A W O R L D B A N K C O U N T R Y S T U D Y Restoring Restoring Fiscal Restoring Fiscal Discipline for Poverty Reduction in Peru is part of the World Bank Country Study series. These reports are published with the approval of the subject government to communicate the results Discipline for Poverty of the Bank's work on the economic and related conditions of mem- ber countries to governments and to the development community. This report was done in joint collaboration with the Inter-American Reduction in Peru Development Bank. Fiscal Since inauguration in July 2001, President Toledo has proposed a number of important reforms critical to tackling the development challenges facing Peru. The efforts of the Toledo administration to A Public Expenditure Review reinvigorate the momentum for reform are facing great social pres- Discipline sures for short-term, populist measures. Peru has an urgent need to build momentum for "second-generation" reforms, especially those in the area of public expenditures, which would provide longer-term sustainability to the government's strategy of poverty reduction and growth. Despite its initial progress, Peru's chances of meeting its develop- ment goals in the medium term will depend on completion of a pub- for lic expenditure management agenda summarized in the seven main themes of this report: restoring fiscal discipline, reorienting the Poverty budget toward pro-poor expenditure, improving the efficiency of pub- lic expenditure, addressing the promises and risks of decentraliza- tion, upgrading the civil service, improving governance and reducing corruption, and fine-tuning mining taxation and environmental poli- cies. This study distills the critical lessons and challenges that are rel- evant for policymakers in Peru, Latin America, and other countries Reduction embarking on comprehensive public expenditure reforms. World Bank Country Studies are available individually or by subscrip- tion, both in print and on-line. in Peru THE INTER-AMERICAN DEVELOPMENT BANK 1300 New York Ave., NW Washington, DC 20577 USA Telephone: 202 623-1000 Internet: www.iadb.org ISBN 0-8213-5447-7 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org THE WORLD BANK THE INTER-AMERICAN DEVELOPMENT BANK A W O R L D B A N K C O U N T R Y S T U D Y Restoring Fiscal Discipline for Poverty Reduction in Peru A Public Expenditure Review THE WORLD BANK THE INTER-AMERICAN DEVELOPMENT BANK Washington, D.C. Washington, D.C. Copyright © 2003 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing: June 2003 1 2 3 4 05 04 03 World Bank Country Studies are among the many reports originally prepared for internal use as part of the continuing analysis by the Bank of the economic and related conditions of its develop- ing member countries and to facilitate its dialogues with the governments. This report was pre- pared in joint collaboration with the Inter-American Development Bank. Some of the reports are published in this series with the least possible delay for the use of governments, and the academic, business, financial, and development communities. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to journal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and normally will grant permission for use. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use, is granted by the World Bank, provided that the appropriate fee is paid. Please contact the Copyright Clearance Center before photocopying items. Copyright Clearance Center, Inc. 222 Rosewood Drive Danvers, MA 01923, U.S.A. Tel: 978-750-8400 · Fax: 978-750-4470. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the World Bank at the address above, or faxed to 202-522-2422. ISBN: 0-8213-5447-7 eISBN: 0-8213-5448-5 ISSN: 0253-2123 Cover Art: Fernando De Szyslo (Peru, 1925), Duino-Nueve (Orrantia/92), 1992, oil on canvas, 99.06 cm × 99.06 cm, Inter-American Development Bank Art Collection, Washington, DC, cour- tesy of the IDB Photo Library Unit. Photo of art: Gregory R. Staley Photography. Library of Congress Cataloging-in-Publication Data Restoring fiscal discipline for poverty reduction in Peru: a public expenditure review. p. cm. -- (A World Bank country study) Includes bibliographical references. ISBN 0-8213-5447-7 1. Government spending policy--Peru. 2. Economic assistance, Domestic--Peru. 3. Poverty-- Government policy--Peru. I. World Bank. II. Series. HJ7742.R48 2003 336.3 9 0985--dc21 2003050056 CONTENTS Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xiii Abbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xv I. Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Rationale and Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Main Findings and Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 II. Peru's Fiscal Challenges and Vulnerabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Long-Run Fiscal Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 The Current Fiscal Disequilibria, 1999­2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 The Macroeconomic Consistent Expenditure Envelope, 2002­03 . . . . . . . . . . . . . . . . .23 The Fiscal Adjustment Effort. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 Revamping the Fiscal Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 III. Reorienting the Budget Toward Pro-Poor Expenditure . . . . . . . . . . . . . . . . . . . . .41 Modernizing Financial Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 Upgrading the Budget Management System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46 Building a Medium-Term Expenditure Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Protecting the Budget as a Medium-Term Social Policy . . . . . . . . . . . . . . . . . . . . . . . . .57 Budget Transparency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 IV. Improving the Efficiency of Public Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . .65 Does Higher Government Expenditure Buy Better Results in Education and Health Care? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 Tracing Leakages of Public Funds in Peru-- A Public Expenditure Tracking Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70 Enhancing Targeting of Social Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88 V. Addressing the Promises and Risks of Decentralization . . . . . . . . . . . . . . . . . . . . .95 The Decision to Decentralize: Promises and Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 The Deconcentration of Central Government Spending and Transfers to the Municipalities: Intergovernmental Fiscal Relations . . . . . . . . . . . . . . . . .98 The Emerging Legal Framework for Decentralization in Peru . . . . . . . . . . . . . . . . . . .102 Entering a Decentralization Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104 Decentralization in the Education Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 Decentralization in the Health Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 VI. Upgrading the Civil Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115 The Size and Employment Conditions of the Public Sector . . . . . . . . . . . . . . . . . . . . .116 Reforming the Legal and Institutional Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . .123 Additional Measures for Improving Human Resource Administration . . . . . . . . . . . . .127 The Education Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .127 iii iv CONTENTS VII. Improving Governance and Reducing Corruption . . . . . . . . . . . . . . . . . . . . . . . .133 The Costs of Weak Governance and Corruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138 Impact on Public Service Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139 A Policy Agenda for Improving Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141 VIII. Fine-Tuning Fiscal and Environmental Mining Policies . . . . . . . . . . . . . . . . . . .147 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148 The Peruvian Taxation Regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148 Comparison of Peru's Tax System to Minerals Tax Systems in Selected Countries . . . . .148 Improving Mining Environment Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .165 Annex A. A Public Debt Sustainability Analysis for Peru . . . . . . . . . . . . . . . . . . . . . . .169 Annex B. The Fiscal Effort Required for a Sustained Structural Deficit . . . . . . . . . . .193 Annex C. Questionnaire on Public Financial Management . . . . . . . . . . . . . . . . . . . . . .199 Annex D. A Public Expenditure Tracking Survey: Methodological Issues . . . . . . . . . .207 Annex E. Public Enterprise Reform in Peru: Introducing Management Contracts at FONAFE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217 Annex F. A Methodology for Functional Reviews and for Analyzing Functional Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221 Annex G. List of Background Papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .223 Statistical Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .225 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .281 Map of Peru . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .285 LIST OF TABLES 2.1 Key Macroeconomic Variables, 1969­89 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 2.2 Elasticity to GDP, 1970­2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 2.3 Elasticities and Growth Volatilities for Peru, 1990­2000 . . . . . . . . . . . . . . . . . . . . . . .20 2.4 Key Economic Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 2.5 Recommendations for Tax Reform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 2.6 Social Expenditures are Pro- or Countercyclical . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35 3.1 Peru: Status of Budget Management Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 3.2 Composition of CG Budget Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 3.3 Revised Projection of the 2002 Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52 3.4 Composition of Social Expenditure 1999­02 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 3.5 Comparing Multiyear Budgets by Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56 3.6 Budget for Protected Social Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 3.7 Budget Transparency Ratings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61 4.1.a Regression Results for Education Indicators: Linear Regressions . . . . . . . . . . . . . . . .67 4.1.b Regression Results for Health Indicators: Log-Log Regressions . . . . . . . . . . . . . . . . .68 4.2 Peru: Efficiency of Public Expenditure in Reaching Social Outcomes, 1990­98 . . . . .69 4.3 Per Capita Transfers to Municipalities in 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75 CONTENTS v 4.4 Municipalities that Do Not Know the Arrival Day of the Transfer . . . . . . . . . . . . . . . .76 4.5 Municipalities that Are Subject to CG Supervision . . . . . . . . . . . . . . . . . . . . . . . . . . .77 4.6 Leakages in FONCOMUN and Canon Minero . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77 4.7 Fraction of Transfers Used for Current Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . .78 4.8 Beneficiary Households that Received Training/Information . . . . . . . . . . . . . . . . . . .80 4.9 VDL Transfer Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81 4.10 Leakage Stage 3: Municipality to Local Committees . . . . . . . . . . . . . . . . . . . . . . . . . .82 4.11 Worst Offenders, Leakage Stage 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83 4.12 Leakage Stage 4: Vaso de Leche Program (at household level, in percent) . . . . . . . . .85 4.13 Leakage Stage 5: Vaso de Leche Program (at beneficiary household level) . . . . . . . . .86 4.14 Vaso de Leche Leakages (in percent) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87 4.15 Correlation Expenditure and Poverty by Department . . . . . . . . . . . . . . . . . . . . . . . . .89 4.16 Household Access to Social Programs by Poverty Level, 2000 . . . . . . . . . . . . . . . . . .90 4.17 Targeting by Food Assistance, Health, and Education Programs, 2000 . . . . . . . . . . . .92 5.1 Budgeted Expenditure for the Public Sector for FY02 . . . . . . . . . . . . . . . . . . . . . . . . .99 5.2 Expenditure Coefficients by CTAR, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100 5.3 Expenditure by Department, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101 6.1 Number of Employees and the Average Monthly Wage Bill . . . . . . . . . . . . . . . . . . .118 6.2 Urban Public and Private Sector Employment and Average Monthly Earnings, 1997­2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .120 6.3 Dispersion of Salaries Charged to the PUP by Occupational Group and Salary Grade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122 6.4 Maximum, Average and Minimum Earnings by Occupational Group and Regimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123 6.5 Hourly Earnings by Occupational Groups and Areas, Third Quarter, 1997 (Nuevos Soles) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129 8.1 Mineral Taxes in Peru and Other Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 8.2 Comparative Measures of Profitability and Effective Tax Rate for a Model Base Metal Mine in Selected Jurisdictions . . . . . . . . . . . . . . . . . . . . . . . . . . .153 8.3 Comparative Economic Measures for a Model Gold Metal Mine in Selected Jurisdictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154 8.4 Copper Model: Peru's Tax System Sensitivity to Price and Cost Changes . . . . . . . . .154 8.5 Availability of Tax Stability in Selected Jurisdictions . . . . . . . . . . . . . . . . . . . . . . . . .155 8.6 Tax System Sensitivity to Income Tax Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .156 8.7 Income Tax Rates Applied to Mining Projects in Selected Jurisdictions . . . . . . . . . .157 8.8 Depreciation Applied to Typical Mining Equipment in Selected Jurisdictions . . . . . .158 8.9 Tax Sensitivity to Building Depreciation Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . .158 8.10 Loss Carry Forward/Back Policy in Selected Jurisdictions . . . . . . . . . . . . . . . . . . . .159 8.11 Tax Sensitivity to Loss Carry Forward Time Limit . . . . . . . . . . . . . . . . . . . . . . . . . .159 8.12 Tax Sensitivity to Annual Allowed Closure Deduction Over the Mine Life . . . . . . . .160 8.13 Tax Sensitivity to Reinvestment Allowance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .161 vi CONTENTS 8.14 Presence of Mineral Royalty Tax Systems in Selected Jurisdictions . . . . . . . . . . . . . .162 8.15 Tax System Sensitivity to a Royalty Tax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .163 8.16 Tax System Sensitivity to Import Duty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .163 8.17 Tax Sensitivity to Eliminating Import Duty and Imposing a Royalty . . . . . . . . . . . . .164 A.1 Current Macroeconomic and Financial Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . .171 A.2 Debt Dynamics Non-Financial Public Sector: Scenario With No External Shocks . . .172 A.3 Scenario Without Shocks (Scenario A) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176 A.4 Scenario With No Shocks (Scenario B) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .178 A.5 Scenario With No Shocks (Scenario C) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180 A.6 Real Exchange Rate Adjustment in the Event of a "Sudden Stop" . . . . . . . . . . . . . .182 A.7 Public Sector Debt Sustainability With and With No Shocks . . . . . . . . . . . . . . . . . . .185 A.8 Scenarios With Shocks (10% depreciation in real exchange rate and adjustment for contingent liabilities) (Scenario A) . . . . . . . . . . . . . . . . . . . . . . . . . . .186 A.9 Scenarios With Shocks (10% depreciation in real exchange rate and adjustment for contingent liabilities) (Scenario B) . . . . . . . . . . . . . . . . . . . . . . . . . . .188 A.10 Scenarios With Shocks (10% depreciation in real exchange rate and adjustment for contingent liabilities) (Scenario C) . . . . . . . . . . . . . . . . . . . . . . . . . .190 B.1 Required Fiscal Effort for Debt Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .196 D.1 Total Transfers to Municipalities in 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .210 D.2 Volatility of Transfers to Municipalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211 E.1 Public Enterprises Target Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218 SA.1 Peru: Combined Public Sector Operations (millions of nuevo soles) . . . . . . . . . . . . .226 SA.2 Peru: Combined Public Sector Operations (percent of GDP) . . . . . . . . . . . . . . . . . .227 SA.3 Peru: Combined Public Sector Operations (millions of 1994 soles) . . . . . . . . . . . . .228 SA.4 Peru: Central Government Operations (millions of nuevo soles) . . . . . . . . . . . . . . . .229 SA.5 Peru: Central Government Operations (percent of GDP) . . . . . . . . . . . . . . . . . . . . .231 SA.6 Peru: Central Government Operations (millions of 1994 soles) . . . . . . . . . . . . . . . .233 SA.7a Peru: Structure of Central Government's Fixed Capital Formation (millions of nuevo soles) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .235 SA.7b Peru: Structure of Central Government's Fixed Capital Formation (percent of total) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .236 SA.8a Peru: Operations of the Non-Financial Public Sector (millions of nuevo soles) . . . . .237 SA.8b Peru: Operations of the Non-Financial Public Sector (percent of GDP) . . . . . . . . . .238 SA.8c Peru: Operations of the Non-Financial Public Sector (millions of 1994 soles) . . . . .239 SA.9 Peru: Local Government Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .240 SA.10 Peru: Operations of the Non-Financial State Enterprises (millions of nuevo soles) . .241 SA.11 Peru: Operations of the Rest of the Central Government (millions of nuevo soles) . .242 SA.12a Peru: Functional Classification of Central Government Budget Expenditures (millions of nuevo soles) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .243 SA.12b Peru: Functional Classification of Central Government Budget Expenditures (percent of GDP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .244 SA.13 Peru: Central Government Spending by Ministry or Institution . . . . . . . . . . . . . . . .245 CONTENTS vii SA.14 Peru: Social Spending by Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246 SA.15 Peru: Composition of Social Expenditures (millions of nuevo soles) . . . . . . . . . . . . .247 SA.16 Peru: PESEM: Projected Social Spending by Institution, 2002­06 . . . . . . . . . . . . . .249 SA.17 Peru: Pledges at the 2002 Consultative Group for Peru . . . . . . . . . . . . . . . . . . . . . .250 SA.18 Peru: Privatizations and Concessions, 2001­04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251 SA.19 Peru: Current Revenues by Awarded Concessions (millions of US$) . . . . . . . . . . . .252 SA.20 Peru: Comparative Performance in Social Indicators . . . . . . . . . . . . . . . . . . . . . . . . .252 SA.21 Peru: Main Tax Breaks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253 SA.22 Peru: Alternative Measures for Tax Reform by Sector or Region . . . . . . . . . . . . . . . .253 SA.23a Latin America General Government's Tax Revenues, 1998 (percent of GDP) . . . . .254 SA.23b Latin American Economies: Outline of Fiscal Policy Rules . . . . . . . . . . . . . . . . . . . .255 SA.24 Peru: Elasticities of Social Spending to Total Spending of the Consolidated Public Sector in Peru (1997­2002) . . . . . . . . . . . . . . . . . . . . . . . . . . .256 SA.25 Peru: Degree of Transparency in the Publication of Institutional Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257 SA.26 Peru: Published Information of the Budget in 2002 . . . . . . . . . . . . . . . . . . . . . . . . .257 SA.27 Peru: Composite Governance Indicators, International Comparisons, 1998 . . . . . . .258 SA.28 Volatility in Latin America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259 SA.29 Economic Distribution of General Government Revenue and Expenditure, Selected Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .260 SA.30 Peru: Distribution of Social Investment by Different Social Programs and by Decile of Poverty Severity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .261 SA.31a Peru: Average Monthly Per Capital Income of Households, 2000 (US$) . . . . . . . . .262 SA.31b Peru: Household Access to Social Programs by Income Per Capita Deciles, 2000 (percent of total) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262 SA.31c Peru: Household Access to Social Programs by Income Per Capita Deciles, 2000 (percent of subtotal) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262 SA.32 Peru: Mistargeting by Types of Food Assistance, 2000 . . . . . . . . . . . . . . . . . . . . . . .263 SA.33 Peru: Mistargeting by Types of Health Programs, 2000 . . . . . . . . . . . . . . . . . . . . . .263 SA.34 Peru: Mistargeting by Types of Education Programs, 2000 . . . . . . . . . . . . . . . . . . . .264 SA.35 Peru: Evolution of Household Access to Social Programs, 1998­2000 . . . . . . . . . . .264 SA.36 Peru: Annual Private Expenditure on Education by Per Capita Income Decile, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .265 SA.37 Peru: Annual Expenditure on Health by Per Capita Income Decile . . . . . . . . . . . . .266 SA.38 Peru: Definitions of the Consolidated Public Sector (CPS) . . . . . . . . . . . . . . . . . . . .266 SA.39 Peru: Central Government National and Regional Budget Composition, 2002 . . . .267 SA.40 Peru: Main Revenue Sources for Municipalites, 2002 . . . . . . . . . . . . . . . . . . . . . . . .268 SA.41 Peru: Departmental Budget­Education and Health by Expenditure Category, 2002 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .269 SA.42 Peru Indexes on Corruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .270 SA.43 Peru: Corruption in the Judiciary, National and Municipal Agencies . . . . . . . . . . . .272 SA.44 Peru: Expenditure by Department, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274 viii CONTENTS SA.45 Peru: Expenditure Coefficients by Department, 2001 . . . . . . . . . . . . . . . . . . . . . . . .275 SA.46 Peru: Decentralization in LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276 SA.47 Peru: Principal Revenue Sources of Subnational Governments . . . . . . . . . . . . . . . . .276 SA.48 Peru: Summary of Mining Taxation Authority for Selected Taxes and Fees . . . . . . . .277 SA.49 Peru: Mining Fiscal Methods and their Amenability to Fiscal Decentralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .279 SA.50 Total Transfers in Soles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .279 LIST OF FIGURES 2.1 NFPS Fiscal and Primary Balances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 2.2 Tax Revenue Composition, 1970­2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 2.3 Procyclical Fiscal Stance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 2.4 Volatility in Fiscal Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 2.5 Fiscal Trend Over Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 2.6 LAC: VAT Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 2.7 Central Government's Expenditure Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 2.8 LAC: Capital Expenditure (1998) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 2.9 Structure of Central Government's Fixed Capital Formation . . . . . . . . . . . . . . . . . . .32 2.10 Social and Public Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 2.11 Growth Rates in Social and Public Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35 2.12 Actual and Potential Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 2.13 Actual Primary Balance (% of GDP) and Structural Primary Balance (% of potential GDP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 3.1 Comparison of Central Government and NFPS Balance . . . . . . . . . . . . . . . . . . . . . . .53 3.2 Share of Social Expenditures to GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54 3.3 Sectoral Budget Allocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54 3.4 Index of Budget Transparency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62 3.5 Percent of Government Entities Publishing Information . . . . . . . . . . . . . . . . . . . . . . .63 4.1 Lorenz Curves for Intergovernmental Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74 4.2 How the Milk Disappears in the Vaso de Leche Program . . . . . . . . . . . . . . . . . . . . . .86 4.3 Lorenz Curves in Selected Social Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90 4.4 Lorenz Curves for Health Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91 4.5 Lorenz Curves for Education Expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91 4.6 Access to Social Programs, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92 5.1 Municipal Revenues, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .102 5.2 Central Government Transfers to Municipals, 2001 . . . . . . . . . . . . . . . . . . . . . . . . .102 6.1 International Government Employment Comparison . . . . . . . . . . . . . . . . . . . . . . . .116 6.2 Corruption in Public Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117 6.3 Employment and Average Earnings of the Private and Public Sector . . . . . . . . . . . . .121 6.4 Permanent and Contractual Average Earnings (Nuevos Soles) . . . . . . . . . . . . . . . . .124 CONTENTS ix 7.1 Peru's Governance Indicators, 1998­2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .134 7.2 Obstacles to Doing Business . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .135 7.3 Perceived Level of Corruption in Selected Agencies . . . . . . . . . . . . . . . . . . . . . . . . .135 7.4 Bribes by Type of Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137 7.5 Corruption in Public Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138 7.6 Average Percent of Income Paid in Bribes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139 7.7 Indexes of Service Quality and Bribery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .140 7.8 Quality of Public Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .140 7.9 Quality of Health and Education Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141 7.10 Provision of Public Services--Local Governments vs. Other Government Agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142 7.11 National and Municipal Agencies Are Ridden by Different Types of Corruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142 7.12 Public Official's View on Reforms in Public Sector . . . . . . . . . . . . . . . . . . . . . . . . . .144 A.1 Scenario with No Shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .175 A.2 Scenario with Shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .182 A.3 Peru Debt Dynamics: Comparison of Scenarios With and With No Shocks . . . . . . .184 D.1 FONCOMUN Transfer Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .212 LIST OF BOXES 2.1 The Fiscal Prudence and Transparency Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22 2.2 Tax Regime and Proposed Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25 2.3 Peru: Evaluation of the Transparency Law 2000­01 . . . . . . . . . . . . . . . . . . . . . . . . . .36 3.1 Main IGR/PEM Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 3.2 The Integrated Financial Management System (SIAF) of Peru . . . . . . . . . . . . . . . . . .43 3.3 Introducing the SIAF at the Ministries of Defense and National Security (Interior) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 3.4 Peru: Principal Stages of the Budget Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47 3.5 Peru: Institutionalizing a Virtual Poverty Fund: A Look at Best-Practice Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 4.1 MEXICO: Integrating Social Programs for Improved Efficiency-- The Case of PROGRESA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93 5.1 Lessons from Decentralization Experiences in Latin America . . . . . . . . . . . . . . . . . . .97 5.2 The Decentralization Drive in Education in LAC: Some Cases . . . . . . . . . . . . . . . . .108 5.3 Decentralization of the Health Sector in LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111 6.1 The Sequencing of Civil Service Reform Measures . . . . . . . . . . . . . . . . . . . . . . . . . .130 7.1 Regulation of Infrastructure in Peru . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136 7.2 A Multipronged Strategy for Addressing Corruption and Improving Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 8.1 Peru: Description of the Existing Mining Fiscal System . . . . . . . . . . . . . . . . . . . . . . .149 8.2 Creating Mining Stabilization Funds: The Ancash Model . . . . . . . . . . . . . . . . . . . . .165 ABSTRACT S ince his inauguration in July 2001, President Toledo has proposed to take actions in the areas of macroeconomic stabilization; reopening of country's access to international financial markets; budget modernization and State decentralization; social policy; revamping of the armed forces, the police, and internal security services; and consolidation of Peru's democracy and social stability through consensus-building and transparent and participatory mechanisms. These are all critical development challenges faced by Peru. The momentum for reform, which was strong in the early 1990s, ran out of steam in the late 1990s, and genuine efforts by the new administration to re- invigorate it are facing great social pressure for short-term, populist measures. There is an urgent need to build momentum for "second-generation" reforms, especially those in public expenditure, which would provide longer-term sustainability to the government's strategy of poverty reduction and growth. Despite its initial progress, Peru is not likely to meet these development goals in the near term. Achieving success in the medium term will depend on completing a public expenditure manage- ment agenda summarized in the seven main themes of this report: Restoring fiscal discipline Reorienting the budget toward pro-poor expenditure Improving the efficiency of public expenditure Addressing the promises and risks of decentralization Upgrading the civil service Improving governance and reducing corruption Fine-tuning mining taxation and environmental policies. How to address these themes is clear. Significant primary surpluses are fundamental to achiev- ing sustainable fiscal and debt-management policies, and are thus a top priority. Shifting expendi- ture to pro-poor priorities is essential for achieving desired social outcomes in the medium term. Making the most effective use of scarce resources is needed for a country in the midst of fiscal retrenchment. Fiscal discipline also implies resolving fundamental questions of the decentralization process, particularly, what functions must be decentralized, and with what resources, timing, and desired outcomes? An ill-designed decentralization process, especially if developed through brand- new regional governments, has considerable potential for engendering severe fiscal disarray and macroeconomic destabilization. Thus there is a need to proceed very cautiously. The government of Peru also needs to upgrade its human capital. Such reform would con- tribute to improving governance and fighting corruption, thus providing more productive and less expensive delivery of public service. Finally, fine-tuning mining taxation and transfers and mining- related environmental policies is needed not only to strengthen the global competitiveness of one of the main exports of Peru through a stable tax regime, but to enhance budgetary transparency and minimize environmental damage. This report is the product of the analytical work of a large number of experts, both Peruvian and non-Peruvian. They distill critical lessons and challenges that are relevant for policymakers in Peru, in Latin America, and in countries that are embarking on comprehensive public expenditure reform. Thus, it is a major contribution to our knowledge and thinking in this area. xi ACKNOWLEDGMENTS T his report is a joint effort of the World Bank and the Inter-American Development Bank. José R. López-Cálix (WB-LCSPE) and Alberto Melo (IDB-RE3/OD5) are the authors and project managers. The report is an outgrowth of a collective effort including partial contributions and background papers by Osvaldo Schenone (consultant­section on taxation, Chapter II); Carlos Oliva (consultant­section on fiscal patterns, Chapter II); Nelson Shack and Oscar Pajuelo (General Budget Director and Accountant General of Peru­Annex C); Bruno Barletti and Lourdes Cueva (SIAF Coordinator and assistant­section on transparency, Chapter III); Rossana Mostajo (consultant­section on protected programs, Chapter III); Lorena Alcazar, Erik Wachtenheim, Mitchel Seligson (consultants­section on public expenditure tracking survey, Chapter IV and Annex D); Gillette Hall (WB­Box on PROGRESA, Chapter IV), Juvenal Diaz (FONCODES­ section on targeting, Chapter IV); Richard Moore (consultant­section on decentralization of social sectors, Chapter V and Annex F); Eduardo Fernandez (consultant­municipal debt, Chapter V); Peter Gregory (consultant­civil service, Chapter VI); Francesca Recanatini and Daniel Kauffman (WB­governance and corruption, Chapter VII); José Luis Guasch (WB­box on regulatory agen- cies, Chapter VII); James Otto (consultant­mining taxation, Chapter VIII) and Felix Remy (WB­environment, Chapter VIII); Elaine A. Tinsley (WB­Statistical Appendix and thorough analytical and research assistance); Rashmi Shankar (consultant­research assistance on Annexes A and B); and Gabriel Ortiz de Zevallos (consultant­Annex E). Peer reviewers were Ritva Reinikka and William Dorotinsky (both WB), who provided very helpful and detailed input. The report also received numerous comments from Andrew Wolfe and Mauricio Villafuerte (IMF), Guillermo Perry, Mauricio Carrizosa, Sara Guerschanik Calvo, Yasuhiko Matsuda, Gillette Hall, Pierre Werbrouck, Quentin Wodon, Keta Ruiz, Marianne Fay, Harold Adelman, Norbert Schady, Evangeline Javier, Daniel Cotlear, Livia Benavides (all from the WB), and Gonzalo Afcha, Xavier Comas, Rafael de la Cruz, Gonzalo Deustua, Jaime Fernandez, Amanda Glassman, Susan Kolodin, Jorge Lamas, and Jaime Enrique Vargas (all from the IDB). It also greatly benefited from the field survey work of Instituto Apoyo in Peru, which did a magnificent job. Hazel Vargas pro- vided the main administrative and production support and editorial assistance. Diane Stamm was the principal language editor. Also providing production support at various stages were Silvia Marquina de León, Rosalía Rushton, Marianella Rivadeneira and Patricia Chacón Holt. Ana Maria Arteaga and Ana Maria Angulo provided excellent support to field research in Lima. The report benefited from the active guidance and support of Isabel Guerrero (WB­Director), Ana-Maria Arriagada (WB­LCSHD Sector Director), Mauricio Carrizosa (WB­LCSPE Sector Manager), Ernesto May (WB­LCSPR Director), and Ana Maria Rodriguez (IDB, RE3/OD5 Chief). The team would like to express its sincere gratitude for the excellent cooperation provided by Peruvian officials throughout the process, especially during the team visits in Peru, and including the main mission in late November 2001. In particular, the team would like to thank Mr. Roberto Dañino, President of the Council of Ministers; Pedro Pablo Kuczynski, Minister; Kurt Burneo, Patricia Teuillet, and Fernando Zavala, Viceministers of Economy and Finance; and Oscar Pajuelo, Accountant General, for their enthusiastic support. Many thanks also go to Veronica Zavala, Nelson Shack, Oscar Blanco, Fernando Lituma, Bruno Barletti, Carlos Ricse, Lourdes Cueva, Waldo Mendoza, Carmen Negron, Javier Abugattas, Lander Aleman, Jorge Estrella, Luis Paz, Pedro Francke, Beatriz Merino and Gustavo Perochena, Gilberto Moncada, and Reynaldo Bringas who provided timely inputs and responded to numerous requests. The report has been subject to early and multiple consultation and dissemination efforts in Lima, Peru. In December 2001, at the closing of the main mission, draft preliminary conclusions were discussed with Peruvian officials. In March 2002, a detailed proposal on the revamping of xiii xiv ACKNOWLEDGMENTS the Fiscal Prudence and Transparency Law was presented in a Ministry of Economy and Finance seminar on Fiscal Rules. A month later, a framework for decentralization was submitted in the FONCODES/Presidency joint seminar on the Decentralization of the Social Sectors. Finally, in August 2002, the full draft report was openly discussed with officials, think tanks, the media, donors, the private sector, and civil society representatives in a seminar on Fiscal Discipline, Pub- lic Expenditure Efficiency, and Decentralization jointly organized with MEF and IDB. Prelimi- nary findings of this report and background papers have also been instrumental in the elaboration of components of the World Bank's Second Programmatic Social Reform Loan and the Country Assistance Strategy for Peru. ABBREVIATIONS AND ACRONYMS 2SLS Two-stage least squares ACE Community educational associations ADEs Areas of Educational Development AEs Areas of Execution BCRP Central Bank of Peru BN Banco de la Nación (Treasury) CEC Consejos Escolares Consultivos CETICOS Centers for Export, Transportation, Industry, Commercialization and Services CG Central Government CIF Cost, insurance & freight CLAS Local Committees of Health Administration CONAM Consejo Nacional del Ambiente (National Environmental Council) CONASEV Comisión Nacional Supervisora de Empresas y Valores CONITE National Commission for Foreign Investment and Technology CONSUCODE Consejo Superior de Contrataciones y Adquisiciones del Estado, (Superior Council of Contractings and Procurements from the State) COOPOP National Office for Popular Cooperation COPRI Commission for Promotion of Private Investment CPS Consolidated public sector CTARs Transitory Councils of Regional Administration DGAES Dirección General de Asuntos Económicos y Sociales, (General Office of Economic and Social Affairs) DGFP General Directorate of Fiscal Policy DGSP Dirección General de Salud Pública DISAs Regional Directorates of Health DNPP Dirección Nacional del Presupuesto Publico (National Directorate of the Public Budget) DREs Regional Directorates of Education EBFs Extra-Budgetary Funds EDUCO Educación Comunitaria EGASA Empresa de Gas Sociedad Anonima ENAHO Encuesta Nacional de Hogares (National Household Survey) ESSALUD Seguro Social de Salud (Health Insurance) ETR Effective tax rate EUs Executing Units FEF Fondo de Estabilización Fiscal FGT2 Foster, Greer, and Thorbecke 2 FIDE Intergovernmental Decentralization Fund FLIRB Front-loaded interest reduction Brady bonds FONAFE National Fund for Financing Government Enterprise Activity FONAHPU National Savings Fund FONAVI Fondo Nacional de Vivienda FONCODES Fondo Nacional de Compensación y Desarrollo Social FONCOMUN Fondo de Compensación Municipal FONCOR Regional Compensation Fund FSF Fiscal Stabilization Fund xv xvi ABBREVIATIONS AND ACRONYMS FY Fiscal year GDE General Directorate of the Environment GDP Gross domestic product GFS Government Financial Statistics GG General government GOP Government of Peru HIPCs Highly indebted poor countries IBRD International Bank for Reconstruction and Development IDB, IADB Inter-American Development Bank IES Extraordinary Solidarity Tax (Impuesto Extraordinario de Solidaridad) IGR Institutional and Governance Review IGV Value added tax IMF International Monetary Fund INA Iniciativa Nacional Anti-Corrupción INADE Instituto Nacional de Desarrollo (National Institute for Development) INEI National Institute for Statistics and Information IRR Internal rates of return ISC Selective tax on consumption or excise tax ISPs Institutos superior pedagógicas JNE Junta Nacional Electoral LAC Latin America and the Caribbean LIBOR London Interbank Offered Rate LPTF Ley de Prudencia y Transparencia Fiscal MAB Multiyear Budget MED Ministry of Education MEF Ministry of Economy and Finance MEM Ministry of Energy and Mines MINDEF Ministry of Defense MINSA Ministry of Health MMM Marco Macroeconómico Multianual (Multiyear Macroeconomic Framework) Mt/day Metric tons per day MTBF Medium-term budgeting framework MTEF Medium-term expenditure framework NA Not applicable NFPS Non-financial public sector NGOs Nongovernmental organizations NPS Nonpersonal services NS Nuevos Soles ODI Investment office OECD Organization for Economic Cooperation and Development OLS Ordinary least squares ONP Oficina de Normalización Provisional (Pensions) OSIPTEL Organismo Superior de Inversión Privada en Telecomunicaciones OSITRAN Organismo Superior de la Inversión en Infraestructura de Transporte de Uso Publico OUs Operational Units PAAG Programa de Administración de Acuerdos de Gestión PAF Poverty Action Fund PARSALUD Programa de Apoyo a la Reforma del Sector Salud (Health Sector Reform Program) PCM Presidencia del Consejo de Ministros (Premier's Office) ABBREVIATIONS AND ACRONYMS xvii PDI Institutional Development Projects PEAP Poverty Eradication Action Plan PEM Public expenditure management PER Public Expenditure Review PESEM Plan Estratégico Sectorial Multianual (Multiyear Strategic Sector Plans) PETS Public Expenditure Tracking Surveys PPP Purchasing Power Parity PPS Probability Proportionate to Size PROGRESA Programa de Educación, Salud y Alimentación PRONAA Programa Nacional de Asistencia Alimentaria (National Food Assistance Program) PRONAMACHCS Proyecto Nacional de Mantenimiento de Cuencas Hidrográficas y Conservación de Suelos (National Project for the Management of Water Sheds) PSPs Protected Social Programs PSRL Programmatic social reform loan PUP Planilla Única de Pagos (Unified Payroll) RENIEC Registro Nacional de Identificación y Estado Civil RER Real exchange rate SEAL South-West Electricity Society SEDAPAL Lima's Water and Sewerage Service SENSICO National Service for Construction Industry Training SEPs Sector Strategy Plans SIAFITO A simplified SIAF SIAF-SP Sistema Integrado de Administración Financiera del Sector Publico (Integrated Financial Management System) SIMSIP Simulations for Social Indicators and Poverty SIS-SMI Integral and Mother and Child Health Insurances SMI Seguro Materno-Infantil (Mother and Child Health Insurance) SPC Sector Publico Consolidado SUNAT Superintendencia Nacional de Administración Tributaria (Superintendency for Tax Administration) SUR Seemingly unrelated regression UEs Unidades Ejecutoras UIT Unidades de Ingreso Tributario USEs Unidades de Servicios Educativos (Educational Service Units) USIS Unidades de Seguro Integral de Salud VAT Value added tax VDL Vaso de Leche CURRENCY EQUIVALENTS: US$1.00 = NS 3.61 (Nuevo Sol exchange rate as of 1 October 2002) GOVERNMENT FISCAL YEAR: January 1­December 31 WEIGHTS AND MEASURES: Metric System 1 EXECUTIVE SUMMARY A Comprehensive Reform Agenda of Public Expenditure Management for Peru A successful economic policy needs two inputs: sound technical design and strong political support to bring it forward. A common mistake is to believe that both aspects are substitutes rather than complementary one to each other. In practice, this leads to wrongly believe that political power may offset poor technical design permanently. Juan Carlos de Pablo "Cómo Fracasa Rotundamente un Ministro de Economía, 1988" Rationale and Organization For the World Bank and the Inter-American Development Bank (IDB), it is a welcome privilege to provide the incoming Administration of a member country with joint and comprehensive diag- noses and policy recommendations in those areas that contribute most to enhancing the client's development prospects.1 One of these key areas is public expenditure management, which is the purpose of this report. As the Toledo Administration enters its second year in office, we hope that the contents of this Public Expenditure Review (PER) will be useful for Peru to deal with some of the formidable challenges the country faces in its development agenda. This chapter presents a synthesis of Peru's public expenditure reform agenda from the Banks' perspective. It is based on, and distills, the thematic chapters that make up this report. Chapters II­IV examine the core functions of public expenditure management: macro fiscal aggregates, resource allocations to strategic sectors, and microeconomic efficiency of public spending. Chapters V­VIII explore selected themes, such as the decentralization of public administration and the social sectors; civil service reform; governance and corruption; and fiscal and environmental mining issues. These chapters are, in turn, supported by 15 topic-specific background papers, including an innovative public expenditure tracking survey on municipal transfers, and numerous and valuable comments provided by Peruvian officials throughout dissemination activities. 1. Several sector studies for Peru were published in the last few years by both institutions in the areas of education, health, private sector participation and infrastructure, institutional governance, financial manage- ment, procurement, and country strategy. However, the last World Bank Public Expenditure Review was produced in 1994. 1 2 WORLD BANK COUNTRY STUDY We do not attempt to provide full coverage of all areas affected by public expenditure. Express- ing the strategic coincidence between the World Bank and the IDB around the idea that poverty reduction is the paramount objective of their assistance efforts, the report deliberately focuses on social sectors spending and leaves aside infrastructure and other sectors, which have been the sub- ject of more specific sector studies in previous years. This report was finalized in June 2002, and discussed with Peruvian officials in mid-August 2002. Policy developments that have taken place after that time are not reflected in it. Main Findings and Recommendations At the onset of the new century, Peru faces substantial development challenges. Significant achieve- ments in terms of growth, poverty reduction, and improvements in social indicators up to 1997 have been partially set back by ensuing persistent slowdowns in economic activity, fiscal indiscipline, politi- cal uncertainty, institutional weaknesses, and poor governance. The momentum for reform, which used to be strong in the early 1990s, ran out of steam in the late 1990s, and genuine efforts by the new Peruvian officials to reinvigorate it are facing strong social pressure for short-term, populist mea- sures.2 Populism might bring palliatives, but is deleterious to the macroeconomic fundamentals of the country and, ultimately, to the standard of living. There is, rather, an urgent need to restore fiscal discipline and build momentum for "second-generation" reforms, including those in public expendi- ture, which would provide longer-term sustainability to the strategy of poverty reduction. The new Administration inherited a fiscal system under severe stress. Since 1999, Peru's fiscal stance has been fragile due to a combination of adverse external shocks, expansive fiscal policies, elections, and other unfavorable domestic developments accompanying the transition to the post- Fujimori legacy. Led by a primary deficit that for the first time in more than a decade has become negative since 1999, the combined fiscal deficit significantly increased from 0.4 percent of GDP in 1997 to 3.2 percent of GDP in 2000, only to be slightly reduced to an estimated 2.5 percent of GDP in 2001. Low, volatile, and procyclical tax revenue combined with a rigid public expenditure facing popular demands complicate efforts to finance an ambitious poverty-reduction agenda, while achieving fiscal adjustment. Despite a multiyear pro-poor budgeting framework, social spending remains underfunded, diverted through multiple leakages, and mistargeted. The concentration of budget formulation authority in Lima coupled with discretional and decentralized budget execution through Transitory Councils of Regional Administration (CTARs) reduces the impact of social expenditure and the efficiency of service delivery. Since his arrival in July 2001, President Toledo has proposed taking action in the areas of macroeconomic stabilization; reopening of country's access to international financial markets; bud- get modernization and State decentralization; social policy; revamping of the armed forces, police and internal security services; and consolidation of Peru's democracy and social stability through consensus-building and transparent and participatory mechanisms. Despite its initial progress, Peru is not likely to meet these goals in the near term. Achieving success in the medium term will rather depend to a great extent on how quickly it restores fiscal discipline, shifts scarce spending to the government's priorities, and ensures an efficient delivery of public services under a very complex and politically driven decentralization process. We summarize the public expenditure management agenda in seven main messages: Restoring fiscal discipline Reorienting the budget toward pro-poor expenditure Improving the efficiency of public expenditure Addressing the promises and risks of decentralization 2. In this report, populism is defined as a conception and practice of economic policy that favors unsustain- able expansionary fiscal and monetary polices as the main engine of GDP growth and employment creation. For a discussion of the concept, see Dornbusch and Edwards (1991). RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 3 Upgrading the civil service Improving governance and reducing corruption Fine-tuning mining taxation and environmental policies. The storyline behind these seven themes is a bit long to tell, but is simple. First, significant primary surpluses are the inescapable premise for sustainable fiscal and debt-management policies. They are also needed to lay the basis for viable countercyclical fiscal and social policies in the future. Second, shifting expenditure to pro-poor priorities is essential for achieving desired social outcomes in the medium term. Budget reorientation requires not only enhancing transparency of budget cuts and reallocations during implementation, but reviewing inertial and in-year seasonal spending. Improving budget protection mechanisms of priority social programs is useful to consolidate an effec- tive countercyclical social policy. Third, making most effective use of scarce resources is essential for a country in the midst of fiscal retrenchment. This implies reducing anti-poor disparities in regional budget allocations; eliminating duplications in social and basic infrastructure projects; minimizing leaks of public funds, particularly in food supplementary programs and at the bottom level of service delivery and, if feasible, considering their gradual downsizing or replacement by cash-transfer pro- grams; and improving targeting of resources, especially outside Lima, for most social programs, espe- cially those addressing the most vulnerable. Fourth, it is essential promptly to resolve fundamental questions of the decentralization process: in particular, what functions must be decentralized? And with what timing and outcomes in mind? An ill-designed decentralization process, especially if devel- oped through brand-new regional governments, has considerable potential for provoking severe fiscal disarray and macroeconomic destabilization. Thus there is a need to proceed very cautiously. Fifth, the government needs to upgrade its human capital in charge of service delivery. The civil service regime is in total disarray and needs to be overhauled. An accurate and dependable public registry of civil servants must be created. The myriad of salary supplements and benefits in cash and in kind now prevailing must be consolidated and replaced by a more rational salary system. Sixth, improving gov- ernance and fighting corruption is essential not only to provide a better and less expensive public ser- vice delivery, but to fight income inequality and facilitate business activity and entrepreneurship in Peru. In general, local governments perform much worse than national agencies in the quality of delivery of services. In addition, there are more corruption and governance problems at the local level. Seventh, fine-tuning mining taxation, mining-canon transfers, and mining-related environmen- tal policies is needed not only to strengthen the global competitiveness of one of the main export sources of Peru through a stable tax regime, but to enhance transparency of transfers to local govern- ments and minimize environmental damage. Conflicts between mining companies and the local pop- ulations will thereby be minimized, thus contributing to improving the quality of mining investment activities, the transparency of intergovernmental relations, and the relationship between foreign investors and local communities. Around the seven organizing themes of this report revolves a set of multiple and detailed diag- noses and policy recommendations, with some suggestions for their prioritization. The reader should, however, keep in mind several limitations. First, there is no magic formula for achieving results. What is described here is essentially based on Banks' best-practice experiences elsewhere in Latin America and the rest of the world, and on their informed views of Peru. Second, proposed sequencing of reforms is itself subject to political constraints. The World Bank and the IDB remain ready to assist the Government of Peru in implementing the PER's agenda. Finally, it is important to add that implementing these reforms can be achieved only through a new contract between the government and citizens, which would elevate standards for government accountability, transparency, and civil society participation and oversight. Restoring Fiscal Discipline Fiscal sustainability is perhaps the most important challenge of the current Administration. Peru needs to strengthen its structural fiscal position, not only because failure to do so jeopardizes the 4 WORLD BANK COUNTRY STUDY macro framework in the medium term, but also because significant contingent liabilities might arise from pensions, the financial sector, and subnational government--registered or nonregistered-- debts, and because sudden slowdowns or stops in external financing due to financial contagion from external (or regional) turbulence are not to be excluded in the coming years. Fiscal deficits are explained by a combination of both structural and cyclical factors. On the structural side, they fea- ture procyclical and volatile patterns, mainly determined by the behavior of tax revenue and public expenditure. Over the last 30 years, Peru's non-financial public sector (NFPS) deficit experienced five cycles: 1970­79, 1980­85, 1986­91, 1992­96, and 1997­01, and averaged a high 5.8 percent of GDP. This pattern resulted from long periods of fiscal expansion, followed by increasingly shorter periods of fiscal adjustment, particularly before the early 1990s. The main component of the deficit was the Central Government imbalance, which represented about three fourths of the average NFPS deficit. Most fiscal variables are procyclical: they have significant positive correlation coefficients with the growth rate, and the cyclical components of the fiscal deficit and GDP growth are almost per- fectly correlated. In addition, fiscal variables are extremely volatile, and this goes beyond their close correlation to GDP growth. Peru's growth rate volatility, measured by the standard deviation, is 33 percent higher than the average for Latin American and Caribbean (LAC) countries, which is itself twice as high as the average in industrial economies. The volatility of fiscal variables, though, is much higher than the volatility of GDP growth: almost six times in the case of tax revenues; five times in the case of public salaries and transfers; more than seven times in the case of capital investments (the critical budget adjustment variable); and 11 times in the case of fiscal deficits. Tax revenue volatility is also explained by continuous changes in tax policy and administration, the hyperinfla- tion of the late 1980s, and elections in the late 1990s. Expenditure volatility is explained by erratic policies affecting the public payroll, transfers, and capital investments. Recent disequilibria are explained by expansive fiscal policies, and correcting them is a priority. Since 1997, trends in CG real revenue and real public expenditure have decreased and increased, respectively. In 2001, the tax ratio fell below 12 percent of GDP, a dismal level by international standards, and among the lowest in Latin America. As a result, primary balances have become nega- tive since 1999, and fiscal deficits have deteriorated. Resulting public debt has increased as a per- centage of GDP from 42.7 percent in 1998 to a high 46.5 percent in 2001, even though significant amounts of privatization proceeds have been used to finance expenditure and repay public debt. Although the public debt is mostly long term and owed to external creditors, the domestic and short-term components of it have grown at a rapid pace in the last several years. In this connection, it must be borne in mind that domestic debt is exposed to currency risk and rapidly increasing debt service, which contributes to tightening of scarce resources in a shrinking budget. Had Peru kept government tax revenue and expenditure constant in real terms during 1998­2001, it would have preserved a sound fiscal deficit below 1 percent of GDP. What needs to be done to achieve fiscal discipline and, hence, macro sustainability? Six main policies are essential. First, and most important, Peru needs a comprehensive tax reform. This reform should aim at improving the tax system's equity, neutrality, simplicity, and revenue-yielding capacity. Reform should mainly be based on a combination of measures to broaden the tax base and enhance tax administration. Proposed measures to broaden the tax base suggest removing key tax exemptions (be it sector- or region-specific, or on interest and capital gains, or the exemption on import duties for Centers for Export, Transportation, Industry, Commercialization and Services [CETICOS]); introducing a presumptive income tax method for the self-employed and a slightly wider three-level income tax rate structure for individuals; and unifying the consumption selective tax for vehicles. By complementing these measures with the introduction of a royalty tax on mining (offset by the elimination of the sector's import duties), and/or eventual modification of workers' participation, and the implementation of a major phased improvement in tax administration, the government could raise additional resources approaching 2 percent of GDP within two to three years. The fact that the set of measures proposed does not modify the present tax structure and RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 5 does not establish new taxes, except in the case of the royalty tax, makes it easier to gain political consensus for its enactment. The other pillar of tax reform is a marked improvement of tax admin- istration by SUNAT, the collecting agency. In general, any improvement results from a combina- tion of enhanced taxpayer services (with upgraded technology support), better auditing systems that focus on tax evaders, and better human resource staffing. The second area where a significant effort is required is public expenditure. Because capital investment continues to be the residual adjusting item in the budget, Peru now has one of the lowest capital expenditure ratios in LAC: less than 2.5 percent of GDP in 2001, and decreasing. Hence, expenditure reform should aim at reverting the present trend toward an increased current expenditure and a decreased capital investment. Excessive current spending could be cut in the area of defense and internal security in the short term (0.6 percent of GDP in 2002), and in wages and pensions in the medium term, supported by social security and civil service reform. If the proposed increase in fiscal revenue materializes, then capital investment, especially in the social sectors, should be raised accord- ingly, thus contributing to filling the present gap that Peru has in social expenditure--1.2 percent of GDP in education and 0.2 percent of GDP in health--with respect to LAC averages. A third area of policy reform is public debt sustainability. The public debt-to-GDP ratio of 46.5 percent in 2001 is too high for a country the public debt of which is highly exposed to cur- rency risk, and which faces potential contingent liabilities in dollars in the financial sector. Lowering such ratio to more sustainable levels is essential to face potential risks in the next decade. Assuming a conservative baseline scenario combining moderate growth and low inflation, a slight increase in external interest rates, decreasing, but still significant mixed concessional­multilateral financing, and a fiscal correction lowering the NFPS deficit from 2.5 percent of GDP in 2001 to -1.4 percent of GDP in 2010, would bring the public debt-to-GDP ratio to a benchmark 36 percent of GDP at the end of the decade. A fourth area of policy reform relates to fiscal rules through the revamping of the Law of Fiscal Prudence and Transparency. Because the Law is not a substitute, but a complementary tool to build credibility of a sound fiscal program, a prior condition for its official revamping requires, in the near term, thorough compliance with the benchmarks of the 2002­03 macroeconomic program, sup- ported by an IMF Stand-By. Moreover, we do not expect a significant countercyclical policy as long as growth remains moderate and fiscal revenue is stagnant. So, revamping of the Law should proceed gradually, focusing on early credibility gains; prior institutional strengthening required by the Law's technical redesign and monitoring; consensus-building exercises led by Peruvian officials; virtual monitoring indicators and reporting tools, including those for the Comptroller's and Congress's oversight; and introducing a set of sanctions for noncompliance with numerical benchmarks and pro- cedures. The design of a set of fiscal rules for incoming subnational governments should impose hard budget constraints on them and justifies upgrading the incoming Law to one of fiscal responsibility. In addition, in the medium term an effective countercyclical fiscal policy could be based on a struc- tural primary balance benchmark. Such policy would require making automatic fiscal savings in booms and dis-savings in downturns, measured by the differential between estimated actual and structural balance. Under the baseline scenario described above, achieving a more sustainable public debt-to-GDP ratio by 2004­10 would entail a structural primary surplus ranging around 1 percent of GDP (roughly equivalent to around a 1 percent of GDO nonstructural deficit) during such years. Future savings would strengthen the role of the Fiscal Stabilization Fund. A fifth area of policy reform concerns contingent fiscal liabilities. Peru must start making trans- parent budgetary provisions to deal with their eventual materialization, which would make projected restoration of fiscal discipline a more difficult task. This report reviews three main sources of contin- gent risk: bank loans, pension payments, and municipal debt. The Banks' broad estimates of the potential fiscal exposure to a systemic banking crisis is 4 percent of GDP, which is considered in the lower bound of comparable--mostly double-digit--world standards. Sound macroeconomic poli- cies, close supervision of the financial sector, and reduced exposure to external financial contagion can minimize this risk. In addition, pension reform has generated two nonbanking debts with the 6 WORLD BANK COUNTRY STUDY private sector, which are actually paid through annual transfers. Official transfers to social security represent 0.6 percent of GDP in 2002, and conservative projections indicate they will increase to about 1.0 percent of GDP per year during 2005­07. Pension reform should minimize this liability. Finally, registered municipal debt is not a threat to macroeconomic stability, but it is significant and explosive, and some local governments will become insolvent and illiquid. These local governments will have severe difficulties assuming additional responsibilities with decentralization due to their financial constraints. Nationwide, total registered debt represents a third of the corresponding national debt (current national liabilities considered as short-term debt). In the short term, identify- ing nonsolvent municipalities and developing a plan for the gradual repayment of their registered debt, especially on pensions, is essential before these districts assume further responsibilities under the forthcoming decentralization process. In addition, the Fiscal Prudence and Transparency Law should put strict limits on new external and domestic debt, and prohibit subnational governments from acquiring contingent debts. Finally, a sixth area concerns public debt financing. Peru must complete its deficit-financing plan by meeting its privatization targets in the nearest term, sustaining its prudent external debt management strategy--including enhanced access to international capital markets--and developing a market for long-term domestic public debt, issued in soles and inflation-indexed. Reorienting the Budget Toward Pro-Poor Expenditure The task of putting the economy on a sustainable, rapid, and poverty-reducing growth path must be to be carried out with the help of modern public expenditure management. Promising steps have been taken by Peru over the past few years. They include (a) creation of an Integrated Financial Management System (SIAF) in 1998; (b) strengthening of accounting institutions and procedures, compatible with world standards, applied to the entire public sector since 1999; (c) development of a website Portal de Transparencia in 2000, which includes defense and national security spending for the first time starting in 2002; (d) passage of the Law of Fiscal Prudence and Transparency in 1999; (e) introduction of performance management contracts for State-owned enterprises under the National Fund for Financing Government Enterprise Activity (FONAFE) in 1999; (f) adoption of multiple medium-term expenditure frameworks starting in 2000, based on three-year Macro- economic Frameworks and five-year Multiyear Budget (MAB) plans, and carried out through a par- ticipatory methodology for multiyear budgeting formulation starting in 2002; and (g) adoption of budgetary provisions for a set of Protected Social Programs (PSPs) as the main countercyclical social policy starting in 2000. The trend toward pro-poor spending and enhanced accountability and transparency seems irreversible, but consolidating these achievements in the medium term means solving five critical institutional weaknesses: (a) excessive budget rigidity, (b) an incomplete SIAF, (c) a disconnection between multiyear and annual budgeting, (d) a fragile institutional framework surrounding its countercyclical social policy, and (e) inadequate and un-user-unfriendly budget monitoring and reporting that affects public perceptions about the real degree of budget transparency. Following a budget tracking test of the status of public expenditure management (PEM) jointly developed by the World Bank and the IMF, Peru scores low in the area of budget formula- tion, average in the area of budget execution, and high in the area of budget reporting other than external. In the first area, the assessment recommends that budget classification be made fully con- sistent with the Government Financial Statistics (GFS) definition of General Government, and the quality of budget projections (especially in function and sector composition) be upgraded if social priorities are to be sustained in the medium term. The major constraint for budget formulation, though, is inertial expenditure. Peru ranks high among countries with the most inflexible budget not only in LAC, but worldwide. In addition to Brazil, India, and a few highly indebted poor countries (HIPCs), which have above 90 percent of their total expenditure fixed for particular functions in 2002, Peru has, respectively, 91 percent and 86 percent of ordinary and total budget resources assigned to inertial commitments. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 7 Budget rigidities originate from different sources, including wage and benefits considered by the Constitution as inalienable rights; permanent contracts hidden as nonpersonal services (no less than 40 percent of the total budget is allocated to such services, according to Peruvian officials' esti- mates); an inertial payroll under artificially prolonged investment projects; earmarked transfers to municipalities and the private sector, such as the Vaso de Leche and Comedores Populares programs; and an earmarked use of the budget "contingency reserve" to fund the March school bonus, and July and December salary supplements (aguinaldos). Peru needs to reverse such a pervasive pattern. PER recommendations in this area include (a) an immediate general hiring freeze under the nonper- sonal services category accompanied by strict supervision, on a case-by-case basis, of short-term renewal of existing temporary contracts, upon their prior revision; (b) an urgent review of present policies regarding pension and wage benefits in a comprehensive civil service reform; (c) streamlin- ing of funds assigned to inefficient social programs--like Vaso de Leche and Comedores Populares; and (d) making fully transparent the use of the contingency reserve, to let it play its true role, which is to fill unpredicted in-year budget needs, thus preventing it from filling recurrent needs. SIAF core subsystems are incomplete. The lack of a budget module, in particular, is a severe handicap. Contrary to similar systems in Latin America, Peru's SIAF is neither a budget formula- tion tool nor a budget assignment tool, but fundamentally a virtual payment system. As a result, budget formulation and in-year modifications remain little known externally, centralized and decided at the Ministry of Economy and Finance's (MEF's) discretion, and only known ex post, once they are monthly registered in SIAF. Hence, a user-friendly budget module integrated in SIAF is an urgently needed policy tool to help assess consistency of budget proposals with macroeconomic conditions in early stages; facilitate monitoring of in-year budget modifications, consistent with inter- and intrasector budget priorities in a participatory and transparent way; and allow Congress' legally mandated oversight of the budget. SIAF also requires a multiyear budget submodule too; that is, one with capacity to track Government of Peru (GOP) compli- ance with multiyear budgeting priorities and facilitate production of mandated biannual reports on compliance with MAB targets. Finally, other critical missing modules are the ones on human resources management, national assets, procurement, the national investment system, contingent liabilities risk, and regional and municipal financial management. The human resource module is essential for civil service reform. The one on national assets should have due registry of state assets. The module on procurement, the starting point for e-government, is critical for anti- corruption efforts, the most vulnerable area of governance. The module on public investment should help prioritize multiyear investment plans. The module on contingent liabilities risk would help register them and make proper budgetary provisions, while the one on municipal management would prepare SIAF support for an accelerated scheme of fiscal decentralization, as future regional and municipal governments require a tailor-made SIAF, the development of which should take many years. The budget-tracking test also found that multiyear projections are disconnected from annual budget priorities and modifications, their quality requires significant improvement, and their moni- toring requires an upgraded and regular reporting system. To such purpose, the creation of a bud- get module and refinement of the existing set of financial and physical indicators in SIAF should facilitate a user-friendly tracking of pro-poor expenditure and an effective results-based control sys- tem in budget execution. Complementary external reporting of budget execution, multiyear bud- geting, and audits should be frequent, timely, and comprehensive. Regarding reporting, the SIAF subsystem to track financial and physical performance indicators, now being developed, requires a thorough conceptual review before completion, because it is a critical tool for an outcome-oriented budget, and support from monitoring targets and performance management contracts. The present subsystem has too many useless and outdated indicators, and the use of the indicators is poor. The development of new SIAF modules requires detailed prioritization and permanent train- ing. Peruvian officials have good reason to consider creating a university degree in the management of SIAF systems. Besides, associated systems could be developed outside the SIAF, like a fiscal 8 WORLD BANK COUNTRY STUDY cadastre, or eventually be connected to a broader e-government strategy (like e-procurement or a geo-referenced expenditure map at the district level, for targeting purposes). These shortcomings have not prevented an important shift in the budget toward pro-poor expenditure in Peru. In spite of the recession and resulting fiscal adjustment, the share of social expenditure to GDP grew from 3.9 percent of GDP in 1990 to 7.0 in 1999, and this share has remained steadfast until 2002. The fact that the share of social expenditure to GDP has not declined is a positive outcome in a fiscally retrenched economy where total expenditure decreased from 19 percent of GDP to 17.8 percent of GDP during 1999­2002 (about one seventh in dollar terms). Likewise, significant inter- and intrasectoral pro-poor spending reallocation began to take place between 2000 and 2002. The budget for defense and national security has decreased from 2.9 to 2.1 percent of GDP; the budget for education, health, and social assistance has increased from 5.3 to 5.6 percent of GDP; and the Seguro Social de Salud (Health Insurance, ESSALUD) expenditure (nontargeted public health insurance administration) has decreased from 1.5 to 1.3 percent of GDP. In addition, the share of poverty-reduction programs (consisting of both programs with universal coverage spending and programs based on targeted spending) in social expenditure increased from 57 percent in 2000 to 60 percent in 2002. Since 2000, poverty bud- geting has also included an innovative effort to protect key social programs (PSPs). This approach has consisted of assigning global budget allocations to nonsalary current and capital spending grouped in 11 programs and distributed among eight social and basic infrastructure sectors. In 2001, budget resources equivalent to 12.7 percent of the ordinary budget and 1.9 percent of GDP were protected and executed. In 2002, the government decided to change its PSPs policy from merely budget protection when a downturn occurs to sustained spending floors needed to reach millennium development goals. This was accompanied by improvement of their focus and monitoring mechanisms, and full transparency in their execution. The ongoing restructuring of PSPs includes expenditure allocation shifts in broad sector functions--education, health and sanitation, and social assistance; a reduced number (from 11 to 6) of now core PSPs, grouped under three sectors: education (pre-primary, pri- mary, and secondary), health (collective and individual) and social assistance, a reduced budget as the amount of aggregate earmarked expenditure to this revised set of PSP represents 1.2 percent of GDP in 2002, down from the previous 1.9 percent of GDP in 2001 (this reduction of 0.7 percent of GDP makes room for increased budget flexibility); a set of monitoring indicators for specific pro- jects grouped under the selected PSPs; and, in a few projects, definition of some performance man- agement contracts to improve their efficiency (for example, the Integral and Mother and Child Maternal Insurances, SIS-SMI). To introduce further flexibility, the PSP budget also includes escape clauses to make provisions for budget underexecution or a major budgetary shortfall at the sector function level; and internal reallocations from underperforming to overperforming programs or pro- jects, but maintaining global spending floors at the program and project levels. To institutionalize and make its countercyclical social policy more transparent, Peru should enhance its de facto "vir- tual" protected budget policy following the example of Uganda. An enhanced mechanism would add two major features: the development of a permanent institutional framework in charge of PSPs policy, and the introduction of oversight capacity by subnational governments and civil society orga- nizations, and of transparent monitoring and evaluation indicators, into PSPs implementation. Improving the Efficiency of Public Services Delivery The efficiency of Peruvian public expenditure is below the Latin American average for health, but above average for education, and in relation to world averages, the results are similar. However, these results should be viewed with caution. Indeed, Peru's level of efficiency in reaching health outcomes is below both the Latin American and world averages. This is particularly true in malnutrition­height, under-five mortality, and life expectancy. The range of values found suggests that Peru could and should do much better in health outcomes. And although outcome indicators for Peru's net primary enrollment are on a par with other Latin American countries, and net secondary enrollment is much RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 9 above the LAC average, its educational shortcomings are too severe in terms of the low quality of education services and poor teacher qualifications, especially in rural areas. Two major sources of inefficiencies are leakages of public resources and poor targeting of social programs, which ultimately contribute to poor governance and a corruption-prone management. A highly decentralized budget execution such as Peru's raises the possibility of significant leakages in the flow of resources, thus reducing the extent to which GOP expenditure is able to produce better outcomes. By applying an innovative Public Expenditure Tracking Survey (PETS), it is possible to show that there are substantial leakages in the resources that municipalities first receive under the Vaso de Leche program and later transfer to the ultimate beneficiaries through several intermediaries. The actual amount received by final beneficiaries appears much lower than the initial amount sent from the Central Government (CG): on average only 29 cents of each dol- lar initially transferred by the CG actually reach the intended beneficiaries! Surprisingly, the survey indicates that the leak is much higher at the bottom levels (Vaso de Leche committees and house- holds) than at the top levels (CG and municipalities leaks), which demonstrates not only signifi- cant improvements in the official channeling of resources by the SIAF/MEF, but challenges the predominant view that local private organizations are more accountable in managing public resources than official ones. Leakages in municipality-level transfers are compounded by the gen- eralized lack of audit controls, poor transparency, and volatility. Finally, leaks most affect the poorest urban and provincial municipalities, and their level appears similar among districts of dif- ferent sizes and distances to the province. Peruvian officials should pursue a twin strategy. In the short term, any effort leading to improving supervision, accountability, and transparency of the flow of resources should be pur- sued. This includes a regular and user-friendly publication of detailed transfers to municipalities by SIAF, capacity training at the local level in SIAF reporting, proper registration of beneficiaries, and strengthened budget and auditing procedures by the Comptroller's office. A mini-Stabilization Fund for canon minero transfers could contribute to minimizing their annual volatility. In the medium term, however, authorities should consider replacing these programs with a voluntarily adopted cash-transfer ("grants") program, following the successful model of the Programa de Educación, Salud y Alimentación (PROGRESA) program in Mexico. Under such a scheme, grants would be provided to mothers not only for food supplements, but also for basic education and preventive health, which, following a proper registration of beneficiaries, should also improve targeting of resources on beneficiaries at risk. Poor targeting is another important shortcoming of the multiplicity of social programs in Peru. The general level of spending on targeted extreme poverty-reduction programs increased from 1.1 percent to 1.7 percent of GDP between 1993 and 1998, but then remained constant until 2002, which appears adequate due to the fiscal situation and program inefficiencies, and despite pressures to increase it. Poor targeting, however, deviates part of these outlays toward the non-poor. Targeting rates in social programs vary tremendously. Past findings indicate that with the exception of the Fondo Nacional de Compensación y Desarrollo Social (FONCODES) and the Programa Nacional de Asistencia Alimentaria (National Food Assistance Program, PRONAA), which have had about a 50 percent success rate, most targeted social programs have had rates reaching below 40 percent of the intended poorest. More recent findings on the food and nutrition programs con- firm this picture. Targeting among the three main national programs--Desayunos Escolares, Vaso de Leche, and Comedores Populares, also reflects prevailing weaknesses: during 1998­2000, Desayunos Escolares had the highest success rate at 70 percent, followed by Vaso de Leche and Comedores Popu- lares, which had lower rates of 64 and 54 percent, respectively. As a program becomes better tar- geted, the differential between its rural and urban success rates decreases: Desayunos Escolares had a lower range (77 percent compared to 56 percent) than Vaso de Leche (77 percent compared to 50 percent) and Comedores Populares (72 percent compared to 37 percent). Finally, programs of temporary employment, like A Trabajar Urbano, have not started on the right foot, since they offer too high a wage rate to attract workers in the poorest deciles. 10 WORLD BANK COUNTRY STUDY Peruvian officials should make intensive use of excellent targeting tools, especially the poverty map, already in their hands. The myriad of multiple social programs in Peru should be ranked, and those with substantial mistargeting should be redesigned, merged, or closed. Fiscal savings gener- ated from such an exercise should be used to increase coverage of more effective programs. Target- ing criteria should be made more explicit and focused on extremely poor populations. In the case of A Trabajar Urbano, the immediate solution is to decrease the wage rate paid. A similar policy in A Trabajar Rural is not possible, since the individual income earned by a rural worker is often shared with other members of his or her community, and so the self-targeting mechanism does not work in the same way as it does with urban workers. In both cases, though, authorities should be aware of the countercyclical and short-term nature of these employment-generation programs: once the economy recovers its dynamism with high rates of growth, their budget resources should be phased out. Addressing the Promises and Risks of Decentralization In the past decade, Peru has made irreversible progress toward redefining the role of the State, but this agenda is unfinished. In the early-to mid-1990s, the size and scope of the public sector were reduced; the nature of government shifted from ownership and intervention to market regulation and provision of social services. However, in the mid-to-late-1990s the State reform process lost steam, and key tasks were not even initiated. Prominent among pending reforms set aside by the authoritarian leadership of the 1990s was the decentralization of the State. The current government has decided to reinvigorate the State modernization and decentraliza- tion agenda. This decision must be seen as part of a broader, positive trend toward democracy in the region. However, State decentralization is pregnant with both potentialities and risks. If appropriately carried out, it can, indeed, increase both the efficiency of the delivery of government services and the responsiveness of government to citizens' needs. Nonetheless, decentralization also entails serious macroeconomic risks. First, CG deficits can worsen if governments are unable to reduce expenditures or increase revenues to make room to accommodate the fiscal cost of additional functions and responsibilities transferred to subnational governments. The government may find itself decentraliz- ing both revenues and expenditures, but unable to reduce existing levels of CG spending in those functions that have been decentralized. Second, attempts to eliminate rising deficits through higher taxes can fail or take longer than expected to materialize. Third, decentralization can also hamper a government's ability to respond to economic shocks because it reduces central control over aggregate public sector revenues and expenditures. Fourth, subnational government debt policies and indebt- edness behavior can undermine overall fiscal discipline. Subnational governments may end up bor- rowing excessively in the expectation that the CG will bail them out. In sum, decentralization has the potential to be a destabilizing factor for already deteriorated Peruvian public finances and, ultimately, for the country's overall macroeconomic framework. The Latin American experience suggests, moreover, that decentralization should not be carried out with- out a coherent overall strategy and without prior technical and institutional preparation at the sub- national levels of government. In the case of Peru, the fact that municipal governments are heavily dependent on central transfers justifies caution in the transfer of responsibilities to them. Taking into account that the degree of satisfaction in the provision of public services by local governments appears to be low, gradualness in the transfer of functions and a sizeable effort to strengthen the administrative capacity of subnational governments prior to the transfer of responsibilities seem to be necessary conditions of any reasonable decentralization strategy. Putting the process of decentralization on a sound path is a critical public policy challenge for the Toledo Administration. While Peru's initial efforts toward modernization and decentralization offer substantial promise and, for the most part, are headed in the right direction, the latent risk that fiscal discipline may turn out to be a casualty of the decentralization process will continue to be high until a clear framework for intergovernmental public finances is finally enacted and enforced. The guiding criterion for such a framework has to be the establishment of hard budget constraints for RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 11 the expenditure and--external and domestic--indebtedness decisions of subnational governments. The immediate agenda includes issues of paramount importance, such as the alternative options for intergovernmental fiscal relations, the budget and indebtedness constraints on subnational govern- ments, the possible functions and competencies to be decentralized, the characteristics and funding sources of the future Regional Compensation Fund and Intergovernmental Fund for Decentraliza- tion, and any possible changes to the transfers scheme for municipalities. It is also imperative that the hard budget constraints for regional and local governments here advocated be an integral part of the future revamped Law of Fiscal Prudence and Transparency. Finally, a complete diagnosis of the financial status of municipalities, including its pension liabilities and other registered and non- registered debt, and a plan for both strengthening municipal tax collection and restructuring local government debt, are essential before these are assigned new expenditure powers. Upgrading the Civil Service Civil service reform is the missing link of the outcome- and poverty-reducing-oriented moderniza- tion of the state in Peru, and is in desperate need of a major overhaul. The past decade witnessed a severe deterioration of Peru's civil service. Hyperinflation of the late 1980s and the beginning of the 1990s sharply reduced the real earnings of civil servants. A poorly conceived downsizing of the CG work force in the early 1990s, which, in the end, amounted to a very modest reduction in the over- all public sector, led, nonetheless, to the departure of the most capable civil servants. There is an inherited disregard for the provisions of the existing civil service law and its associated regulations. Civil servants are employed under a variety of legal regimens with a wide dispersion in salaries for similar tasks. The existing formal salary structure has not been revised in over a decade, so base salaries are increased through resorting to a multiplicity of supplements and benefits in cash and in kind. With the proclamation 10 years ago of a ban on hiring into permanent positions, institutions have resorted to hiring on a contractual basis, a process that is not subject to any required proce- dures for recruitment and selection, and that has severe fiscal implications. All these factors have contributed to the existing legacy of a work force that is poorly prepared, poorly trained, poorly supervised, and hence of very low productivity. Consequently, there is an urgent need for a thorough reform to reestablish some semblance of control over the wage bill and to promote the creation of a more efficient government. The current civil service law has major weaknesses that argue for its replacement by a new framework designed to promote a higher level of efficiency in the delivery of government services. Any initiative designed to improve the performance of government must include a major over- haul in the institutional framework that governs public sector employment and the management of its human resources. Four actions are particularly relevant. First, the creation of a Civil Service Reform Commission should precede the establishment of an autonomous Directorate to administer the civil service. Second, developing an SIAF module on human resource management would greatly simplify registry of public servants and would be instrumental in the task of simplifying the remunera- tion package from the myriad of present benefits to a cash-equivalent amount. Third, introducing pilot performance contracts in the health sector seems appropriate, given its recent successful experi- ence with Seguro Materno-Infantil (Mother and Child Health Insurance, SMI), which makes it a nat- ural candidate for later mainstreaming of such practice in the rest of the public sector. Finally, a new law and reglamento should place emphasis on merit and specify determinants for compensation. The education sector also faces serious problems with its civil service. There is general dissatisfac- tion with the quality of instruction and with the allocation of faculty resources among schools. The Ministry of Education is taking steps to improve the quality of instruction in the country's schools of pedagogy. However, the profession continues to attract individuals whose test scores rank at the bot- tom of the distribution of students entering institutions of higher learning. Improvement in the qual- ity of instruction is likely to require a number of reforms. While the current level of salaries appears to be more than adequate to attract a surplus of candidates with minimal qualifications, the level and earnings trajectory over time are inadequate to attract more promising individuals. In addition, both 12 WORLD BANK COUNTRY STUDY the Ley Magisterial and the Ley General de Educación are too rigid for most forms of decentralization and should be revised. Reforms must focus on four aspects. On the institutional side, the current institutional morass in the education sector (financially depending on regional bodies, and normatively depending on sector ministries) should be unified under the decentralization design. On the recruiting policy, new entrants to the profession should be required to qualify by passing a rigorous examination. Success- ful candidates would be hired at a higher salary not much different from the current entry wage, but should be able to look forward to salary increases similar to those accruing to persons with similar skills and abilities in the private sector as long as performance is evaluated and meets defined stan- dards. On the hiring levels, implementation of a new remuneration policy would be facilitated by moving to a system of contracting of teaching personnel in place of the currently generalized system of permanent tenure upon appointment. This would also have obvious fiscal savings. On the legal side, the new Ley del Profesorado and the Reglamento de la Ley del Profesorado should be based on teaching performance and effective attendance, rather than on a range of factors such as seniority, number of training courses attended, papers presented at professional meetings, and so forth, none of which necessarily have any impact on the quality of teaching. Improving Governance and Reducing Corruption Peru's picture in matters of governance is mixed. In the World Bank's scoring system for governance indicators, Peru's scores in government effectiveness and regulatory quality have dropped significantly, even as the country moved from an authoritarian regime to one with a more vocal and active opposi- tion. The score in corruption has remained approaching the LAC average. Peru's corruption appears particularly high in procurement contracts of public works. To offset such a gloomy picture, Peru's scores in voice and accountability, political stability, and rule of law have improved, thus reflecting the political opening that is taking place. Significant challenges remain in these areas, however. Reducing corruption and improving governance requires a multipronged strategy that addresses the various forms of corruption--from state capture to administrative corruption, enhances political accountability and transparency, promotes a competitive private sector, strengthens institutional restraints, improves public sector management, and embraces civil society participation. It is more than just a matter of law enforcement. The evidence in this report also identified the support of public officials for civil society oversight and merit-based civil service reform as anticorruption mech- anisms. Public officials also favor strengthening public administration through simpler administrative procedures. They are also in favor of the introduction of a system of checks and balances and the oversight of public sector activities by the civil society. The government first needs to build credibility with respect to its intentions to fight corruption. In the survey, there is a significant discrepancy between users and public officials in their perception of the government's commitment to fighting corruption. More than half of users, versus only 20 percent of public officials, reported that the government is not committed at all. This lack of trust is keenly felt among low-income users of public services. Conversely, the fact that a high pro- portion of public officials report preparedness in the public sector to fight corruption is an encour- aging sign. The government needs to make the fight against corruption a top priority. The church and the media should be potential contributors in the fight. As a complement, the government needs to continue enhancing transparency, especially through SIAF. Transparency is synonymous with openness. Introducing it requires that the government pro- vide the public--civil society organizations, the media, or anyone else who is interested--sufficient information about the budget and project activities, to serve as an effective check on abuses by gov- ernment and public officials. Enacting a Freedom of Information Law that requires governments to provide information to the public unless there is a valid reason for the information to remain secret (such as a threat to national security) would be welcome. The government needs to be accountable and reduce the potential gains for "captor" firms and politicians alike. Government officials should consider issuing an Ethics Code Law for public RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 13 service, with clear prohibitions of conflicts of interest, and effective enforcement of its regulations on both political and civil service positions. This is an essential corruption-deterrence tool. They should also demand transparency of the financing of political parties and campaigns, making clear the links, both explicit and implicit, between politicians and the interests that support them. Man- dated and publicized detailed reports of the finances of all political organizations, and identification of contributors and beneficiaries, are tools to reduce the sources of state capture. Another possible measure would include banning the use of State resources for incumbents' political campaigns, limiting the amount that can be spent on political campaigns, providing public funding, and pro- hibiting certain types of entities from contributing to political campaigns. The government needs to promote a sound business environment by examining and simplify- ing the number of regulations and inspections. Clear rules defining when a regulation is appropri- ate, and sound analysis of the impact of regulations on firms accompanied by a constant dialogue with the business community, should be required to make regulatory reform sustainable. Finally, a high-level steering committee, supported by a professional secretariat, can help bring together representatives of government bodies, both central and local, to develop the specific action plans to implement the governance strategy. By explicitly including representatives from outside government, such committees can further build credibility while mobilizing an important ally for reducing corruption. Fine-Tuning Mining Taxation and Environmental Policies Given its size in the economy and social network, the mining sector is essential for public finance, structural reform, and governance in Peru. The mining industry generates more than 45 percent of export earnings and contributes to 5 to 7 percent of GDP. It employs over one-quarter million people, and its tax exemptions account for about 0.4 percent of GDP, which is nonnegligible. Key structural and transparency issues have negatively affected the competitiveness and attractiveness of mining activities with foreign investors and local society, particularly with the local communities since the late 1990s. The mining reform brought about the effective opening of the sector to direct foreign investment and the shift of the role of the State from owner­operator to lessor­regulator. This was accomplished by reforming the legal and institutional framework, privatizing the mining state-owned operations, and initiating an environmental management program. While the first three reforms, legal, institutional, and privatization, have been widely consid- ered successful, the sectoral approach of the environmental reform has proven to be controversial and is considered in need of adjustment. The perception of inadequate environmental control and of widespread conflict between mining companies and local populations is endangering the contri- bution and stability of a key sector of Peru's economy. In addition, the highly centralized and nontransparent nature of transfers from the CG has created an issue of credibility with the local populations. Although the Mining Law establishes that 40 percent of the income tax generated by the mining operations--the canon minero--should go back to the municipalities (20 percent) and the regions (20 percent), and the report's findings show that its overall redistribution has been pro- gressive, there is a widespread perception that the local communities are not benefiting from it. Furthermore, the environmental information that the Ministry of Energy and Mines (MEM) has received from the mining companies has not been adequately released to the public, resulting in a feeling of mistrust against mining activities. Loss of competitiveness due to the taxation system is also an issue. To the extent that it affects rates of return on investment, the effective tax rate in a given country is a major factor influencing the decision of foreign investors about where to invest their capital. The report shows empirical evidence that supports the view that the current Peruvian mining tax system is among the more globally competitive systems in the world, and provides an appropriate balance between investor needs and government needs. However, there is room for improvement. Addressing major issues in taxation, environmental regulations, and the transparency of canon minero transfers is very important for preserving international competitiveness. Main tax policy 14 WORLD BANK COUNTRY STUDY changes proposed are (a) introducing a 3-percent royalty tax combined with the elimination of import duties and modifications to worker participation; (b) adjusting by a premium increase of 5 percent all future stability agreements; (c) extending the carry-loss forward time limit beyond five years; (d) enhancing the system to register tax deduction of mine contributions to local com- munities and infrastructure; and (e) publishing canon minero monthly transfers to municipalities in SIAF, while providing training at the local level. Agreements on roles and responsibilities in the environmental provisions are needed, accompanied by a new regulatory framework of environmen- tal management, a new designated national authority with no conflicting interests and integrated with regional and local authorities; and with local participation--nongovernmental organizations (NGOs) included--in local environmental monitoring and consultation mechanisms, to allow them to protect community interests. Given the volatility of fiscal transfers, the creation of small mining stabilization funds to offset temporary declines in fiscal revenue, and simulating the suc- cessful experience of the Ancash Model, are not to be excluded. Assembling All the Pieces of the Agenda Peru's policymakers are well aware of the difficult fiscal situation and the need for discipline. The agenda described above is large and requires prioritization and sequencing, possibly starting with tax reform and a modern and institutionalized pro-poor budgeting, followed by efficiency improvements in social programs and civil service reform--both critical inputs to consolidate expenditure management reform--and then completed by the decentralization, governance, and anticorruption programs, and by fine-tuned environment and taxation policies in the mining sector. Short-lived aggregate demand boosts, like those advocated by populism, will only delay needed reforms. Consistent choices have to be made and some efforts will likely have to be post- poned. Most of the PER's agenda, however, is feasible provided that there is not only commitment of the Executive, but also support in the Legislature. Peruvian officials are well aware of this. The 2001 budget presentation to Congress broadly covered each of the three aspects of public expendi- ture performance. It singled out the paramount need to restore fiscal discipline to preserve price stability, to readjust the pattern of expenditures toward priority social areas, and to improve service delivery with efficiency gains. This report argues that for fiscal policy to become a powerful tool for Peru's development, designing an outcome-managed, poverty-oriented, and consensus-built bud- geting strategy is necessary. Both the World Bank and the IDB are ready to continue assisting authorities in marshalling such efforts. II PERU'S FISCAL CHALLENGES AND VULNERABILITIES At the onset of the new century, Peru faces daunting fiscal challenges due to a combination of adverse external shocks, expansive fiscal policies, elections, and other unfavorable domestic developments accom- panying the transition to the post-Fujimori legacy. During the early-to-mid-1990s, volatile and pro- cyclical expenditure accompanying good times made the country's response to negative shocks more difficult to manage and vulnerable to a prolonged slowdown, like the ones that have occurred since the end of the decade. Since 1998, the fiscal stance has deteriorated, with successive negative primary deficits, in part due to a resilient decline in tax revenue, which reached a decade-low 11.9 percent of GDP in 2001. The ensuing combined fiscal deficit significantly increased from 0.4 percent of GDP in 1997 to 3.2 percent of GDP in 2000, only to be slightly reduced to an estimated 2.5 percent of GDP in 2001. Rigid and expansionary expenditure has also undermined fiscal stability. Ending privatization proceeds and high and dollarized public debt offers limited room for financing deficits in the years to come. The prolonged economic slowdown and political uncertainty resulting from external and domestic factors, including populism, also put additional stress on the fiscal balances. The fiscal rules enshrined in the Fiscal Prudence and Transparency Law of 1999 have not been fully met. In any case, they are no substitute for sound policy, and require thorough revamping before they become a tool for consolidating credibility. This chapter analyzes Peru's main fiscal challenges and vulnerabilities with respect to restoring fiscal discipline, including tax policy measures and new rules to be considered in a revamped Law. Long-Run Fiscal Trends Long-run trends in Peruvian fiscal policy can be broadly split into two contrasting periods: before and after the 1990s.1 The average non-financial public sector (NFPS) fiscal deficit was a staggering 5.4 percent of GDP during 1970­90, and a much more moderate 2.4 percent of GDP in the 1990s. Such outcomes resulted from long periods of fiscal expansion, followed by increasingly 1. This section mainly draws from Oliva (2002). 15 16 WORLD BANK COUNTRY STUDY FIGURE 2.1: NFPS FISCAL AND PRIMARY BALANCES (% of GDP) 8% 4% 0% -4% -8% -12% 1970 1975 1980 1985 1990 1995 2000 Fiscal balance Primary balance Source: MEF, IMF and World Bank staff estimates. shorter periods of fiscal adjustment, particularly before the early 1990s. The main component of the deficit was the Central Government (CG) imbalance, which represented about three fourths of the average NFPS deficit. The first period reflected high deficits resulting from persistent macro- economic disorders, negative primary surpluses with either short-lived or half-hearted efforts at fiscal adjustment. The lower deficits of the second period came from continuous primary surpluses that became the norm rather than the exception (Figure 2.1). The impact of a severely deteriorated economy during 1970­90 left a long-lasting memory among Peruvians about the paramount need to preserve stabilization. As GDP growth deteriorated after 1975, the tax ratio ranged from 8 to 16 percent of GDP, with the most striking decline occurring throughout the late 1980s, reflecting the hyperinflation crisis (Figure 2.2). As a result, the average fiscal deficit increased to 6.0 percent of GDP and stayed high until the end of the decade. Such high deficits were financed by resorting to either domestic financing, leading to hyperinflation levels in the late 1980s, or to huge increases in external indebtedness, with the ratio of public external debt to GDP increasing from about 14 percent in the mid-1970s to over 56 per- cent in the late 1980s (Table 2.1). The governments' heavy reliance on debt led to an explosive ris- ing share of interest payments of total CG expenditure from 5 percent in 1970 to 40 percent in FIGURE 2.2: TAX REVENUE COMPOSITION, 1970­2001 Percent of GDP 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 1970 1975 1980 1985 1990 1995 2000 Income Imports ISC IGV Other Source: MEF, IMF and World Bank staff estimates. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 17 TABLE 2.1: KEY MACROECONOMIC VARIABLES 1969­89 Fiscal Balance Average Inflation Public External Period (% GDP) Rate (%) GDP Growth (%) Debt/GDP (%) 1969­75 -3.9 10.6 5.5 14.1 1976­79 -6.8 48.6 0.7 32.3 1980­85 -6.0 94.2 0.5 41.5 1986­89 -5.9 445.8 -1.5 56.2 Source: World Bank staff estimates based on MEF and IMF data. 1990. Only after 1990 was the share of interest payments dramatically reduced, following the restoration of normal relations between Peru and the international financial community, prompt payment of arrears to multilateral institutions, and debt restructuring agreements with Paris Club lenders, and other bilateral and private creditors.2 By 2001, the share of interest payments of total expenditure had been reduced to 12 percent (Table SA.4). Long-run changes in the tax revenue structure are relevant. The three most important tax categories3 were the income tax, the general sales tax (VAT), and the excise tax (or selective tax on consumption, ISC) (Figure 2.2). Four subperiods can be identified according to which of these three categories dominates. In the first period (1970­75), the income tax matched the VAT; in the second period (1975­85), it was the VAT that played the leading role; in the third period (1985­93) the ISC prevailed; and in the fourth period (1994­2000), the VAT recovered its major significance. The declining trend of the income tax--which fell from a dominant share of 35 percent in 1970 to 23 percent in 2000--and the exceptional major role that the excise tax played during 1984­92-- were the outcome of the excessive number of income tax and VAT deductions and exemptions, and the serious deficiencies in tax administration. Long-run trends in public expenditure are determined by two structural factors: the role of the public sector in the economy, and output performance. Three subperiods can be distin- guished. In the first one (1970­78), total expenditure exhibited an increasing trend, rising from 15 percent to about 20 percent of GDP, as explained by the steady expansion of the role of the State in the economy during the Velasco government and a strong economic expansion.4 The average ratio of public investment to total investment stood at a high 27.0 percent5 (Paredes and Sachs 1991). The second subperiod (1979­91) featured the opposite trend--that is, CG expen- diture declined as a share of GDP, which reflected the severe deterioration of tax collection and the economy. The third subperiod (1991­2001) reverted to an accelerated expenditure expan- sion: from the deep trough of 14.5 percent of GDP in 1991, CG expenditure increased to an average of 17.6 percent of GDP during 1999­2001, close to the high levels reached during the era of strong state intervention, pushed by current spending rather than by capital investment (Table SA.5). 2. Relations between Peru and the international financial community deteriorated throughout the 1980s. According to Larrain and Sachs (1991), "Peru first partially suspended its debt service in 1983, but the magni- tude of the suspension was small. In 1984, renegotiations failed and the country started to accumulate more significant arrears. These increased dramatically after 1985, when President Alan Garcia announced the govern- ment's unilateral decision to limit the service of foreign debt to 10 percent of export revenues, a restriction that would apply in principle to all foreign creditors." 3. Taxes on foreign trade transactions were excluded from the comparison. 4. Peru's economy grew an average annual rate of 5.5 percent between 1969 and 1975 (Table 2.1). 5. Compare with the value of the same indicator in a recent period (1999­2001), which was just 18 percent (Table SA.2). The difference was the result, of course, of the privatization in the 1990s of about 80 percent of the large number of State enterprises that the Velasco government bequeathed to its successors. 18 WORLD BANK COUNTRY STUDY TABLE 2.2: ELASTICITY TO GDP, 1970­2000 Variable Estimated Coefficient Total tax revenues 0.998 Income tax 0.972 Import tax (tariffs) 0.972 Total VAT 1.007 Internal VAT 0.999 Imports VAT 1.024 Total ISC 1.010 Gasoline ISC 1.032 Other ISC 0.992 Total non-financial expend. 0.962 Current 0.965 Wage and salaries 0.926 Goods and services 0.995 Transfers 1.001 Capital 0.944 Investments 0.971 Source: MEF, IMF and World Bank staff estimates. Fiscal variables were highly procyclical before and during the 1990s. Regression analysis over the entire period shows that all tax and expenditure categories exhibit an almost unitary elasticity with respect to GDP (Table 2.2); that is, they were highly procyclical.6 Most fiscal variables also have significant correlation coefficients (above 0.5) with the growth rate (Arias and others 2000), and the cyclical components of the primary balance and GDP growth are almost perfectly corre- lated (Figure 2.3). Fiscal variables were extremely volatile in both periods. The volatility of fiscal variables is associ- ated to Peru's growth rate volatility (measured by the standard deviation). In the 1990s, it was 33 percent higher than its average for Latin America and the Caribbean (LAC), itself twice as high as in industrial economies (De Ferranti and others 2000). Unstable international export prices and terms of trade, financial shocks, and limited openness in the financial system explain GDP growth volatility. The volatility of fiscal variables, however, appears much higher than the one for GDP growth during 1970­1996: almost six times in the case of tax revenues, five times in the case of pub- lic salaries and transfers, more than seven times in the case of capital investments, and 11 times in the case of fiscal deficits (Figure 2.4). In the 1990s, tax revenue volatility is explained by continuous changes to tax policy and administration, the external environment,7 the legacy of hyperinflation of the late 1980s, and elections in the late 1990s. For its part, public expenditure volatility is explained by erratic policies affecting capital investment, debt payments, the public payroll, and transfers: It reflects the continuous use of capital outlays as the typical procyclical adjustment variable (elasticity 6. Regressions are estimated with a simple ordinary least squares (OLS) model: nominal revenue and expenditure from each category was the dependent variable and nominal GDP was the explanatory variable. Variables for 1970­2000 were in logarithms and adjusted by a first-order autocorrelation coefficient to take into account serial autocorrelation. R-square exceeded 0.99 and the t-statistics indicated that the coefficients were different from zero at a significance level of 99 percent. 7. Tax revenue volatility originates in the external sector in a twofold sense: while during 1970­90 the share of export taxes on total revenue was significant, which made tax collection vulnerable to Peruvian com- modity-export prices in the world markets, in the 1990s, the Peruvian economy became sensitive to capital inflows surging from the opening of the capital account. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 19 FIGURE 2.3: PROCYCLICAL FISCAL STANCE (billions of 1994 soles) 13 8 3 -2 -7 -12 1970 1975 1980 1985 1990 1995 2000 cyclical component of primary balance cyclical component of output Source: World Bank staff estimates. of almost 2), the high variability of debt interest flows (elasticity of 7.1), and the swinging patterns of wage policy (it had an elasticity of -1.7) (Table 2.3). The biggest policy mistake (and lesson) of the early 1990s is not to have taken the opportunity to save the transitory additional revenue generated with the economic boom. Since 1993, strong capital inflows dominated Peru's economic performance and brought about a strong economic boom. During 1993­97, GDP grew at an average annual rate of 7.0 percent. Instead of developing a countercyclical fiscal policy, the expansive fiscal stance adopted magnified the boom and increased the risks associated with it. Hence, the overall fiscal deficit increased to 3.0 percent of GDP in 1995, prompting the current account deficit in the balance of payments to widen from 5.7 percent in 1994 to 7.7 percent of GDP in 1995, which made the country's external position more vulnerable. In 1996, the government had no choice but to undertake a severe fiscal adjustment: the fiscal deficit FIGURE 2.4: VOLATILITY IN FISCAL VARIABLES 1.0 NFPS 0.8 CG Deficit Volatility 0.6 Y K VAT MT Absolute0.4 IN W T 0.2 3 5 7 9 11 13 15 Relative Volatility* TR - tax revenue VAT - value added YT - income tax W - wages expenditure MT - import tax IN - interest expenditure KX - capital expenditure *Relative is ratio of standard deviations wrt GDP Source: World Bank staff estimates 20 WORLD BANK COUNTRY STUDY TABLE 2.3: ELASTICITIES AND GROWTH VOLATILITIES FOR PERU (1990­2000)a Relative to % of Total Absolute Growth Revenue/ Volatility Volatilityb Expenditure Elasticity t-ratio GDP 5.0 1.0 100.0 Central Government Revenue 8.3 1.7 100.0 1.6 15.3 Expenditure 15.7 3.2 100.0 1.0 5.5 Current Revenue 7.9 1.6 98.8 1.5 14.5 Tax Revenue 8.0 1.6 87.8 1.4 11.6 Non-Financial Expenditure 16.8 3.4 80.6 1.7 14.5 Current Expenditure 14.8 3.0 61.9 1.7 13.0 Capital Expenditure 32.0 6.4 18.7 1.9 3.6 Interest Expenditure 27.2 5.5 19.4 7.1 3.9 Wages 16.4 3.3 17.0 -1.7 -5.3 Subsidies and Current Transfers 10.2 2.1 31.8 1.2 11.7 Fiscal Balance 13.6 2.7 Primary Balance 20.3 4.1 a. Volatilities are computed as standard deviations over a 10-year period in annual percentage rates of growth. b. Volatility in the rate of growth of the fiscal variable divided by GDP growth volatility. Source: World Bank staff estimates based on MEF and IMF data. was reduced to 1.1 percent of GDP in 1996, and to an average of 0.3 percent of GDP during 1997­98, a situation of virtual balance. The Current Fiscal Disequilibria, 1999­2001 In the late 1990s, fiscal difficulties originated from adverse external developments, but were later accompanied by expansionary domestic policies. Financial contagion hit the Peruvian economy in mid-1998 as foreign banks drastically curtailed the availability of credit lines to the Peruvian banks in the wake of the Russian crisis of August 1998. The turmoil also affected the stock market, resulting in portfolio outflows and a sharp drop in stock prices. The reduction of foreign credit lines to the Peruvian banks was especially deleterious, as the banks had relied heavily on them to finance the 1993­97 credit boom. The ensuing credit crunch was a major blow for an economy that was barely recovering from the effects of another shock: the El Niño phenomenon, which struck in the first half of the year. The capital outflows episode thus came to a close, and the well- known vicious circle of credit reduction, decline in output and demand, lower sales, rising diffi- culties for firms in servicing their debts, deterioration of bank loan portfolios, and further credit reduction, set in. In 1998, the economy entered into a recession, with private investment signifi- cantly falling, both in real terms and as a percentage of GDP. Declining economic activity led to lower tax revenue and public expenditure, which also deepened the economic contraction. As a result, the overall fiscal deficit increased to an average of 2.9 percent of GDP during 1999­2001. Three domestic reasons also contributed to this outcome. On the expenditure front, the political business cycle deteriorated the fiscal accounts, as in the run-up to the 2000 presidential elections, the CG expanded current noninterest expenditures from 11.8 percent of GDP in 1998 to 12.7 percent of GDP in 2000. On the revenue side, the strong decline in tax revenue from an average 13.8 percent of GDP during 1995­97 to 12.3 percent during 1999­2001 was mainly associated not with the cycle, because the deviation of the cyclically adjusted balance with respect to the observed primary balance was small, which would suggest that the effect of cyclical factors was RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 21 FIGURE 2.5: FISCAL TRENDS OVER TIME 4.5 20 4.0 18 3.5 3.0 16 Soles 2.5 GDP 14 2.0 % 1994 1.5 12 1.0 Blns 10 0.5 0.0 8 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 CG Deficit, LHS GG Deficit, LHS Real CG revenue Real CG expenditure Source: World Bank staff estimates based on MEF and IMF small, but rather with a structural shift in tax revenue behavior caused by frequent changes to the tax regime.8 Real revenue decreased and real expenditure increased. Fiscal vulnerability becomes more apparent when tax revenue and CG expenditure are considered in real terms. In 1999, tax rev- enue fell to pre-1995 levels (below 12.5 percent), while expenditure rose considerably (about 19 percent of GDP). As a result, the CG (and combined General Government) primary balance has been negative since 1999, and fiscal deficits have surged (Figure 2.5). An increased--mainly short and domestic--public debt (as a percentage of GDP from 42.7 percent in 1998 to about 46.5 percent in 2001) and privatization receipts financed the deficit. Barely 8 percent of public debt is short term, and prospects for improving public debt ratios in the medium term are favor- able. However, public debt is exposed to currency risk. Debt service, which already increased from 3 percent of GDP in 1998 to a projected 3.8 percent of GDP in 2002, will contribute to tightening scarce resources in a shrinking budget. Had Peru kept government tax and expendi- ture constant in real terms during 1998­2001, it would have preserved a fiscal deficit below 1 percent of GDP. Peru did not fully comply with its Fiscal Prudence and Transparency Law of 1999. The Law established a set of procedural and numerical fiscal rules (Box 2.1), and established five numerical targets for (a) the overall deficit; (b) the difference between the inflation rate and the annual growth rate of consolidated General Government's non-financial expenditures; (c) the total debt increase of the non-financial public sector; (d) the General Government's non-financial expendi- tures during the first seven months of an election year9 as a percentage of budgeted non-financial expenditures of the General Government for the entire year; and (e) the overall deficit in the first half of each election year as a percentage of the budgeted deficit for the year. Government compli- ance with numerical targets was mixed, but the main target referring to the overall deficit was not 8. In a simple framework where tax revenues are a function of income, the cyclical effect can be seen as a movement along the revenue curve in response to a lower level of income. The deterioration of the tax sys- tem, by contrast, should be seen as a downward shift of the entire tax revenue curve, entailing lower revenues for any given level of income than in the situation prior to the deterioration. Our quantitative estimates indi- cate that the effect of the business cycle on the primary deficit during 1999­2001 was a worsening of just 0.1 to 0.2 percent of GDP. 9. Due to well-known political events in Peru, both 2000 and 2001 turned out to be election years. 22 WORLD BANK COUNTRY STUDY BOX 2.1: THE FISCAL PRUDENCE AND TRANSPARENCY LAW The Fiscal Prudence and Transparency Law was approved in December 1999 (Law No. 27245) and its regulations issued in April 2000 (Supreme Decree No. 039-2000-EF). It was suspended in December 2001 (Law No. 27577). The Fiscal Stabilization Fund was slightly amended through the 2002 Budget Law (Law No. 27573), which has three components: Numerical rules include (a) a zero fiscal deficit in general (over the cycle), but below 2 percent of GDP in 2000, 1.5 percent in 2001, and 1 percent in 2002; (b) an increase of the non-financial general government spending below 2 percent in real terms; (c) electoral expenditure ceilings on the non-financial expenditure below 60 per- cent in the first seven months, and a public sector deficit below 50 percent in the first semester. These rules have two exceptions allowing for a fiscal deficit below 2 percent of GDP: one for emergency situations and another for recessions. Procedural rules require the introduction of transparency criteria established in the IMF Code of Fiscal transparency from 2002, including (a) a Transparency Decree of Acceso Ciudadano, detailing periodic information to be pub- lished in the Portal de Transparencia Económica; (b) a multiyear budgeting framework; (c) a mandated pre-election fiscal report; (d) an annual review of contingent fiscal risks and tax expenditure to be published in the annual bud- get; and (e) a quarterly report on budget expansions and their implication over the macroeconomic framework. The Fiscal Stabilization Fund was created from 75 percent of privatization proceeds, 50 percent of concessions, and excess revenue (above 0.3 percent its average of the past the years). Its resources should be mainly dedicated to poverty-alleviation programs, and activated when current revenue is at least 0.3 percent of GDP below its aver- age in the last three years. met, and the Law was suspended in December 2001. Although a case could be made that this was due to revenue-growth slowdown-related factors beyond the control of Peruvian officials, the fact is that President Fujimori's late Administration weakened the fiscal balances with new tax exemp- tions and electoral spending. In addition, the increase of non-financial expenditures in 2000 and 2001 built on their already high level reached in 1999.10 The present level of public debt needs to be reduced to more sustainable levels. Three years in a row of fiscal disequilibrium has led to an increase in public indebtedness and a reduction of long- term public savings. The accumulated deficit during 1999­2001 amounted to 8.8 percent of GDP. Its financing came from domestic credit (4.6 percent), external credit (2.0 percent), and depletion of privatization proceeds (2.2 percent). The stock of total public debt increased from an already high 42.7 percent of GDP in 1998 to 45.8 percent of GDP at the end of 2001. While the net pub- lic external debt ratio increased only slightly,11 the domestic public debt ratio rose from 6.3 percent of GDP to 9.8 percent of GDP. (All domestic debt is in U.S. dollars.) External debt represents around 70 percent of total public debt and has moderate risk, because virtually all of it is long term. However, its service rapidly increased from 3 to 4 percent of GDP between 1998 and 2002, amounting to 23 percent of total expenditure in 2002, which is a considerable burden on tight fis- cal balances. Two additional vulnerabilities are the foreign-currency risk exposure of the domestic debt and the rollover risk to changes in foreign investor confidence. Peru reopened its access to international markets with a US$500 million, 10-year bond placement combined with a retirement of Brady Bonds amounting to US$930 million in 2002, which lowered its debt servicing costs, although at the expense of a shorter maturity. Given the low tax revenue situation and expenditure pressures, Peru requires not only a fiscal adjustment, but also a prudent debt management strategy. 10. In terms of GDP, combined general government non-financial expenditures in 1999 were the highest for the entire decade of the 1990s (18.9 percent of GDP). 11. From 34.7 percent of GDP in March of 1999 to 35.1 percent in December 2001. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 23 The Macroeconomic Consistent Expenditure Envelope, 2002­03 The medium-term outlook for the economy is favorable. Since the focus of this chapter is on analyz- ing expenditure trends, the starting point for such analysis is the determination of an aggregate level of spending that is consistent with medium-term macroeconomic objectives (that is, a sustainable fiscal deficit and public debt ratio) (Table 2.4). A base-case scenario assumes that real growth of about 4 percent a year and low inflation inside the 2 to 3 percent range a year would be achieved on the basis of the recovery of the mining sector, increased productivity, and increased private sector investment and capital inflows attracted by sound macroeconomic policies, political stability, and the privatization program. It also assumes the combination of an increased export capacity, continuous implementation of structural reforms, a managed floating exchange rate regime, and a successful fis- cal adjustment, supported by the IMF program. Mining production should benefit from the first full-year operation of the Antamina project. Structural change should lead to completion of the pri- vatization program and improved regulations of private sector activities, developing public sector modernization and decentralization, strengthening financial supervision, and developing pension and social sector reform policies. In sum, under a low inflation and moderate growth scenario, pre- sent fiscal imbalances are expected to be gradually corrected provided the Government of Peru (GOP) adopts an urgent tax reform; cuts unnecessary outlays; protects expenditure shifted toward social priority needs; minimizes the cost of potentially significant contingent liabilities, like those from pensions, and maintains a prudent debt management strategy. This macroeconomic scenario is entirely consistent in the short term with the stand-by arrange- ment agreed with the IMF. However, significant constraints remain, including limited policy credi- bility following a long period of political instability and a poor GOP track record in complying with fiscal rules; policy uncertainty due to continuous changes in the tax regime and regulations that affect prospects for attracting foreign investment; fiscal adjustment delays due to Congress inherit- ing populist-oriented and spending-prone fiscal decentralization measures; weak domestic consensus about further pension reform and the GOP's privatization plans; and a poor rule of law due to an inefficient judicial system and poorly managed local government institutions. Recommendations: Authorities should: Redouble efforts to build credibility before the Law is revamped Offset several measures introduced by the transition government that had negative implica- tions for the fiscal balances with revenue resulting from tax reform Set rules to prevent decentralization from increasing transfers to regional and local govern- ments that are not offset by similar improvements in revenue collection Stabilize the regulatory framework for private sector activities and strengthen the rule of law to diminish uncertainty that erodes the authority of rules and institutions that are essential to promote domestic and attract foreign investment. Peru's most pressing fiscal challenge is to obtain primary surpluses, because negative primary fiscal balances are unsustainable. Recommendations: In the short term, under the base case, the mid-term revised adjustment program aims at a gradual reduction in the combined fiscal deficit to 2.3 percent in 2002 and 1.9 percent of GDP in 2003. Most of the adjustment would be achieved through cuts in public spending, while reorienting and protecting priority social programs; and tax reform, combining tax policy measures, changes to the exemptions regime, and an improved tax administration (Box 2.2; Table 2.5). This scenario also assumes that the privatization of energy enterprises and the granting of a few concessions will materialize during 2002­03, averaging a minimum 24 WORLD BANK COUNTRY STUDY 2­3 3.0 19.2 16.4 16.7 2003e 2.5- 1.7 2.2 1.9 0.3 2.4 9.5 0.3 14.1 12.4 14.1 1.9- 1.9- 0.5 4.5 40.9 3.5­4.0 13,205 3.7 2.5 3.1 2002e 18.8 15.7 16.5 2.3- 2.0 2.0 1.7 0.3 2.4 13.9 11.9 14.4 10.0 0.4- 2.3- 2.3- 0.6 3.8 44.9 200,083 0.2 2.0 3.2 18.3 15.1 16.5 2.1- 2.2 2.1 1.9 0.2 2.3 14.1 11.9 14.7 10.3 0.6- 2.8- 2.5- 0.6 3.9 46.5 2001p 189,532 3.1 3.8 4.0 2000 20.1 16.2 17.3 3.0- 2.7 2.2 1.9 0.3 2.3 14.7 12.0 14.9 10.4 0.5- 2.7- 3.2- 0.8 3.8 45.9 186,756 A, Annex in 0.9 3.5 4.8 1999 21.7 16.8 18.0 3.7- 2.0 2.1 2.0 0.1 2.8 9.7 14.5 12.5 14.6 1.0- 3.1- 3.1- 0.8 3.9 payments. 48.0 174,719 1­6 CG-IES Tables of and and 1998 0.5- 7.3 4.5 23.5 19.0 17.3 6.4- 1.9 1.9 1.8 0.1 2.0 9.8 0.8 15.7 13.8 13.7 1.1- 0.8- 0.5 3.4 42.7 166,514 data, Bank enterprises 6.7 8.5 4.4 1997 23.8 19.4 18.2 5.2- 1.5 1.8 1.7 0.1 3.1 8.3 0.9 15.6 14.1 13.1 0.8- 0.1- 0.9 3.3 31.9 Central public 157,274 on of based assets on 2.5 4.3 1996 11.5 22.5 18.2 16.4 6.1- 1.8 2.4 2.3 0.1 2.9 8.6 1.0 15.7 13.9 13.9 1.4- 1.0- 3.9 4.1 45.2 SA.5, tax 136,929 of Table net is with NDICATORSI GDP) GDP) revenue of of (% (% debt debt NS) differences Tax CONOMIC Investment GDP) estimates. E of service/GDP Balance minor data. staff EY (millions average) Fixed (% investment Savings external domestic K debt/GDP has Indicators on on debt Bank CPI Growth investment revenue Transfers 2.4: (% Account Government Revenue Expenditure Balance Balance sector sect. table Domestic National revenue Balance Debt Nominal preliminary World GDP Macro Interest Interest Privatization This on ABLE Government Private Tax Nontax Interest Current Consumption Public Public GDP T Real Inflation Key Gross Gross Current Central Current Current Primary Overall NFPS NFPS Memo: Note: based Source: RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 25 BOX 2.2: TAX REGIME AND PROPOSED CHANGES Income tax on individuals. This tax is levied on four classes of income: real estate, other capital assets, self- employed labor, and wage labor. The maximum marginal tax rate for net income greater than 54 Unidades de Ingreso Tributario (1 UIT = about US$828) is 30 percent. The marginal rate for net income between 7 UITs and 54 UITs is 15 percent. Labor income below 7 UITs is exempted. Dividends are not taxed. The interest income from savings with the national financial system and capital gains from the sale of securities and real estate are exempted. During 1997­2000, revenue from the income tax on individuals was, on average, 1.0 percent of GDP (Kopits and others 2000; Schenone 2001). Over 75 percent of this revenue comes from wage-earners' income and only about 10 percent from the self-employed. This is tantamount to the existence of a high degree of hori- zontal inequity. The situation is worsened by underreporting of income from independent professionals' ser- vices and rents from real estate. The considerable number of exemptions also blunts the intended progressivity of this tax category, which detracts from the system's claim to vertical equity. The exemption on interest income from bank deposits is 0.1 percent of GDP (Kopits and others 2000; and Schenone 2001). Recommendations: (a) Elimination of exemptions on all forms of capital gains including the sale of securities and real estate, on interest income, and other income from financial assets; (b) introduction of a more gradual structure of marginal tax rates (for example, with three rates, 10, 20, and 30 percent) instead of the current two-rate structure; and (c) introduction of presumptive-income methods for self-employed taxpayers. Income tax on corporate income. The taxable base for the corporate income tax is the worldwide income of enterprises legally regarded as domiciled in Peru. The rate is 20 percent. Exemptions are numerous and include interest income from savings with the national financial system; interest income from registered instru- ments issued by corporations; capital gains from the sale of securities; "native" agriculture and/or transformation or processing of products described as "native," and/or alternative crops in the Amazon region; enterprises operating in the Centers for Export, Transformation, Industry, Commercialization, and Services (CETICOS); and income from agricultural enterprises with sales under 50 UIT. Rate reductions and other tax benefits are agricul- ture (a rate of 15 percent); income of enterprises engaged in agriculture, aquaculture, fisheries, tourism, and manufacturing activities associated with the processing, transformation, and marketing of commodities from the aforementioned activities, the processing of forestry products and forest extraction (a rate of 5 percent if the business is located in the departments of Loreto and Madre de Dios, and in some districts of the Ucayali depart- ment, or 10 percent if it is located anywhere else in the Amazon region); accelerated depreciation is allowed in mining, hydrocarbons, tourism, agriculture and agribusiness; and tax discount for reinvestment in mining, educa- tion, and manufacturing industry, and for enterprises located in the Amazon region. The effects of this type of tax benefits are well known: the neutrality of the tax system is impaired and the revenue-yielding capacity is dam- aged. Collection from the tax on corporate income fell from 2.2 percent of GDP in the first quarter of 1999 to 1.4 percent in the fourth quarter of 2001. The fiscal cost of exemptions on deposits of corporations in the bank- ing system is 0.2 percent of GDP. Moreover, in the case of foreign firms, a part of the benefits granted ends up as a transfer to the treasuries of foreign governments where the firms in question have their headquarters. In many cases, these types of tax incentives are not instrumental in attaining the desired investment objectives and lack transparency because it is very difficult to quantify the benefits given to a particular sector or region of a country. Recommendations: (a) An across-the board elimination of sector- and region-specific exemptions is desirable; (b) alternatively, if the authorities want to subsidize specific productive sectors and/or regions of the country, explicit transfers should be included in the national budget; and (c) exemptions on all forms of capital gains from portfolio investments must be eliminated. Import tariffs. Tariffs are levied on imports of goods on the basis of their CIF value. There are two basic tariff levels: a 12 percent general tariff (applicable to 5,780 customs items); and a special tariff of 20 percent (applica- ble to 752 items). In addition, there are two types of surcharges: one on agricultural products (surcharge rate of 5 percent) and one on meat products (surcharge rate of 10 percent). Only one exemption raises serious (continued) 26 WORLD BANK COUNTRY STUDY BOX 2.2: TAX REGIME AND PROPOSED CHANGES (CONTINUED) concern: the exemption from tariffs for foreign goods introduced into the CETICOS that do not pay import duties and can be reexported. This exemption impairs the horizontal equity of the tax system in that it provides for unequal treatment of enterprises that are otherwise equal just because they are located in different places. Recommendation: The CETICOS regime must be abolished. The Extraordinary Solidarity Tax. This is a tax on firms' payroll. The taxable base is the wage compensa- tion to workers classified in tax code categories 4 and 5, excluding national holidays and Christmas bonuses. It has long been recognized as a burden on the competitiveness of enterprises and is being phased out. The cur- rent government reduced its rate from 5 percent to 2 percent. The government is seeking congressional approval for its complete elimination. Recommendation: The government's decision to eliminate the Extraordinary Solidarity Tax is adequate. The General Sales Tax. The general sales tax (VAT) has a rate of 18 percent and is levied on the domestic sales of movable goods, the provision or use of services in Peru, construction contracts, the first sale of struc- tures by their builders, and imports. The following goods are exempted: fish and seafood, cochinilla, bull semen, whole raw milk, potatoes and other root vegetables, dried vegetables, other vegetables and fruits, cereal seeds, rice, raw coffee, tea, cocoa, unprocessed tobacco, wools and pelts, raw cotton, nonmonetary gold, vehicles for diplomatic use, books for educational institutions, and the first sale of structures valued at less than 35 UITs. The following services are exempted: financial services, public passenger transportation (except for air trans- port), international cargo transport, live shows declared of cultural value by the National Cultural Institute, sale of beverages and food at snack bars and universities, construction and repair of foreign ships, interest on trans- ferable securities issued by enterprises pursuant to the Securities Market Law, life insurance polices, postal ser- vices intended to complete such services originating abroad, and exports of goods and services. Apart from the aforementioned list of exemptions, there are a few special regimes for the VAT. First, the Hydrocarbons Act exempts the import of goods and inputs required in the exploration phase of oil contracts, and the sale of petrol, natural gas, and derivatives, for internal consumption, in the departments of Loreto, Madre de Dios, and Ucayali. The Amazon Investment Promotion Act of 1999 exempts from the VAT the sale of all goods and ser- vices for internal consumption and construction contracts in the Amazon region. Furthermore, it allows a fiscal credit of 25 percent or 50 percent on the VAT for sales to the rest of the country. The Agriculture Act exempts from the VAT agricultural producers the annual sales of which are below 50 UIT. The first sale of homes below US$35,000 is also exempted. There is, finally, a special rate for rice (5 percent). In analyzing the fiscal impact of the VAT exemptions we must distinguish between the general exemptions and the special regimes. Defenders of the general exemptions contend that, by selecting items of popular consumption, they introduce a desirable element of progressivity in this tax and hence a positive redistributive effect. It has been demonstrated, however, that, to the extent that they include items that are also consumed by the rich, exemp- tions to the VAT are, at best, powerless to generate positive redistributive effects and, at worst, deepen the negative redistributive effects that are an inherent feature of this tax. For their part, the special regimes damage the VAT's revenue-yielding capacity. In the case of the Amazon-region exemptions, they not only deprive the national treasury of revenue in the Amazon region itself, but open the door for VAT evasion on the part of tax- payers from regions other than the Amazon region through the simple expedient of doing their shopping in the latter region. Finally, the exemption for agricultural producers the annual sales of which are below 50 UIT breaks its universality and neutrality. Recommendations: (a) VAT exemptions for the Amazon region must be eliminated; (b) VAT exemptions granted by the Hydrocarbons Law must be eliminated; and (c) the 50-UIT threshold for agricultural producers must be eliminated. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 27 TABLE 2.5: RECOMMENDATIONS FOR TAX REFORM Recommended Measure Revenue Impact (% of GDP) Income Tax Elimination of exemptions on interest income and capital gains 0.27 Introduction of a three-level structure of rates for individuals 0.10 Elimination of sector- and region-specific exemptions 0.60 Value Added Tax Elimination of VAT exemptions for Amazon region 0.25 Elimination of exemptions granted by Hydrocarbons Act 0.10 Elimination of the 50-UIT threshold for agriculture 0.30 Selective Tax on Consumption Elimination of ISC exemptions for the Amazon region1 ­ Elimination of exemption on fuels used by power industry2 ­ Unification of ISC for vehicles at 25-percent rate 0.10 Import Tariffs Elimination of the CETICOS regime 0.13 Special Regime for Mining Introduce a system of mining royalties3 0.10 Extraordinary Solidarity Tax (EST) Elimination of the EST -0.20 TOTAL IMPACT 1.65 1. The elimination of both the VAT and ISC exemptions for the Amazon region is estimated to have a total revenue impact of 0.10 percent of GDP. This is the reason why no revenue impact figure is entered in this row of the table. 2. The revenue impact of this measure as a percentage of GDP is very small. This is the reason no figure is entered in this row. 3. The revenue impact would depend on the characteristics of the royalties regime (see Chapter VIII). Source: Schenone (2002) and World Bank staff estimates. of 0.6 percent of GDP a year. Under these assumptions, continuous implementation of a managed floating exchange rate and an open trade regime, the external current account deficit is projected to be about 2.5 in 2003, and the international reserve position is pro- jected to gradually decline to comfortable levels (7.5 months of imports by 2005). In the medium term, our estimates of the potential output and structural fiscal balance under the base-case scenario indicate that, in order to stabilize debt indicators, the present deficit in the non-financial public sector primary balance should rapidly shift to a surplus of 0.9 percent of GDP by 2004. The consolidation of this effort would require NFPS primary surpluses to reach between 1.0 and 1.3 percent of GDP during the second half of the decade (Annex B). Public debt ratios should decrease to more sustainable levels in the medium term and debt cur- rency risk should be reduced. A public debt ratio of 46.5 percent of GDP is too high, especially for a country the balance sheet of which contains a significant currency risk, which exposes it to exter- nal shocks. Such risk originates from the fact that most of Peru's public debt (external and domes- tic) is in dollars, while the government's earnings are mostly in soles (trade taxes are the only revenue that is dollar-linked). Existing dollar-denominated debt also actually puts authorities in the 28 WORLD BANK COUNTRY STUDY position of being against their own commitment to let their own currency float freely, thus press- ing them to continue intervening to avoid its depreciation. Recommendations: Under a baseline scenario, a debt sustainability analysis estimates the target public debt-to- GDP ratio to be reached in 2010 is 36.2 (Annex A, Table A.4). Such ratio would require a fiscal correlation in the structural primary surplus of 1.6 percent of GDP (cyclically adjusted) between 2001 and 2010 (Annex B, Table B.1). Such estimates take into account the debt exchange of the Peruvian Brady bonds with private creditors in 2002 and its slightly short- ened maturities. If sustained fiscal efforts are promptly developed, the public-debt-to-GDP ratio is projected to improve substantially from about 46.5 percent in 2001 to 39.3 percent in 2005, and then decrease until its benchmark ratio in 2010. The resulting path toward sus- tainability in the public external debt service, however, is a longer one. As a percentage to GDP, debt service should increase from about 3.8 to 6.0 percent between 2001 and 2005, and then decrease to 4.6 percent of GDP in 2010. Under the extreme case of a modified baseline scenario with shocks derived from the impact of a sudden financing stop (a real exchange rate depreciation and the materialization of con- tingent liabilities contained in short-term debt), the target public-debt-to-GDP ratio to be reached in 2010 would increase to 40.8 percent. Such ratio would require a fiscal correction in the structural primary surplus of 2.4 percent of GDP (cyclically adjusted) between 2001 and 2010 (Annex A, Table A.4). On public domestic debt. Developing a market for domestic public debt issued in soles would contribute to reducing its currency risk and bringing more credibility to the infla- tion-targeting policy. Despite being small,12 an inflation-indexed public debt in local cur- rency, hopefully placed beyond the short term, would also contribute to reducing inflationary surprises. On public external debt. The initial focus of the new Peruvian officials has been on reduc- ing rollover costs, widening maturities, and refinancing debt with Paris Club creditors. Increased coordination between the MEF and the Central Bank of Peru (BCRP), and strengthened management of the maturity and currency composition of external debt, would facilitate debt servicing. Approval of a General Debt Law in 2002 is expected to deal with debt management issues. Peru should start making provisions for contingent liabilities. These could materialize in the near future, thus making more difficult projected recovery of budget discipline. There are at least three main sources of contingent risk that could become a government liability: nonperforming bank loans, pension payments, and municipal debt (not counting those that could originate from Banco Agrario or public enterprises loan guarantees). The World Bank's broad estimates of the potential fiscal exposure of a banking crisis is 4 per- cent of GDP, which is considered in the lower bound of comparable--mostly double-digit-- world standards. Moreover, official estimates put nonperforming loans at about 10 percent of the Bank's portfolio in 2001, but such a ratio appears to be stabilizing in 2002. The pension system is a potentially significant fiscal liability. Pension reform has generated two nonbanking debts with the private sector that are actually paid through annual trans- 12. Internal debt was very small during most of the 1990s, because the GOP associated it to the decades of large deficits and inflation. Bonds were reissued in 1999 to support a bank merger, the purchase of the bad asset portfolio of a bank taken over by the government, and two bond-for-loan swap programs. In 2000 and 2001, new bond issues were made under the Financial System Consolidation Program. The BCRP has strict limits on purchasing government debt. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 29 fers: one is to cover the pension of public employees retiring under the public pension sys- tem, and the other are bonos de reconocimiento of those who already migrated to the private pension system. Official transfers to social security represent 0.6 percent of GDP in 2002, and conservative projections indicate they might increase to about 0.6 to 1.0 percent of GDP per year during 2005­07 (IMF data, and Arias and others 2000). Registered municipal debt is not a threat to macroeconomic stability, but this is not an indi- cation that the problem is not very serious at the local level. Indeed, it is significant and explosive, because some local governments are likely to become nonsolvent and illiquid, like those of Arequipa and La Victoria (Fernandez 2002). These two local governments will have severe difficulties assuming additional responsibilities with decentralization due to financial constraints. Nationwide, total registered debt represents a third of the correspond- ing national debt (current national liabilities considered as short-term debt). Nonregistered municipal debt adds up to such a burden. Recommendations: Peru should start making budgetary provisions for contingent liabilities, so a first step would be to make a comprehensive assessment of all potential sources and quantify them. Financial contingent liabilities should be monitored through an adequate prudential super- vision. Those eventually arising from a sudden stop of external financing due to contagion, can only be avoided with sound macroeconomic management (Annex A). Reducing contingent liabilities in pensions will likely require more structural solutions. Identifying nonsolvent municipalities and their eventually nonregistered short-term debt, and developing a plan for the gradual repayment of their registered debt are essential steps before districts assume further responsibilities under the forthcoming decentralization process. The revamped Fiscal Prudence and Transparency Law should continue preventing govern- ment officials from resorting to the Central Bank for financing of its own expenditure, and should add specific prohibition to incoming subnational governments (or ceilings) on acquiring future debts (Chapter V). There are also additional external and domestic risks to this scenario that would make Peru's growth potential uncertain, increase credit risk, reduce financing needed for a sustained social policy, or weaken the institutional leadership on social reform and fiscal decentralization. On the external front, a prolonged slowdown of the global economy, adverse regional developments, or unexpected shocks in the international capital markets could lead to a significant decline in external demand for Peruvian exports, or in mineral prices, and/or to private capital outflows. This could reduce export dynamism and growth, destabilize the domestic currency, put strong pressure on both the public and corporate sectors, which are heavily indebted in foreign currency, and impose additional stress on the banking sector. On the domestic front, the new Administration and Congress could fail to agree on the implementation of the necessary policy measures needed for an effective fiscal adjustment and a financially sound State reform and decentralization process. Lack of consensus could also delay or revert the comprehensive legislative reform of the tax system to be introduced in 2002. Authorities should be supported in their efforts to: Build a contingent financing plan based on (a) the prompt materialization of external donor financing to the incoming Poverty-Reduction Strategy, (b) increased and enhanced access to international capital markets by further placement of sovereign bonds, and (c) the com- pletion of the privatization process. Develop consensus-building mechanisms on fiscal issues, aimed at reaching a common, shared, medium-term vision with the private sector and civil society. Such consensus should be adequately reflected in the revamping of the Law. 30 WORLD BANK COUNTRY STUDY FIGURE 2.6: LAC VAT EFFECTIVENESS 80 Dominican Republic 70 Chile 60 50 Ecuador Peru 40 Venezuela Brazil Effectiveness 30 Colombia 20 Argentina 10 5 10 15 20 25 Rate Source: Schenone (2002) and IMF (2000). The Fiscal Adjustment Effort Fiscal adjustment requires a certain combination of expenditure-reduction-cum-revenue-increase. In 2002, Peru's required revenue increase will not likely result from a rapid recovery of the economy, because the cyclical component of the fiscal balance is very small (Figure 2.13). Besides, economic recovery has had limited impact on tax collection due to the prolonged debate that took place in Congress on the desagio tributario (a quasi tax amnesty) and the granting of additional income tax exemptions. This points out the need for a combination of expenditure cuts with structural remedies on the revenue side. A long-lasting fiscal adjustment requires measures to (a) strengthen the tax regime and administration; and (b) revert the trend toward increased current outlays, while curbing inefficient expenditure. Strengthening the Tax Regime Continuous changes in the tax regime are weakening the successful tax reform of the 1990s. In 1993, Peru carried out a well-designed reform of its tax system: the reform aimed at simplifying the tax sys- tem, broadening the tax base, and increasing revenue yield (IMF 2000). The number of tax cate- gories was reduced from almost 70 to 4 basic categories: the general sales tax (VAT), the income tax on individuals and corporations, the selective tax on consumption (an excise tax), and tariffs on imported goods. The income tax brackets were reduced from 5 to 3, and business income was taxed at a flat rate of 30 percent. Dividend income was excluded from the tax base and the net-wealth tax was eliminated. The minimum income tax, based on gross assets, was maintained. The base for the value added and excise taxes was broadened and most exemptions eliminated. Later, in 1994, the base for the income and value added taxes were broadened further.13 However, since 1996 a long series of exemptions and other preferential regime measures was introduced, which severely eroded tax collection capacity. Despite having a VAT collection effectiveness ratio that is average in LAC terms--about 50 percent (Figure 2.6)--the main tax breaks represent a minimum of 2 percent of GDP of direct tax losses (Table SA.21). Strengthening the tax regime requires a set of specific actions 13. In the institutional domain, the Superintendency for Tax Administration (SUNAT) was created with effective administrative and budgetary autonomy. Personnel policies were reformed to get highly qualified staff and endow them with up-to-date computational equipment. The tax code was modified to enhance SUNAT's ability to audit taxpayers, increase fines and penalties, simplify procedures, and strengthen tax courts. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 31 FIGURE 2.7: CENTRAL GOVERNMENT'S EXPENDITURE COMPOSITION 80% 60% expenditures 40% total of 20% % 0% 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 Interest Current expenditures Capital expenditures Source: MEF, IMF and World Bank estimates. to improve the equity, neutrality, simplicity, and revenue-yielding capacity (Box 2.2). A specific set of ensuing tax policy recommendations could generate an estimated potential revenue-generating capacity of 1.65 percent of GDP (Table 2.5).14 Reverting Expenditure Trends Current expenditure is increasing explosively. Since the mid-1990s, Peru's current expenditure has been increasing at an accelerated pace. It rose from about 50 percent of total CG expenditure in 1990 to almost 73 percent in the early 2000s (85 percent for the CG in 2001). These very high lev- els by Peruvian standards were characteristic of strong State intervention in the 1970s (Figure 2.7). One could argue that the levels of current and capital expenditure are broadly in line with LAC regional averages (Table SA.29); but what is worrisome is its increasing trend and present composi- tion. In 2001, as a share of total expenditure, Peru's 39 percent wage bill was second highest in LAC, and as a share of GDP, Peru's 6.6 percent figure was higher than the 6.3 percent LAC aver- age. These ratios suggest an increasing sense of overstaffing and do not include additional "services" hired as temporary contracts under goods and services categories (Chapter VI). It also suggests a high level of budget rigidity (Chapter III). Such a rising trend would not have been possible unless capital expenditure and interest payments had not decreased their share below 15 percent. Contrary to the early 1990s, interest payments have been below the region's average. For its part, Peru's capi- tal expenditure (3 percent of GDP) is among the lowest in LAC (Figure 2.8) and keeps decreasing (see Table SA.5). Indeed the budget priority in the 1990s was not to increase the capital stock. Recommendation: The proposed comprehensive reform of the State should examine proposals to combine a gradual reduction of current expenditure, especially wages and salaries, while developing a more qualified and efficient labor force under a civil service reform (Chapter V). Public investment remains low and only recently has given priority to the social sectors. In the 1990s, capital expenditure was more volatile and procyclical than all the other expenditures cate- 14. Both proposals, the tax regime changes and the set of tax policy recommendations, are based on Schenone (2002). 32 WORLD BANK COUNTRY STUDY FIGURE 2.8: LAC: CAPITAL EXPENDITURE (1998) (% of GDP) Guyana Nicaragua Belize Dominican Republic Jamaica Suriname Chile Ecuador Honduras Venezuela Barbados Bolivia Guatemala Panama Paraquay Peru El Salvador Mexico Uruguay Bahamas Costa Rica Trinidad & Tobago Argentina Haiti 0 3 6 9 12 15 18 Source: World Bank staff estimates. gories (Table 2.3). In addition, the average level of CG capital expenditures as a percentage of GDP was below 2.8 percent in recession years and 4.0 percent in boom years. These outcomes reflect the fact that, to a great extent, public investment continues to be the residual item in the budget--that is, the one that bears the brunt of fiscal adjustment efforts. The composition of fixed capital formation has suffered important changes. Whereas investment in eco- nomic infrastructure played a dominant but declining share as a percentage of total capital invest- ment, falling from 72 percent in 1990 to 54 percent in 2000, the share of investment in social infrastructure increased from 12 percent to 19 percent of total expenditure (Figure 2.9). The dis- aggregated analysis of the internal decomposition of investment in economic infrastructure shows that, by far, the biggest decline was in the share of public investment devoted to agriculture, which more than halved, from 46 percent to 19 percent of total capital formation, whereas the energy FIGURE 2.9: STRUCTURE OF CENTRAL GOVERNMENT'S FIXED CAPITAL FORMATION 1991 2000 Multi- Multi- sector sector General 11% 12% 6% General 15% Social 12% Economic Economic Social 54% 71% 19% Source: MEF. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 33 and mining sector increased its share from almost nil in 1990 to 10 percent in 2000. For its part, the transport and communication sector essentially maintained its share at roughly a quarter of total capital formation (Table SA.7b). The doubling share of investment in social infrastructure is a very positive trend inherited from the 1990s. This effort is also reflected in the pattern of total social expenditure, which seemingly rose from 3.9 percent of GDP in 1990 to 6.9 percent in 2000, and remained steady until 2002 despite the economic slowdown.15 Recommendation: Peru has the capacity to gradually fill its financial gap in the social sectors by raising its his- torically underfunded education and health expenditure ratios of 2.9 and 3 percent of GDP, respectively, in 2000, to at least the average LAC ratios of 4.1 and 3.2 percent of GDP, respectively, in the next five years. The observed increase in fixed capital formation of the "general sectors"--mainly including general public services like military expenditures--is a matter of concern. As a percentage of total capital investment, such share almost tripled from 5.6 percent to 14.8 percent, reflecting, to a con- siderable extent, the leading role assigned to military expenditures within total public expenditures. In a similar vein, during 1998­2000, defense and national security expenditure averaged 2.9 per- cent of GDP. Such a relatively high level of military expenditure no longer seems to be justified by the needs of a country in the aftermath of the successful peace process with Ecuador, and by a country now far removed from the civil conflict that afflicted it in the mid-1990s. Moreover, such expenditure turned out to be prone to corruption, evidenced by lack of any control. Recommendations: Authorities should redress this situation by: Reducing resources devoted to military outlays to the budgeted 2.1 percent of GDP in 2002, and shifting part of those savings to social expenditure; and Completing the application of SIAF subsystems to military expenditure to make it more transparent and accountable. Developing effective countercyclical fiscal and social policies remains a priority, especially in good times. Both policies are, in general, complementary. The most significant increases in total (and social) spending took place in the early-to-mid-1990s, when growth was high. Then, in the late 1990s, real social spending remained fairly constant even during the economic downturn of 1998 and, later, in 2001 (Figures 2.10 and 2.11). How expansive total and social expenditures were despite fiscal contractions is an empirical question. The percent changes and point elasticities of total and social expenditures were negative in 1998 (Table 2.6). During 1991­2002, the level of social expenditure was not protected compared to fiscal adjustments. The elasticity of all categories of social expenditure was positive and equal to or higher than unity--which reveals their procyclical behavior (Table 2.6). These findings are all supportive of procyclical policies. However, on one hand, Peruvian officials have developed the basic institutional framework required for a countercyclical fiscal policy with the creation of the Fiscal Stabilization Fund as part of the Law of Fiscal Prudence and Transparency. Short of resources, however, the Fund has played a rather marginal countercyclical role. On the other hand, since 2001, and in the midst of another economic downturn, Peru's government opted for a countercyclical social policy, as can be inferred from the nonnegative 15. Authorities indicated that such increase could partly be due to a reclassification of line items grouped under social expenditure. 34 WORLD BANK COUNTRY STUDY FIGURE 2.10: SOCIAL AND PUBLIC SPENDING (1994 min NS) 20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 91 92 93 94 95 96 97 98 99 00 01 02 Social Spending Public Spending Source: MEF. growth of real social spending, when compared to the actual decrease in real total expendi- ture (Figure 2.11). In a similar pattern, point elasticities for social, health, nutrition, justice, and social security expenditure in 2001 were also negative, indicating that in a year of fiscal contraction, expenditure in these sectors actually increased.16 These results for fiscal contrac- tion appear robust to period changes. Revamping the Fiscal Rules Peru needs a fiscal rule to help stabilize its fiscal balances in the medium term, reduce its volatile and procyclical fiscal deficit bias, and consolidate credibility. As mentioned, Peru is under consider- able fiscal stress, with only a mild countercyclical social policy in the early 2000s. Procyclical and deficit-biased fiscal policies can be explained by the combination of faulty measures, especially dur- ing election years, fiscal conservatism, and weak institutions to attempt (and succeed) in running surpluses during the economic boom of the mid-1990s. If respected, a fiscal rule might strengthen credibility because it reverses political incentives for achieving fiscal stabilization in favor of disci- plined compliance, promotes accountability and transparency when supported by sanctions and open access to information, and tends to eliminate the deficit bias if supported by countercyclical, eventually automatic, stabilizers. Peru's Law of Fiscal Prudence and Transparency is one of the most comprehensive in Latin America (Table SA.23b). It contains numerical and procedural rules, and a Fiscal Stabilization Fund (Box 2.3). Compliance with the Law was mixed. The main numerical target was not met, and compliance with procedural rules was also unsatisfactory, especially with respect to required reporting. How- ever, a multiyear budgeting framework has been established, supported by a Portal de Transparen- 16. Elasticities were calculated such that ei = Xi/X, where ei is the elasticity of expenditure in the i sector, Xi is the percentage change in real expenditure in the ith sector, and X equals the percentage change in total real expenditure. The higher the magnitude of the elasticity, the greater the responsiveness of that category of expenditure to variations in total expenditure. Sectors that are favored in periods of fiscal contraction of total government expenditure have elasticities of less than unity, that is, they would be expected to experience reductions of a proportionately smaller magnitude than the total expenditure cutback, or even real increases if the elasticity is found negative. An elasticity of exactly equal to zero means that the sector is completely pro- tected, and those greater than unity suggest that this is a relevant category of expenditure. RESTORING FISCAL DISCIPLINE FOR POVERTY REDUCTION IN PERU 35 FIGURE 2.11: GROWTH RATES IN SOCIAL AND PUBLIC SPENDING l 40 Rea 30 or 20 Log 10 in Spending 0 -10 Change -20 92 93 94 95 96 97 98 99 00 01 02 Social Spending Public Spending Source: MEF. cia, but all mandated reports are missing (Table 3.1): the semiannual progress report of the multi- year framework, the annual report on contingent risk and tax expenditure, and the quarterly report on budget expansions (Valderrama 2002). Peru faces four broad dilemmas in revamping its Law. First, adding flexibility versus credibility for a more rigid rule would imply higher credibility, if the probability of fulfilling it is high; second, looking for more simplicity versus complexity, since a simpler rule would allow an easier understand- ing by the public; third, adding more procedural versus numerical rules, since numerical rules seem easier to understand, but harder to comply with; and fourth, opening for participation versus confi- dentiality in the revamping process of the Law, because a consensus-built rule would allow for enhanced credibility, but would delay its approval. Two LAC best practices seem particularly appealing for revamping Peruvian Fiscal Law: Brazil and Chile. On one hand Brazil has a complex, still credible Fiscal Responsibility Law that was built under an extremely participatory process and that combines a mix of procedural-cum- numerical-based rules. It is not clear whether adherence to such rigid targets fully eliminates TABLE 2.6: SOCIAL EXPENDITURES ARE PRO- OR COUNTERCYCLICAL?a Percent Point Point Elasticity Change in Elasticity Elasticity (1991­2002)b 1998 (1998) (2001) Total Expenditures . . . -0.28 . . . . . . Social Expenditures 1.19 -1.50 -8.32 -0.17 Education 0.99 -0.66 5.01 1.78 Health 1.15 -5.58 -1.87 -5.83 Nutrition 2.35 0.81 13.74 -1.13 Justice 1.75 -0.39 58.38 -3.2 Social Security 1.47 -0.75 -17.01 -3.34 Other Social Investment 1.66 -2.75 -61.87 2.32 a. Elasticities are with respect to total real expenditure of the consolidated public sector. b. Significant at 1 percent. Source: World Bank staff estimates. 36 WORLD BANK COUNTRY STUDY BOX 2.3: PERU: EVALUATION OF THE TRANSPARENCY LAW 2000­01 2000 2001 Law Actual Law Observed CPS deficit as % of GDP 2.0% 3.2% 1.5% 2.3% Difference between annual 2.0% 0.6% 2.0% -5.0% growth rate of general government's non-financial expenditures and inflation, less than 2%. CPS debt increase