3.7) 2.7 4 6 6 5 8 0 13 0 PRIVATE CREDITORS Intf,o.t (S) 13 5 16.2 13.1 11 3 13.1 9.3 9 2 7 8 9 9 Meur,ty (yjere) 8 a 9.1 10.2 a 9 9.4 9 1 8 1 8 1 10 4 Grace Pe'i;0d boeer) 4.0 3.3 3 3 2.4 4.8 3 8 2 0 1 5 3 9 0-0nt El,e.nt (S) (16.7) (27.4) (14.8) (6.0) (16.1) 1 0 1 a 6 7 (1 0) ..contins on no-t Pigs -~~~~~~~~~~~~~~~~~~~~~~~~ I -271- TABLE A.30: PUBLIC AND PUBLICLY-GUARANTEED DEBT, 1980-88 (US$ millons) (conclusion) 1980 1481 1982 1988 1984 196 196M 1987 1988 RAJOR ECONOMIC AOW CATES (U S. o mILLr0") Crea Nationsl Prodwet (04) 231,874 252,475 257,290 1941,US 199,473 218.644 238,683 314,.42 373,158 E.,orte of 0. & S. (Xcs) 23,280 26.9S4 25.475 24,343 30,213 29.86s 25,131 28,396 36,564 Imports of C * S (XcS) 36.250 36.873 39,77s 31.206 3o.5as 29,774 30,652 80.64 32.120 International Roam"** (REs) 6,675 7,480 8,997 4,582 11.961 11,610 6.784 7,477 8.458 PRINCIPAL RATIOS Totel EŁttnel DOGt EDT/XQS (U) 304.8 299.2 392.8 400.S 848.8 366.9 448.0 43.2 813.4 EDT/OW (U) 30.6 31.9 36.8 50.1 52.8 48.2 41.7 39.4 30S7 REs/EDT (5) 9.7 9.8 4.3 4.7 11.S 11.1 6.0 6.0 7.4 RU/Nm (MONHs) 2.8 2.3 1.2 1.7 4.7 4.7 2.7 2.9 3.2 P,blie and Publicly Grareantoed Oebt OO/XOS (S) 246.2 242.3 316.2 831.8 295.2 S80.7 891.0 373.4 277.2 DO2/OP (U) 24.7 25.9 28.8 41.5 44.7 41.8 86.6 88.7 27.2 TDS/XQS (U) U.6 67.1 72.1 47.0 38.9 81.8 38.8 34.6 41.9 TOS/Ge (U) 5.7 6.1 6.6 8.9 8.9 4.8 8.6 8.1 4.1 INT/XOS (S) 27 2 29.6 40.0 81.8 28.9 24.7 28.0 22.8 82.0 iKr/Ge (S) 2.7 3.2 8.6 4.0 8.6 3.3 2.5 2.0 3.1 RES/DOD (6) 12.0 11.5 8,4 8.7 13.4 12.8 6.9 7.1 8.3 P,oportion rf D0 buraed Debt Conceasioal D0D 1,791.0 1,680.0 1,740.0 1,476.0 1,SU.0 2,e42.0 4e87.0 4,223.0 3,262.0 Variabl. Rat. DOD 41,402 0 so,284.0 58,549.0 e4.260.0 72,S11.C 69,e48.0 '0718.0 72,55t.0 74,371.0 U.. of IF Cr.dit (USt1 a,) 0.0 0.0 850.2 2,644.3 4,188.4 4,619.3 4S01.2 3.976,0 3,333.0 Purct,ae. 0 0 0.0 850.6 2166.8 1787.8 0.0 0.0 0.0 491.0 R o 0.0 0.0 0.0 0.0 0.0 68.6 616.2 1,149.0 929 0 Source IBM/RDoS -272- TABLE A.31: PUBLIC AND PRIVATE DEBT, 1980-88 (USS millions) CC Ile. 1i82 1983 1i84 IC895 lA86 1 ,87 :, C'AL DEEr OVTSTAP0CNC C."8'RSED Z,Y TCTAL 0EaT 57,3.1 65,323 74,221 80,648 89,197 90705 98,255 16,Z"21 1':355 P,bl-c/P,bf.c's C0.. 40.'C6 45.53: 51,097 50,136 69,893 73,528 83,614 91,587 8H 84: P,,-ete . n-g-,8te.d 16.6,5 i9.792 23,124 21,5:.2 11,304 17,177 .4,641 14,434 I: 5.4 0ISPURSe4ENTS 11,428 16:.26 16,026 10.362 11,623 2,467 3,440 3,6:5 6. VA PRINCIPAL REPAYMe14TS 6.8.30 7,412 7,534 3,705 4,529 3,749 3.946 4,572 S. C3 NE'T FLOWS 4.518 8,714 6,492 6,657 7,004 (1.282) (506) (957) 1I0QC !NTEREST PAYFIETS 7.856 10,347 11,544 9,396 9,049 9,030 7,743 7,527 :3 2:3 NEr TRANSFERS (3,258) (1,633) (3.052) (2,739) (1,455) (10.312) (8,249) (8,484) (:2 :22 TC'AL DET SRVICE 14.686 17.759 19,078 13.101 13,578 12,779 11,689 12.099 : 3.6 So.,4 ItPRO/DRs