Achievements and Challenges of Fiscal Decentralization Lessonsfrom Mexico EDITED BY MARCELO M. GIUGALE AND STEVEN B. WEBB 20563 _ 4 May 2000 -~~~w 1> _Ti *i= A' , U+, ,_+ Achievements and Challenges of Fiscal Decentralization Lessons from Mexico Edited by Marcelo M. Giugale and Steven B. Webb THE WORLD BANK WASHINGTON, D.C. ©2000 The International Bank for Reconstuction and Development/THE WORLD BANK 1818 H Street N.W. Washington, D.C. 20433 All rights reserved Manufactured in the United States of America First printing May 2000 The cover illustration is Pareja (1990) by Rodrigo Pimentel, reproduced courtesy of the artist and Chac-Mool Gallery, Los Angeles. Photograph by Javier Hinojosa. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the acccuracy of the data included in this publication and accepts no responsibility for any conse- quence of their use. The material in this publication is copyrighted. The World Bank encourages dis- semination of its work and will normally grant permission to reproduce portions of the work promptly. 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 directly to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax a 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 Office of the Publisher, World Bank, at the address above or faxed to 202-522-2422. Library of Congress Cataloging-in-Publication data has been applied for. Contents Acknowledgements ............................................. ix List of Acronyms ................................................ xi Preface .................................................... xv Olivier Lafourcade Overview .................................................... 1 Marcelo Giugale, Vinh Nguyen, Fernando Rojas, and Steven B. Webb Institutions for Decentralization .............................. 4 Spending and Service Delivery .............................. 11 Subnational Taxes and Revenue .................. 15 Intergovernmental Transfers .................. 19 Subnational Borrowing and Debt Management .23 Options and Agenda for Action .29 Medium- to Long-Term Agenda: 2001-2006 .33 Perspectiva General .39 Marcelo Giugale, Vinh Nguyen, Fernando Rojas, y Steven B. Webb Las Instituciones de la Descentralizaci6n .42 Gasto y Prestaci6n de Servicios .51 Impuestos e Ingresos Subnacionales .56 Transferencias intergubernamentales .60 Endeudamiento y Administraci6n de la Deuda Subnacional .... 64 Opciones y Agenda de Trabajo .71 Agenda de mediano a largo plazo: 2001-2006 .76 Chapter 1: Principles of Decentralization .85 Thomas Courchene, Jorge Martinez-Vazquez, Charles E. McLure, Jr., and Steven B. Webb Different Federal Models .85 Market-Preserving Federalism .86 The Federal-State Fiscal Accord .87 iii IV FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Spending and Service-Delivery Assignment ................... 88 Common Problems with Expenditure Assignments: Lessons from International Experience ........................ 91 Tax Assignments ........................................... 94 Transfers and System Integration ........................... 104 Debt and Borrowing ....................................... 111 Interlinkages ............................................. 115 Appendix: Intergovernmental Transfers in Developed Federations ............................................... 116 Chapter 2: Historical Forces: Geographic and Political .............. 123 Thomas Courchene, Alberto Diaz-Cayeros, and Steven B. Webb Federal Concentration of Revenue .......................... 123 Rise of Democracy and Opposition Parties ................... 128 Transfer-Led Decentralization .............................. 129 Inequality and Globalization ............................... 131 Chapter 3: Assignment of Spending Responsibilities and Service Delivery ................................................. 139 Enrique Cabrero Mendoza and Jorge Martinez-Vazquez Current Assignments ...................................... 143 Problems with Expenditure Assignments .................... 151 Decentralization in Specific Sectors .......................... 153 Institutions of Coordination ................................ 166 Administrative Capacity and Democratic Governance ......... 168 Toward a More Efficient Expenditure Decentralization ........ 170 Improvement of Administrative Capacity and Democratic Governance (Medium and Long Term) and Use of Asymmetric Assignments in the Interim (Short Term) .................. 174 Chapter 4: Tax Assignment ...................................... 177 Alberto Diaz-Cayeros and Charles E. McLure, Jr. Potential Sources of State Tax Revenues ..................... 177 Potential Sources of Municipal Revenues .................... 194 Constraints in the Constitution ............................. 195 Summary: Long- and Short-Run Reassignments of Taxes ...... 196 Chapter 5: Transfers and the Nature of the Mexican Federation ...... 200 Thomas Courchene and Alberto D{az-Cayeros Principles for Intergovernmental Transfers ................... 200 Mexican Constitution and Fiscal Federalism .................. 202 CONTENTS V Current Intergovernmental Transfer Programs ............... 206 Vertical and Horizontal Balance ............................ 216 The Evolution of Mexican Federalism ....................... 224 Alternatives for the Future ................................. 226 Chapter 6: Subnational Borrowing and Debt Management .......... 237 Marcelo Giugale, Fausto Hernandez Trillo, and Joao C. Oliveira Subnational Debt .......................................... 237 Fiscal Imbalances and Debt Accumulation ................... 241 Subnational Borrowing: The Experience Up to 2000 ........... 247 A New Reform Initiative ................................... 251 Notes ..................................................... 261 Bibliography .................................................. 271 Author Biographies ............................................ 279 List of Tables 0.1 Fiscal Magnitudes: Federal and State Governments, 1997 . . .2 0.1 Magnitudes Fiscales: Gobiernos federal y estatales, 1997 .. .40 1.1 Attributes, Strengths, and Weaknesses of Various Techniques of Tax Assignment ...................................... 101 1.2 Institutional Arrangements to Set and Keep Hard Budget Constraints on States .................................. 112 2.1 (0 Convergence in Mexico ............................. 134 3.1 Mexico's Expenditure Assignment ..................... 140 3.2 Expenditure by Level of Government in Mexico, 1989-96 (Percent) ............................................ 143 3.3 National Education Expenditure as a Percentage of GDP, by Source of Funds, 1990-97 .............................. 144 3.4 Distribution of Programmable Expenditure by Level of Subnational Government in Mexico, 1980-96 (Percent) .... 145 3.5 Ramo 26 as a Percentage of Municipal Gross Expenditure, 1989-95 ............................................. 149 3.6 Horizontal Disparities in Expenditures Per Capita, 1997 . .150 3.7 State Expenses as a Percentage of Federal Education Expenses, 1989-94 ............................................. 157 4.1 Increase in Revenues from Increase in Payroll Tax (1996 Thousands of Pesos) .................................. 181 4.2 Increase in Own Revenues from Reassignment of Major Excises by State ........................................... 185 VI FISCAL DECENTRALIZATION: LESSONS FROM MEXICO 4.3 Current and Proposed Assignments in Mexico ........... 198 4.4 Increase in Own Revenues from Reassignment of Various Taxes, by State (Thousands of Pesos 1996) ..................... 199 5.1 Federal Funds Transferred to States and Municipalities (In Ascending Order of Federal Accountability) ............. 209 5.2 Summary Statistics of State Shares of IPF, 1960-93 ........ 217 5.3 Total Transfers to Subnational Governments in Mexico, 1999 (Ranked in Ascending Order of Federal Conditionality) . . 220 5.4 Transfers and Own Revenue as Percentage of Total Subnational Resources ........................................... 223 5.5 Fiscal Cost of Compensation ........................... 234 6.1 Mexico: Total Debt, 1994-98 (Millions of Pesos) .......... 238 6.2 Mexico: Total Debt, 1994-97 (Debt/Revenue Ratio in Descending Order of Total Debt) ....................... 242 List of Figures 0.1 Composition of Total Subnational Revenues, 1999 ......... 17 0.2 Mexico: States' Primary Deficit, 1989-97 .................. 25 0.1 Estructura del total de ingresos subnacionales, 1999 ....... 58 0.2 Mexico: Deficit Primario De Los Estados 1989-1997 ........ 68 2.1 Subnational Tax and Spending Shares in Argentina, Brazil, and Mexico (Shares of GDP) ............................... 130 2.2 Differences in Per Capita Gross Domestic Product (By State) ............................................ 132 2.3 Index of Illiteracy ..................................... 133 2.4 Regional Convergence, 1960-88 ........................ 135 2.5 Regional Divergence, 1988-96 .......................... 136 2.6 Regional GDP Versus Poverty Index .................... 137 3.1 State Revenue from Own Sources, Earmarked Transfers, and Participations (Constant Prices, 1993 = 100) .............. 146 3.2 Composition of State Expenditures, 1989-96 (Constant Prices, 1993 = 100) ........................................... 147 5.1 Composition of the Federal Government Budget ......... 206 5.2 Total Per Capita Revenue Shares by State, 1999 .......... 211 5.3 Where Ramo 33 Came From ........................... 212 5.4 Total Per Capita Ramo 33 (DF Includes Ramo 25) ........ 213 5.5 Education Transfers Complement State Expenditure ..... 214 5.6 State Own Revenue ................................... 218 5.7 Municipal Own Revenue .............................. 219 5.8 State Funds .......................................... 221 5.9 Municipal Funds ..................................... 222 5.10 Total Local Funds by State (Revenues and Transfers) ..... 224 CONTENTS VII 5.11 Medium-Term Simulation (Modest Tax Reassignment Proposal, Revenue- Sharing Formulas Unchanged) ................ 231 5.12 Long-Term Simulation (Revenue Sharing Transformed into Derivation Principle Own Revenue) .................... 233 6.1 Mexico: Subnational Governments' Total Outstanding Debt (at 1996 Prices) .......................................... 239 6.2 State's Real Debt, 1994-98 ............................. 240 6.3 Mexico: Aggregate Subnational Governments' Fiscal Deficit, 1989-97 ............................................. 243 6.4 Jalisco: Fiscal Deficit, 1989-97 .244 6.5 State of Mexico: Fiscal Deficit, 1989-97 .244 6.6 Nuevo Leon: Fiscal Deficit, 1989-97 .245 6.7 Sonora: Fiscal Deficit, 1989-97 .245 6.8 Federal District: Fiscal Deficit, 1989-97 .246 6.9 Sinaloa: Fiscal Deficit, 1989-97 .246 List of Boxes 0.1 Coordination of Decentralization: Lessons in Latin America .5 0.2 International Experience with Differential Treatment of Municipalities .9 0.3 Education and Health .................................. 12 0.1 Coordinaci6n de la descentralizaci6n: Lecciones de America Latina ....................................... 44 0.2 Experiencia internacional en el tratamiento diferencial de municipios ........................................ 49 0.3 Educaci6n y Salud .................................... 52 3.1 Mexico's Education System ............................ 154 3.2 Impact of Alternative Distribution Formulas for Education Funds ............................................... 159 3.3 Mexico's Health System ............................... 161 Acknowledgements THis VOLUME IS THE RESULT of a team effort and, as such, it has benefited from an array of invaluable contributions. Our thanks are therefore due to a large number of people. First, the chapters' authors, who have provided not just material of outstanding technical quality but a remarkable com- mitment to enriching the debate about decentralization in Mexico. We consider ourselves fortunate to share this book with Enrique Cabrero Mendoza, Thomas Courchene, Alberto Diaz-Cayeros, Fausto Hernandez Trillo, Jorge Martinez-Vazquez, Charles E. McLure, Jr., Vinh Nguyen, Joao C. Oliveira, and Fernando Rojas. While this book reflects the authors' own views (and not necessarily the views of the World Bank, its Board of Executive Directors or its member countries), its production was institutionally housed at the World Bank. We thus benefited enormously from the general guidance of Olivier Lafourcade (Director of the Mexico Country Department) and Guillermo Perry (Chief Economist for the Latin America and Caribbean Region), and from the aus- pices of the office of David de Ferranti (Vice-President for Latin America and the Caribbean Region). We are also thankful to Adolfo Brizzi, Richard Clifford, Julio Cordoba, Jozef Draaisma, Stephen Everhart, Fernando Montes-Negret and Eduardo Velez for their illuminating comments on earlier versions of the various chapters, and to Robert Duval-Hernandez and Eric Zuckerman, for their quantitative and analytical assistance. Finally, we are especially grateful to Michael Geller, who superbly man- aged the production and support team. This team included: Peggy O'Donnell, our planning and budget advisor; Elizabeth Forsythe and Diane Stamm, the principal language editors; Maria del Carmen Navarrete Angeles, who contributed to the Spanish translation of the overview; Christopher Neal and Lee Morrison who, together with Paola Scalabrin and Jeong Hyun You, spearheaded the publicity and printing efforts; Cynthia Masson-Barrero and Liliana Wiesner who helped shepherd the project ix X FISCAL DECENTRALIZATION: LESSONS FROM MEXICO along at various stages; and Esthella Provas at Chac-Mool Gallery in Los Angeles who identified the art of Rodrigo Pimentel for use on the cover of this volume. Our thanks to all of them. Marcelo M. Giugale and Steven B. Webb Washington D.C. and Mexico City April 17, 2000 Acronyms CAPFCE Comite Administrador del Programa Federal de Construcci6n de Escuelas CEDEMUN Centro de Desarrollo Municipal CGC Commonwealth Grants Commission CIDAC Centro de Investigaci6n para el Desarrollo, A.C. CNA Comisi6n Nacional del Agua CNS Consejo Nacional de Salud CONAFE Consejo Nacional de Fomento Educativo COPLADE Comites de Planeaci6n para el Desarrollo del Estado COPLADEMUN Comites de Planeaci6n para el Desarrollo Municipal COPRE Comisi6n Presidencial para la Reforma del Estado CVAT Compensating value added tax DIF Desarrollo Integral de la Familia EDUCO Educaci6n con Participaci6n de la Communidad FAEA Fondo de Aportaciones para la Educaci6n de los Adultos FAEB Fondo de Aportaciones para la Educaci6n Basica y Normal FAETA Fondo de Aportaciones para la Educaci6n Tecnol6gica y de los Adultos FAFM Fondo de Aportaciones para el Fortalecimiento de los Municipios FAISE Fondo de Infraestructura Social Estatal FAISM Fondo de Aportaciones para Infraestructura Social Municipal FAM Fondo de Aportaciones Multiples FASP Fondo de Aportaciones para la Seguridad Pudblica FASSA Fondo de Aportaciones para los Servicios de Salud FFM Fideicomiso de Fomento Minero FGP Fondo General de Participaciones FISM Fondo para la Infraestructura Social Municipal xi XII FISCAL DECENTRALIZATION: LESSONS FROM MEXICO GDP Gross domestic product IEPS Impueso Especial sobre Producci6n y Servicio IMF International Monetary Fund IMSS- Instituto Mexicano del Seguro Social COPLAMAR Coordinaci6n General del Plan Nacional de Zonas Deprimidas y Grupos Marginados INDETEC Instituto para el Desarrollo Tecnico de las Haciendas Publicas INE Instituto Nacional de Ecologia INEA Instituto Nacional para la Educaci6n de los Adultos INEGI Instituto Nacional de Estadistica Geografia e Informatica IPF Inversi6n Publica Federal Instituto Nacional de Estadisticas Geografia y Investigaci ISAN Impuesto Sobre Autom6biles Nuevos ISIM Impuesto Sobre Ingresos Mercantiles ISSSTE Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado IVA Impuesto al Valor Agregado (Value Added Tax) MARINA Secretaria de Marina NAFTA North American Free Trade Agreement NEP Nueva Estrategia de Programci6n OCDE Organizacion para la Cooperaci6n y el Desarrollo Econ6mico OECD Organization for Economic Cooperation and Development OPD Organismos publicos de descentralizaci6n PAN Partido Acci6n Nacional PAO Planes operativos anuales PARE Programa para Abatir el Rezago Educativo PEMEX Mexico's nationalized oil company PRD Partido de la Revoluci6n Democratica PRI Partido Revoluci6nario Institucional PROCAMPO Programa de Apoyos Directos al Campo PRODEI Programa para el Desarrollo de la Educaci6n Inicial PROGRESA Programa de Educaci6n, Salud y Alimentaci6n PROMAP Programa de Mejoramiento de la Administraci6n Puiblica PRONASOL Programa Nacional de Solidaridad RFP Recaudaci6n Federal Participable RST Retail sales tax SECODAM Secretaria de Controlaria y Desarrollo Administrativo ACRONYMS XLII SEDENA Secretaria de la Defense Nacional SEDESOL Secretariat de Desarrollo Social SHCP Secretaria de Hacienda y Credito Piblico SNCF Sistema Nacional de Coordinaci6n Fiscal (Pacto Fiscal) SSA Secretaria de Salud VAT Value Added Tax Preface Olivier Lafourcade* DEMOCRATIZATION, DECENTRALIZATION, DEVELOPMENT. These three sequen- tial forces have swept the world over the last decade and have redrawn the maps of politics, power, and prosperity. Undoubtedly, they have been fos- tered and accelerated by globalization (financial and commercial) and by the information technology revolution. Understanding those forces of change (despite their speed and complexity) is of the utmost importance for, in the end, their value will be measured by one simple yardstick-their impact on people's quality of life, especially among the poor. Modern Mexico has been fully engaged in, and is thus a rich case study of, the democratization, decentralization, and development trio. In recent years, enhanced political competition has redistributed decision-making across the three levels of government (federal, state, and municipal) and has made it more accountable to the average citizen. It has also given sub- national governments a renewed role as economic agents. The taxation, spending, borrowing, and institutions of Mexican states and municipali- ties are now increasingly (albeit unevenly) under the rigor of market dis- cipline. The combined, closer scrutiny of voters and financiers is creating a new incentive framework for policymakers-a framework where neces- sary reforms become both inescapable and, more importantly, a perceived source of potential reward. This book is the product of the analytical work of a large number of experts, Mexican and foreign. They here document Mexico's decentral- ization experience; conceptualize its main trends, policies, and options; and bring it into the light of international comparison. They distill critical * Director, Mexico Country Department, the World Bank. As with other chapters in this book, the opinions expressed here are the author's own, and do not necessarily reflect the views of the World Bank, its Board of Executive Directors, or its member countries. xv xvi FISCAL DECENTRALIZATION: LESSONS FROM MEXICO lessons and challenges that are of relevance for Mexico, for Latin America and, generally, for countries that are embarking on far reaching decen- tralization efforts. This renders the volume a major contribution to our knowledge and thinking in this area. And a timely one: for decentraliza- tion is an irreversible process that is likely to continue occupying policy- makers for years to come. Overview Marcelo Giugale, Vinh Nguyen, Fernando Rojas, and Steven B. Webb ALTHOUGH THE PHENOMENON OF "DECENTRALIZATION" is worldwide and its meaning, implementation, and effects vary from country to country, the experiences point to three lessons applicable anywhere. First, decentralization often has major effects on concerns ranging from macroeconomic stability to poverty alleviation, the provision of social ser- vices, and the quality of governance (Burki, Dillinger, and Perry 1999). For this reason, a country decentralizing its public sector needs to develop a coherent decentralization strategy. Decentralization, properly done, can have many economic benefits as well, producing greater efficiency, responsive- ness, and accountability in the delivery of services desired by the local pop- ulation. Improperly done, decentralization can have undesirable conse- quences, leading to macroeconomic disequilibrium, exacerbating regional differences and conflicts, or reducing the quality and quantity of public ser- vices. Paying adequate attention to the interactions of various policy areas can avert many problems with decentralization. Although there is no one right path to decentralization, and determining the appropriate extent and nature of decentralization is a political decision for the country, a number of common mistakes can be avoided. For example, it is always a mistake to decentralize revenues without decentralizing a corresponding set of responsibilities, and it is usually wrong to give exactly the same revenue and responsibilities to small and large municipal governments. Second, because decentralization everywhere is an ongoing, evolving activity, a successful strategy requires adequate institutional infrastructure to develop, monitor, and implement the decentralization policy. The neces- sary infrastructure includes legal and regulatory frameworks, organizations for coordination, and capacity-building programs. Even the best-planned strategies do not sustain themselves without such institutional support. In particular, experience everywhere shows the need to develop and publi- cize accurate, complete, and trustworthy information on all aspects of the 1 2 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO decentralization process. This reduces the confusion and political and bureaucratic conflicts that invariably accompany any major change. Third, the Latin American experience with decentralization shows that abrupt, across-the-board efforts have generally failed. Countries have had to go back to incremental decentralization, with differentiated rules accord- ing to different management capacity of territorial units. This was the experience with the Colombian laws of 1986, the Brazilian constitution of 1988, the Venezuelan Organic Law of Decentralization in 1988, and the Bolivian Popular Participation and Decentralization laws of 1994 and 1996. These lessons apply to Mexico. The country is progressing toward more autonomy, fiscal responsibility, and accountability at subnational levels of government. States now spend close to half as much as the federal gov- ernment (see table 0.1). The municipal share of public spending has also grown, although this trend is not as fully documented. This process is being driven by heightened political competition at all levels of government and by the desire of the federal government to include decentralization in its program for broadening political participation. Each country must, however, develop its own strategy of decentraliza- tion and its own particular institutional infrastructure in accordance with its history, its objectives, and the constraints it faces. The models of decen- tralization found in the world range along a spectrum. One end, exempli- fied in Canada, may be labeled full legislative federalism in the sense that the main source of funds for subnational governments is their own taxes, and they have almost complete autonomy as to how they spend them. Canada carries provincial autonomy so far as to have no federal constraints Table 0.1. Fiscal Magnitudes: Federal and State Governments, 1997 Percent of GDP Federal States* Own spending 11.5 4.9 1991 8.4 3.0 1996 7.7 4.9 Own revenue 15.8 1.0 Disposable revenue 13.2 5.4 Primary balance -1.3 0.0 Overall balance -4.9 -0.1 Debt service 18.4 0.3 Debt stock 31.0 1.5 ' Tabasco and Tamaulipuas not included; no data. Source: SHCP and authors' estimates. OVERVIEW 3 at all on provincial borrowing, leaving the regulatory task entirely to the capital market, although the provinces impose strict controls on municipal borrowing. The regional inequalities arising from differences in the tax base in Canada are partly equalized through federal transfers. Switzerland and the United States follow variants of this model, though with less equaliza- tion (none in the United States), less uniform state tax policies, and more detailed federal attempts to influence state spending. In Latin America, Brazil's decentralization is closest to this end of the spectrum, with the larg- er states raising much of their own revenue and designing their own social sector programs. But the federal government in Brazil has frequently got- ten involved, often disruptively, in the personnel policies and debt man- agement of the states. The model at the other end of the spectrum-administrative federal- ism-is found in Germany and Australia. Transfers are the major source of states' revenue, and federal (or joint) policies guide most subnational expenditures. Regional equalization policies, largely implemented through transfers, are very strong. Moreover, tight central controls are imposed on subnational borrowing. Colombia and Venezuela follow this model in Latin America, although the equalization component is largely absent. Varying combinations of the features mentioned-revenue indepen- dence, expenditure autonomy, debt autonomy, and equalization-may be found in other countries. In Argentina, for example, transfers are more important than in Brazil, but states have more autonomy in controlling per- sonnel costs and the market plays a larger role in debt management. Mexico could, in principle, mix and match these characteristics as it wish- es. Given that Mexico has already made the major decisions about decen- tralization of spending responsibility and subnational borrowing, the next critical decision concerns the importance of transfers and the extent to which the tax system should be decentralized. This will go hand in hand with the decision about the degree to which federal transfers will com- pensate for the regional disparities that would otherwise follow from tax decentralization. Another important decision, which may require revision as the transfer system evolves, is the extent to which the center will guide, monitor, or control the details of expenditures (for example, through finan- cial market regulations or conditions of transfer). The institutional struc- tures mentioned earlier will be needed to attain consensus on these policy objectives and to implement and sustain them. Mexico has already started down the road to decentralization, with its own model being constructed in Mexico, as it must be. Some actions already taken seem desirable and should be sustained and strengthened, such as the end of discretionary transfers and support for more uniform and com- prehensible state accounting and public reporting. Others seem less desir- able. This book is a collection of papers that draws on worldwide experiences 4 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO with decentralization in order to focus on what is needed to develop a coherent and sustainable decentralization strategy in Mexico, and the nec- essary institutional underpinnings for it. In this overview, we distill the main messages that the authors deliver in those papers. The discussion here is organized around five themes: (1) institutions for decentralization, (2) spending and service delivery, (3) sub- national taxes and revenue, (4) intergovernmental transfers, and (5) sub- national borrowing and debt management. A vision of Mexico's path and options in decentralization then emerges and is articulated for the short, medium, and long term. Institutions for Decentralization The key institutional issues in decentralization are reinforcing confidence in decentralization among key stakeholders, improving intergovernmen- tal coordination and conflict resolution, strengthening and democratizing the process for control and accountability, coordinating the multilevel budget process, regularizing the distinctions between municipalities with different administrative capacities, implementing cooperative capacity building, and dealing with the peculiarities of the Federal District's rela- tionship with its neighbors and the federal government. Intergovernmental Coordination International experience shows the importance, and the difficulties, of col- lective institutions that steer the decentralization process and serve as forums for negotiations and preemptive conflict resolution (see box 0.1). Developing such institutions is particularly challenging for Mexico, because of its trial-and-error approach to decentralization and the resultant politi- cal, regional, intersectoral, and even intrasectoral fragmentation. Thus far Mexico has improved its intergovernmental and intersectoral coordination largely through top-down mechanisms, such as conditional transfers and federal monitoring. Now it is necessary to create a consensus-building mechanism that is trusted by all parties and that takes care of the follow- ing critical tasks: * Proposing a long-term view of decentralization, intermediate goals, and ongoing adjustments; * Promoting forums for discussing and negotiating intergovernmental fiscal arrangements and preventing interjurisdictional conflict; * Producing an annual report on progress and current issues in decen- tralization, keeping federal agencies and subnational governments alerted to their responsibilities within the decentralization process; OVERVIEW 5 Box 0.1. Coordination of Decentralization: Lessons in Latin America The Latin American experience with coordination of the decentralization process indicates the following lessons: 1. Horizontal, intersecretary agreements (national level) or conferences of governors, mayors, or finance secretaries (subnational level) have had only modest effects, generally restricted to a few specific policies. Broader scope agreements seem to require presidential support and definite com- mitment from the Ministry of Finance. 2. Presidential Commissions may be more effective, but their sustainabili- ty depends on the continuity of presidential support, as indicated by the Comisi6n Presidencial para la Reforma del Estado (COPRE) experience in Venezuela. 3. Presidential support for a lower-level agency, such as the Bolivian expe- rience with the Sub-secretarfa de Participaci6n Popular, is usually short-lived; in Bolivia it lasted only while it had unrestricted presidential backing. 4. A single agency at the ministerial level, like the Ministry of Governmnent or the Interior, or the Ministry of the Presidency, has usually failed, as the Ministry of Finance or the planning agency became pivotal for the advance of decentralization. 5. A special unit at the presidential level, like the Secretaria de la Presidencia, usually lacks effective coordination capacity vis-a-vis sector ministries, and especially the Ministry of Finance. 6. Experiences where no agency is primarily responsible, as in Chile and Colombia, illustrate that the Ministry of Finance or the national planning office usually takes over when the process of decentralization is not clear- ly headed by another agency. * Providing reliable information for the public discussion of federal, state, and local fiscal and financial policy, and evaluating the design and exchange of information; * Designing, collecting, and disseminating key indicators for the imple- mentation of decentralization, monitoring the changes of revenue and expenditure responsibilities, and evaluating the capacity of subna- tional governments to assume new responsibilities. To help address these needs, in early 2000 the Mexican federal govern- ment established a Decentralization Committee within the Secretaria de Hacienda y Credito Publico (SHCP), with a Technical Secretariat, and chaired 6 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO on a rotating basis by the three undersecretaries-Ingresos, Egresos, and Credito Publico. This interagency committee will play an important role in coordinating intergovernmental fiscal relations and preparing a revision of the fiscal pact and a new Ley de Coordinaci6n Hacendaria (dealing with revenues, transfers, and responsibilities at all levels of government). The Secretariat will be responsible for technical analysis of the evolution of the decentralization process in order to guide federal policy on the matter. The unit will design and test indicators for monitoring intergovernmental fis- cal relations and decentralization as a whole; collect and publish subna- tional fiscal statistics; and disseminate periodic analysis, promote public debate, and maintain communication with congresses, their accounting offices (Contaduria Mayor de Hacienda de la Honorable Camara de Diputados), and the Secretaria de Controlarfa y Desarrollo Administrativo (SECODAM). The committee will put together and coordinate the decentralization efforts cur- rently developed by each of the three undersecretaries, and coordinate the various technical assistance and subnational strengthening programs offered by SHCP, Banobras, and other federal institutions. The committee will not be formally institutionalized during the Zedillo administration, leaving it to the next administration to decide whether and how to do this. Prevention and Resolution of Conflicts Sound intergovernmental coordination will require specific institutions for negotiations and conflict resolution, because effective decentralization requires more than the design of revenues, transfers, borrowing, and responsibilities. No system can foresee and specify everything in advance. Issues and disputes arise that need to be settled, and the method for doing this has important incentive effects for how the system actuallv works. Like other decentralizing countries, Mexico has a growing number of interjurisdictional conflicts. The courts need to strengthen their capacity to handle constitutional and administrative disputes that arise in connection with decentralization. The country could also develop alternative means for resolving disputes, because court procedures tend to be expensive, time-consuming, and often unnecessary. Some conflicts are being averted or resolved through informal negotiations and willingness to compromise. The lack of institutionalized alternate channels for resolving disputes often makes the outcome a question of which side has greater political power, and this undermines the credibility of the rules. It would help if the non- judicial means to resolve interjurisdictional conflicts were institutionalized and if conciliatory experts and judges were trained in the principles and operating rules of decentralization. In the health sector, the National Health Council (Consejo Nacional de Salud) resolves many disputes, and this insti- tution could be a model for other sectors with major decentralization issues, such as education, water, road transport, and environment. OVERVIEW 7 Democratic Accountability and Control Effective controls are both a prerequisite and an essential component of strengthening management capacity, autonomy, and accountability at sub- national levels. Because the impetus for decentralization in Mexico has come not only from the federal government, but also from democratic competi- tion in Congress and subnational governments, it is important to have sys- tematic public disclosure of all government accounts, opening the poten- tial for decentralized democratic control. In a decentralized Mexico, institutional development toward effective controls should capitalize on the strengths of the existing mechanisms. Mexico has accumulated significant experience and technical expertise at SECODAM and some of the state comptrollers (in the executive branch). The congressional accounting offices, independent units of increasingly plu- ralist congresses, are emerging as some of the more trusted institutions in the country. The Homologaci6n Program of SHCP also tries to standardize the states' financial reporting. Besides these official institutions, some states and federal agencies now conduct external audits through specialized firms. In the Mexican model, controls are a matter of intergovernmental coor- dination. The congressional accounting office dictates guidelines for the accounting offices of state legislatures to follow. To strengthen the politi- cal autonomy and credibility of these institutions, Mexico might consider the models of the U.S. General Accounting Office, which Congress endows with considerable autonomous investigative authority, or the parliamen- tary finance committees in some British commonwealth countries, which the Constitution establishes under the chairmanship of the opposition party. On the executive side in Mexico, SECODAM coordinates with state comptrollers and assists them with knowledge-management programs. The states are legally responsible for fiscal control of municipalities. Although in principle each level of government should have its own independent con- trols, the federal level has organized arrangements that provide the basis for strengthening controls at all levels during the transition to a more decentralized system. Budget Coordination At present, budget coordination in Mexico involves primarily the harmo- nization of spending priorities at different levels of government. Mechanisms are different for federal-state budget coordination and for state-municipal coordination, and for federal-municipal coordination. Federal-state budget coordination operates through the federal budget process and other mechanisms. States used to submit annual sectoral oper- ating plans (planes operativos anuales, POAs) that were then adjusted and incorporated into the federal sectoral plan submitted to SHCP. This process was partially replaced in the Fiscal Coordination Law by automatic formula 8 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO mechanisms, but the federal ministries of education and health still require states to prepare some sort of POAs for the sector budget. Once the plan has been submitted, SHCP further filters the sector POAs and submits the budget bill to Congress. Although the present budget preparation mecha- nism ensures some exchange of information and opportunities for recon- ciliation of federal and state priorities, it still falls short of explicit and trans- parent criteria for allocating federal resources across the states. Besides preparation of the federal budget, the regular federal-state coor- dinating meetings (the quarterly meetings between SHCP, both the Income and Expense departments, and the finance ministries of the states) and the federal Fiscal Pact are also contexts for coordinating the budget. Revenue coordination and tax harmonization between the federal level and the states are achieved mainly through the Fiscal Pact and the corresponding general revenue transfer of nearly 22 percent of the recaudaci6nfederal par- ticipable (the participaciones).1 However, since the fiscal year is the calendar year for all levels of government, the federal budget data go to states and municipalities only at the very end of the previous fiscal year or the begin- ning of the new year, which makes subnational budgeting increasingly dif- ficult because federal budgets and transfers are less predictable and are approved later in December. Further budget coordination takes place through the growing number of pari-passu programs, agreements for sector matching grants, and ear- marked transfers for sector decentralization, such as the Educational Federalization Program, now the first and largest fund of Ramo 33.2 Extraordinary transfers for specific investment purposes may also be seen as coordination mechanisms.3 In fact, extraordinary transfers have stimu- lated the agreements for uniform accounting and reporting between the fed- eral government and the states (the Homologaci6n Program). State-municipal budget coordination involves state-level approval of municipal budgets and borrowing.4 In theory, municipal budgets have to conform with the state development plan, primarily for investment, but this is rarely practiced. Special agreements between states and municipalities can provide an additional opportunity to reconcile the spending priorities of these two levels and to coordinate revenue-raising goals and instruments.5 Federal-municipal budget coordination, even federal-state-municipal coordination, is also pursued through the transformation of Ramo 26 (bud- get category 26) into Ramo 33. Formulas already specify most of the allo- cation of the special funds of Ramo 33-primarily, but not exclusively, sec- tor funds. Therefore, Ramo 33 is in practice a mechanism by which the federal level (Congress and executive branch) guarantees the territorial allo- cation of resources according to federal priorities. Accounting and report- ing for Ramo 33 resources are still deficient, however, and neither stimulates intergovernmental trust nor allows for monitoring or impact evaluation.6 OVERVIEW 9 The current mechanisms for budget coordination-mostly spending- hinder subnational autonomy and initiative. Rather than stimulating more efficient allocation of resources in the territories, they tend to increase fed- eral control of the allocation of resources, even as states are granted more political responsibility for them. As a result, these mechanisms confuse the distribution of responsibilities among levels of government and reduce the accountability of each level. In addition, these mechanisms do not help true bilateral spending coordination because they are mainly formal budget and planning tools.7 Categorizing Municipalities The diversity of state and local governments needs to be matched with dif- ferentiated federal rules toward local governments (see box 0.2 for inter- national experience). There is wide variation in the scale, administrative capacity, tax bases, and poverty levels of states and municipalities in Mexico, but the legal framework for decentralization does not recognize these differences. At present, federal rules do not categorize municipalities by size and capacity to determine differential spending authority, tax authority, borrowing authority, reporting requirements, or eligibility for technical assistance. The details of the problem and the appropriate solu- tion may be different for municipalities and for states. Box 0.2 International Experience with Differential Treatment of Municipalities Experience in Latin America shows that, when responsibilities of subnational units are not explicitly differentiated according to effective fiscal and/or man- agement capacity, de facto differentiation takes place, often in ad hoc, chaotic ways. It is therefore advisable to explicitly differentiate among subnational units while at the same time establishing the standards to move from one category to another. The Colombian process of certification is an example of flexible cat- egorization that stimulates regional and local governments to qualify themselves to assume responsibilities in education and health. To be certified, depart- ments have to demonstrate to the national government that they can assume the new responsibilities in health and education-that they have developed the required planning, financing, monitoring, and reporting capacity. After a department has been certified, its municipalities may apply to the department for certification. In another example of categorization and upgrading, Spain has gradually increased the number and responsibilities of the autonomias-areas like Catalonia with special status-and expanded their responsibilities. 10 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Mexico has over 2,400 municipalities, spanning a wide range, from cities with several million inhabitants, policymakers with high levels of formal education, and computerized fiscal accounts, to small towns with popula- tions under 5,000, leaders with little formal education, and informal accounting practices at best. Only 25 to 40 percent of the municipalities mea- sure up to standard indicators of capacity: having a budget planning unit, an evaluation unit, and computerized accounts; having and using an inter- nal administrative code; having regulations for the cadastre;8 and raising more than half of their own resources. It seems unrealistic to have the same rules for all municipalities. In practice, states often demand less than the legal requirement from small and poor municipalities. Still, small munic- ipalities doing an appropriate amount of reporting for their size should not be in a state of noncompliance. Large municipalities frequently do more than the minimum to satisfy the requirements of their own management or of their private sector creditors. Still, some larger cities have inadequate administration, even though they meet the letter of the law. The fiscal and administrative capacity of Mexico's 32 states also varies widely. For example, as of 1997, only three states had a fiscal coordination law, a budget, accounting, and public spending law, and a public debt law; seven states had none of these laws. All states now have these laws, although their effectiveness varies in part because of their newness. Targeted transfers for capacity-building programs, such as those in the 2000 budget, could help to bring states up to a uniform standard. As state and municipal governments become more democratic and more accountable, citizens will demand better reporting from their governments. Capacity Building All stakeholders agree on the need to improve the technical and adminis- trative capacity of subnational governments. For example, technical short- comings deter proper accounting and reporting of the Ramo 33 transfers, particularly in municipalities with low fiscal and management capacity. There is less agreement, however, on the best way to strengthen subnational capacity. Federally run programs for strengthening subnational adminis- trative capacity, such as the Honiologaci6n Program, have had limited impact. Some federal funds for institutional development at the municipal level come from the option each municipality has of using 2 percent of one Ramo 33 fund (municipal infrastructure) for this purpose. For small munic- ipalities, however, which need capacity building the most, the amount comes to only a few hundred pesos-too little and too scattered for Mexico's 2,400 municipalities. The state of Puebla has developed a series of agree- ments with its municipalities to concentrate and augment these resources at the state level and to provide a program of capacity building, based on a survey of local needs. The federal government is also planning new pro- grams for supporting capacity building at state and local levels. There will OVERVIEW 11 be pressure on these governments in order to improve their access to cred- it markets when these are reformed as described below. Civil service reform at subnational levels is needed and perhaps inevitable, but it will have limited success or impact while the old incen- tives remain in place. State and local government officials, like federal government officials, respond to the incentives with which they are faced. If those incentives discourage initiative and reward inefficiency, it should come as no surprise to find unaccountable and inefficient subnational administrations. Solving the problem of limited administrative capacity in some jurisdictions will require altering the federally controlled personnel incentive system to allow honest, well-trained people to pursue an attrac- tive career in subnational government. Federal District Relations The Federal District faces particular challenges in its relationship with the federal government and with the neighboring states. For example, the Federal District has to have its budget and borrowing limits approved annually in the federal Congress; it does not get some of the Ramo 33 funds; nor is it responsible for basic education. The Federal District believes that it subsidizes its neighbors via support for the metro train system, while its neighbors feel that they subsidize the Federal District via water sector investments. The rapid transformation of the political and fiscal sta- tus of the Federal District, and the possible creation of municipalities with- in it, have made a review and reform of its vertical and horizontal inter- governmental relations necessary. International experience, such as from the Sao Paulo metropolitan area in Brazil, may help to elucidate the range of options. Spending and Service Delivery Decentralization is expected to provide many economic benefits in the areas of spending and service delivery. These benefits include choosing the mix of public sector activities that best suit the taste and needs of citizens in a local area, providing the services in a more cost-effective way by adapting the method of delivery to local circumstances, and allowing cit- izens to express more directly their concerns about service provision. The benefits will also depend on strengthening the link between subnational spending choices and self-taxing decisions. Realizing these benefits in Mexico will depend on the actions of local authorities and citizens, but the federal government can help by creating an environment that gives states and municipalities more incentive to take responsibility for their own pro- grams and finances. This is not a matter of just changing rules, but also of creating a culture of devolution. It will take time. The agenda here sets out some possible steps, which will build on those taken in the past decade. 12 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Two features dominate the current assignment of expenditure respon- sibilities. First, there are many concurrent obligations at all three levels, especially between the federal and state governments for important services such as education, health, or social assistance. Second, few responsibilities are assigned to the municipal level, especially in areas with important benefits, such as education or health (see box 0.3). There appear to be five main problems with the current allocation of decentralized spending responsibilities: (a) inadequatefiscal autonomy for Box 0.3. Education and Health Education The 1992 Acuerdo Nacional para la Modernizaci6n de la Educaci6n Bdsica was a pio- neer step that established a limited decentralization precedent. According to the Acuerdo, the federal government remains responsible for general policies and standards (normative and policymaking functions), teacher training and terri- torial distribution, production of textbooks, evaluation and monitoring, and the provision of financial resources needed to ensure proper coverage and quality of the educational system. In 1998, Ramo 33 complemented the Acuerdo Nacional by guaranteeing trans- fers and introducing greater transparency in resource allocation for education. However, inasmuch as Ramo 33 enshrined inertia in territorial allocation of resources for education, it may be considered more a tool for deconcentration than a decentralization measure that enhances efficiency. Indeed, Ramo 33 brings no incentives for improving subnational performance or mobilizing additional local resources for education, which are two goals of the decentral- ization processes. Health Decentralization in the health sector started earlier and proceeded more grad- ually, allowing states and municipalities more time to learn. Six features of the system today make it work relatively well: 1. Resource allocation from the federal level to states according to well-rec- ognized criteria; 2. Fundability for states and sanitary jurisdictions in the use of health sec- tor funds; 3. Careful and effective transfer of human resources from the federal level to the states, negotiated with the national union; 4. Complete transfer of infrastructure, goods, and equipment to states; 5. Municipal cofinancing of new infrastructure and participation in planning; 6. Consejo Nacional de Salud has resolved many conflicts and monitored the process of decentralization. OVERVIEW 13 states, given their political and economic power and the scope of service delivery responsibilities delegated to them; (b) inefficient and nontrans- parent basis for allocating sectoral transfers; (c) remaining ambiguities in the assignment of expenditure responsibilities and failure to take account of dis- parities between municipalities in their capacity; (d) unclear responsibility and therefore generally inadequate funding for pensions (elaborated in the debt and borrowing section); and (e) lack of effective procedures for dispute reso- lution (discussed above). State Autonomy to Increase Efficiency Although more resources pass through their books than ever before, states in some ways have less fiscal autonomy now than they had a decade ago. With the exception of the Fund for Strengthening State Finances and the Program of Support for Strengthening the Federal Entities, most of the resources they receive from the federal level (including from shared rev- enues) are effectively earmarked either for transfers to municipalities or social sector salaries, especially schoolteachers and university professors with federally mandated pay scales, or they are claimed by the terms of matching grant programs (pari passu). This situation is inconsistent with the growing economic and political power of the states. The Fund for Strengthening State Finances was set up within Ramo 23 in the wake of the 1995 crisis and disbursed money on the condition of a state's fulfilling a fiscal adjustment program agreed with SHCP. Although the agreements for these funds were not transparent or by uniform formula, they represented an advance of the earlier pattern of the President per- sonally allocating major resources to states, also via Ramo 23. In 1998-99, as the fiscal reform agreements ended, the fund was phased out, with a residual going to a national disaster relief fund with transparent rules of access. In 1999, a few incidents, especially with Chihuahua and Nuevo Leon, demonstrated power of states to demand ad hoc resource transfers-debt relief or federal spending on state-priority projects. In the budget for 2000, the Program of Support for Strengthening the Federal Entities was creat- ed in Ramo 23, but with Congress allocating the funds to states, not at exec- utive discretion. It remains to be seen how much the nominal earmarking of the fund will be binding, or whether the funds will be fungible and rel- atively free for the states to spend. But freedom to allocate funds will not be enough. A more fundamental way to increase autonomy of the states would be to give them greater control over spending, especially for per- sonnel, and over taxes, as discussed below. Efficient and Equitable Basisfor Sectoral Transfers Transfers for education and health are allocated mainly on the basis of the number of staff and physical sites (schools, clinics, hospitals). In the educa- tion sector, staffing in 1992 is the base from which allocations are adjusted 14 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO at the margin for inflation and population growth, and sometimes increased further by special request. Although this system does not explicitly penal- ize states that have controlled costs since 1992, neither does it penalize those that fail to improve efficiency, and it actually hurts states that improve stu- dent attendance rates, in the sense that they have fewer resources per attending student. In health, the allocation is still based mainly on the cur- rent number of establishments and staffing, but the system has improved since 1996. Now, requests by the states for incremental health resources are evaluated by SHCP and the Ministry of Health on the basis of demo- graphic and health data. It is understandable that the initial values of the transfers matched previous cost levels, because the states had neither the authority nor the time and money to change staffing quickly. But now it makes sense to move to a system with better incentives. Evidence from a number of countries indicates that formulas for sectoral transfers are more equitable and encourage efficiency better if they are based on a capitation formula, with adjustments made for population density, age, and gender. Moving toward such a system, as the social security system is starting to do, would improve the efficiency and equity of transfers. Assignment of Responsibilities The broad assignment of responsibilities by sectors in Mexico is reasonable: education and health to the states, local streets and sanitation to munici- palities, and other sectors to the federal government. The problems are in the details. Concurrent obligations at all three levels and especially between the federal and state governments in important services such as education, health, or social assistance often leave functions like maintenance, regula- tion, and inspection without any level taking adequate responsibility. Responsibility is ill-defined for roads, water, and higher education. The water issue is especially problematic, because it has so many dimensions in terms of users and geography. A separate study on water is needed to clarify the issues, only some of which relate to decentralization. Although some states have shown strong interest in local environmental issues where their involvement would make sense, the mandate and funding remain largely at the federal level. A pilot program in the 2000 budget offers some federal funding to encourage states to develop their environmental sector capacity and cost recovery. Federal involvement in personnel issues remains problematic. The municipal level has few assigned responsibilities, especially in social sectors. Individual schools and hospitals have little autonomy or community involvement, although international experiences in Colombia (the New School Program), Nicaragua (the Participatory Fund for Maintenance of Social Infrastructure), and El Salvador (the EDUCO Program) indicate that this part of decentralization does the most for per- formance on various indicators: attendance, test scores, and cost control. OVERVIEW 15 The thorny problem of nationally negotiated teachers' salaries probably cannot be solved with direct confrontation and will not need to be. The fun- damental problem is not that their pay is too high. As Mexico's private sec- tor grows and its labor market gets more integrated domestically and more closely linked with the rest of the world, the most serious problem will be that a national salary does not attract qualified teachers in the areas with the fastest private sector growth and the most international linkages. These areas will have to pay more or witness a decline in quality, because only less-qualified personnel will stay or they will spend more of their time and energy in outside employment. If the federal government cannot afford to pay for wage increases nationwide, it may need to give local governments and voters the option to raise more resources locally to improve pay and working conditions above the national (minimum) standard. At the other end of the spectrum, isolated areas are already having trouble attracting enough teachers that meet the usual national standard, even though the national salary is high compared to local wages. The federal transfer sys- tem may need to provide extra support there. Faced with these economic realities, the teachers may see the necessity and benefit of having their union focus less on negotiating national raises and more on setting standard negotiating frameworks for local branches. Separating the assignment of responsibilities for maintenance and oper- ation of infrastructure facilities from responsibility for capital investment has been a persistent problem in Mexico. For example, municipalities are in charge of maintaining school buildings, while the federal and state gov- ernments carry out most capital investments. This dichotomy often leads to insufficient levels of both maintenance and infrastructure investment. Each level of government can blame the other for not doing its part, and each has an incentive to refrain from using its own resources, with the expectation that the other levels will give more. Rigid separation of con- struction, maintenance, and operation decisions has led to notorious inef- ficiencies in allocation and production, as evidenced by all Latin American countries that transferred building responsibilities to specialized funds while keeping operation under traditional sector ministries and leaving maintenance responsibilities unspecified. Subnational Taxes and Revenue Subnational governments in Mexico should raise more taxes for three rea- sons. First, the public sector needs more tax revenue for when revenues from oil decline, not only temporarily because of low oil prices but also per- manently as a share of gross domestic product (GDP) because of econom- ic growth and diversity. Second, to give citizens more control over the size of public spending within their jurisdictions, subnational governments 16 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO need to have adequate taxing authority. If they want to spend more money in an area, they can raise taxes; if they decide to save money on a program, they can reduce taxes. This involves, primarily, having control over mar- ginal (incremental) revenue-what a subnational government can affect by its own actions, especially by changing tax rates, but also by imposing new taxes or repealing old ones, by changing the tax base, and by varying administrative effort. Third, and related to the second, states that borrow need to have some source of funds to repay their debts. If they experience an unexpected adverse shock, relative to their fiscal plan at the time of bor- rowing, they need to be able to raise additional revenue (or cut costs) by enough to sustain their debt service. States typically have their own payroll taxes, annual taxes on automo- biles, and fees. Currently they are legally prohibited from taxing interstate trade and some excises. Taxes represent only about 4.5 percent of total rev- enues of states (other than the Federal District). Payroll taxes are the most important, levied in 23 of the 32 states. Rates range from 0.5 to 4 percent, with most rates in the range of 1 to 2 percent of payrolls (18 states). Only one state has a rate above 2 percent. The definitions of taxpayers are gen- erally similar, but tax bases and exemptions differ substantially across states. The structure of these taxes is not consistent with the bases for the federal payroll taxes that finance social security and housing; making them more consistent would reduce the costs of enforcement and of compliance (and audit) for firms operating in more than one state. The federal personal income tax has a different, narrower, base than the payroll taxes. Moreover, each state administers its own payroll tax, following its own procedures, which are independent of those used for federal payroll taxes. Inconsistent tax bases and duplicate administration are wasteful practices for both tax- payers and administrators. The administration and, for most states, the rate of the payroll tax, both offer substantial opportunities to increase states' own revenues. Most states seem to lack the incentive, however, because they have too little control over how they spend money and relatively too much opportunity to negotiate discretionary transfers, at least prior to 1999. Most of their resources, however, come from formula-based transfers, mostly unconditional. See figure 0.1, which is illustrative and based on prelimi- nary estimates for 1999, as described in Chapter 5. Municipal governments are on average only slightly less dependent than the states on revenue sharing and transfers. In the aggregate, munic- ipal governments receive about two-thirds of total net revenues from these sources. This pattern differs markedly across and within states. The incen- tive structure in revenue-sharing arrangements partly accounts for low rates of tax collection. At least the larger municipal governments would benefit from having access to additional sources of revenue and from improved incentives. The main sources of revenue for municipalities are taxes on OVERVIEW 17 Figure 0.1. Composition of Total Subnational Revenues, 1999 Pesos per capita 7,000 * Own revenue U Unconditional B Conditional 6,000 5,000 4,000 3,000 2,000 1,000 l lu_li i WI I 1 1 1 I < < t ° - ° u 2 m ° °°8b~~~~~~~~~~~~~~~~~~N Estados clasificados indice de pobreza Futentc: Courchene y Diaz (capitulo 5). base (o bases) gravable definida por el gobierno federal y administrada por ellos, solos o en forma mancomunada con los estados. En este esquema, los gobiernos subnacionales ejercerian el importante derecho prioritario- desde el punto de vista politico-de elegir tasas impositivas, al mismo tiempo que evitan la mayor parte de la complejidad, las desigualdades, la exportaci6n de impuestos y las distorsiones por ubicaci6n que generan una legislaci6n y administraci6n subnacionales independientes. Es decir, los recargos tributarios combinan la sencillez de la participaci6n en los impuestos con la ventaja de permitir que los gobiernos subnacionales deci- dan sus tasas impositivas. Canada utiliza este sistema en forma conside- rable. Los recargos de ese tipo podrian aplicarse a impuestos al consumo, al impuesto sobre la renta de acuerdo a la residencia, o incluso al impuesto al valor agregado (gravado en los destinos de venta). Esto ultimo podria lograrse con un impuesto al valor agregado compensatorio. Los estados auin podrian imponer impuestos sobre la n6mina, preferentemente con una base PERSPECTIVA GENERAL 59 unificada. Para que cualquier impuesto concatenado de ese tipo funcione para los estados en Mexico, el nivel federal necesitaria mejorar el disefno y aplicaci6n de su impuesto correspondiente, el impuesto al valor agregado, o el impuesto sobre la renta. Los impuestos al consumo, por ejemplo sobre bebidas alcoh6licas o productos del tabaco, son buenos candidatos para asignaci6n a los estados, como se hace en muchos paises. Los gravamenes estatales al consumo deben ser impuestos por (o a favor de) los estados donde se hace el consumo, no por los estados donde se producen o se importan los productos en cuesti6n. Como los impuestos al consumo son relativamente visibles, ayu- darian a garantizar la responsabilidad de los funcionarios estatales. El go- bierno federal podria imponer minimos, que las tasas estatales al consumo no podrian rebasar, a fin de evitar una "carrera para llegar al punto mas bajo" provocada por la competencia entre los estados al buscar atraer la venta de productos que saben son pensados para pasar de contrabando a otros esta- dos, y proteger a los comerciantes locales contra la competencia de pro- ductos introducidos fraudulentamente de estados con bajos impuestos. Los estados que apoyan a sus municipios administrando el impuesto pre- dial-recopilando informaci6n para el catastro, y actualizandolo-podrian negociar alguna participaci6n en los ingresos con los municipios, con f6r- mulas que remuneren a los estados por su apoyo. En Colombia existen acuerdos similares, donde el Departamento de Antioquia maneja un impuesto para usos muiltiples destinado a sus municipios. En El Salvador, los municipios mas grandes y asociaciones municipales mantienen la re- gistraci6n y las cuentas corrientes para los municipios mas pequenios. En Mexico, el gobierno federal podria apoyar acuerdos de este tipo patroci- nando programas experimentales y divulgando las lecciones que han deja- do las experiencias nacionales y extranjeras. Como un enfoque practico para simulaci6n y analisis de politica, es util tomar en cuenta las medidas de reforma que serian neutrales en cuanto a ingresos, sustituyendo las transferencias del gobierno central por ingresos de los estados e impuestos municipales. Este tipo de medidas (a) reducirian los impuestos federales para proporcionar un "espacio fiscal" que los go- biernos estatales y municipales podrian o no ocupar, a su elecci6n; (b) reducirian las transferencias totales a los estados en la cantidad en que dis- minuyan los ingresos federales; (c) asignarian impuestos a gobiernos estatales y municipales que, en promedio, podrian sustituir las transfe- rencias perdidas si se gravara a tasas similares;10 y (d) se ajustarian las trans- ferencias para mantener niveles de financiamiento para gobiernos estatales especificos, si utilizan el espacio fiscal proporcionado. Sin embargo, seria conveniente que Mexico disminuya la dependencia del sector publico agre- gado de los ingresos petr6leros. En ese contexto, seria uitil que parte del esfuerzo fiscal nacional ocurra en los niveles subnacionales. 60 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Transferencias intergubernamentales Las transferencias intergubernamentales cumplen diversas funciones en los sistemas federales. Primero, equilibran la interacci6n entre las necesidades de gasto y el ingreso de los gobiernos subnacionales. Segundo, integran el federalismo fiscal y las dimensiones social y polItica de las federaciones. Por ejemplo, establecer mas condiciones a las transferencias gubernamentales es sefial de un aumento en la centralizaci6n de la federaci6n y viceversa; menos condiciones significa mas descentralizaci6n. Los enfoques alterna- tivos a la cuesti6n de las transferencias ofrecen la flexibilidad de generar un grado deseado de equidad horizontal en todos los gobiemos subna- cionales y pueden incorporar incentivos importantes aumentando la responsabilidad o estimulando el esfuerzo fiscal. Tambien son el principal metodo para llevar a cabo los prop6sitos de la sociedad en lo que concierne a la naturaleza del Pacto Fiscal y al grado de igualdad de oportunidades entre los ciudadanos que viven en diferentes estados. Mexico ha experimentado un sorprendente grado de descentralizaci6n en los ultimos afhos, facilitado por un gran aumento en el nuimero y diver- sidad de transferencias. Aunque podrian reconsiderarse aspectos selec- cionados de estos nuevos planes de transferencias, la revoluci6n en gasto y transferencias en el federalismo fiscal mexicano es impresionante. Las dos principales categorias de transferencias son las llamadas parti- cipaciones y aportaciones. Las participaciones eran originalmente ingresos de estados y municipios cuya recaudaci6n se deleg6 al nivel federal mediante el Pacto Fiscal por cuestiones de eficiencia tributaria. De acuerdo con la ley, la federaci6n s6lo recauda esos impuestos y los distribuye a sus duenios. En la practica, el gobierno federal redacta la f6rmula para la distribuci6n de estos fondos y los aumenta con ingresos de fuentes federales, como los del petr6leo, asi que son diferentes a las participaciones de impuestos (donde las rentas recabadas en un estado permanecen ahi) y mAs bien son como un programa de transferencias de tipo general de participaci6n de ingre- sos. La mayorfa de estas transferencias se expiden bajo el Ramo 28. Las aportaciones se concibieron como dinero federal destinado a pagar obliga- ciones que eran previamente federales, y que es transferido a los estados y municipios junto con esos compromisos (por ejemplo, educaci6n y salud). Estos fondos, que anteriormente pertenecian al Ramo 26, ahora proceden a traves del Ramo 33. Las transferencias-a los estados-de participaciones de ingresos den- tro del Ramo 28 (sin contar los ramos que pasan directamente a los munici- pios o fondos del Ramo 33 destinados a maestros y trabajadores del sector salud) casi sextuplicaron los ingresos propios de los estados en 1996. Las transferencias estatales y federales a los municipios conforme al Ramo 28 casi duplicaron los ingresos propios de los municipios. Incluyendo el Ramo 33, los ingresos totales de los estados representaban cerca del 6 por ciento PERSPECTIVA GENERAL 61 del producto interno bruto y las rentas totales de los municipios alrededor del 1.5 por ciento. Esos porcentajes se incrementaron de forma casi continua hasta el presente. Tambien hay fondos del Ramo 33 que no se relacionan con ingresos ante- riores de los estados ni con obligaciones previas del gobierno central. En otro tiempo, hasta cierto punto eran a discreci6n del presidente, pero despu6s todos se dedicaron al Fondo para Fortalecer las Finanzas del Estado que estados especificos negociaban con secretarias federales, la mayoria de las veces la Secretaria de Hacienda y Cr6dito Puiblico. En 1998 y 1999, las transferencias discrecionales conforme al Ramo 23 cayeron a menos de una d6cima parte de su valor anterior a 1995. En el presupuesto del anio 2000, las transferencias a discreci6n del ejecutivo terminaron com- pletamente. Los presupuestos de las diferentes secretarias de estado tambien incluyen transferencias importantes. Algunas van a los estados como fondos de contrapartida conforme a acuerdos estado por estado (descritos en la sec- ci6n institucional). Estos son muy importantes en los sectores de la salud, el vial y el educativo. Las diferentes secretarias tambien invierten recursos directamente para proyectos en determinados estados, que pueden ser importantes transferencias implicitas. Estos dos tipos de transferencias tradicionalmente se negociaban con los gobernadores, sin hacer del conocimiento publico los terminos de la operaci6n. En el afno 2000 la Secretaria de Hacienda y Credito Piblico publicara las reglas oficiales para acceso a los programas paralelos (pari-passu) y el porcentaje de dona- ciones de contrapartida necesarias para todos los principales programas. En Mexico, al igual que en muchos paises, los prop6sitos de la sociedad en cuanto a las transferencias cambian continuamente y son poco claros. En el presupuesto en 1999, habia por lo menos 10 programas de transfe- rencias muy importantes, contra 3 en 1997. En el presupuesto del afio 2000, el Congreso agreg6 otra transferencia importante en el Programa de Apoyo para Fortalecer las Entidades Federativas, que es asignada por el Congreso en la ley del presupuesto en vez que a discreci6n del ejecutivo. Cada pro- grama aborda multiples objetivos, y cada objetivo se consigna en diversos programas. Esto hace innecesariamente dificil para los gobiernos subna- cionales calcular cuanto dinero recibiran; y para el gobierno federal deter- minar cuan bien esta atacando sus objetivos el sistema de transferencias. La constituci6n y la ley de participaci6n de ingresos no dan fines especifi- cos a las transferencias, pero su modelo actual revela algo sobre sus inten- ciones. Los principales objetivos de las transferencias en M6xico parecen ser: (a) dejar a los estados participar en el mayor potencial del gobierno federal para recaudar ingresos; (b) subvencionar la prestaci6n de servicios de los gobiernos subnacionales que conllevan externalidades nacionales (salud basica y educaci6n); (c) consolidar la autonomia de los municipios, y (d) proporcionar mas recursos a estados con un elevado indice de 62 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO pobreza. Un quinto posible objetivo seria compensar a los estados con una base impositiva baja per capita, como hace Canada. Las transferencias mexi- canas no atacan este objetivo de manera explicita, aunque lo hacen en forma parcial mediante transferencias seleccionadas para combatir la pobreza, ya que existe una correlaci6n negativa entre esta y el tamanio de la base impositiva per capita. En cuanto al primer objetivo, el gobierno federal hace transferencias a los estados y municipios bajo el Ramo 28 a fin de reforzar su capacidad de recaudaci6n, que es d6bil comparada con sus mandatos de gasto. Parece que los estados reciben considerablemente mas 6rdenes que fondos, y parece que los municipios reciben mAs fondos que 6rdenes (sobre todo los pequefios). Cerca de la mitad de las transferencias para los estados son fon- dos para fines especificos bajo el Ramo 33; la mayoria de las veces los fon- dos no condicionados proceden del Ramo 28. Ademas, parte de los fondos "no condicionados" tienen que usarse para emparejar fondos de otros pro- gramas federales o para pagar sueldos decretados federalmente que no cubre el Ramo 33. La poblaci6n del estado es un factor importante en las f6rmulas para algunos fondos (Fondo General de Participaciones y Fortamun), y unos cuantos (sobre todo el Fondo de Fomento Municipal, Tenencia y Ramo 28 Participaci6n de Impuestos sobre el Consumo) procu- ran compartir los ingresos con los estados en base a su origen. En el pasa- do, cuando los estados podIan argumentar que las rentas eran inadecuadas, el gobierno federal a menudo aumentaba las transferencias, aunque este planteamiento puede estar llegando a su fin. Estas transferencias para un prop6sito determinado no eran automaticas ni ajenas a la politica. En cam- bio, para los municipios un poco mas de la mitad de las transferencias fe- derales (al pasar intactas por las cuentas estatales) son participaciones de ingresos no condicionadas y mas de la mitad del resto (transferencias del Ramo 33) tampoco son condicionadas. En cuanto al segundo objetivo, muchas transferencias federales se de- signan a sectores especificos, algunos, como la educaci6n y la salud, aparentemente porque tienen externalidades a nivel nacional. La selecci6n de transferencias por sector a los municipios (secciones 3-7 del Ramo 33) parece ser un plan de transici6n que busca lanzar servicios municipales descentralizados con el mismo nivel de financiamiento que tenian conforme al sistema anterior. En cuanto al cuarto objetivo, las f6rmulas de los fon- dos para obras de infraestructura en el Ramo 33 incluyen importantes fac- tores destinados a combatir la pobreza. Cuando se suman todos los programas, estado por estado, la cantidad de recursos disponibles per capita para los estados (el total de sus propios ingresos y todos los diferentes programas de transferencias) varia en forma considerable (la mayoria de $2.500 a $4.500 pesos mexicanos, estimados para 1999, con tres estados recibiendo casi $6.000 pesos). Sin embargo, los PERSPECTIVA GENERAL 63 recursos per capita aproximadamente son los mismos en promedio para los estados con niveles de pobreza medios y elevados; y los estados mas ricos (con bajos niveles de pobreza) s6lo tienen un poco mas de los recursos promedio. Los elevados ingresos per capita en estados ricos con poca pobreza se contrarrestan en promedio por las transferencias destinadas a combatir la pobreza que se hacen a los estados mas pobres. Esta uniformi- dad de recursos gubernamentales per capita en todos los niveles de ingre- sos puede reflejar un valor social basico en Mexico. La igualdad nominal en todos los niveles de ingreso corresponde a cierta progresion en termi- nos reales, porque el costo de los servicios piublicos generalmente es menos costoso en los estados mas pobres, entendiendose que los mismos pesos per capita compran mAs. El actual sistema de transferencias tiene cuatro importantes problemas: (a) es demasiado complejo para lograr un conjunto de prop6sitos coherente. Por lo tanto, el gobierno federal distribuye mAs recursos de los necesarios para alcanzar sus objetivos, y el trato que se da a los estados no es equitativo. Los principales beneficiarios de tal complejidad son los estados que obtienen mas recursos de los que recibirian de acuerdo a un sistema sencillo y trans- parente; (b) al fundamentar muchas transferencias en valores iniciales o de reposici6n, en vez de hacerlo en base a costos per capita o por posibles be- neficiarios, es injusto y no estimula la eficiencia; (c) por lo menos hasta 1999, la disponibilidad de transferencias para prop6sitos determinados desti- nadas a estados politicamente favorecidos debilitaba los incentivos para administrar bien el gasto y aumentar los ingresos; y (d) los requisitos de asignaci6n y las donaciones obligatorias de contrapartida dejan a los esta- dos con poca autonomia para aumentar su eficiencia o ajustar las partidas de acuerdo a sus necesidades. Otros paises latinoamericanos tienen problemas similares. Las f6rmu- las de distribuci6n de Colombia para la participaci6n de ingresos general son muy complejas para transmitir seniales transparentes e incentivos efi- caces a los gobiernos subnacionales. Las f6rmulas incluyen cuatro criterios (per capita, por unidad territorial, esfuerzo fiscal, indice de pobreza) ademas de combinarlos. Algunos de estos aplican a indicadores que no son facil- mente asequibles a los gobiernos subnacionales. En consecuencia, los go- biernos subnacionales no comprenden del todo las intenciones del nivel nacional, ni reaccionan oportunamente a las sefiales e incentivos. La forma en que el sistema federal mexicano evolucione en si mismo determinara el futuro del subsistema de transferencias-gastos-impuestos. Existe una gama de posibles modelos. Con el modelo de tipo canadiense- federalismo legislativo total-las transferencias se caracterizarian por: * Un programa de igualamiento para garantizar que todos los estados tengan cierto nivel minimo aceptable de ingresos no condicionados 64 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO (es decir, ingresos propios mas igualaci6n). Este podria ser el nivel del promedio nacional, por ejemplo. * Transferencias condicionadas basadas en f6rmulas, pero s6lo donde hav evidentes externalidades nacionales. Probablemente estos serian iguales per cdpita en una medida considerable, ya que el equilibrio horizontal en todos los estados se corregirIa conforme al programa de igualamiento. * Condiciones vinculadas con esta transferencia como un conjunto de principios socioecon6micos y relacionadas a areas de gasto como la educaci6n y la salud. La idea seria proporcionar considerable arbitrio a los gobiernos subnacionales en la aplicaci6n de estos principios. * Un "acuerdo" federal-estatal o un pacto fiscal sobre la union sociecon6mica, con abundantes estipulaciones para solucionar con- troversias. Una versi6n de ese tipo de acuerdo extenderia los meca- nismos para la apelaci6n y resoluci6n de controversias a los ciu- dadanos mexicanos. Con el modelo de descentralizaci6n aleman-federalismo administra- tivo-un porcentaje considerable de los recursos estatales de fuentes propias vendria a trav6s de la participaci6n de ingresos, pero aun perdu- rarIa un importante desequilibrio vertical en la federaci6n. Este podria atacarse mediante un conjunto de transferencias condicionadas disefiadas para corregir los desequilibrios horizontal y vertical a la vez. Las condi- ciones para estos subsidios serIan de dos tipos. Primero, la legislaci6n con- cerniente al area del gasto podria ser federal delegando su aplicacion a los gobiernos subnacionales o; segundo, las donaciones mismas podrIan ori- tarse para areas especificas de gasto con condiciones asociadas. Con el tiempo, estas condiciones podrIan ser menos estrictas a fin de permitir mas autonomia subnacional en el frente de gastos, pero siempre habrIa mas ingresos bajo control federal que en el otro modelo en que los gobiemos sub- nacionales tienen acceso a rentas propias no condicionadas. El modelo de federalismo administrativo representaria una evoluci6n del modelo me- xicano actual, pero tambien podrIa ser una transici6n oportuna hacia un fede- ralismo legislativo, si es por lo que Mexico se decide a la larga. Endeudamiento y Administraci6n de la Deuda Subnacional La deuda subnacional en Mexico no ha sido un problema macroecon6mi- co nacional, pero podria llegar a serlo. Mas significativo, la carga de la deuda ha sido un problema fiscal para algunos estados (la deuda municipal es PERSPECTIVA GENERAL 65 pequenta, excepto para el Distrito Federal, que de hecho es un estado), y la forma en que el gobierno federal trat6 la deuda subnacional en el pasado ha creado incentivos inadecuados para los asuntos fiscales de los estados. A fin de que los estados tengan un incentivo para controlar costos y aumen- tar rentas-una meta que se han fijado los gobiernos federal y estatales- es importante que ni los estados ni los prestatarios esperen que el nivel fe- deral los saque de apuros. De lo contrario, los emprestitos se convertirian en un instrumento mediante el cual los estados puedan obtener recursos federales extras, transferidos a ellos directamente o a sus acreedores. En muchos casos, esta ha sido la practica en Mexico. (Todos los estados reci- bieron rescates financieros a raiz de la crisis econ6mica de 1995, y desde entonces se ha sacado de apuros a unos cuantos estados al vencer las garantias sobre proyectos de grandes obras de infraestructura.) Cambiar a una nueva practica sin rescates financieros federales requerira no solo cambiar las reglas, sino tambien garantizar que los desafios planteados a las nuevas reglas no sean abrumadores, sobre todo durante la transici6n, y que estas cuenten con suficiente respaldo politico. La mayor parte de los emprestitos de los estados y municipios se ha respaldado con las participaciones estatales corno garantias. El gobierno federal (Secretaria de Hacienda y Credito Piblico) manej6 el proceso de garantias, asi que los bancos y sus instituciones reguladoras trataron la deuda practicamente como si fuera sin riesgos. Ademas, el gobierno fede- ral no siempre fue riguroso disminuyendo las participaciones de algunos estados cuando los acreedores ejercian sus derechos sobre la garantia, lo que significa que el resto de la federaci6n tenia que pagar. Por lo tanto, ni los estados ni sus acreedores se preocupaban mucho sobre la capacidad para cancelar obligaciones. Esto contribuy6 a la crisis de deuda estatal durante la crisis general de 1995. El gobierno federal ha sido la uinica parte interesada que tiene motivos s6lidos para preocuparse sobre la verdadera solvencia de los Estados; en ocasiones se ha negado a registrar y garanti- zar sus deudas; pero en otras, los factores politicos o de otro tipo permi- tieron el endeudamiento excesivo. Despues de la serie de rescates financieros de 1995, para evitar que volvieran a suceder, la Secretaria de Hacienda y Credito Piublico trat6 de terminar su parte en cuanto a garantizar deuda local y estatal; pero los esta- dos y los bancos comerciales-ain traumatizados por la crisis-no se pusieron de acuerdo en planes de fideicomiso aceptables para ambas partes para manejar el proceso de garantias. Asi que la Secretaria de Hacienda y Credito Publico convino, de manera provisional, aceptar la comisi6n de los estados de servir de fideicomiso para operaciones de endeudamiento que ellos-la SHCP-habian autorizado. En el afno 2000, terminara esta funci6n de la SHCP, y en la actualidad esta trabajando con los bancos y los estados 66 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO para garantizar que se aplique un marco normativo para fideicomisos estrictamente privados a fin de garantizar la deuda de los Estados con las participaciones, si eso es lo que quieren las partes. De los muchos problemas politicos de imponer una energica restricci6n presupuestaria a los estados, el mas dificil surge a causa de la cuantiosa pro- porci6n del gasto estatal dedicada a pagar por salarios. Esto crea una fuerte presi6n politica para usar rescates financieros federales y evitar recortes con- siderables en el gasto en vez de pagar intereses de la deuda. Resolver el problema no s6lo requerira la firme determinaci6n del gobierno federal para evitar rescates fiancieros, sino tambien mas autoridad fiscal y respon- sabilidad politica por parte de los estados para recaudar mas rentas (para pago de intereses de la deuda o de salarios), asi como control de costos. En otras palabras, habra una interdependencia entre restricciones presupues- tarias que motiven la responsabilidad fiscal por parte de los estados y su verdadera autonomia fiscal, algo que hara politicamente realistas dichas restricciones. La deuda subnacional se elev6 de $27 mil millones de pesos mexicanos en 1994 a $71,6 mil millones en 1998. La crisis financiera del bienio 1994-1995 y el aumento en las tasas de interes incrementaron la deuda real pendiente de los estados, pero 6sta se redujo de manera considerable por el paquete de rescate financiero puesto en marcha por el gobierno federal en el periodo 1996-1997. En 1997, la deuda subnacional apenas nlegaba a cerca del 6 por ciento del total de la deuda publica y al 2 por ciento del pro- ducto interno bruto (PIB). A diferencia de otros paises, en Argentina la deuda subnacional era de alrededor del 5 por ciento del producto interno bruto; casi 20 por ciento en Canada, y un poco mas por encima de eso en Brasil. Gran parte del endeudamiento subnacional en Mexico ha sido con Banobras, un banco de fomento de propiedad federal. Los estados con altos coeficientes de endeudamiento y una reiterada historia de necesitar rescates fiancieros han seguido recibiendo prestamos de Banobras, pero presunta- mente esto sera un problema minimo con la introducci6n de practicas de cr6dito mas rigurosas y el final de las garantias federales implicitas. El costo de los fondos de Banobras es superior al de los bancos comerciales (porque Banobras no puede recibir dep6sitos), que han estado captando una por- ci6n cada vez mayor de los negocios subnacionales. Ninguna de las deu- das es de tipo externo porque la constituci6n prohibe a los estados tomar prestamos en moneda extranjera o de acreedores extranjeros. Desde 1995, los empr6stitos de la mayoria de los estados han sido mi- nimos o cero; algunos han liquidado su deuda. Sin embargo, unos cuantos han adquirido muchos compromisos y la deuda es una carga para ellos porque tienen muy poco ingreso disponible para pago de intereses. La pro- porci6n del stock de deuda con respecto a los ingresos disponibles, definidos como impuestos propios mAs transferencias no condicionadas, PERSPECTIVA GENERAL 67 comprende desde un maximo de 1.8 (en Sonora) a un minimo de 0.02 (en Hidalgo). Aunque el Distrito Federal tiene la deuda total mas alta, cuenta entre los menos endeudados en relaci6n con el ingreso disponible porque recauda impuestos considerables. Los ocho estados mas endeudados son Baja California Sur, Jalisco, el Estado de M6xico, Nuevo Le6n, Queretaro, Quintana Roo, Sinaloa y Sonora, todos con una proporci6n de obligaciones con respecto a sus ingresos disponibles superior a 1.11 Los mismos estados son los mas endeudados ademas de Baja California Norte, si se usa otra definici6n de ingresos disponibles: ingresos totales menos pagos de salarios (descentralizaci6n de personal, del sector educativo y de salud) y deducien- do las transferencias a los municipios. La principal diferencia es que el coe- ficiente de endeudamiento es mayor para estados como Nuevo Le6n que utiliza gran parte de sus propias rentas para pagar salarios. Segin esta medida, 12 estados tenian stocks de deuda superiores a sus ingresos disponibles en 1997. La deuda de los gobiernos subnacionales de Mexico son considerable- mente menores que sus deficits fiscales acumulados anteriores. El gobier- no federal ha cubierto repetidas veces una parte considerable del d6ficit fis- cal de los estados, mediante transferencias extraordinarias discrecionales (para cubrir aumentos salariales no previstos con anticipaci6n, desarrollo de inversiones, y otros conceptos) y otras formas de rescates financieros como el de 1995 y transferencias para prop6sitos especificos para dis- minuir y reprogramar la deuda. El saldo primario de los gobiernos sub- nacionales en M6xico fue positivo durante el bienio 1995-1997, a diferen- cia de los deficits en el periodo 1992-1994 (consulte la grafica 2). Sin embargo, esto se debi6 principalmente a un aumento de transferencias extraordinarias del nivel federal. Si se descuentan estas transferencias extraordinarias, los estados aiun tenian un d6ficit primario en 1995-1997 y, por supuesto, un deficit general mas cuantioso. La Secretaria de Hacienda y Credito Publico realizaba acuerdos anuales de ajuste fiscal con cada uno de los estados (excepto el Distrito Federal) a causa de los rescates financieros llevados a cabo despues de la crisis de 1995. El programa federal mas importante para estas transferencias (el de Fortalecimiento Financiero de los Estados en el Ramo 23) fue excluido del presupuesto para el afio 1999, y otras partidas del Ramo 23 se redujeron mucho. En el futuro, sera impor- tante observar c6mo se las arreglaran los estados y si el gobiemo federal mantendra su compromiso de terminar con los rescates a los estados. Existen tambien importantes pasivos contingentes de los gobiernos sub- nacionales, como las pensiones y los planes de servicios medicos para sus empleados. Ademas, el Distrito Federal, los estados y los municipios pro- porcionan garantias sobre prestamos a sus respectivas dependencias descentralizadas y empresas puiblicas. Existe registro consolidado de estos pasivos s6lo cuando se contabilizaron para utilizar las participaciones del 68 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Grafica 0.2. Mexico: Deficit Primario De Los Estados 1989-1997 Precios de 1997 (pesos) 4.000.000 2.000.000 !. - 0 1990 \991. 1992 1993 1994 .1995 1996 1997 2.000.000 1989 4.000.000 6.000.000 . 8.000.000 10.000.000 12.000.000 Anio - * 4. *BAL. PRIMARIO REAL -mom BAL. PRIM. REAL (Excepto extraordinarias transferencias) Fuentes: SHCP y cifras estimadas por el Banco Mundial. estado en cuesti6n como garantia. En unos cuantos casos los incumpli- mientos de las entidades descentralizadas fueron lo suficientemente cuan- tiosos para inducir un rescate federal, como ocurri6 con el metro de Monterrey y la carretera de cuota en Coahuila. Para la mayorfa de los lugares, es probable que las companias de agua sean deudores morosos. Los sistemas de pensi6n de los empleados publicos de los estados no estan, en su mayorfa capitalizados y pocos estados tienen capacidad fiscal para operarlos atn como sistemas de reparto ("pay-as-you-go"). Nuevo Le6n reform6 su sistema de pensiones de manera independiente, pero fue una reforma incompleta y sigue siendo una excepci6n. Un estimado par- cial de la deuda contingente de los estados por concepto de pensiones lle- gaba a $167 mil millones de pesos mexicanos en 1997 (cerca del 6 por cien- to del producto interno bruto del pais).12 Entre los 25 estados con suficiente informaci6n para calcular en que afio se agotarian las reservas de sus sis- temas, y tomando en cuenta los porcentajes actuales de aportaciones y otros conceptos, en 1998 cinco estados ya tenian un d6ficit, y se proyectaba que siete mAs podrian quebrar durante el periodo 1999-2002. La mayoria de los gobiernos estatales considera los pasivos por concepto de pensiones como un legado de las practicas anteriores a 1990, cuando los gobiernos estatales estaban, en la pratica, bajo el dominio absoluto del gobiemo federal. Por consiguiente, los gobiernos estatales consideran las pensiones como una obligaci6n implicita del gobierno federal. Al no haber funciones definidas, PERSPECTIVA GENERAL 69 se proporciona un incentivo a ambas partes para que no hagan los tramites pertinentes, con la esperanza de que la falta de tramites hard que, al final, el otro cargue con mas parte del costo. El problema se pasa por alto, pero este crece rapido con la maduraci6n de la fuerza de trabajo. Ademas, cuan- do una deuda considerable de este tipo se vuelve inmanente, se disminuyen los incentivos para que un estado maneje prudentemente los demas asun- tos fiscales. Las experiencias de Argentina y Brasil muestran la necesidad absoluta de atender a las pensiones estatales. Brasil aim tiene que llegar a una res- oluci6n, aunque ya esta analizandose una de manera diligente; Argentina ha puesto bajo la autoridad del gobierno federal la mayor parte de la responsabilidad. Quiza sea mejor separar el disenio de un sistema futuro, donde existe una gama de opciones, de la elaboraci6n de procedimientos para atender los pasivos por pensiones que se formaron en el pasado. Esa elaboraci6n podria ser responsabilidad federal, por lo menos la parte para los trabajadores contratados originalmente por el gobierno federal. En todo caso, la responsabilidad federal necesita limites claros. En el futuro, son los estados quienes deberian encargarse del sistema; de acuerdo a la experiencia, un plan de capitalizaci6n con aportaciones definidas ("defined contribution") tiene mas posibilidades de ser sustentable. En el pasado, las politicas de endeudamiento subnacional de M6xico evolucionaron en un contexto en el que las relaciones fiscales interguber- namentales podian mantenerse encarriladas mediante transferencias para prop6sitos especificos y negociaciones entre integrantes en la misma je- rarquia del partido en el poder. Ahora que los estados y los municipios son mas independientes desde el punto de vista polItico, y en muchos casos son gobernados por partidos de la oposici6n a nivel nacional, el desafio ha con- sistido en dirigir la transici6n hacia un sistema mas estricto y transparente para administrar el endeudamiento subnacional. De lo contrario, los pro- blemas por la deuda podrian debilitar el proceso de descentralizaci6n y, en el peor de los panoramas, llegar a representar un riesgo macroecon6mico para el pais en general. El buen funcionamiento fiscal de los estados parece ser politicamente popular, lo que es una suerte, porque gran parte del pro- ceso de reforma exigira que los gobiernos estatales tomen medidas al respecto asi como una mayor vigilancia por parte del electorado. El go- biemo federal esta fomentando esto al poner en vigor reglas de divulgaci6n financiera (como una condici6n para el registro de la deuda), demostran- do que los gobiemos son entidades piublicas que usan el dinero de los con- tribuyentes, no operaciones privadas con algin derecho a la privacidad. Aunque el anterior sistema por lo menos impedia una crisis macro- econ6mica grave a causa de la deuda subnacional, ya han cambiado algu- nas cosas, la mas importante de las cuales es la mayor independencia politi- ca de los estados y la eliminaci6n de hecho de las transferencias 70 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO extraordinarias. Por consiguiente, hay una transici6n en marcha que requiere mas cambios para que sea sustentable. A partir del 31 de marzo del anio 2000, el gobierno federal dej6 de acep- tar mandatos de los estados para garantizar sus prestamos con participaciones. Por lo tanto, los estados y sus acreedores tendran que instituir un fidei- comiso privado o tomar medidas equivalentes que no incluyan una garan- tfa implicita del erario federal. La Secretaria de Hacienda y Credito Publico ya no participarA de manera directa en el proceso de endeudamiento sub- nacional. En vez de eso, unicamente hard publica una serie de principios generales para explicar cuales serian los planes de fideicomiso o de otro tipo de garantfas adecuados desde un punto de vista legal y financiero; aunque los estados y sus acreedores tendran la opci6n de acordar el plan de garan- tias de cualquier tipo que mas les convenga. A fin de que la transici6n al nuevo sistema sin mandatos sea lo menos complicada posible, durante el primer trimestre del afno 2000 el gobierno federal acept6 mandatos s6lo para estados que estaban ya tramitando una calificaci6n crediticia con recono- cidas firmas calificadoras de valores (consulte mas adelante) De manera simultanea, nuevas regulaciones bancarias (y sus repercu- siones para determinar el costo del credito) crearon recientemente incen- tivos hacia un mejor control financiero y programas de emprestitos mas efi- cientes y mas transparentes por parte de los estados y municipios, y hacia evaluaciones de riesgo subnacional mas acertadas y sin riesgo moral por parte de sus acreedores bancarios. En primer lugar, en 1999 se derog6-ofi- cialmente y en la practica-el Ilamado regimen de excepci6n por medio del cual todo el credito subnacional estaba exento de los requisitos normales para la constituci6n de reservas y los limites de concentraci6n de riesgos. En segundo, los creditos bancarios para estados y municipios estan ahora sujetos a un coeficiente de ponderaci6n punitivo de riesgo del capital (150 por ciento) si no se registran ante la Secretarfa de Hacienda y Credito Piublico, proceso que a su vez esta condicionado a que el prestatario est6 al corriente en la publicaci6n de sus informes fiscales y de deudas y en sus obligaciones de pago con los bancos de fomento del gobierno. En tercero, las nuevas regulaciones introdujeron una ponderaci6n diferencial del ries- go de capital que vincula los requisitos de capitalizaci6n para instrumen- tos de deuda subnacional (y, por consiguiente, la determinaci6n de su costo) con la calificaci6n crediticia del prestatario, conforme al avaluo de dos calificaciones crediticias actuales en moneda nacional, hechas en escala internacional, publicadas y efectuadas por firmas calificadoras de credito conocidas en el ambito internacional.13 Siempre existe cierto peligro de que los estados solo contraten a las agencias calificadoras que les dan las cali- ficaciones favorables, pero al exigirse dos avaliuos de reconocidas califi- cadoras internacionales se atenuara este riesgo. Al usar una escala inter- nacional se aumentara la objetividad y fiabilidad de las calificaciones, ya PERSPECTIVA GENERAL 71 que las agencias calificadores sabrdn que el avaluo que hagan de cualquier estado o ciudad mexicana sera comparado con los que hagan en otras partes del mundo. Como los estados no pueden solicitar prestamos al exte- rior ni en moneda extranjera, no existe riesgo cambiario y por lo tanto es adecuado que las calificaciones sean en moneda nacional. Para ser completo, el endurecimiento de los limites presupuestarios y la reducci6n de riesgo moral requieren un nuevo marco de credito para los bancos puiblicos de fomento porque no tiene el mismo control empresa- rial ni las presiones del mercado que los bancos comerciales. Por lo tanto, cuando los bancos de fomento (especificamente Banobras) otorguen pr6s- tamos a gobiernos estatales y municipales tambien se les exigira que cum- plan con los mismos limites de concentraci6n de credito y la ponderaci6n diferencial del riesgo de capital que los bancos comerciales. Ademas de eso, como asunto de politica empresarial, el gobierno federal (propietario de los bancos de fomento) permitird que los bancos de fomento otorguen presta- mos s6lo a los gobiernos subnacionales que requieran ponderaciones del riesgo de capital inferiores al 100 por ciento.14 Para evitar el coeficiente de ponderaci6n punitivo de riesgo de capital del 150 por ciento, los gobiernos estatales tambien tendran que registrar su deuda ante la Secretaria de Hacienda y Credito Piblico, para lo cual deberan estar al corriente en la publicaci6n de sus informes fiscales y de deuda y en sus obligaciones de pago con los bancos de fomento. De hecho, el segundo de estos requisitos impedira que Banobras capitalice intereses atrasados via nuevos pr6stamos. Las nuevas politicas aplicadas a finales de 1999 y en el afho 2000 estan encaminando a Mexico hacia un enfoque del endeudamiento estatal regi- do por el mercado, que excluye la posibilidad de garantias y rescates fe- derales e incluye calificaciones crediticias internacionales que determinen la ponderaci6n de riesgo de capital y, por consiguiente, el costo y disponi- bilidad del credito bancario para los estados. El gobierno federal tiene una funci6n importante que desempefiar, impulsando y facilitando mejores nor- mas contables y la divulgaci6n de informacion por parte de los estados y municipios, asi como el funcionamiento prudente de los bancos desde el punto de vista financiero. La experiencia en Argentina, Colombia y Estados Unidos muestra c6mo la conducta fiscalmente responsable de los estados y sus acreedores exige la regulaci6n y supervisi6n adecuada del sis- tema bancario. Opciones y Agenda de Trabajo Es util pensar en la agenda de politicas para la descentralizaci6n no s6lo como una lista de medidas para el corto, el mediano y el largo plazo, sino tambien como una serie de etapas que podrian poner en marcha varias 72 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO lineas de reforma. Ya se han realizado algunos cambios importantes, como dar fin a las transferencias discrecionales y a la practica del gobierno fe- deral de retener pagos de intereses de deuda de los estados en favor de los bancos. Proseguir con estas medidas es una parte importante de la agen- da de corto plazo. La experiencia internacional sefiala unos cuantos principios que podrian guiar el desarrollo de politicas para el mediano y el largo plazo: * La autoridad y responsabilidad para la prestaci6n de servicios debe recaer en el mayor nivel local que sea compatible con la eficiencia. * En cada sector debe especificarse con claridad la responsabilidad para desempefiar funciones esenciales, y el nivel de gobiemo respon- sable debe tener acceso a recursos adecuados, incluyendo bases gra- vables, para financiar el servicio * Los gobiernos estatales y locales deben recaudar sus propios recur- sos tanto como sea viable en terminos de eficiencia y equidad, y deben hacerlo de manera que los usuarios paguen los costos de los servicios publicos que consumen. * Los gobiernos deben poder solicitar prestamos si, y s6lo si, son capaces de, y se les exige, pagar los intereses y reembolsar el capital de la deuda completamente. • Los gobiernos y las dependencias que ejerzan el gasto deben entre- gar cuentas claras y transparentes, no s6lo al nivel inmediato supe- rior sino tambi6n, lo mas importante, al p6blico. * El proceso democratico, no uno de tipo tecnico, debe determinar el grado de igualamiento en la distribuci6n de recursos a los estados (y municipios). * Las nuevas perspectivas que se sustenten en estos principios en algunos casos requeriran importantes cambios legislativos y de otros tipos en las funciones y responsabilidades de los gobiernos. M6xico enfrenta algunas disyuntivas importantes sobre el tipo de sistema federal que tendra. El debate publico continuo serfa provechoso, incluso si no se Ilega a una elecci6n definida del sistema. En muchos paises, la opci6n ha surgido de manera paulatina a partir de muchas elecciones de menor trascendencia. Sin embargo, es importante comprender que la decision de avanzar en una cierta dimensi6n-por ejemplo, hacia un mayor uso de los sitemas tributarios subnacionales-afecta la forma en que la descentra- lizaci6n progresara en otras dimensiones, por ejemplo: el total de recursos fiscales, la igualdad de recursos publicos per capita a traves de los estados, la autonomfa local en ejercer el gasto, y el control del endeudamiento. Hoy en dia, los gobiernos subnacionales tienen sistemas tributarios d6biles y dependen de las transferencias del gobierno central, supeditan- dolos al control federal en cuanto a gasto y endeudamiento. M6xico esta siguiendo un modelo cercano al extremo aleman del espectro. Aunque hay PERSPECTIVA GENERAL 73 muchas desigualdades en el financiamiento para los estados, en su ma- yoria se deben a factores hist6ricos y politicos, no a diferencias en la fuerza econ6mica de las economias estatales. Una diversidad de reformas con- solidarfa las instituciones y pondria en claro las reglas, ademas de simpli- ficarlas. A un plazo mas largo, el sistema puede encaminarse-en parte-hacia el federalismo legislativo, el modelo canadiense. Esto ocurrirfa si las trans- ferencias no crecen tan rapido como la demanda de gasto en servicios que se han delegado a gobiernos subnacionales; sobre todo en cuanto a edu- caci6n, salud y carreteras. Los estados, entonces, encabezados por los que cuenten con las mejores bases gravables, bien podrian optar por recaudar mas ingresos propios. En principio, los estados ya pueden hacer eso, pero s6lo lo haran en la practica cuando obtengan mas control sobre sus egre- sos y si las transferencias federales no castigan el esfuerzo fiscal de los Estados. Esto requiere no s6lo terminar con las transferencias negociables, sino tambien que los estados sean mas aut6nomos. Tomar medidas en este sentido no necesariamente implica una reducci6n en los recursos para los estados mas pobres con menos base impositiva; en realidad esos estados probablemente salgan ganando. Sin embargo, los estados ricos pueden incrementar sus fondos mas rapido, aumentando asi la desigualdad de los recursos per capita de los gobiernos locales, a menos que haya una redis- tribuci6n de transferencias mas decisiva que en la actualidad. Ademas de eso, el mercado politico y el mercado financiero cobrarian mas influencia sobre el gasto, los impuestos y las decisiones de endeudamiento que el go- bierno central. Gestiones recientes: 1999-2000 Las recientes medidas tomadas por el gobierno federal para perfeccionar el proceso de descentralizaci6n se concentraron en tres objetivos funda- mentales: (a) imponer estrictas limitaciones presupuestarias a los recursos federales que se proporcionan a estados y municipios; (b) reducir el ries- go moral en los emprestitos subnacionales; y (c) aumentar la transparen- cia, la eficiencia y la calidad de la rendici6n de cuentas en la adminis- traci6n financiera y fiscal de los niveles inferiores de gobierno y del proceso de descentralizaci6n en general. Estas medidas seran valiosas indepen- dientemente del curso que siga el sistema en el largo plazo; y ayudaran a crear los incentivos y la capacidad para tomar las medidas subsecuentes en el mediano plazo, como reformar el sistema fiscal y de transferencias y reconsiderar las asignaciones de responsabilidades por el gasto. Las limitaciones presupuestarias estrictas tendran verdaderamente lugar a medida que reglas claras remplacen a la discreci6n en materia de gasto, transferencias y endeudamiento subnacional.15 El gobierno federal esta terminando con todas las transferencias discrecionales en efectivo a esta- dos y municipios. Todos los programas paralelos (pari-passu) importantes 74 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO funcionaran con reglas transparentes y estaran a disposici6n de todos los estados que cumplan con los requisitos conforme a los criterios publicados. La distribuci6n estado por estado del gasto federal para proyectos locales en educaci6n y transporte se publicarA en el anfo 2000, a fin de aumentar mas la transparencia en el uso de los recursos federales. Las reglas que gobernaran el financiamiento del gobierno federal a los estados complementan los incentivos dirigidos a aumentar la eficiencia y el rigor de mercado en el endeudamiento subnacional, segun se describen lineas arriba. Esto hard mas estrictas las limitaciones presupuestarias y, a la vez, reducira el riesgo moral. Aunque hay pocas posibilidades de que el problema de los sistemas de pensiones de los estados se resuelva antes del ano 2001, es importante Ile- var este asunto a la mesa de discusiones y empezar a estudiarlo. Para este prop6sito, la publicaci6n de un estudio de las obligaciones estatales por con- cepto de jubilaci6n, incluyendo cifras estimadas para cada estado, parece ser una medida sensata. Este sera un paso fundamental en el proceso de esclarecer los limites de la responsabilidad federal, disminuyendo la impre- si6n existente de que el gobierno central le ha expedido un cheque en blan- co a los estados para cubrir las obligaciones de jubilaci6n, y en un momen- to dado idear normas comunes para distribuir dicha carga. El tercer objetivo de las medidas iniciales en la agenda de descentra- lizaci6n consiste en aumentar la transparencia en la administraci6n fiscal descentralizada, como un medio para aumentar la responsabilidad puibli- ca (y la rendici6n de cuentas) tanto del gobierno federal como de los estatales. En el presupuesto del afno 2000, el gobierno federal aument6 la responsabilidad de los gobiernos estatales en cuanto a los gastos financia- dos con recursos federales. En forma experimental, las secretarfas fe- derales y los estados celebraran acuerdos a partir de programas de dona- ciones de contrapartida, lo que sera indicativo de que en el fuLturo se imputara responsabilidad a los gobiernos subnacionales por los fondos que reciban para sectores especificos gracias a la descentralizaci6n. Como primer paso, se estan redactando y ejecutando los convenios de desem- pento-de tipo experimental-para dos sectores: salud y medio ambiente. Estos parecen particularmente adecuados para el programa piloto de con- venios de desempefno, porque disfrutan de una trayectoria uniforme de reformas, con fuerte consenso en el pais y agendas de descentralizaci6n ya avanzadas. El nuevo Comite de Descentralizaci6n, antes descrito, tambien aumentara la transparencia; sobre todo si su Secretaria Tecnica hace ase- quibles al publico los anAlisis e informes respectivos. A fin de crear un proceso de desarrollo institucional mas congruente para la descentralizaci6n, ademAs de los mencionados Comite de Descentralizaci6n y su Secretaria Tecnica, se contara con un programa de capacitaci6n y fortalecimiento institucional para municipios y estados que estara a disposici6n de quien lo solicite, que incluye cursos en los siguientes PERSPECTIVA GENERAL 75 temas: contabilidad, presupuesto, supervisi6n, presentaci6n de informes, auditoria y contrataci6n de bienes o servicios. Dicho programa sera respal- dado directamente por una linea de credito del Banco Interamericano de Desarrollo por $400 millones de d6lares, autorizada a finales de 1999, que aprovechara el trabajo federal ya existente (por ejemplo, la homologaci6n de las practicas contables) y los esfuerzos estatales (como en Puebla y la Universidad de Monterrey). El gobiemo federal tambien puede respaldar el perfeccionamiento de las practicas contables y de presentaci6n de informes a nivel local de diversas maneras: (a) estableciendo normas minimas para categorias diferentes de municipios de acuerdo al volumen total de transferencias recibidas. A los estados no se les exigiria diferenciar entre municipios, pero las normas fe- derales minimas instarfan de hecho a los estados a diferenciar en una forma que sea adecuada a sus situaciones; (b) apoyando las labores de la Contaduria Mayor de Hacienda de la Honorable Camara de Diputados para establecer principios generales comunes para el Ramo 33 y otros gastos sub- nacionales financiados por transferencias federales; (c) buscando de ma- nera diligente sinergias entre PROMAP (Programa de Mejoramiento de la Administraci6n Pbhlica), otros programas similares a nivel federal y las con- tadurias mayores de los estados. La Secretaria de Hacienda y Credito Publico podria documentar las practicas para cubrir los costos en que incurran los estados para administrar el Ramo 33 (preparaci6n de proyec- tos, presupuesto, contabilidad y presentaci6n de informes); (d) evaluando el efecto de las leyes y practicas de los estados que hayan tomado iniciati- vas para supervisar de manera mas estrecha el Ramo 33; (d) evaluando las repercusiones de auditorias externas realizadas por firmas especializadas dondequiera que existan; (f) proporcionando informaci6n y reconocimien- to a organizaciones creibles de la sociedad civil para que mantengan y divulguen sus propios indicadores del desempefno de los gobiemos sub- nacionales; y (g) dando, de manera experimental, incentivos no moneta- rios a los estados que cumplan pronta y plenamente con los acuerdos inter- gubernamentales pactados conforme al Programa de Homologaci6n. Los primeros incentivos pueden constar de reconocimiento p6blico y divul- gaci6n de las practicas adecuadas. FORTALECER LA CAPACIDAD DE LOS GOBIERNOS SUBNACIONALES. La docu- mentaci6n, el reconocimiento y la divulgaci6n de practicas eficaces tienen un poderoso efecto en terminos de demostraci6n y aprendizaje; sobre todo en los primeros afios de la descentralizaci6n. A fin de estimular la capa- citaci6n, la Secretaria de Hacienda y Credito Publico podria reunir y divul- gar las mejores practicas en el manejo del gasto local realizado con recursos del Ramo 33 desde 1998 (afno en que este empez6), promover la capacitaci6n de municipio a municipio en base a practicas intraestatales eficaces en la distribuci6n de los fondos del Ramo 33, y comparar los beneficios de la 76 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO capacitaci6n horizontal y la asistencia tecnica entre municipios con los programas de capacitaci6n federales o estatales. Los cursos de capacitaci6n podrian incluir identificar prioridades, preparar prespuestos y proyectos, llevar a cabo la supervisi6n y la evaluaci6n, y redactar contratos para la adquisici6n de bienes o servicios. A fin de fomentar el desarrollo de capaci- dad de los municipios, el gobierno federal podria aumentar el total de fon- dos del Ramo 33 que pueden usarse para este concepto e invitar a estados y municipios a seguir el ejemplo de Puebla, en cuanto a concentrar y aumentar sus fondos para capacitaci6n a nivel estatal, con objeto de desa- rrollar economias de escala. MEJORAR EL PROCESO DE PREPARACION DEL PRESUPUESTO. Preparar y presen- tar el presupuesto es una oportunidad significativa para que el gobierno federal de a conocer y demuestre su compromiso de asegurar la trans- parencia fiscal de los gobiemos subnacionales y unidades sectoriales. Como experimento y mientras cumple con el programa establecido legalmente para la preparaci6n y aprobaci6n del presupuesto, el gobierno federal podria proporcionar estimados preliminares (y contingentes) de transfe- rencias propuestas para los estados y municipios, de suerte que los gob- iemos subnacionales puedan formular planes de presupuesto provision- ales antes del principio del afio fiscal. ESCLARECER LAS RELACIONES INTERGUBERNAMENTALES CON EL DISTRITO FEDERAL. A fin de reducir los conflictos existentes y futuros, el gobierno fe- deral podria poner en claro las reglas de las relaciones fiscales interguber- namentales para el Distrito Federal. Estas incluirian los criterios tecnicos para reorganizar las transferencias verticales y horizontales relacionadas con el Distrito Federal y las medidas institucionales en vigor para la nego- ciaci6n y resoluci6n de controversias (las transferencias verticales son del nivel federal al Distrito Federal, y las horizontales del Distrito Federal a los estados y municipios aledanios). Las medidas propuestas necesitaran ser consecuentes con la autonomia del Distrito Federal respecto a ingresos, endeudamiento y responsabilidades para ejercer el gasto. Quiza sea nece- sario elaborar un estudio para respaldar esta propuesta, aprovechando la experiencia internacional en transferencias intergubernamentales y los arreglos institucionales de otras areas metropolitanas (Buenos Aires, Santiago y Sao Paulo en cuanto a America Latina; y otras ciudades en Asia Oriental, Europa y America del Norte). Agenda de mediano a largo plazo: 2001-2006 La agenda de reformas para la descentralizaci6n del mediano al largo plazo incluye reconsiderar el pacto fiscal entre los niveles federales y estatales-el disefio conjunto de asignaciones tributarias y transferencias- PERSPECTIVA GENERAL 77 y pasar revista a la asignaci6n de autoridad para tomar decisiones rela- cionadas con los egresos. Sera necesario continuar el desarrollo institucional de diferentes maneras. El gobierno nacional tambien necesitara preservar las reformas que realiz6 en el bienio 1999-2000 al regimen regulatorio para el endeudamiento subnacional, de suerte que los estados reformen uni- formemente sus correspondientes leyes y practicas. Algunos estados, incluyendo siete cuyos gobernadores suscribieron la Declaraci6n de San Lazaro en 1999, se opusieron energicamente al actual sistema de transferencias e impuestos exigiendo una participaci6n mayor en los ingresos que se recaban en sus territorios. Como ya no hay recursos presupuestarios de tipo federal para transferencias de caja a discreci6n del ejecutivo, la soluci6n inmediata ha sido aumentar el gasto federal para las necesidades locales -transferencias en especie, de hecho y para prop6si- tos determinados- y la reactivaci6n del Fondo para Fortalecer las Finanzas del Estado, con partidas establecidas por el Congreso. Casi todas las partes involucradas se beneficiarfan de una soluci6n mas transparente y uni- forme, pero seria dificil acordar que soluci6n en particular. Atacar varios problemas al mismo tiempo, sobre todo las transferencias y los impuestos, permitira que haya mas dimensiones para solucionarlos y, por consiguiente, haria mas probable la soluci6n. Impuestos subnacionales Los estados y municipios necesitan tener considerable control sobre sus ingresos al margen, a fin de que tengan incentivos adecuados para un gasto eficiente y que sus finanzas resistan las sacudidas econ6micas. INTRODUCIR NUEVOS IMPUESTOS, APOYARSE EN LOS IMPUESTOS FEDERALES, ESCOGER UNA 0 DOS ENTRE VARIAS POSIBILIDADES. El porcentaje del impuesto federal sobre el ingreso de las personas fisicas podria reducirse para per- mitir que los estados fijen sus propios impuestos suplementarios. Luego podria haber una reducci6n correspondiente en las transferencias fe- derales o en la participaci6n de ingresos fiscales. Algunas participaciones procedentes de impuestos al consumo sobre combustible para autom6viles podrian transferirse del gobierno federal a los gobiernos de los estados donde se realiza el consumo. Tambi6n podria considerarse un desplaza- miento de ingresos por impuestos al consumo de bebidas alcoh6licas y taba- co para ser aplicado via impuestos suplementarios estatales. En cuanto a los impuestos generales al consumo, podrian identificarse los que encajen mejor para convertirse en impuestos estatales, dentro de los marcos cons- titucional y normativo existentes. Luego, la experimentaci6n podria con- tinuar con uno o algunos de los impuestos al consumo mas prometedores mediante acuerdos federales y estatales. Por ultimo, la Secretaria de Hacienda y Credito Piublico podria tambi6n evaluar, adaptar y promover su dupli- caci6n. El gobierno federal podria tambien trasladar a los gobiemos estatales 78 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO las rentas generadas por un punto porcentual dado del impuesto al valor agregado, empleando el impuesto al valor agregado dual-federal/estatal- y el impuesto al valor agregado compensatorio para ventas interestatales. Un control fiscal federal mas firme seria un requisito previo para cualquiera de estos. RACIONALIZAR EL IMPUESTO SOBRE NOMINAS. El gobierno federal podrfa dar incentivos a los estados para homologar la base para el impuesto sobre n6minas. Una forma podria ser participar en los costos de recaudaci6n si los estados empiezan a usar la misma base que el impuesto federal del seguro social. El asunto mas dificil respecto a los impuestos sobre n6mi- nas (y predial) es encontrar los incentivos adecuados para estimular a esta- dos y municipios para que aumenten sus recaudaciones a niveles mas ele- vados pero que sigan siendo compatibles con la eficiencia tributaria, la incidencia equitativa y la limitaci6n de distorsiones fiscales en la asignacion tributaria de recursos. ADMINISTRACION TRIBUTARIA. Debe fortalecerse la formaci6n de catastros, la actualizaci6n y avahlos del impuesto predial. Las exenciones (incluso las intergubernamentales) deben limitarse o eliminarse para ampliar la base impositiva y hacer que los impuestos sean mas neutrales. Las cuentas de gastos fiscales (via exenciones) podrian conservarse y divulgarse, con mo- delos preparados por el gobierno federal. Deben fortalecerse las auditorias y la recaudaci6n de los impuestos estatales sobre n6minas y los municipales de tipo predial. Para aumentar la recaudacion de impuestos, Mexico podria crear una direcci6n nacional de impuestos, que recaudaria los impuestos federales y estatales con bases iguales o similares. Es obvio que esto se volveria mas pertinente a medida que los estados homologuen el impuesto sobre n6minas y (se les permita que) tengan sus impuestos apoyados en los impuestos federales. APOYO A LOS ESFUERZOS LOCALES PARA EL IMPUESTO PREDIAL. Seria util realizar un estudio para identificar y evaluar los incentivos que llevan a los estados a prestar apoyo al impuesto municipal sobre bienes raices. Podrian fomentarse acuerdos estatales y municipales para permitir que los estados recobren parte de los costos administrativos del impuesto predial o com- partan parte de las rentas adicionales producidas por una mejor adminis- traci6n. Se necesitan nuevos incentivos y acuerdos estatales-municipales para estimular la cooperaci6n para un mayor esfuerzo fiscal local. El go- bierno federal podria considerar un programa para reconocer y premiar a los municipios con los aumentos mas altos en rentas por ingreso predial y a los estados que ofrezcan el apoyo mas eficaz para recaudar esos impuestos. Seria util explicar el efecto del denominado ejido (propiedad comunal de la tierra) sobre el registro de la propiedad y el catastro fiscal. PERSPECTIVA GENERAL 79 La SecretariaTenica para el Comite de Descentalizaci6n podria recabar datos sobre la administraci6n y las rentas generadas por el impuesto pre- dial y evaluarlos. Tomando en cuenta la experiencia que vaya surgiendo, el gobierno fe- deral podria estimular a los estados para que instituyan reglas por las cuales los municipios mismos puedan administrar con eficiencia y del todo el impuesto predial, y que determinen qu6 municipios se beneficia- rfan de los distintos grados de apoyo estatal para formar y actualizar el cat- astro, mantener las cuentas al corriente, facturar y cobrar. Evaluar el efec- to y eficiencia del Programa 100 Ciudades, hasta donde implique apoyo para la formaci6n y administraci6n del catastro, ayudaria a desarrollar la capacidad de las ciudades en comparaci6n a municipios predominante- mente rurales. Las comparaciones podrian incluir coeficientes de rentabil- idad y la distribuci6n de ingresos centro-periferia. Seria tambi6n uitil contar con estudios t6cnicos y empiricos para (a) identificar y evaluar las razones del descenso relativo durante los ultimos 20 anios en la administraci6n del impuesto predial en algunos lugares (por ejemplo, Monterrey), y el relativo fortalecimiento en otros sitios (como el Distrito Federal); los estudios deberian arrojar propuestas para efectuar reformas de politicas y replicar las practicas que hayan resultado eficaces; y (b) calcular el efecto de los cambios en las tasas del impuesto a n6minas sobre las rentas fiscales y las economias de los estados-empleo, produc- tividad laboral y competitividad. Transferencias intergubernamentales En coordinaci6n con la agenda de reformas tributarias, algunas de las transferencias con contrapartida (pari-passu) y tal vez hasta las participa- ciones podrian reducirse, junto con la disminuci6n de tasas impositivas fe- derales (por ejemplo, el impuesto sobre el ingreso de las personas fisicas o el impuesto al valor agregado) y la delegacion del derecho-a los estados- para cobrar recargos sobre los impuestos. A1 mediano plazo, es necesario contar con un sistema mas sencillo y mas transparente que garantice recursos iguales per capita (o cualquier otro obje- tivo acordado por los gobiernos nacional y subnacionales) y un tratamien- to mas equitativo a los estados con niveles de pobreza similares. La medi- da de equidad debe ser las transferencias totales per capita mas las rentas que podrian obtenerse de las bases gravables del estado en cuesti6n. Sera necesario respetar las prioridades de los Estados y permitir que estos alcan- cen estandares equivalentes de diferentes maneras. Las f6rmulas para transferencias no deben sancionar a los que aumenten sus propias rentas. A fin de simplificar el sistema de transferencias, todos los programas especiales con fines especificos, como el Plan de Seguridad Publica o el Programa 100 Ciudades, podrian incluirse en el Ramo 33. Al agrupar todos los programas especiales bajo un solo rubro se facilitaria la formulaci6n de 80 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO politicas y se avanzaria hacia un sistema de transferencias mas sencillo, mas transparente y con mayor autonomia en los niveles inferiores de gobierno. Egresos y prestacion de servicios El gobierno federal podria poner en claro cuales son las responsabilidades por sector entre los niveles de gobierno, coordinando las responsabili- dades en cuanto a inversi6n y mantenimiento, asi como contrataci6n de empleados y sueldos. Se podria aconsejar a los estados que ajusten los suel- dos y el numero de maestros y personal del sector salud de acuerdo a las condiciones de demanda locales. Esto ya esta realizAndose en Areas de bajos ingresos en Oaxaca y podria extenderse a otros estados. A los estados que tengan una base impositva adecuada y demanda adicional de educaci6n, se les podria permitir que aumenten los sueldos de los maestros por enci- ma del nivel nacional, pagando esa diferencia extra de sus propios impuestos. Tambien podrian tomar iniciativas en las Areas de indicadores de desempenxo, sueldos, promoci6n e incentivos para mejorar la adminis- traci6n del sector publico estatal. AUMENTAR Los INCENTIVOS HACIA LA EFICIENCIA. Algunas reformas especi- ficas por sector aumentarian la eficiencia de los incentivos. En educaci6n, un nuevo sistema de transferencias podria tomar las partidas existentes como un minimo nominal y hacer avanzar los aumentos (o congelamien- tos) hacia una f6rmula con base en el numero de nifios en edad escolar, con- forme al censo y a proyecciones transparentes a partir de este. A la larga, la f6rmula federal trataria de reflejar el actual nivel de asistencia a los edu- candos, pero ese tipo de incentivos probablemente debe dejarse a los esta- dos. En el sector salud, con las partidas existentes como un minimo nomi- nal, los incrementos deben seguir desplazando las partidas hacia una f6rmula que se sustente en algun indicador de demanda posible, como poblaci6n ajustada por sexo y edad. La experiencia internacional demues- tra que debe estimularse a los estados para que deleguen autoridad y responsabilidad publica a las unidades que proporcionan los servicios directamente, como es el caso de hospitales y escuelas. La Secretaria de Hacienda y Credito Publico podrfa evaluar las diver- sas legislaciones estatales, las practicas de facto, asi como los manuales que regulan los incentivos hacia la autonomia municipal y la distribuci6n efi- ciente del Ramo 33. La Secretaria de Hacienda y Credito Publico y la Secretaria de Contraloria y Desarrollo Administrativo podrian comparar las legislaciones estatales y las practicas respecto al Ramo 33 con los meto- dos del Programa de Mejoramiento de la Administraci6n Publica y el nuevo programa, enfocado a resultados, denominado Nueva Estrategia de Programaci6n; y ofrecer recomendaciones a fin de que se impartan cursos para ensefiar las nuevas tecnicas de programaci6n a los Comit6s de Planeaci6n para el Desarrollo del Estado y los Comit6s de Planeacion para PERSPECTIVA GENERAL 81 el Desarrollo Municipal (comites de programaci6n para el desarrollo estatal y municipal respectivamente). (Las siglas en espafiol de los organismos y programas antes citados se mencionan a continuaci6n, en el mismo orden en que aparecen: SHCP, SECODAM, PROMAP, NEP, COPLADEs y COPLADEMUNs) El estudio podria comparar aquellos estados en que el COPLADE es responsable de aplicar un programa con los estados en que no tiene esa responsabilidad. ESCLARECER LAS RESPONSABILIDADES POR PENSIONES CONFORME A UNA BASE SUSTENTABLE. Es necesario contar con un acuerdo entre los gobiemos fe-deral y estatales para poner en claro cudl es el nivel de gobierno responsable de las pensiones para los empleados publicos subnacionales. Quiza seria necesario que el gobierno federal reconozca explicitamente que cierta parte de una linea base de las pensiones de los estados no es responsabilidad de estos uiltimos (como por los ex empleados federales y los que han rebasa- do los 50 anos de edad) en recompensa porque el estado en cuesti6n emprenda reformas fiscales, como pagar los intereses de obligaciones sobrantes, entre ellas el seguro social. 0 bien, como se hizo en Argentina, el gobierno federal podria ofrecer hacerse cargo de los planes de pensiones de los estados (para reformarlos), junto con la recaudaci6n de contribu- ciones. El estudio inicial arrojaria medidas especificas para hacer explicito el monto y naturaleza de las obligaciones de jubilaci6n. Institucionesfundamentales para la descentralizaci6n MEJORAR LA RESOLUCION DE CONTROVERSIAS. A fin de resolver los problemas que surgen incluso en los sistemas mejor disefiados y ofrecer a los partici- pantes un incentivo para que encuentren soluciones en las que colaboren todos los involucrados, es necesario contar con mejores instituciones para resolver las controversias. Crear estas instituciones llevara tiempo, pero es deseable empezar a hacerlo pronto. Podrian promoverse comisiones sec- toriales para areas importantes como salud y educaci6n. A mediano plazo- con las lecciones aprendidas gracias a la experiencia en cada sector-podria instituirse una comisi6n general gubernamental para que atienda los asun- tos y apelaciones de todo el sistema que le envien las comisiones sectoria- les. Una vez que los modelos federal-estatal funcionen, los estados podrian crear sus propias comisiones municipales-estatales para atender los asun- tos entre los municipios, y entre ellos y el estado. MEJORAR LA INFORMACION Y LAS NORMAS CONTABLES. Es importante que se sigan rindiendo cuentas claras en cada nivel de gobierno; la presentaci6n de informes municipales a los estados y la supervisi6n de partidas por parte del estado. En la actualidad, algunos estados imponen sanciones a los municipios por atrasarse en los informes suspendi6ndoles las transferencias, y esta practica podria vincularse al proceso de documentacion y evaluaci6n. 82 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO La evaluaci6n de practicas ya existentes tambi6n podria diferenciar los controles ejecutivos de los que ejercen las contadurias mayores de cuentas de los congresos de los estados. La Secretaria de Hacienda y Credito Publico podria evaluar si el modelo que esta preparando la Contaduria Mayor de la Uni6n estimula de manera adecuada una distribuci6n eficiente y, mas adelante, formular y publicar las recomendaciones que considere pertinentes. Tambi6n es importante establecer un sistema contable uniforme e intercambios de informaci6n sectorial e intersectorial, tanto de los nive- les superiores a los inferiores como en sentido inverso. MEJORAR LA COORDINACION DENTRO Y ENTRE LOS DIFERENTES NIVELES DE GOBIERNO. Es posible que el gobierno federal quiera promover alianzas municipales para (a) obtener economias de escala (en la contrataci6n de bienes o servicios, servicios especializados, equipo compartido, costos administrativos); (b) identificar soluciones fiscales o tecnicas de otro tipo mediante esfuerzos subregionales coordinados que emprendan de manera conjunta dos o mas municipios (recolecci6n de basura, distribuci6n inter- jurisdiccional de comisiones para control de la contaminaci6n); (c) identi- ficar y evaluar experiencias en curso; y (d) emprender proyectos experi- mentales de corto plazo que puedan acometerse y evaluarse pronto. Podrian tambien crearse foros intersectoriales, interestatales e intermu- nicipales para elaborar el presupuesto de manera integral para una regi6n o subregi6n dada. Aunque al principio sean informales o para prop6sitos especificos, estos foros podrian sentar las bases para presupuestos mas flexi- bles y para la elaboraci6n integral en el futuro de presupuestos por regi6n. DAR TRATO DlFERENTE A LOS MUNICIPIOS ACORDE A SU CAPACIDAD. Es impor- tante establecer un conjunto de criterios para certificar que los municipios cuentan con la capacidad necesaria para aplicar ciertos programas o tomar prestamos. Los requisitos de informes financieros deben determinarse segun el tamafno fiscal de los municipios, que se trate o no de la capital de un estado, y si quiere asumir responsabilidad y autoridad adicionales. En conclusi6n, la agenda que aquf se contempla incluye seis objetivos principales: 1. Las politicas sectoriales e intersectoriales de descentralizaci6n necesi- tan desarrollarse dentro de un esquema estrategico mas coherente. Esto ayudaria a obtener una asignaci6n mas clara de responsabili- dades, mas autonomia subnacional y estrategias de ejecuci6n mas efi- caces. 2. La capacidad administrativa de los gobiernos estatales necesita con- solidarse, sobre todo a nivel municipal. PERSPECTIVA GENERAL 83 3. La capacidad de los estados para fijar impuestos y endeudarse nece- sita ampliarse de manera consistente y sostenible para permitirles enfrentar sus nuevas responsabilidades. 4. A fin de tener incentivos adecuados para una autonomia fiscal respon- sable, los estados necesitan enfrentar estrictas limitaciones pre- supuestarias del gobierno federal en cuanto a transferencias, deuda y responsabilidades por la prestaci6n de servicios. 5. Necesitan perfeccionarse las practicas contables en todos los niveles subnacionales de gobierno, ya que la coordinaci6n y supervisi6n intergubernamental del proceso de descentralizaci6n requiere el inter- cambio de informaci6n pertinente y andloga. Esa informaci6n tiene que compartirse entre los gobiernos y el electorado, quienes son los defensores fundamentales de la reforma en una democracia. 6. Es necesario contar con procesos para crear consenso y con mecanis- mos para resolver controversias y conflictos prioritarios, como una forma de ayudar a que los circulos cada vez mayores de actores politicos y econ6micos participen en el dinamico proceso de la descentralizaci6n. Aunque las politicas aqui esbozadas pueden ser dificiles desde un punto de vista politico y econ6mico, todas las partes interesadas en Mdxico estan conscientes de los costos y riesgos que implica no tomar medidas cohe- rentes. Estos costos incluirian la subutilizaci6n de la capacidad tributaria subnacional, un gasto ineficiente debido a la falta de claridad para rendir cuentas y de controles sobre la prestaci6n de servicios, y un endeudamiento subnacional potencialmente explosivo, sobre todo surgiendo de deuda contingente y obligaciones de jubilaci6n. 1 Principles of Decentralization Thomas Courchene, Jorge Martinez-Vazquez, Charles E. McLure, Jr., and Steven B. Webb STUDIES OF FISCAL DECENTRALIZATION in many countries, some of size similar to Mexico, have led economists, political scientists, and policy- makers to agree on some useful principles for policy design. First, politi- cal and cultural considerations take priority in determining the appropri- ate model, and many models can work. The next section outlines several possibilities. Within the political and cultural constraints, the economic characteristics of the model need to be optimized, and a section on market- preserving federalism provides four economic criteria for performance. While no model is perfect, and "working" means having ways to com- pensate for weaknesses, there are certain decentralization-design strategies that can improve performance, and other design flaws that need to be avoided. The bulk of this chapter addresses these design issues in four areas: spending, taxation, transfers, and debt management. Different Federal Models At least three models or theories of federalism are relevant to Mexico and, indeed, are partially realized. The first model-"classical" federalism- assigns expenditure and taxation among the levels of government, with financial shortfalls accommodated by a set of conditional and unconditional transfers. The second model relates to the direct federal-municipal relationship that exists not only in practice in Mexico, but in the constitution as well. Brazil's national constitution also recognizes in detail the municipal level. In some federal systems (such as Canada), municipalities are creatures of the provinces or states and are not even mentioned in the constitution. The United States' Constitution also makes no mention of the municipal level, but this is taken to mean that, while the states define and govern the 85 86 FISCAL DECENTRALIZATION: LFSSONS FROM MEXICO municipal level, the federal government can have all kinds of direct pro- grams for cities and other substate entities. The third model-administrative federalism-is the German model where most legislation is federal, but most implementation and adminis- tration (even in the areas of federal jurisdiction) are conducted at the lInder (state) level. As they currently exist in Mexico, expenditures on education would resonate well with this model. The appendix to this chapter elabo- rates on the cases of Australia, Canada, Germany, and the United States. These three models could coexist in long-term equilibrium, as they do now in Mexico. For example, expenditures on some set of national (pan- Mexican) public goods could be legislated federally but implemented by the states. In this context, federally imposed conditions would comprise a set of principles, rather than specific and binding conditions, This would provide some degree of state autonomy in determining how best to satis- fy these principles. Likewise, in any long-term scenario, municipalities could retain their constitutional status and their financial links with the fed- eral government (via pass-through transfers). Given the apparent ascen- dancy of the states as economic and political actors, however, a more like- ly scenario might be one in which conditions are placed on state transfers regarding how these transfers are redirected to the municipalities. Aspects of classical federalism will surely play a key role in any federal system in Mexico. There is considerable scope for creative options. For example, con- ditional transfers can be rationalized, formula-driven, and designed so that states can take into account their own preferences, cultures, and eco- nomic requirements in satisfying the conditions. Moreover, the Ramo 28 revenue-sharing unconditional transfers could give way to more subna- tional taxation, and the mix between conditional and unconditional trans- fers could shift toward the unconditional as a result of pressures for fur- ther decentralization. The message is twofold. First, Mexico has many avenues along which to pursue decentralization. Second, the transitional status quo embodies a variety of structural components that could, with creative applications, gen- erate a more decentralized federation. Building or rebuilding a federal model has many objectives, including cultural and political ones on which this book does not dwell (Riker 1964; Stepan 1997). The economic objectives are more than fiscal-the focus of this book-and have to do with provid- ing a better context for economic development. Market-Preserving Federalism What Weingast (1995) and McKinnon (1998) refer to as "market-preserv- ing federalism" provides a useful summary of the immediate economic objectives of decentralization. According to McKinnon, market-preserving federalism comprises four key components: PRINCIPLES OF DECENTRALIZATION 87 1. Monetary separation. State governments cannot own or control com- mercial banks. 2. Fiscal separation. State governments do not have access to discre- tionary or additional central government financing to cover state deficits. 3. Freedom of interstate commerce. Goods, services, people, firms, and cap- ital are allowed to move freely across state lines. 4. Unrestricted public choice. States are allowed to design and deliver alternative bundles of public goods and services and to finance them by alternative means of taxation. The first two principles ensure that there are no bailouts. States can, of course, still borrow on capital markets, but credit-rating agencies and financial markets generally are more vigilant if they realize that the states face hard budget constraints. Unrestricted public choice is possible only if the first three principles are in place. In this market-preserving context, the exercise of competitive federalism enhances efficiency and welfare. The Federal-State Fiscal Accord To create and sustain the conditions for market-reserving federalism, fed- eral and state governments may sign a federal-state fiscal accord preserv- ing and promoting the socioeconomic union as the federation decentralizes. The specific elements can vary considerably, but the provisions generally include the following: * Principles to ensure the free flow of goods, services, people, firms, and capital across state borders. - Provisions to ensure that people who move across state boundaries continue to have uninterrupted access to basic services such as health and education. * A code of tax conduct to ensure that states do not discriminate against residents of other states. - A principle similar to the "national treatment" provision under the North American Free Trade Agreement (NAFTA), referred to in the fiscal accord as "state treatment," meaning that each state agrees to apply its policies equally to all residents of the state, even if they have migrated from other states. * Commitment of the federal government to these provisions, except where the federal constitution dictates otherwise. These are the core elements in any fiscal accord. But the fiscal accord could be made more encompassing. For example, it could incorporate principles or conditions associated with the system of intergovernmental 88 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO grants. As in NAFTA, there would have to be mechanisms for resolving disputes, replete with sanctions (such as withholding part of the grant). The dispute-resolution process should be timely and transparent. "Opting in" Provisions Mexico is a complex federation. The constituent states differ in their cul- tural heritage, level of economic development, and relationship with the outside world. It would be surprising if all states viewed the prospects of decentralization in the same way. In some policy or program areas, it may be appropriate to provide for "opting in" provisions. If a state does not sup- port the details of tax decentralization, for example, it need not embrace the new system immediately. Instead, it could continue to follow the existing system and to receive revenues from grants rather than from taxes. Opting in can provide considerable flexibility, and states would have the option of switching to the new system at any time. Although there will be some asymmetry of treatment, the transfer would be designed to ensure that, overall, all states are treated in a comparable manner. Regardless of the model followed, Mexico will need to improve the allocation of spending, the system of taxation, the system of transfers, and the management of subnational borrowing and debt. The rest of this chap- ter addresses these four topics. Spending and Service-Delivery Assignment Decisions about which level of government has responsibility for a partic- ular public service must be judged by how well the assignment achieves the basic criteria of fiscal performance, namely, efficiency, income redis- tribution, and macroeconomic stability. These objectives sometimes con- flict, and the government has to determine the tradeoffs. COST AND PRODUCER EFFICIENCY. Efficiency means satisfying the needs and preferences of taxpayers at the lowest possible cost. Services should be pro- vided by the lowest level of government compatible with the size of ben- efit area associated with them. Local officials are more familiar with the needs and preferences of individuals residing in their jurisdiction. They are also likely to be more politically accountable, and therefore responsive to, these needs and preferences. Because these needs and preferences are not uniform throughout the country, it is unlikely that the central government will be able to deliver more meaningful services for a given budget. Lower- level governments, especially if they are democratically accountable, are better able than central government to match expenditures with preferences and needs, making decentralization a means of increasing consumer effi- ciency within government budgets. Moreover, efficiency is enhanced if con- PRINCIPLES OF DECENTRALIZATION 89 sumption benefits are linked to the costs of provision via fees, service charges, or local taxes. A second type of economic efficiency, producer efficiency, also needs to be considered. Because there may be significant economies of scale, it may be more cost-efficient to produce certain public services on a larger scale at a higher level of government. Producer efficiency also reflects the skills and qualifications of government officials, which may be better at the higi- er levels of government, although not always. Better accountability, com- petition, and scope for innovation may render the delivery of services by lower-level government more cost-effective than delivery by central government. The optimal degree of decentralization depends on the nature of the ser- vice. For example, the benefit area is clearly the local community for sani- tation services, but the national territory for air traffic control. On the other hand, leaving the supply of public services with wider benefit areas to smaller units of government is likely to result in the inefficient underpro- vision of services, as when, for example, a tertiary hospital providing regional services is financed by a single municipality. Where there is wide variation in the size and capacity of local government, as in Mexico, it may be optimal to have various options for the extent of local fiscal control, with objective criteria and appropriate financing arrangements corresponding to each. INCOME REDISTRIBUTION AND MACROECONOMIC STABILITY. Public expendi- tures for equity or income equalization, such as social welfare or low- income housing, should be primarily the domain of central government. Local or regional governments cannot sustain independent programs of this nature because doing so would attract the needy from other areas, while taxing (potentially mobile) residents more heavily. In addition, the need for redistribution and social welfare assistance may be particularly high with- in subnational governments that are themselves relatively poor. Although policy formulation and funding should be the responsibility of central government, implementation can be left to local governments that may have better information on the identity and needs of the poor. However, it is not uncommon for subnational governments to legislate and fund their own social welfare policies. This is sustainable as long as these policies fit a national or regional norm or pattern. Expenditures undertaken to stabilize the economy, such as massive investment or unemployment compensation, are ascribed to the central government. The economies of regional and local governments are too open and dependent on those of the rest of the country for these governments to be able to affect employment or aggregate demand in a meaningful way. 90 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO LEGISLATION, FINANCING, AND DELIVERY OF PUBLIC SERVICES. The provision of any public service involves three functions. The first is the legislation of principles and norms. Often, the public function is limited to this activity alone, as is the case with environmental laws and regulations, regulations for the use of airwaves, or traffic laws. The second element is financing. The characteristic of a public good legitimizes the funding of the service out of a general pool of funds raised through taxes. In some cases, funds can be earmarked from some particular service, and these funds may come from cost-recovery fees associated with the service. The last function is the pro- duction or delivery of the service. At times, this activity may be undertak- en by the private sector. The design of a clear and efficient system of expenditure assignment should address each of these functions. Different levels of government may handle different functions for any expenditure area. For example, the cen- tral government legislates and regulates the content and standards of gen- eral basic education (curriculum content, teacher qualifications and train- ing, and so on), the state finances the education system, and the local school district runs the schools. Efficiency means making the best use of available resources, and accountability means ensuring that government officials are responsive to their constituents. Concentrating regulation, financing, and delivery of a public service in a single level of government usually enhances accountability, but this is not always possible or desirable. National interests and goals may dictate that national standards be stated and controlled by the central government. The lack of adequate resources or considerations of equity may dictate that funding be provided by the cen- tral government and that services be delivered or produced at the local level. This may be the case for many social welfare programs. No SINGLE-BEST ASSIGNMENT. These principles should guide the assign- ment of expenditure responsibilities, but they are not definitive for all times and all places. Some public services, such as primary education and primary health services, may be local in nature because of the size of their benefit area; but, because of their relevance to welfare and income redis- tribution, they may also be considered a concurrent responsibility of the regional or central governments. Economic efficiency calls for greater autonomy of subnational governments in expenditure and taxing decisions, but this may lead to fiscal disparities among subnational governments that are unacceptable in a given country. There are other important trade- offs in the implementation of intergovernmental fiscal relations. Achieving a more equal distribution of resources among subnational governments is a worthy objective. But to redistribute resources to poorer regions, central authorities need to tax better-off regions more heavily. Placing higher tax burdens on better-off regions limits their ability to grow. Different coun- PRINCIPLES OF DECENTRALIZATION 91 tries and governments make different choices in the tradeoff between redistribution and growth. Even in a single country, the "best" expenditure assignments will change over time, responding to changes in costs and tech- nology or in citizens' preferences. It is often argued that an explicit assignment of expenditure responsi- bilities is not possible because some responsibilities will always be con- current among different levels of government. This argument is fallacious. Although many types of expenditure responsibilities should be shared among different levels of government, failure to have a concrete and explic- it assignment leads to instability and friction, if not open conflict, in inter- governmental relations. Worse, it leads to the inefficient provision of pub- lic services. Common Problems with Expenditure Assignments: Lessons from International Experience LACK OF CLEAR DELINEATION BETWEEN THE PUBLIC AND PRIVATE SECTORS. Often subnational governments and, of course, central governments get involved in the financing and provision of private market activities. This hinders economic development because subnational governments tend to compete unfairly with private entities. Involvement in the private sector by local authorities creates conflicts of interest and opportunities for corrup- tion and fraud, and this undermines the appropriate regulatory functions of both central and subnational governments. Ultimately, public funds in private activities decreases the amount of funds available for the provision of public services. The best that subnational governments can do to pro- mote economic development in their jurisdictions is to provide the best pos- sible level of public services, capital infrastructure and, where appropriate, other regulation. LACK OF FORMAL ASSIGNMENT. The country's constitution often assigns expenditure responsibilities among different levels of government, but the assignment is usually not specific. Expenditure responsibilities should be specified mostly in the law, not the constitution, providing specificity as well as flexibility. The lack of clear, formal assignment of responsibilities tends to destabilize intergovernmental relations. This was the case early on during the transitions in Kazakhstan, Russia, and Ukraine. What confuses expenditure assignments in practice is that even though responsibility for implementation or delivery may be assigned explicitly, responsibility is not assigned for designing policy, imposing norms over expenditure functions, or financing those expendi- ture responsibilities. 92 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO LACK OF AN INSTITUTIONAL MECHANISM FOR COORDINATION AND CONFLICT RESOLUTION. No assignment of expenditure responsibilities can cover all the possibilities and contingencies in a legal document. It is important to intro- duce well-developed institutions of cooperation and coordination among different levels of government that can discuss and resolve disagreements. Conflicts involving expenditure assignment and inefficiencies in service delivery are more likely where the different levels of government are involved in the same sector, but there is little communication and exchange between them. When multiple levels of government are involved in most sectors, governments need broad and formal coordination institutions, as with "cooperative federalism" in Germany. In the United States, the pat- tern of assigning responsibilities varies widely from sector to sector and from state to state, so the only sector coordination is done by technocrats in some areas where there is a clear need, such as highways and law enforcement. Somewhere in between, Australia, Canada, and New Zealand use formal gatherings of politicians and bureaucrats to discuss mutually important fiscal issues. MISALIGNMENT OF CAPITAL AND RECURRENT EXPENDITURE ASSIGNMENTS. In general, capital expenditure responsibilities should be assigned among the different levels of government in the same manner as recurrent expen- ditures. Decentralization in both capital investment decisions and recurrent expenditures increases the efficiency associated with being closer to the needs and preferences of taxpayers and improves government account- ability and responsiveness. Furthermore, maintaining infrastructure at efficient levels in general requires local ownership. Responsibility for cap- ital infrastructure should be placed at the level of government in charge of delivering specific services, including the operation and maintenance of facilities. In this way, only capital infrastructure facilities that are desired by subnational governments will be built, encouraging their maintenance and repair. To assign capital expenditure responsibilities properly, it is nec- essary to address the issue of long-term financing. Subnational governments need to be allowed to borrow from banks or in capital markets, and that borrowing needs to be regulated either directly or through the market. Borrowing for justified, long-lived infrastructure is both efficient and equitable.1 Most countries only allow borrowing for capital investment purposes. MISMATCH OF FINANCING CAPACITY AND MINIMUM LEVELS OF SERVICE. Do subnational governments have enough resources to provide some mini- mum level of service in the functions assigned to them in areas of nation- al interest (education, health, social welfare, and so forth)? The concept of a minimal level of service, in contrast to what people of a subnational area PRINCIPLES OF DECENTRALIZATION 93 can or choose to afford, has operational meaning only when the wider soci- ety has an interest in assuring such a minimum. In substance, the issue is whether it is desirable to introduce mechanisms that force or encourage sub- national governments to spend (some of) their funds in a particular way in order to provide some minimum standard of service (provided the funds or the sources of funds are present). One approach, commonly followed, is for the central government to rely on conditional grants or matching grants to induce subnational govern- ments to provide the minimum level of service. This is another example of how expenditure assignment directly relates to the design of other elements of fiscal decentralization, in this case, the transfer system. To protect the principle of subnational autonomy-the core operational concept of fiscal decentralization-minimum standards and similar restrictions on subna- tional governments should be used sparingly. ONE SIZE OF MANDATE DoEs NOT Frr ALL. Achieving the optimal scale in local service provision is an issue in many transition and developing countries, especially when the constitution or the law treats all municipalities the same. Municipalities with small populations cannot take advantage of minimum economies of scale. To improve efficiency, small municipalities could be consolidated into larger ones; associations of municipalities or spe- cial jurisdictions could be created; local governments could deliver services through private contractors; or the state or national government could provide the service in small municipalities, with appropriate finance. Asymmetric decentralization is often a practical solution, providing options and incentives by giving more responsibility and resources to entities that demonstrate satisfactory administrative and fiscal capacity. It is important to recognize, however, that regional and municipal governments may be underdeveloped precisely because they have never been given responsi- bilities or the resources with which to carry them out. Small and poor municipalities often need technical assistance to help achieve adequate capacity. UNFUNDED EXPENDITURE MANDATES. Unfunded mandates-the imposition of expenditure requirements on subnational governments without ade- quate funding by central or federal authorities-are a common problem in countries with decentralized fiscal systems. Unfunded mandates take dif- ferent forms, including the setting of wages at the central level, mandato- ry increases in pension payments, and the setting of standards of provision for services. As expected, subnational and central governments tend to have quite different views of what is and is not an unfunded mandate. Central governments often argue that subnational governments receive funds through general funding sources (for example, revenue sharing or general 94 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO transfers). Of course, subnational governments argue that general rev- enues should remain at the discretion of subnational governments. Prohibiting unfunded mandates by law is not always respected in prac- tice, and some may be desirable and acceptable to subnational governments (such as equal opportunity for citizens or environmental standards with cross-regional effects). Universal or blanket reimbursement policies, inter- national experience shows, can easily create an administrative nightmare. An important step is to clarify expenditure assignments as much as possi- ble. Other solutions include requiring supermajority votes in Congress to impose central mandates without funding, introducing mechanisms of self-restraint in Congress, establishing closer working partnerships among the various levels of government, and introducing categorical grants to fund large mandates. TIMING. The assignment of responsibility for expenditures needs to precede or at least coincide with the design of other pieces of a decentralized svs- tem. Failing to do this is a common mistake. Many countries of Latin America assigned revenues to subnational governments and put transfers in place before transferring functional competencies from the central gov- ernment to subnational governments. This approach produced weak sub- national governments and fiscally overburdened central governments, which in many cases continued to take on most expenditure responsibili- ties with fewer resources. Tax Assignments Definition and Purpose Other things being equal, the jurisdiction responsible for spending should also be responsible for raising the requisite revenues. Although this is a favorite principle of federalism scholars, in reality it seldom occurs. Most federations have significant vertical imbalances where a level with more revenue sources than spending responsibility makes vertical transfers to a level with revenues less than responsive. The challenge is to design inter- governmental transfers with an eye toward ensuring that, at the margin, states view transfers as they do their own revenue. For example, transfers should not be open-ended. Tax assignment considers which level of government should tax what (the tax base) and how. Possible techniques of tax assignment include independent legislation and administration by subnational governments, surcharges on the tax base of a higher level of government, and tax shar- ing. (Revenue sharing is also considered, even though it is not a form of tax assignment.) The purpose of tax assignment is to provide subnational gov- PRINCIPLES OF DECENTRALIZATION 95 ernments with revenues that they control and thus to decentralize the con- trol of public spending. It is important to distinguish between two concepts. Own revenues belong to subnational governments by law. They may include revenue-sharing funds as well as shared taxes and taxes levied by or for the subnational gov- ernment. (Whether borrowing is considered a source of revenue depends on the context. For the present discussion it is not.) For decentralization to be viable, subnational governments must have their own revenues that are adequate to cover their current expenditures, including debt service. Marginal (incremental) own revenues are revenues that subnational govern- ments can affect by their own actions, especially by changing tax rates, but also by imposing new taxes or repealing old ones, by changing the tax base, and by varying administrative effort. Access to marginal revenues is key to fiscal decentralization because it gives subnational governments control over the size of public spending within their jurisdictions. Objectives and Constraints Although the overriding objective of tax assignment is to increase the access of subnational governments to marginal own revenues, not all ways of increasing access are equally satisfactory. Subsidiary objectives and con- straints inevitably influence the means used to increase access. DESIRABLE ATTRIBUTES OF SUBNATIONAL TAXATION. Decentralization works best when taxes and the benefits of public spending (or costs incurred by governments) are closely related. In the extreme case of user charges and fees, payments act almost like prices for private goods, in that citizens get what they pay for and pay for what they get. The implication for tax assign- ment is that revenues from taxes closely related to identifiable benefits or public expenditures should be levied by the level of government that pro- vides the benefits or incurs the costs. For example, revenues from taxes on motor fuels are intended to reflect the benefits of using roads and highways, and therefore should go to the governments that finance their construction and maintenance; revenues from taxes on alcoholic beverages and tobac- co products should go to the governments that incur the health-related pub- lic costs occasioned by consumption of those products. This raises a subsidiary question: should revenues from benefit-related charges and taxes be assigned (earmarked) to finance the corresponding activity, or should they be part of the general revenue of the level of gov- ernment that provides the services in question? Optimal resource alloca- tion requires that prices (or taxes intended to serve as quasi-prices) equal the marginal cost of providing services. Marginal cost pricing may provide more or less revenue than is needed to finance activities (depending on 96 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO whether average cost is rising or falling), and funds that are earmarked may be trapped in suboptimal places, as when gasoline taxes are used to fund freeways, instead of mass transit in the United States. User charges, fees, and taxes related closely to benefits are most likely to be feasible for financing services that, although provided publicly, exhib- it an important characteristic of private goods: the feasibility and likelihood of excluding those who do not pay. Toll roads cannot be used without pay- ing for access. Those who do not pay motor fuel taxes cannot use roads and highways, even if access is not restricted. In theory it would be possible to deny access to education and health care, but this is rarely done for social reasons, especially in the case of primary education and basic health care. Inherent in the nature of public goods is the inability to charge those who benefit from them. Thus much of public spending provides generalized ben- efits that must be financed by taxes that are only loosely related to the bene- fits of public services: general sales taxes, income taxes, and propertv taxes. When production and consumption do not occur in the same jurisdic- tion (because goods are traded between jurisdictions) or individuals do not earn income where they live (because they commute between jurisdic- tions or have investments in other jurisdictions), an important two-stage question arises in the attempt to identify beneficiaries of generalized pub- lic services. First, are the benefits of public services provided primarily to businesses or to individuals? And, second, when such benefits are provided to individuals, are they more closely related to (a) consumption or the res- idence of taxpayers, or to (b) production or the source of income? If gen- eralized benefits are provided primarily to business (and are thus more closely related to production than to consumption), an origin-based sales tax or a commercial property tax might be appropriate.2 If such benefits are provided to individuals and are more closely related to consumption or res- idence of the taxpayer, a destination-based sales tax, a residence-based indi- vidual income tax, or a residential property tax might closely reflect the ben- efits of public services. But if such benefits are related more closely to the earning of income, a source-based individual income tax would be more appropriate. Tax assignment must recognize administrative reality. Some taxes that are attractive in theory may not be administratively feasible or may require inordinate amounts of administrative resources, something Mexico cannot afford. Subnational governments ordinarily should employ debt financing only for capital projects; they should not borrow for current expenditures. If this rule is to be respected, subnational governments need stable sources of own revenue (expenditures of subnational governments are assumed to be rel- atively stable.) Therefore, relatively volatile sources of revenues, such as the corporate income tax and taxes on natural resources, should not be assigned to subnational governments. PRINCIPLES OF DECENTRALIZATION 97 UNDESIRABLE ATTRIBIUTES OF SUBNATIONAL TAXATION. If the benefits of free markets are to be realized, subnational governments should not impose taxes that distort the location of economic activity. Taxes levied at the ori- gin of production or the source of income are more likely to distort loca- tional decisions than are taxes levied at the point of consumption or resi- dence, unless they reflect benefits of services provided to businesses or to employees at their place of employment. Governments often wish to levy taxes on business that exceed the value of benefits provided to (or costs incurred on behalf of) business. This may be done because taxing business is more popular than taxing people (or the products they buy) or because it is thought that taxes on business can be exported to nonresidents of the taxing jurisdiction. Taxes on business that exceed the value of benefits provided are likely to distort the location of eco- nomic activity, unless they are levied at uniform rates throughout the country. But taxes that must be levied at uniform rates do not provide a mar- ginal source of revenue for any one subnational government, and thus do not further fiscal autonomy. Moreover, agreements between jurisdictions to fix tax rates (or the imposition of uniform rates by a higher level of gov- ernment) represent a form of cartelization that encourages overexpansion of the public sector and inefficiency. Competition among subnational gov- ernments both prevents excess taxation of business and impedes ineffi- ciency in government. Some taxes can be exported to nonresidents if they are imposed by (or for) subnational governments. The best examples of exported taxes are taxes on business that cannot be shifted to consumers or to labor and thus are borne by owners of business, many of whom may be nonresidents. (The view that taxes on business will automatically be shifted to consumers, many of whom are nonresidents, is generally incorrect. Exporting to nonresident con- sumers is likely to occur only when the taxing jurisdiction dominates the national market for the product taxed.) In the absence of untaxed compe- tition (for example, from imports), excises imposed on products by the pro- ducing jurisdiction could be exported to consumers throughout the nation. Tax exporting is generally undesirable. Not only does it impose unfair burdens on nonresidents, but it also cheapens the provision of public ser- vices in the taxing jurisdiction, encouraging overconsumption. An impor- tant exception involves taxation that is closely related to the benefits of a public service (such as taxes on motor fuels that finance roads used by non- residents); in this case, failure to export taxes to users is undesirable. Again, this points to the superiority of destination-based sales taxes and resi- dence-based income taxes over production-based sales taxes and source- based income taxes. The former are much less likely to be exported than the latter. Except in rare cases, subnational governments within a nation differ sub- stantially in their inherent fiscal capacity. Fiscal capacity may be measured 98 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO by applying a representative tax system-that is, by calculating how much revenue each state would receive, per capita, if it levied the taxes typical of all states. Although such measurement depends on the structure of the typical state tax system, it is unlikely to be misleading. Some states are sim- ply richer or poorer than others, as measured, for example, by per capita income or per capita GDP. But there may be other explanations. For exam- ple, some activities (for example, tourism) are easier to tax than others. Particularly important is the presence of geographically concentrated, high-value natural resources that can be taxed by subnational govern- ments. Where horizontalfiscal disparities are important, equalization grants may be appropriate to allow all states to provide comparable levels of ser- vices with comparable levels of tax effort. (Tax effort can be measured by comparing actual tax collections with collections under the representative tax system. Subnational governments that collect substantially more than under the representative tax system are exerting high tax effort; those that collect substantially less are exerting low tax effort.) OTHER CONSIDERATIONS. Responsibility for income redistribution via taxes should be lodged primarily with the central government. Attempts by subnational governments to redistribute income are likely to be futile, for example, because taxes intended to be borne by capital are likely to be shift- ed to labor and to owners of land. Moreover, they are likely to distort the location of economic activity or to result in tax exporting. Public spending should be lodged at the lowest possible level of gov- ernment to avoid spillovers of benefits between jurisdictions, realize economies of scale, and maximize the fiscal autonomy of those affected by public spending. This principle has a counterpart in the field of tax assign- ment, because of the seemingly inevitable existence of vertical fiscal imbal- ance. Higher levels of government can effectively administer virtually any tax that lower levels can administer, but the converse is not true; some forms of taxation do not lend themselves to implementation by or for lower lev- els of government. Moreover, some taxes are not appropriate for local gov- ernments, even if they are administratively feasible; customs duties are per- haps the best example. To minimize vertical fiscal imbalance, it is desirable to assign to subnational governments all taxes that are administratively fea- sible and appropriate for them. In judging administrative feasibility and appropriateness, one should consider the use of subnational surcharges on the tax base of a higher level of government and independent legislation and administration by the subnational government. It is often suggested that subnational governments should be rewarded for exerting high tax effort, for example, by providing matching grants that reflect tax effort. In general there is little analytical support for this propo- sition. Unless there are good reasons to believe that tax effort would oth- PRINCIPLES OF DECENTRALIZATION 99 erwise be artificially low, it is generally better to assume a neutral stance toward subnational tax effort. To overcome inertia, however, it may be desirable to encourage tax effort during the transition to a decentralized system. Assignment of Taxes to Cover Generalized Benefits Subnational governments should, to the extent possible, rely on taxes, fees, and charges that are related to benefits received. Using taxes to cover the generalized benefits of public spending has the following implications for tax assignment. First, if, as suggested, most of the public services provid- ed by subnational governments are provided to individuals, not busi- nesses, taxation that is likely to be borne directly by individuals or shifted to them by businesses in a predictable manner is preferable to taxation of business, including taxes that may be shifted to individuals, but in an unpredictable manner. This implies that a state individual income tax or a state value added tax is preferable to a state corporate income tax. If ben- efits are provided where people live, not where they work, destination- based indirect taxes and residence-based income taxes are likely to be preferable to origin-based indirect taxes and source-based income taxes. Finally, if subnational taxes are seen as payment for services providing gen- eralized benefits, coverage should be as broad as possible, given concerns for equity; subnational taxes should be "mass taxes," not taxes paid by a select few. Because the progressive individual income tax is the tax mechanism of choice for both income redistribution and endogenous macroeconomic stabilization, it should be reserved primarily for use by the central gov- ernment. This does not mean, however, that subnational governments should not also rely on the individual income tax. A flat-rate individual income tax may provide a satisfactory surrogate for benefit-related taxa- tion. Under certain circumstances, a flat-rate subnational tax can be imposed on the same base as a graduated-rate national tax.3 The assignment of taxes on natural resources raises difficult philosoph- ical and political questions as well as economic issues (see McLure 1994). For the most part, economic arguments favor assigning taxes on econom- ically important natural resources to the national government, especially if the resources are geographically concentrated in only some taxing juris- dictions. (Taxes on common low-value and ubiquitous resources such as sand and gravel can be assigned to subnational governments.) Because rev- enues tend to be unstable, taxes on natural resources are ill-suited for use by subnational governments, which generally need stable sources of rev- enue. Moreover, subnational governments may not be in as good a posi- tion as national governments to impose taxes on economic rents, general- ly agreed to be the most satisfactory form of tax on natural resources. In 100 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO theory, taxes on economic rent, being a tax on surplus, do not affect pro- duction decisions; the pattern of production that maximizes before-tax income also maximizes after-tax income. Taxes on production encourage inefficient exploitation (high-grading) by discouraging production that would cover marginal costs, but not the tax. Property taxes on deposits of natural resources are even worse, because they encourage uneconomical- ly rapid exploitation in order to avoid future taxes. If effective ownership of natural resources is in private hands, subna- tional taxes may be exported to nonresident owners. Effective ownership may be in private hands even if the state owns the natural resources. This may occur, for example, when resources are exploited under concessions, subject to terms that are not adjusted for changes in taxation. Finally, if important natural resources are distributed unequally across a nation, sub- national taxation may produce or aggravate horizontal fiscal disparities. Besides raising questions of equity-probably the most important reason to assign taxation of natural resources to the federal government-such dis- parities may induce economically inefficient allocation of private resources to jurisdictions with low tax rates or high levels of public services. These inefficiencies are likely to be relatively small, except in extreme cases, as in Alaska, where the combination of vast natural resources and a small pop- ulation allows the financing of generous public services (including annu- al per capita grants) with low tax rates (see Mieszkowski and Toder 1983). Table 1.1 summarizes the current wisdom on tax assignment. Philosophical arguments are more ambiguous. Advocates of taxation by subnational governments commonly argue that natural resources and the right to tax them are part of the "heritage" of such jurisdictions. But it can equally be argued that the resources and tax revenues are the heritage of the entire nation. The resolution of this issue calls into question the concept of nationhood as it is understood in particular countries. All things con- sidered, it seems appropriate to assign most taxes on important natural resources to the federal government. Since the Mexican constitution makes this assignment, taxes on natural resources are not discussed further here. Methods of Revenue Assignment This section evaluates four methods of revenue assignment, including tax assignment: independent legislation and administration, tax and revenue sharing, state surcharges, and dual administration of surcharges. INDEPENDENT LEGISLATION AND ADMINISTRATION. A system of tax assignment based on independent legislation and administration provides maximum state sovereignty, but this sovereignty is bought at a high price-a price that Mexico can ill-afford and need not pay. State taxes in the United States illus- trate these problems. Inconsistent state laws and administrative practices PRINCIPLES OF DECENTRALIZATION 101 Table 1.1. Attributes, Strengths, and Weaknesses of Various Techniques of Tax Assignment Techniques of revenue assignment Subnational surcharge: National base of higher legislation; level Subnational subnational government: legislation and rate and administered Tax Revenue administration administration by higher level sharing sharing Assignment of taxing powers-best practice: Defining Tax Subnational National National National National Base Setting Tax Subnational Subnational Subnational National National Rate(s) Administration Subnational Subnational National National National Criteria of choice: Autonomy Best High: choice High: choice None None of tax rate of tax rate Complexity High Low: tax on Low Low Low local activity Horizontal Potentially Potentially Potentially Potentially Can offset Disparities high high high high horizontal disparities Overall Generally not Generally Poor Current appropriate best practice for Mexico approach Most appropriate method of assigning revenuesfrom major taxes to subnational governments: Excises Best Transition VAT Best Transition Individual Best Income Payroll Tax Best Corporate Income Best Property Good (greater Good (most state control) efficient solution) Natural Resources Best 102 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO raise the costs of compliance and administration and create inequity. Use of source-based taxes such as the corporate income tax distorts the geo- graphic allocation of resources and may result in tax exporting. Even the retail sales tax, which is imposed by 45 states and the District of Columbia, contains an origin-based component that distorts locational decisions, since a substantial amount of revenue (an estimated 40 percent) is derived from taxes levied on sales to businesses instead of sales to households.4 Taxes on natural resources generally take the form of state severance (pro- duction) taxes, income taxes, and local property taxes, all of which are dis- tortionary. (Distortion is mitigated by the fact that the corporate income tax and sales taxes on business inputs are deductible in calculating liabilities under the federal tax.) There is substantial tax exporting, especially by resource-rich states.5 In short, this system is not appropriate for Mexico. TAX AND REVENUE SHARING. In tax sharing the central government directs given fractions of revenues from selected taxes to the various states where the revenue originates. In this approach, widely used in the former Soviet Union, states have essentially no control over the choice of taxes, tax base, tax rates, or tax administration; they are thus fiscally weak.6 Although dif- ferent in appearance, tax sharing closely resembles revenue sharing and grants. Unlike tax sharing, revenue sharing returns funds to subnational governments on the basis of formulas, instead of to the jurisdictions of ori- gin. Like tax sharing, it provides no subnational autonomy over the choice of taxes, the definition of the tax base, the setting of tax rates, and tax administration. The use of formulas does, however, offer the possibility of reducing horizontal disparities among jurisdictions. STATE SURCHARGES. An ideal system for many countries (including Mexico) consists of subnational surcharges levied on a tax base (or tax bases) defined and administered by the central government, perhaps with input from the states. In this scheme, subnational governments would exercise the all- important choice of tax rates, but most of the complexity, inequities, tax exporting, and locational distortions inherent in subnational choice of tax bases and administration of taxes would be avoided. That is, state sur- charges combine simplicity with subnational sovereignty over tax rates. Canada makes substantial use of this type of system. Ideally, surcharges could be imposed on some combination of excise tax, residence-based income tax, and value added tax (levied at the destination of sales). States might impose payroll taxes, but on a uniform base. State surcharges on the corporate income tax may be required in order to raise adequate revenues. Of the three methods of tax assignment, subnational surcharges provide the optimal combination of subnational autonomy and simplicity. PRINCIPLES OF DECENTRALIZATION 103 Subnational governments have autonomy over rates-the most important issue for determining the level of subnational spending-while the central government defines tax bases and administers taxes, minimizing the costs of administration and compliance. By comparison, independent legislation and administration are too complicated, and tax sharing eliminates all subnational autonomy. Since tax surcharges may produce horizontal fis- cal disparities and vertical fiscal imbalance, they can be combined with rev- enue sharing from the federal taxes. Despite the manifest advantages of tax surcharges as a system toward which Mexico should move, the movement could be a gradual one con- sisting of three stages. In the first stage, the present system of tax shar- ing/revenue sharing would be replaced with a system in which revenues from certain taxes are shared with the states on a derivation basis. For this purpose, the determination of derivation for each shared tax should approx- imate as closely as possible the technique that would be used to impose a state surcharge on the federal tax in question. In the second stage, tax shar- ing would be replaced with a system of subnational surcharges in which state tax rates are constrained to be equal. In the third stage, states would be given greater-perhaps complete-control over state surcharge rates. DUAL ADMINISTRATION OF SURCHARGES. It is common for surcharges of sub- national governments to be administered by national governments. This arrangement seems preferable to the alternatives of dual administration, in which the federal government administers federal taxes and state gov- ernments administer surcharges. In some countries, subnational govern- ments administer national taxes, but this is less common. Administration of the value added tax by the German lander and the Canadian province of Quebec are notable examples. This option deserves little discussion in the Mexican context, where it has an unsuccessful history. On the one hand, federal administration of state surcharges raises several concerns: Will the federal government efficiently transfer the revenue collected on behalf of the state? And will the federal and state governments have dif- ferent priorities in the assignment of scarce administrative resources? On the other hand, dual administration raises the specter of costly duplication of administrative efforts, overly burdensome compliance, and the possi- bility that administration will not be consistent between states or between states and the federal government. The risk that the federal government will not pay to state governments the revenue collected on their behalf can easily be addressed by having tax- payers deposit taxes directly in the bank accounts of each government. A greater risk is that taxpayers might meet their obligations to the federal gov- ernment, but not to state governments, and that the federal government would condone this behavior. It is desirable to find ways to prevent this. 104 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Federal and state tax administrations may have quite different priorities. First, the federal government may prefer to concentrate its resources on administering the taxes that yield the greatest marginal revenue for the fed- eral government; these may not be the taxes that yield the greatest marginal revenue for the state.7 Second, in the administration of a particular tax, the federal government will presumably prefer to concentrate its activities on taxpayers whose expected marginal yield of revenue for the federal gov- ernment is highest. Since economic activity is not homogeneous throughout the country, the federal tax administration might pay less attention than state tax administrations to taxpayers in some states. (What might be a "big fish" for the state might be a "small fry" for the federal government.) To the extent that the federal administration ignores taxpayers in poorer states, horizon- tal fiscal disparities are accentuated. Third, the assumption that the federal tax administration allocates its resources rationally may not be valid; the administration's coverage of taxpayers in various parts of the country may vary widely for reasons that have no basis in sound policy. (It may be dif- ficult to hire trained auditors to work in the hinterland; alternatively, cov- erage in some rural areas may be "denser" than in the major cities.) One risk of dual administration is obvious: instead of filing a single tax return with the federal government, the taxpayer would file a return with each state where there is an actual or potential tax liability. Similarly, the return would be potentially liable to audit by the tax authorities of all these governments. Under a system of dual administration, tax adminis- trators in all states must apply the law consistently. Although a single law might, in principle, apply throughout the country, it might not be inter- preted uniformly by state tax administrations. The treatment of interstate transactions is especially important. In a sys- tem with federal tax administration, the flow of information on interstate sales would stay with the federal government.8 By comparison, in a dual system, the information on interstate transactions would have to flow between state administrations. The states could be given a role in the supervision of the federal tax administration, essentially making it a national agency that is accountable to both the federal and state governments, instead of afederal agency that is accountable only to the federal government. This would bring the ben- efits of uniformity, avoid duplicate effort, and respect state priorities. Transfers and System Integration Federal systems are characterized by a formal (constitutional) division of powers between central and subnational levels of government. These pow- ers can be assigned to one or the other order of government (self-rule) or they can be concurrent (shared rule). Theoretically, there could be a PRINCIPLES OF DECENTRALIZATION 105 serendipitous matching of revenue and expenditure responsibilities among the two orders of government such that there would be no need for inter- governmental transfers. However, in virtually all federations, the central government has de facto, if not de jure, capacity to raise revenue in excess of its expenditure responsibilities. This is certainly the case for the Mexican federation, even more so now that state taxation responsibilities have been transferred to the center and significant expenditure responsibilities devolved to the states. As mentioned earlier, in general, two sorts of fiscal balance (or transfer) issues arise in federal systems-vertical fiscal imbalance and horizontal fis- cal imbalance. Vertical fiscal imbalance relates to the allocation (or, rather, misallocation) of revenue-raising capacity relative to expenditure respon- sibilities. The Mexican federal government has revenue-raising capacities well in excess of its expenditure responsibilities. Although a high degree of vertical fiscal imbalance in favor of the federal government can be ratio- nalized on a variety of grounds (such as economies of scale in tax collec- tion) and does exist in some mature federations (such as Australia), a cor- responding set of intergovernmental grants (federal-state and perhaps federal-municipal) is required to address the revenue shortfall of subna- tional governments. Horizontal fiscal imbalance has to do with fiscal capacity across the states. Fiscal capacity has a precise meaning: it relates to standardized per capita revenues, where these are calculated as the product of the all- state average tax rates and commonly defined per capita state tax bases. The intergovernmental transfers associated with alleviating these hori- zontal fiscal imbalances are typically referred to as equalization transfers and programs. Analytical Underpinnings of Intergovernmental Transfers Expenditure responsibilities at the state (subnational government) level usually exceed the states' revenue-raising capacity, lending to both verti- cal and horizontal fiscal imbalances. These imbalances occur for various rea- sons. The most obvious is that they arise from the allocation of taxing and expenditure responsibilities under the constitution. Or they could result from "first-mover" advantage. That is, according to the constitution, the states may have adequate capacity to raise revenue but may have to share the major sources of revenue with the federal government, which has pre- emptively occupied these tax areas. Unless the federal government is will- ing to transfer tax room to the states, there is no way, short of excessively high tax rates, that the states can exercise their taxing authority. Third, it may be more efficient for the federal or central government to collect taxes on behalf of the states. This can also occur for a variety of reasons-inad- equate state-level collection capacity, economies of scale, and the presence 106 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO of spillovers. This last reason is becoming more important: with increas- ing economic integration the optimal jurisdiction for collecting taxes is expanding (that is, the tax bases are becoming more "mobile") relative to the optimal jurisdiction for spending. In order to "internalize" these poten- tial spillovers or externalities, taxes are collected at the national level with an understanding that some or all of the revenues will be returned to the states under a set of transfer arrangements. Whatever the reason, both horizontal and vertical fiscal imbalances exist at the subnational level. What is the role of intergovernmental transfers in ameliorating horizontal fiscal imbalances across the states? ACCOMMODATING HORIZONTAL IMBALANCES. The literature focuses on two cri- teria for assessing horizontal imbalances: fiscal efficiency and fiscal equity. Drawing on Boadway and Hobson (1993), the underlying theoretical model postulates that an individual's "comprehensive income" is the sum of his or her market income, W, and net fiscal benefits (NFBs), defined as the dif- ference between the value of government expenditures and taxes paid. This leads to the following basic equation of labor-market equilibrium: W, + NFBi = Wj + NFB. where subscript i refers to state i and subscript j refers to state j. Thus an individual will be indifferent between living in state i and state j when com- prehensive incomes are the same in both states. But now assume that NFBj > NFBi. Net fiscal benefits in state j can exceed those in state i for a variety of reasons. For example, states with high source-based tax revenues (from, say, resources or corporate income taxes) or with high residence- based revenues (because they are higher-income states) could finance a given level of public goods and services with lower tax rates, which, in turn (and in the absence of capitalization) implies higher net fiscal benefits. Moreover, NFBs could also arise on the expenditure side: it may be more costly to provide a representative bundle of goods and services in some states than in others. These differences in NFBs provide the backdrop for what the theoretical literature refers to as the fiscal efficiency and fiscal equi- ty rationales for equalization, that is, for ameliorating horizontal differences in fiscal capacity. FISCAL EFFICIENCY. Consider fiscal efficiency first. Output maximization requires that W. = Wi, namely that individuals distribute themselves across states so that their market-based marginal products are everywhere iden- tical (that is, in state i and state j). However, if, as assumed, NFB1 > NFBj, then individuals will be willing to move to state j even if this means that W. < Wi, in order to ensure that their comprehensive income (W + NFB) is PRINCIPLES OF DECENTRALIZATION 107 in equilibrium. The result is inefficient or fiscally induced migration. The solution proposed in the literature is to provide a set of intergovernmen- tal transfers (they could be federal-state transfers as in Canada or interstate transfers as in the German federation) in order to ensure that NFB. = NFB., which, in turn, implies that, in equilibrium, Wi = W1. Hence, migration will be efficiency-driven, not NFB-driven. Several caveats are in order. First, since the underlying model focuses on individuals, equalization payments to provinces are a second-best solu- tion. The first-best solution is a set of payments to individuals. But this is typically ruled out on constitutional or practical grounds. Hence, the lit- erature argues that the presence of equalizing grants allows states, in prin- ciple, to equalize NFBs. In practice, this is highly unlikely to occur since it implies that similarly situated individuals in both states are treated simi- larly by taxes and expenditures. Second, the model implicitly assumes zero migration costs. If there are costs to migration, then NFBs can differ up to this cost differential with no deleterious impact on migration effi- ciency. The third caveat is more important. Although the model is gener- al in the sense that it can accommodate any degree of capitalization of NFBs in terms of rents and wages, its typical application tends to argue for full equalization of revenues up and down, as in Australia. This application effectively assumes zero capitalization. Phrased differently, it assumes that an increase in revenues in state i, for example, is equivalent to an increase in NFBs in state i. But this does not hold if the revenue increases become capitalized in wages, rents, and so forth. This is not a defect in the theory, only in the way the theory tends to be applied. Finally, in Canada at least, the emphasis is shifting away from the fiscal efficiency case for equalization to the fiscal equity case, because the efficiency gains may not be very significant in any event (Watson 1986). FISCAL EQUITY. The traditional case for equalizing federal-state transfers in a federation is premised on fiscal equity, which in turn is based on the pub- lic finance concept of horizontal equity-equal treatment of equals. The argument is straightforward: individuals with similar employment incomes (Wi = W.) should, on horizontal equity grounds, also have similar NFBs. Thus "fiscal equity requires that all differences in NFBs be equalized regardless of their source" (Boadway and Hobson 1993, p. 89). Under what has come to be viewed as "broad-based horizontal equity," the underlying assumption is that people who are equally well off in the absence of gov- ernment should also be equally well off in the presence of both levels of gov- ernment. Hence the role for the federal government, under broad-based hor- izontal equity, is to design discriminatory transfers (essentially via equalization) to provincial governments so as to "undo" any differential NFBs that arise because of government policies (both federal and state). This 108 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO approach turns the federation into something comparable to a unitary state (Boadway 1998; Hobson 1998). One can mount a convincing case, how- ever, that the goal of broad-based horizontal equity is the very antithesis of federalism. Nonetheless, the theoretical conception of fiscal equity underpins the comprehensive equalization system in place in Australia, where state rev- enues are leveled upward and downward to an all-state per capita level and then further adjusted to take into account differential state needs and costs of providing the average all-state expenditure bundle. Not all federations accept this approach to fiscal equity. The Canadian equalization system focuses only on revenues and then brings only low-revenue-capacity provinces up to some acceptable standard (it does not bring above-aver- age provinces down to this standard). And the United States does not even have an equalization program, presumably on the grounds that dif- ferential NFBs will be fully capitalized so that there is "nothing to equal- ize." The degree to which federations seek to ameliorate horizontal imbal- ances across their subnational governments appears to be derived from the deeper political values and logic of the federal contract rather than from a simple application of fiscal theory. Vertical Balance and Conditionality This section reviews the theory and practice of vertical fiscal transfers, beginning with a focus on conditionality. Conditional grants come in a variety of forms and rationales. The tra- ditional public finance literature has devoted substantial attention to the "interjurisdictional spillover" rationale for conditionality. An interjuris- dictional spillover (or externality) is said to exist when the activities of the public sector in jurisdiction A provide benefits to the residents of another jurisdiction (Boadway and Hobson 1993, p. 96). In the presence of such externalities, jurisdiction A has an incentive to underspend on the provi- sion of public goods in the presence of such externalities. The greater the proportion of external to "own" benefits, the greater the underprovision will likely be. The obvious way to correct for this is to provide a matching conditional (expenditure-specific) grant where the degree of matching (degree of cost sharing) relates to the degree of interjurisdictional spillover. Although this argument is well grounded in theory, it is rare in practice to find open-ended, shared-cost conditional grants where the degree of cost sharing is predicated on the differences in intergovernmental spillovers. But this externality can be addressed in other ways. For example, the per capi- ta value of conditional grants could vary across subnational jurisdictions in accordance with the degree of spillover. And the various equalization programs in federal systems imply that these moneys are designed to ensure that all states can provide national, average public services to their PRINCIPLES OF DECENTRALIZATION 109 citizens. This can only be an implicit condition, since equalization grants are generally unconditional. The most typical forms of conditional grants link the transfers to spend- ing in specific areas or for specific purposes. These specific-purpose grants are intended to ensure that all individuals, as a right of citizenship, receive minimum levels of certain public goods and services. Presumably, this explains the degree of conditionality embedded in Mexico's Ramo 33 trans- fers, especially for health and education. Although some forms of conditionality may embody problematic incen- tives (such as open-ended, 50-50 shared-cost grants that distort state expen- diture priorities because states are spending "50 percent pesos" in the grant areas and "100 percent pesos" elsewhere), or overly strict conditions that prevent states from achieving the goals of the conditions in more effi- cient ways, conditional grants can also be employed creatively. Several examples come to mind. First, cost sharing can be combined with subna- tional spending priorities at the margin. For example, the cost sharing can be close-ended (intramarginal) rather than open-ended. And the amount of this intramarginal sharing can be different for different states (it even could accommodate the spillover rationale alluded to earlier). At the mar- gin, however, states would be spending 100 percent their own pesos on all spending areas. The United States' experience with shared-cost grants a few decades ago suggests another degree of flexibility. The federal government provided cost-sharing grants for highway construction, where the degree of sharing incorporated considerations relating to the states' fiscal capacities and the costs per mile of highway construction (mountain states required more than plains states, for example). In effect, this brought both fiscal capacity and fiscal needs into the design of conditional grants. Canada's experience with conditional grants presents another perspec- tive. In the 1950s and 1960s, when the Canadian welfare state was in its embryonic stage, the federal government provided generous shared-cost programs in areas such as welfare, health, and education. As these pro- grams became established, they developed receptive and demanding citi- zens in their respective provinces and came to be viewed as an integral part of the Canadian social contract. As this occurred, the Canadian federal gov- ernment could and did gradually abandon the shared-cost and highly con- ditional nature of these grants and converted them into block-funded unconditional programs, with the important proviso that the provinces agreed to a set of pan-Canadian principles in these expenditure areas. In turn, this suggests that there may be an optimal evolution of conditional transfers. Previously centralized federations that embark on a process of decentralization will probably have to resort, initially, to highly conditional grants. However, as the federation evolves and as citizens come to view 110 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO these conditional-grant programs as entitlements, the federation can grad- ually eliminate the conditional nature of these transfers and use instead a set of mutually agreed operating principles. In either case the result is greater subnational autonomy. Since Mexico is only beginning the process of meaningful decentraliza- tion, some of the federal-state transfers will probably have to be conditional. But even at this early stage of decentralization, the design and implemen- tation of conditional grants can be flexible enough to accommodate both conditionality and an important degree of subnational autonomy. Horizontal and Vertical Transfers in Practice Different federations combine transfers in quite different ways. Canada is one of the few countries that compartmentalizes these two types of trans- fers. The federal government first overlays an equalization system on the taxation capacities of the various provinces. This is "pure" horizontal equalization since the equalization payments only go to low-fiscal-capac- ity provinces. Then, with all provinces thus brought up to the equalization standard, the vertical transfers are designed to be equal in per capita terms, subject to a set of pan-Canadian principles on which Ottawa and the provinces have jointly agreed. The Australian system is quite different. Because the degree of subna- tional fiscal imbalance is so large, the unconditional financial adjustment grants overseen by the Commonwealth Grants Commission (CGC) incor- porate both horizontal and vertical fiscal-balance concerns. That is, all Australian states receive CGC grants. One could interpret the per capita level of grants going to the "richest" state, Victoria, as embodying the all-state level of vertical transfers and the additional grants going to the other states as embodying the horizontal component of the grants. Beyond this, Australia has a set of specific-purpose payments (conditional grants) that, in the aggre- gate, exceed the unconditional CGC grants. The continuing importance of conditional grants in Australia reflects the more centralized nature of their federation (or, analytically equivalent, citizen preferences do not differ across states and that the conditions effectively capture these preferences). The German system is different again. The shared taxes are distributed to the hinder according to criteria that embody derivation, equal-per-capi- ta, and equalization principles. The resulting differences in fiscal capacity are then subject to an overarching interlander revenue-sharing pool; rich lander contribute to the pool, and poor lander draw revenues from the pool. This revenue-sharing pool is an exercise in pure horizontal equalization. No magic formula applies to all federations. Moreover, there is ample evi- dence that the approaches to intergovernmental transfers chosen are far from arbitrary; they find their rationale and underlying logic in the deep- er political values of their respective federations. PRINCIPLES OF DECENTRALIZATION 111 Debt and Borrowing Two levels of policy influence borrowing and debt management by state and municipal governments-the policies of subnational governments themselves and the policies of the national government (or the state for municipal government) that set the constraints and incentives for the local governments. The proper policies for the subnational governments are straightfor- ward-they should only borrow to fund an activity (investment) that will yield a rate of return to society above the interest rate and sufficient to ser- vice the debt. A project that generates its own direct revenue is most like- ly to meet these criteria, but it can also be met in an investment like local roads or school construction that is associated with growth of economic activity and tax revenue adequate to cover the debt service. Since the principal of a loan typically comes due before the full benefits of the investment arrive, the stock of debt should be kept small enough so that debt service (interest plus amortization that is not scheduled for refi- nancing) is less than the current balance before debt service. Alternatively, amortization not scheduled for refinancing should be less than the current account balance. To these minimum criteria for borrowing one must add a safety factor, so that even in the face of adverse shocks the local government will main- tain its fiscal independence and thus meet the conditions for market-pre- serving federalism. Then, even after paying scheduled amortization, there is enough current surplus left to fund part of the investment program, that is, without borrowing for all investment. These rules apply to any level of government. In a perfect world, decisionmakers would want to do this to maximize social welfare, or voters could discipline them to act this way. In the real world, there are many problems of information, and decisionmakers have short time horizons, so governments at all levels often borrow in excess of these rules. For local governments, there is an extra complication in that if they overborrow, a higher level of government may bail them out, in effect rewarding their imprudence with extra resources. In other cases, the high- er level of government may impose spending requirements on them or limit their revenue options, making unsustainable deficits almost unavoidable and giving the local government a reason to expect a bailout from above. To deal with such problems, the higher level of government may establish and enforce rules that prevent the local government from borrowing impru- dently. These policies are our central concern here. Research on Latin America and the rest of the world indicates several conditions that are needed to achieve sound management of debt and bor- rowing by states and municipalities (see table 1.2). 112 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 1.2 Institutional Arrangements to Set and Keep Hard Budget Constraints on States Constraint Institutional arrangement Hard budget constraint from the central * Rule-based transfers government to subnational governments * Firm allocation of spending responsibilities Constraints on borrowing * Ex ante constraints * Ex post consequences and resulting incentives * Enforcement of payment by subnational governments * Enforcement of losses on banks with bad loans to uncreditworthv subnational governments, via bank regulation * Independence of Central Bank and regulators Autonomy of subnational government * Ability to control spending and to reduce costs and raise own revenue costs * Ability to increase tax rates and improve enforcement Transparency: timely disclosure of * Regular publication of data information to improve accountability * External auditing of policymakers * Legislative oversight Source: Dillinger, Perry, and Webb 2000. HARD BUDGET CONSTRAINTS. First and foremost, a firm allocation of expen- diture responsibilities is critical for establishing a hard budget constraint for subnational governments. If the central government can effectively delegate functions to subnational governments to go along with the dele- gation of revenue sources, central spending and deficits are more likely to be contained. If this is not the case-because, for example, the constitution or the law mandates resource transfers without allocating equivalent responsibilities-the central government can find itself with a constitutional obligation and political expectation to continue providing some services, even after revenues or tax bases have been turned over to subnational governments with the understanding that they will do the task. Or it may have to resume spending on functions when subnational governments experience a fiscal failure. PRINCIPLES OF DECENTRALIZATION 113 Basing transfers exclusively on clear rules is the other necessary ingre- dient for a hard budget constraint. Wherever there is recourse to significant discretionary transfers, including matching grants, subnational govern- ments will have an incentive to overspend in the expectation that they can get a larger transfer. BORROWING CONSTRAINTS. Although tax and spending policies create fiscal pressures, whether they cause problems for macroeconomic management depends on whether the subnational governments face hard limits on their ability to borrow or to spend more (Ter-Minassian and Craig 1997). Unsustainable deficits are less likely if the central government controls sub- national borrowing ex ante. But how to enforce this in practice is not always clear when the subnational governments have considerable polit- ical autonomy. Pseudo-strict controls could make matters worse if central government approval creates the impression, and perhaps the self-fulfill- ing expectation, that the central government has extended a guarantee. The best controls would mimic the requirements of prudent lenders. To run deficits, subnational government must find a source of financing, which potentially includes contractual borrowing from private foreign or domestic banks (especially banks owned by subnational governments), issuance of domestic or foreign bonds, and the running up of arrears to sup- pliers and personnel. A creditor and the subnational government would only agree to finance unsustainable deficits if both sides expected to gain, most likely through some sort of federal bailout. The bailout can take many forms, including allowing the financial system (implicitly insured by the government) to count as an asset a subnational government debt that is not being serviced. Unsustainable deficits are also less likely if the central gov- ernment credibly commits not to have bailouts, prohibiting explicit bailouts and forcing subnational governments to service their debts, and if regula- tors force creditors to accept the losses implied by any failure of subnational governments to service debt. It is still an open question whether ex ante reg- ulation or ex post enforcement of debt service is more effective in preventing excessive subnational government borrowing. Although conflicts are also possible, both can work together and complement each other. Ex ante con- trols keep the problem from growing so big that it threatens the entire sys- tem, and ex post consequences increase appropriately the concerns of indi- vidual borrowers and lenders. Financing from the central bank often is what loosens the budget constraint for subnational governments, either directly by discounting subnational debt or indirectly by easing the national government's budget constraint or allowing commercial banks to roll over bad subnational debts. Unsustainable deficits are less likely when the central bank (and the bank supervisory agency) is more autonomous and has a strong anti-inflation mandate. 114 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Subnational governments may also accumulate excessive contingent liabilities. This is more likely to happen wherever subnational govern- ments are allowed to run their own pension regimes, when they own banks, and when they make concessions to the private sector without ade- quate regulation from above. AUTONOMY TO REDUCE COSTS AND RAISE REVENUE. The third group of insti- tutions relates to the capacity and autonomy of subnational governments to stay within the budget constraint. The first element is expenditure auton- omy. If subnational governments do not have autonomy over their expen- diture, there is no fiscal decentralization and no macro fiscal problem like- ly to come of it. Giving subnational governments autonomy over spending is, of course, the way in which decentralization can improve efficiency in matching the needs and desires of a diverse population. But to live within a sound budget constraint, subnational governments must have authority to control their costs. Too often central governments keep for themselves decisions (such as determining the wages of teachers and doctors) that crit- ically affect the costs of subnational governments, and a liberal decision may throw subnational governments into deficit. In particular, subnational governments must have the authority to spend less, particularly to cut per- sonnel, salaries, and pension benefits, collectively the largest single item of subnational expenditure, in order to be able to adjust to shocks or contribute to needed fiscal retrenchment. If central rules constrain this authority, it is more difficult to reduce deficits, and expectations of a central government bailout are higher. Thus unsustainable deficits should be less likely if sub- national governments have authority to cut their costs. With fiscal decentralization, subnational governments usually obtain cer- tain tax bases, but for reasons of politics, equity, and efficiency, these bases rarely cover all their expenses. Subnational governments always receive sig- nificant federal transfers. It is commonly believed that subnational gov- ernments will have smaller deficits if they rely more on their own tax bases (and have the power to change tax rates on the margin), because then they have the ability to adjust to shocks by increasing revenue. In addition, rely- ing on one's own resources may strengthen the incentives to control spend- ing. Unsustainable overall public sector deficits are less likely if subnational governments raise much of their own revenue and have enough flexibili- ty to change rates or impose new taxes. TRANSPARENCY. Although increases in competitive democracy accelerated the process of decentralization, sometimes to a pace that is problematic, democracy can contribute to responsible macroeconomic management, as well as to other worthy ends. When congress and opposition parties have access to accurate and up-to-date information about public sector finances, PRINCIPLES OF DECENTRALIZATION 115 they become effective watchdogs and bring the law and public opinion on the side of good management. It is thus important that subnational gov- ernments publish information on fiscal balances, revenues sources, expen- diture composition, borrowing, debt structure, and contingent liabilities. A federal government can impose such requirements on subnational gov- ernments. This works well administratively because the federal government is giving the states transfers, which can be made conditional on trans- parency and reporting requirements. Politically it works, as well, when there is some balance of parties at the federal level, and each side wants to be sure that the other side cannot profit from lack of transparency in the subnational governments. In this context, transparency not only creates pressures for good fiscal management but also acts in a virtuous circle with political competition. Interlinkages While this chapter has dealt separately with the areas of spending, taxes, transfers and borrowing, in practice they are all closely linked. For exam- ple, states can only control their debt if they can control spending and/or borrowing. Further, a federation can maintain a tough budget constraint in its transfers to states only if state fiscal problems do not threaten the nation's financial system or do not result in politically unacceptable short- falls in payments to service providers, like teachers. Or the dissatisfaction of a state with the tax transfers plan can lead it to threaten to withdraw from the fiscal pact if the government does not agree to increase transfers or pay for some expenditures that are normally the state's responsibility. Reforms to any part of the system must thus take account of the rest of the system and often must be accompanied by other reforms. Appendix Intergovernmental Transfers in Developed Federations All federations have a constitution that allocates powers on a self-rule, shared-rule basis. Beyond this, some federations are highly decentralized (Canada), while others are highly centralized (Germany, Australia). Some are legislativefederalisms, where subnational governments have substantial legislative powers especially with respect to expenditures (Canada, Australia, the United States). Germany, however, is a model of an admin- istrativefederalism because virtually all legislative power rests with the cen- ter and almost all implementation and administration is the responsibili- ty of the lander. All federal systems have to address both vertical and horizontal fiscal balance, although the United States, alone among federal systems, has no program of equalization directed toward ameliorating horizontal balances across subnational governments. Accommodating these fiscal imbalances falls to the system of intergovernmental transfers. In this sense, intergov- ernmental grants are the residual component of the expenditure, tax, and transfer assignment nexus, and the nature of these transfers varies marked- ly from federation to federation. Nonetheless, the nature and magnitude of, and incentives within, the system of intergovernmental transfers in each of these federal systems resonate closely with the underlying nature of the federation itself. Far from being merely a residual component of fis- cal federalism, the transfer system tends to embody the values and norms of the citizen-government and intergovernmental relationship consistent with the federation's social and political contract. To be sure, causation may run both ways. If grants are unconditional, the balance of power tilts toward subnational governments, and vice versa. Decentralized federations may dictate that most grants be unconditional. Phrased differently, the design of intergovernmental grants is not inci- dental to the underlying social and economic nature of the federation itself. This appendix examines intergovernmental transfers in a comparative fed- 116 PRINCIPLES OF DECENTRALIZATION 117 eralism context, beginning with the Commonwealth of Australia and then turning to Germany, Canada, and the United States. Australia Australia is not only a highly centralized federation but also a highly egal- itarian nation. For example, welfare payments are designed and delivered from Canberra and, as a result, are identical across the country, in sharp contrast to, say, Canada, where responsibility for welfare is provincial. Wage grids are also essentially uniform across Australia-university pro- fessors are on the same wage grid whether they teach in Perth, Sydney, or Launceston. More relevant for our purposes, the Australian states do not have effec- tive access to broad-based taxes (income taxes, sales taxes), and the taxes that they do levy are being eroded by a combination of global forces and high-court decisions (Courchene 1999). The Australian states are highly dependent on transfers: "Australia has by far the highest degree of verti- cal fiscal imbalance among the major federations in the world" (Walsh 1996, p. 115). However, the system of intergovernmental grants complements Australia's centralization and uniformity nicely. First, more than half of the cash transfers to the states are in the form of conditional grants (specific- purpose transfers, in the Australian context), which, in turn, enhances both centralization and uniformity. Second, Australia's approach to remov- ing horizontal fiscal balances-the financial adjustment grants monitored by the Commonwealth Grants Commission-is the most comprehensive among federal nations. The fiscal equalization principle that guides the Commonwealth Grants Commission is as follows: "Each State should be given the capacity to provide the average standard of State-type public ser- vices, assuming it does so at an average level of operational efficiency and makes an average effort to raise revenues from its own sources" (Commonwealth Grants Commission 1995, p. 1). Operationally, the equalization system works as follows. First, the 19 state revenues (at assumed common tax rates) are equalized both upward and downward to the all-state average. With revenues fully equalized, the 40 or so expenditure categories are then subject to upward and downward equalization to ensure that each state can deliver the average expenditure bundle at an "average level of operational efficiency." The end result of these calculations is a series of per capita "relativities"-that is, revenue- or needs- adjusted ratios relative to the national average. Hence, during 1994-95, for example, Victoria's relativity was 0.85 and Tasmania's was 1.54-that is, Victoria would receive 85 percent of its population share of overall Commonwealth Grants Commission grants, and Tasmania would receive 154 percent. This isfull revenue and expenditure (needs) equalization. 118 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO The Commonwealth Grants Commission is the key institution in Australian fiscal federalism. Its operations are open and increasingly trans- parent, which ensures that the resulting "relativities" are accepted by all Australians. Beyond this, Australia has an effective process of state-com- monwealth coordination to ensure overall macroeconomic cooperation and to preserve and promote the internal socioeconomic union (Courchene 1999). Thus Australia has latched onto a highly egalitarian equalization pro- gram and, more generally, a system of intergovernmental transfers that meshes well with the underlying homogeneity and egalitarian nature of its federation. Germany The German federation is also highly centralized, but in a way quite dif- ferent from Australia. Designated as an administrative federalism, all major tax rates are set centrally, with no variations allowed at the lInder level (although the lInder administer or collect these revenues). Apart from a relatively minor range of 1inder and municipal taxes (taxes on property, cars, and beer as well as fees of various sorts), most lander rev- enue comes from revenue-sharing arrangements with the center. The major shared or joint taxes include corporate and personal income taxes, capital taxes, and the value added tax. Some of this revenue sharing follows the principle of derivation, some of it is equal per capita, and some (especial- ly for the new lander) is based on equalizing principles. Beyond revenue sharing, there is a second and overarching tier-an interlander revenue-sharing pool. The rich lander contribute a share of their per capita revenues in excess of 102 percent of the national average (70 per- cent of per capita revenues between 102 and 110 percent of the national average and 100 percent of any revenues beyond this). The poorer lander then draw from this pool to bring them up to at least 95 percent of the all- lknder average. The guiding principle underlying intergovernmental trans- fers in Germany is the constitutional provision assuring "the uniformity of living conditions." Needs are also taken into account in operations of the interlander revenue-sharing pool, meaning that the "standardized" rev- enues for purposes of the pool incorporate expenditure needs to a degree. Other things being equal, lander with large cities or dense populations are deemed to require more revenue and vice versa. (This is in sharp contrast to the Australian approach in which population scarcity is deemed to require greater expenditures, which, in turn, suggests that the approach to needs equalization in federations is rather subjective). The key institutional/constitutional feature of German fiscal federalism is the upper house or Bundesrat, which is a house of the lInder in that it is made up of direct representatives of the kInder governments. All legisla- tion pertaining to the kInder, including the tax rates on shared taxes, must receive the imprimatur of the Bundesrat. PRINCIPLES OF DECENTRALIZATION 119 Canada In contrast to Germany, the Canadian provinces have no formal role in the operations of the central government: members of the Senate (upper cham- ber) are appointed for "life" (up to age 75) by the government of the day, and, as a result, the-Senate is not a federal chamber in a meaningful sense. Provincial concerns and issues tend to be articulated through the provin- cial legislatures and their premiers. Canada also differs from the typical fed- eration in that there is an explicit and extensive listing of provincial pow- ers under the constitution. Beyond this, other features of the Canadian federation propel it toward decentralization. The province of Quebec, with one-quarter of Canada's population, is linguistically, culturally, and insti- tutionally distinct. Quebec has long advocated states' rights, which in turn has moved Canada toward not only greater decentralization, but greater asymmetry as well. (Quebec has its own, separate, personal income tax, whereas the rest of the provinces piggyback their tax rates on Ottawa's per- sonal income tax system.) The Canadian federation is highly decentralized on both the expendi- ture and tax fronts. For example, the provinces levy their own personal and corporate taxes and their own sales taxes (except for Alberta), and, in gen- eral, control the natural resources within their borders. Hence, the Canadian system of intergovernmental transfers accommodates this decentralization. Focusing first on Canada's equalization program, the constitutional principle is less comprehensive than that in Australia: Parliament and the government of Canada are committed to making equalization payments to ensure that provincial governments have sufficient revenues to provide rea- sonably comparable levels of public services at reasonably comparable tax rates. Canada does this by providing equalization payments to the poorer provinces in order to bring their per capita revenues up to the so-called five- province standard (close to the national-average standard). Unlike Australia, however, the revenues of rich provinces are not "leveled down" to this five-province average, nor does Canada equalize for needs on the expenditure side. These equalization payments are wholly unconditional. More interesting, perhaps, is the Canadian approach to vertical imbal- ance. When introduced in the 1950s and 1960s, vertical balance grants were of the shared-cost, conditional variety. Over time, as the programs that they were associated with became established in the various provinces, the federal government relaxed the conditionality. Currently, all vertical trans- fers are rolled into a single block fund, and the moneys can be spent where the provinces wish. However, a national set of social policy principles con- tinues to guide all provincial spending, especially in the area of health. Because of the decentralized nature of the Canadian federation, Canada has had to engage in creative measures to preserve and promote its socioe- conomic union. For example, all provinces adhere to a mechanism that allocates corporate revenues across provinces for those enterprises that 120 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO operate nationally. And recently (February 1999), the provinces and Ottawa have signed a framework to improve Canada's social union. The key coor- dinating institution has been what federal scholars refer to as executive fed- eralism: the frequent meetings (more than 1,000 annually at last count) of federal and provincial officials (or executives) in areas of mutual concern and interest. More recently, the provinces have mounted their own nation- al institution, the annual premiers conference, which is moving the provinces toward addressing some pan-Canadian goals. Nonetheless, the internal Canadian socioeconomic union remains less fully developed than that of Australia, for example. The United States The federal system in the United States probably suffers least from verti- cal fiscal imbalance, in part because U.S. states engage in a narrower range of activities than do Canadian provinces, for example. What is most fasci- nating about the U.S. approach to intergovernmental transfers is the absence of a formal revenue equalization program, although, on the expenditure side (for example, with respect to defense), regional considerations enter into allocation decisions, as they do in other federations. One view of the U.S. approach is that Americans simply ignore any hor- izontal fiscal imbalances. Another view is that there really are no horizon- tal imbalances since any meaningful differences in per capita revenue across states are capitalized in property values, wages, and rents. Wallace Oates, one of the foremost scholars of U.S. federalism, takes this latter view: [E]xisting fiscal differentials (e.g., varying levels of taxable capac- ity) across jurisdictions will tend, to some extent at least, to be cap- italized into property values so that those who choose to live in fiscally disadvantaged areas are compensated by having to pay lower land rents; from this perspective, horizontal equity under a federal system is, to some degree, self-policing. The need for equalizing grants in a federation is thus questionable. Perhaps it is best to regard their role as a matter of "taste" (Oates 1983, pp. 95-96). It may well be that assuming full (or 100 percent) capitalization of dif- ferences in fiscal capacity is going too far. But so does the existing equal- ization literature, which, in general, assumes zero capitalization. However one comes out on this issue, it is clear that the U.S. intergov- ernmental fiscal relations (or, rather, the lack of such) accord well with the laissez-faire U.S. constitutional rhetoric of "life, liberty, and the pursuit of happiness." PRINCIPLES OF DECENTRALIZATION 121 Recapitulation Intergovernmental transfer arrangements are anything but arbitrary. Indeed, they complement the existing tax and expenditure allocation in a manner that integrates overall fiscal federalism in directions consistent with the implicit or explicit values and norms of the respective federal system. They are, in effect, part and parcel of the constitutional-institutional machin- ery that reflects and embodies the deeper societal values of the federation. To be sure, the intergovernmental arrangements in the various federa- tions represent the status quo. And in most of these federations, challenges are emerging on the transfer-intergovernmental front. For example, the Australians are about to introduce a value added tax, the proceeds of which will be allocated to the states. Whether this will be on a derivation basis or equal-per-capita basis, for example, is not clear. The likelihood is that it will be run through the Commonwealth Grants Commission. In Germany, the richer kInder, such as Bavaria and Baden-Wurtemberg, are upset that they retain too small a share of any revenue increase they gen- erate and have taken the equalization program to the German constitutional court. In Canada, the richer provinces are complaining that the federal gov- ernment is mounting too many poor-province preferences in programs other than equalization. And so on. This is to be expected in the best of times. But with the pervasiveness of the forces of globalization and the revolution in knowledge and informa- tion, all federations are examining their fiscal federalism arrangements in general and the nature of intergovernmental transfers in particular. What this means is that the arrangements (expenditure assignment, tax assign- ment, and intergovernmental transfers) will, on a continual basis, have to find their appropriate role in the deeper political logic and interests of the federation as it evolves. But what the comparative experience reveals is that the nature, magnitude, and incentives within the system of intergovern- mental transfers, independent of the tax and expenditure assignment, can play a critical role in ensuring that the federal-state interface can accom- modate the direction and values that the federation itself is pursuing. Implications for Mexico What implications does the role of intergovernmental transfers in devel- oped federal systems hold for Mexico in its ongoing decentralization? Perhaps the most important lesson is that transfer arrangements can be designed to accommodate and integrate a wide range of expenditure and tax assignments in ways that are consistent with the overall equity and effi- ciency goals of the federation. For example, at the margin the system of transfers can embody incentives that encourage further decentralization. Or they can be tailored to produce the desired degree of fiscal equality 122 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO across the Mexican states. Or by creative use of conditionality, they can ensure that all citizens have access to those public goods and services that ought to attend Mexican citizenship. Beyond this, the transfer system can be designed in ways that will accommodate a smooth transition between the current status quo in tax and expenditure assignment and the longer-term evolution of subnational expenditure and tax devolution. To be sure, there is a sense in which inter- governmental transfers are, at any point in time, the necessary residual ele- ment in the interplay of subnational tax and expenditure assignment. However, and more important, transfers can do much more than merely bridge any vertical and horizontal imbalances resulting from the mix of sub- national expenditure and taxation: they can and should ensure that the entire federal-state fiscal interface is in line with the deeper political and social logic of the federation. In this important sense, they are much more than a residual element in fiscal federalism. 2 Historical Forces: Geographical and Political Thomas Courchene, Alberto Diaz-Cayeros, and Steven B. Webb FISCAL FEDERALISM IS EVOLVING in Mexico. Expenditures have been sig- nificantly devolved over the past few years, and the associated financing has been incorporated in myriad conditional grant programs. Although the pattern of decentralization in Mexico is similar to that observed in other Latin American countries, the speed and depth of the process have gone beyond most expectations (Garman, Haggard, and Willis 1999). How did the system of fiscal coordination and revenue-sharing agree- ments between states and federation in Mexico emerge? Why was such a highly centralized fiscal arrangement reached, notwithstanding centripetal forces for state autonomy and a long tradition of resistance in the "provinces" to imposition from the center? Contrary to conventional accounts, fiscal centralization in Mexico was not a necessary outcome, but a contingent result of political processes. Moreover, the specific revenue- sharing agreements between states and the federal government can be understood as consequences of incremental reforms, rather than of an overarching grand design. Although political actors in the first half of the century well understood the advantages of a certain degree of fiscal decen- tralization, they were unwilling to give up extensive fiscal authority. Federal Concentration of Revenue The basic dilemma in the Mexican political economy at the beginning of the twentieth century was the prevalence of fragmented markets and a weak tax authority, both state and federal. Local taxation was chaotic. Given the financial disarray of both local and the federal governments, public expen- diture was mostly financed by debt and monetary emission. The alca- bala-the colonial tax on the movement of goods across jurisdictions which 123 124 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO inhibited the extension of markets in the nineteenth century-haunted the regional economies of Mexico. National and regional politicians in Mexico attempted to improve the regional flow of market transactions within the federal system, seeking to rationalize and unify state tax systems into a coherent national tax system and a system of tax settlement between lev- els of government. They met in National Tax Conventions in 1925,1933, and 1947 for this purpose. Alberto J. Pani, the charismatic federal finance minister who convoked governors to the First National Tax Convention, expressed the situation in 1925 this way: Each state establishes its own revenue system, the Federal Government maintains its own, and since the objects taxed hap- pen to be the same, since there is no concrete and defined plan for the limitation between the federal capacity and the local capacities to create taxes, since states often rival each other lead- ing into true economic wars and creating, in the name of a sales tax; e.g., true local import duties in order to sustain internal pro- duction taxes that are incorrectly established; the tax becomes increasingly burdensome due to the multiplicity of rates, fines and penalties, increasing the complexity of the system and increasing in a disproportionate and unjustifiable manner the deadweight expenses for revenue collection, surveillance and administration.1 While the first two National Tax Conventions made little progress, the Third in 1947 envisioned a centralized fiscal system that (a) local governments would rely on as exclusive sources of revenue on the property tax and some other minor taxes, eliminating their taxes on trade and industry; (b) states would receive revenue shares from federal excises on natural resources, alcoholic beverages, matches, and so forth, and they would be guaranteed 25 percent of any additional revenue collected through those federal taxes; (c) a national sales tax would be introduced where the rate would be shared between states and the federal government, but it would be administered as a centralized federal tax; (d) the income tax would become exclusively federal, although states would retain some tax authority over taxes with very low yields on agriculture and livestock; (e) the contribuci6n federal would be finally reduced to 5 percent in all states, in order to gradually phase it out during the next years. These proposals were fully accepted by the assembly. During the second half of the twentieth century the federal government consolidated its fiscal centralization by becoming the only level of gov- ernment allowed to levy taxes on foreign trade, natural resources (includ- HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 125 ing all oil and mining rights), banks, insurance institutions, electricity, tobacco, gas, matches, alcohol, forestry, and beer (art. 73-XXIX). Although states theoretically still retained the capacity to levy an income tax, payroll taxes, sales taxes, and other taxes not explicitly stated in article 73, in prac- tice the federal government came to monopolize almost all sources of rev- enue. The federal treasury provided tax revenue shares (participaciones) to the states, which increasingly came to constitute the most important source of local government finance. The system of tax coordination which characterizes the Mexican feder- al system today was the consequence of a regional compromise struck between the federal and state governments. The compromise required that local politicians delegate financial power to the president, in exchange for sources of patronage through the federal bureaucracies, attractive careers in the federal government, and an active involvement of the federal gov- ernment in state development. The arrangement took almost two decades to become stable. The solution was achieved through institutional rules and a very peculiar political organization. The configuration of veto players made it self-enforcing, in the sense that local and national players were bet- ter off with this arrangement, and therefore willing to abide by it. The loser in this arrangement was federalism. [The weakening of the local machines] was coupled with the ever- increasing financial dependence of the formal state governments upon the central authorities, because just as the growing com- plexities of social and economic life called for greater expendi- tures by governmental agencies, the national government was busily preempting most of the major sources of tax revenue for itself. This forced the local officers to go to Mexico City, hat in hand, seeking grants from the national government to satisfy the demands of their constituents (Scott 1959; p. 135). Electoral success compensated the governors for the loss of local finan- cial independence and fiscal initiative, although many Mexicans disap- proved of this tradeoff, since it led to an overwhelming federal government (Casanova 1965). Two parallel developments converged to create the tax system that characterizes intergovernmental relations in Mexico today. The first was the establishment of a complex system of revenue sharing between states and the federal government at the beginning of the 1940s, which gave states unconditional transfers (participaciones) out of the collection of rev- enue from exclusive federal excises. The revenue-sharing system was com- plicated because it established state shares case by case, on each specific tax. The incorporation of all states into the same national sales tax in 1974 was 126 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO achieved through what federal financial authorities called a "prudent strat- egy which quietly but stubbornly achieved its goal" (Secretaria de Hacienda y Credito Piblico 1973, p. 6). In contrast with the Tax Conventions, which had been open debates, the changes of the 1970s were decided through closed bilateral negotiations. The second development was the final implementation of a federal sales tax in 1947 (the Impuesto Sobre Ingresos Mercantiles-ISIM), which states could join if they agreed to coordinate their rate with the federal rate. By the early 1970s all states had joined the system, obtaining most of their local tax collection from this tax, in addition to receiving unconditional tax transfers through the revenue-sharing systems. Both developments-the revenue-sharing system and the unified fed- eral sales tax-came together in 1979 when fiscal relations between states and the federal government were merged into a single system, with the replacement of the ISIM with the Value Added Tax (Impuesto al Valor Agregado-IVA) at the federal level: all states would receive unique revenue shares out of federal tax collection, according to previously agreed formulas. This unification meant that local budgets would be financed mostly through unconditional federal transfers (participaciones) contained in the revenue sharing agreements. After 1980, state governments depended almost com- pletely on federal revenue transfers. While every state had been encouraged to join in the federal sales tax, the introduction of the Value Added Tax in 1979 produced a major redis- tribution of resources among states. On the one hand, the IVA would be collected where value was added, not where sales occurred, since the lat- ter would require that when the tax was paid by the final consumer in one state, there had to be a credit for the tax paid in other states, with a conse- quent redistribution of where revenue is accrued. On the other hand, the success of the IVA required the elimination of some remaining state-level excises, in order to bring about more horizontal equity among regionally dispersed producers. These issues were addressed through the negotiation of the Sistema Nacional de Coordinaci6n Fiscal (SNCF or Pacto Fiscal) among states and the federal government, including the Federal District. The arrangement tied participaciones to explicit formulas that considered pop- ulation, education expenditure, revenue collected in the past, and indica- tors of state performance in tax collection. The revenue-sharing arrangement of the SNCF was a contract between states and the federal government. There was no constitutional provision that forced states to give up their authority over taxation: states belonged to the system by agreeing to withdraw their own taxes and receive partic- ipaciones in exchange. Thus, governors retained the legal power to withdraw from the system. When the system was created, governors also signed HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 127 administrative collaboration agreements, which involved working close- ly with federal authorities on issues of federal tax compliance. The creation of the SNCF was accepted by the states with almost no resis- tance. This was attributable to the political conditions, the timing of the reform, and the state of federal finances. The dominance of the party in office (the PRI) during the late 1970s was overwhelming: President L6pez Portillo ran unopposed in the 1976 election. The reform was also skillful- ly timed around the temporal horizons of governors: during 1980 when the reform would come into effect, most of the incumbent governors would be just finishing their terms, according to the staggered timing of local elec- tions. Finally, the country was in the midst of an oil boom, bringing the fed- eral government a substantial amount of revenue from the windfall gains of the nationalized oil company, PEMEX. Since taxation of natural resources, and oil in particular, was exclusively federal, states were not directly profiting from the expansion in available resources. They did receive more resources and projects through Inversi6n Pu'blica Federal, but those transfers were ultimately controlled by the president and his bureau- cracies. The SNCF offered the opportunity for states to reap part of the ben- efits of the oil boom as unconditional tax transfers, although the arrange- ment made state governments more dependent on the federal government. The formulas have been changed frequently ever since. At first the main ingredient in the formulas was to assure states the same revenue they were previously collecting from local taxes, to ensure participation of all states. Later on, the formulas also reflected some measures of the effort at tax collection, in local taxes, federal IVA collection in each state, or formerly federal taxes transferred to the state administrations collecting them (name- ly, the tax on new cars, Impuesto Sobre Automobiles Nuevos, ISAN). This sug- gests that the system then moved into greater concern for incentive com- patibility and performance. Since the beginning, the formulas also included a complementary fund, which attempted to compensate states that were receiving the least resources. This introduced an equalizing tendency in the shares, which has been further reinforced in recent years by giving a greater weight to population factors. (For discussions of the formulas and their changes, see Chapoy Bonifaz 1992; Arellano 1996; Aguilar 1996; Diaz Cayeros 1995.) Thus, the overall pattern through time has been, abstracting from the subtleties of each individual formula, that at the beginning states received revenue shares much in line with the revenue they were collecting before- hand from their own taxes, their rate in the federal ISIM, and their partici- paciones. That meant, in fact, that poorer states had smaller per capita par- ticipaciones than richer states. It also meant that the oil-producing states received a disproportionate share of resources, because they had previously 128 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO been receiving high participaciones on federal oil taxes. However, as formulas have been changed, there has been a slight tendency toward per capita con- vergence, since poorer states have witnessed larger increases in participa- ciones than richer states, consonant with the larger weight given to popu- lation in the calculation of revenue shares (Diaz Cayeros 1995, pp. 94-95). Rise of Democracy and Opposition Parties The end of Mexico's hegemonic party system has accompanied democra- tization from below. This process, although limited at first, has been extend- ed throughout the country. Local political forces have come to question the concentration of resources and authority at the federal level. Political plu- ralism is driving the decentralization witnessed during the past few years. The dominance or control of the political system by a single political party intrinsically contradicts fiscal decentralization. From the viewpoint of governance, state and municipal governments are democratically elect- ed in Mexico. To a large extent, the tradition of central control within a sin- gle dominant political party led at first to deconcentration-regional admin- istrative units reporting to the center-rather than to decentralization proper, where territorial governments are chosen by and are accountable to the local populace. Both decentralization and deconcentration have accelerated since the Zedillo administration came to power in December 1994. The impetus for decentralization is the result of the need to address the increasing levels of public debt held by some states and the need to improve efficiency and rationalize service delivery in key sectors, such as social welfare, education, health, and transportation. The Mexican government rightly sees decen- tralization of these services as the key to more efficient public expenditures. This has come in contradiction, however, to Mexico's long tradition of centralization and the federal government's de facto control of the tax sys- tem. This has added to the confusion about what is an appropriate strate- gy for decentralization and has had a direct impact on decentralization pol- icy. The lack of a clear vision for sectoral policies, such as the final direction of education and health reform, has added to the state of flux of fiscal decen- tralization in Mexico. Undoubtedly, the two most important events in Mexico's recent politi- cal history are (a) the winning of governorships by opposition parties since 1989, and (b) the loss of the PRI's majority in the federal Chamber of Deputies in 1997. Both of these events have affected the political process- es and debates over Mexican federalism. Baja California was the first state governorship to be won by an opposition party (Partido Acci6n Nacional, PAN). Other PAN victories followed in Chihuahua, Guanajuato, and Jalisco. As of the beginning of 1999, the PAN controlled five governorships HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 129 (Aguascalientes, Baja California, Guanajuato, Jalisco, and Quer6taro), while the other major opposition party, the Partido de la Revoluci6n Democratica (PRD) controlled two more (Baja California Sur and Zacatecas) plus the Federal District. At the municipal level, by the end of 1997, opposition par- ties controlled 28 percent of municipalities, which account for 45 percent of the population. The partisan plurality of local governments has increased the demand for devolving power over resources, to the point where even PRI governors and mayors are advocating decentralization, federalism, and greater local autonomy. The absence of a single-party majority in the federal Congress' Chamber of Deputies since 1997 is particularly important for federalism. All the major reforms tending toward greater decentralization pursued since then, such as the creation of Ramo 33 and the federal transfers (aportaciones) in the Fiscal Coordination Law, have required the support of at least one oppo- sition party, namely the PAN. Any changes that deepen federalism will require partisan coalitions. The lack of a majority party in the lower cham- ber could continue in the near future no matter which party wins the pres- idency in 2000. Moreover, since the Senate does not have authority over the federal budget, the most important debates concerning further devolution of expenditure authority and the size, composition, and allocation of trans- fers will take place in the Chamber of Deputies. By now, all the major parties are committed to federalism and to deep- ening the decentralization process. This was not true some years back, when the PAN was alone in advocating for greater local autonomy, consonant with its regionalist strategy of electoral competition and its stress on the sub- sidiarity principle, that each government responsibility should be assigned to the lowest level practicable. At that time, both the PRI and the PRD were highly centralist in their approach to national problems. But this has changed very quickly, as the presidency has relinquished some of its dom- inance, governors have become heavier political players, and the PRD has won governorships and seen its electoral support spread beyond Mexico City. By 1999 the most important contenders for the presidency had all been governors: Cuauhtemoc Cardenas (Michoacan and the Federal District), Vicente Fox (Guanajuato), Roberto Madrazo (Tabasco), Manuel Bartlett (Puebla), and Francisco Labastida (Sinaloa). Transfer-Led Decentralization In the last ten years Mexico has rapidly decentralized public expenditures, quickly reaching almost the extent of the other large federations of Latin America-Argentina and Brazil-which have a long tradition of decen- tralization. Figure 2.1 shows this decentralization as a movement out along the x-axis, but the closeness to the x-axis also shows the small extent to 130 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 2.1. Subnational Tax and Spending Shares in Argentina, Brazil, and Mexico (Shares of GDP) SN Taxes, share of total public spending 18 16Br9 16 0 ~~~~~~/Locus Brazil '85 14 o 12 10 Argentna'96 8 -/Argentina'85 6 4 / Mexico'85 Mexico '96 2 0 5 10 15 20 25 30 35 40 SN Spending, share of total public spending Source: IMF, Government Financial Statistics; World Bank estimates. which the subnational expenditures are financed through their own rev- enues. The vertical distance from the diagonal line, the locus of fiscal bal- ance along which taxes would equal spending, shows the dependence on transfers (and deficits). The great increase of subnational spending in Mexico over the past 12 years has all been accompanied by increases of transfers, with earmarking for broad sectoral purposes, as Chapters 3 and 5 describe in more detail. Tax authority of states and municipalities has not increased along with expanding responsibilities, and the rules for transfers have not provided incentives to intensify collection efforts. So states in Mexico depend much more heavily on transfers than in Argentina and Brazil. History and politics explain this pattern. The first wave of decentral- ization in the late 1980s and early 1990s came right after the culmination of the centralization of taxes described above. Decentralization was at first really deconcentration, with the center delegating spending and responsi- bility, but keeping control over the priorities for spending. A prime chan- nel for this control was the earmarking and conditions on transfers, con- trol that would be lost if states were raising most of their own money. So HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 131 the center did not delegate tax authority, and the states had little interest in trying to tax more, because it was easier to go to the federal executive for transfers. In the second phase of decentralization, after 1997, the opposition par- ties were flexing their political muscle to get resources for their con- stituencies. The federal government offered some increased tax authority to the states, but in isolation it did not look politically attractive to the states, especially without any rollback of federal taxation. So the second phase of decentralization also came with transfers, but with much looser earmark- ing and linkage to expanded responsibilities. Inequality and Globalization Two initial conditions set the boundaries for the future evolution of fiscal arrangements in Mexico. The first has long been a feature of Mexican fed- eralism, namely the degree of inequality across states. Understanding this inequality is critical to designing a more decentralized federation. The sec- ond is the impact of the North American Free Trade Agreement (NAFTA). As with selected provinces in Canada, some Mexican states are becoming fully integrated economically within the new environment created by NAFTA, and they are demanding more fiscal and economic autonomy than some other provinces can handle (see Courchene 1998 on Canada). Unless an appropriate system for transferring resources and responsibilities is devised, the uneven pattern of regional integration may exacerbate polit- ical and regional tensions. Inequality and Federalism In the nineteenth century, Alexander von Humboldt, the great political economist, labeled Mexico "the country of inequality." Income inequality in Mexico continues to be very high, even by Latin American standards. According to the 1996 income distribution surveys, the Gini coefficient for household inequality was 0.478, where the bottom 10 percent of the pop- ulation had 1.8 percent of the income share, while the top 10 percent gar- nered 42 percent (INEGI 1996). Personal income inequality is also found at the state level, although to a lesser degree. According to the only available survey on income distribution at the state level, carried out by INEGI (the government's statistical office) in 1996, the Gini coefficients of Campeche, Coahuila, the Federal District, Guanajuato, Hidalgo, Jalisco, Mexico, Oaxaca, and Tabasco, where the survey was conducted, ranged from a low of 0.41 for Guanajuato to a high of 0.47 for Campeche (INEGI 1999; an INEGI survey in 1989 indicated a coefficient for Tlaxcala of 0.38). In 1996, in the wake of the economic crisis, 62 percent of the population was below the poverty line, with 30 percent in extreme poverty. Although 132 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO growth since then has probably brought the rates down again, resuming the trend for the decade prior to 1994, the problem of poverty remains acute. Much of this poverty is concentrated in the southern states (Trejo and Jones 1998, p. 72). Inequality among households is also reflected in region- al inequality. Figure 2.2 highlights the differences in per capita gross domestic prod- uct (GDP) by state, according to the official statistics produced by INEGI for 1993. Some states have income levels similar to those in African coun- tries, while others have income levels roughly at par with those of lower- income European countries. There is clearly a north-south difference in per capita GDP, with northern states generally much better off than southern states. Moreover, the regional distribution of GDP does not fully reflect lev- els of individual welfare since some southern states have high GDPs (for example, the oil-producing states, Campeche and Tabasco, on the one hand, and tourist havens, Canctn and Quintana Roo, on the other) that are not reflected in correspondingly high levels of welfare of the general population. Figure 2.3 presents another overview of inequality-the index of illiter- acy across Mexican states. Here, the north-south divide is even more strik- Figure 2.2. Differences in Per Capita Gross Domestic Product (By State) Producto Estatal Bruto por Habitante 1993 (dolares) * 4,480 to 10,200 (7) * 3,390 to 4,480 (8) S_urce Author alculationsfr.2,700 to 3,390 (9) i = g ~~~~~~~~~~~~~~~1,700 to 2,700 (8) Source: Authors' calculations from INEGI data. HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 133 Figure 2.3. Index of Illiteracy Porcentaje de Poblacion Analfabeta 1995 * 16.3 to 26.1 (5) *11.9 to 16.3 (6) 8.B to 11.9 (6) E 5.6 to 8.8 (8) D2.9to5.6(7) Source: Authors' calculations from INEGI data. ing, with the differences reflecting, in part, the concentration of indigenous populations in southern states and the fact that the northern states have taken more advantage of the development opportunities from open inter- national markets. In an era when human capital holds the greatest promise of economic growth and betterment, this evidence of unequal outcomes challenges the claims of some advocates of decentralization that it will improve the respon- siveness of the Mexican public sector to the needs of the population as a whole. In particular, these data speak directly to the citizenship rationale for intergovernmental transfers. Any evolution of Mexican decentralization will have to ensure that future expenditure and transfer arrangements provide an acceptable equality of opportunity across states in key public services such as education and health. Will state per capita GDPs converge, as suggested by the growing liter- ature on economic growth? According to the convergence hypothesis, since poorer states have less capital, any capital they get has a higher mar- 134 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO ginal return there and they tend to grow faster than richer states, and this will eventually lead to the convergence of income levels at the steady states. Within the neoclassical growth theory, once the steady state has been reached, differences in growth rates are attributable to variations in saving rates. There is also a technology argument for convergence, in that poor states have poorer technology and thus greater opportunity to catch-up by adopting state-of-the-art methods. Against this, some argue that the poor have less access to capital and are less able to access new technology. In its mild form, the convergence hypothesis does not require a reduction in the absolute gap between rich and poor, but it does require that poor states grow faster than rich states. Table 2.1 and figure 2.4 provide indicators of convergence among Mexican states. (For definitions and the-basic findings of this literature, see Barro and Sala-i-Martin 1991,1995.) The coefficient of variation and the typ- ical log deviation of per capita GDP suggest that regional inequality in Mexico peaked in absolute terms during the mid-1980s and decreased slightly but remained high in the early-1990s. The 1980-85 divergence (in terms of the coefficient of variation in table 2.1) presumably had its origins in the Mexican oil boom. Convergence is evident over the 1985-88 period, but this catch-up still results in a coefficient of variation higher than that for the earlier years. Arguably, the convergence process is running up against the highly unequal levels of education across states, as reflected in figure 2.1 (For a discussion of these issues, see Diaz-Cayeros 1999, Navarrete 1996, and Alzati 1998.) Table 2.1. U7Convergence in Mexico Standard Coefficient of Typical log Year deviation (a) variation (a,lp) deviation 1970 17.21 .4169 .1762 1975 18.42 .3781 .1618 1980 27.24 .4628 .1731 (22.52)0 (.4181)° 1985 63.94 .8889 .2058 (23.39)0 (.3953)0 1988 34.07 .5413 .1804 (23.12)0 (.4108)° 1993 32.62 .5218 .1901 0 estimates in parenthesis refer to measures excluding the outlier observations (Tabasco in 1980 and Campeche in 1985 and 1988). Source: Authors' calculation with Banco de Mexico and INEGI data. HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 135 Figure 2.4. Regional Convergence, 1960-88 Average annual growth (percent) 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 0 10 20 30 40 50 60 70 80 90 1960 per capita GDP (1980 = 100) Source: Authors' calculations from INEGI data. The convergence data can be placed in a longer-term framework. In fig- ure 2.4 (which focuses on the 1960-88 period), the trend line is in the direc- tion of convergence. However, from 1988 to 1996, this convergence effec- tively disappears (figure 2.5). Arguably, this reflects the move to trade liberalization, introduced on a gradual basis in 1985 and given a substan- tial boost in 1987. It will be important to extend figures 2.1 through 2.4 to incorporate more of the post-NAFTA (post-1995) data. One would expect the NAFTA environment to favor those states that are better-situated geo- graphically and have better-developed infrastructure and human capital. All of these factors favor the northern states. There is a high correlation between per capita GDP and a more accurate indicator of human welfare in each state, given by the Foster-Greer- Thorbecke poverty index (see Mogoll6n 1999). Figure 2.6 reports the Foster- Greer-Thorbecke index for all states calculated with official INEGI wage data for 1995, using a poverty line set at equal to twice the minimum wage and taking into account the depth of poverty. The correlation between the Foster-Greer-Thorbecke and per capita GDP is -0.4965. Except for the out- liers of Campeche and Quintana Roo, which have high per capita GDP, but 136 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 2.5. Regional Divergence, 1988-96 Average annual growth (percent) 4.0 3.0 2.0 * 1.0 *+ -* + 0.0. .0 20.0 40.0 20.0 80.0 100.0 120.0 140.0 160.0 18( .0 -1.0 -2.0 -3.0 1996 per capita GDP (base 1980=100) Source: Authors' calculations from INEGI data. low levels of welfare as measured by poverty, there is a very close link between the two measures. Thus regional inequality in Mexico is high and will likely remain high. Even if convergence were to take place at a rate of 2 percent, as has hap- pened historically in other countries over the long run (Barro and Sala-i- Matin 1991), it would take more than 60 years for all states to arrive at the same level of welfare (estimates in Mexico find a slightly higher rate of con- vergence, but over a shorter period; see Navarrete 1996). Such conver- gence is doubtful without the achievement of similar levels of education- al attainment; in the estimates for Mexico, education has sped up the convergence process. The rift between north and south, which is so obvi- ous in the illiteracy indicators, is probably the greatest challenge facing Mexican federalism in the years to come. North American Integration and Mexican Federalism The opportunities presented by NAFTA might accrue, initially at least, pri- marily to the northern Mexican states, but the implications of NAFTA for Mexican federalism will likely go well beyond this economic dimension. HISTORICAL FORCES: GEOGRAPHIC AND POLITICAL 137 Courchene and Telmer's (1998) study of "region states," focusing on Ontario, Canada, reached some instructive conclusions: * Ontario's economic future progressively lies in the North American market rather than in the Canadian market. Already, Ontario's exports to the United States are more than twice Ontario's exports to the rest of Canada. Indeed, roughly 45 percent of Ontario's GDP is now exported to the United States. * Ontario's policies will be increasingly oriented toward making the province a more attractive location for penetrating the NAFTA mar- ket. * When considering its competitive position, Ontario cares much less about tax rates in its sister provinces, (although they are typically high- er), than it does about tax rates in competitive areas with which it com- petes directly, such as Michigan, Ohio, and New York. * Moreover, Ontario wants to secure enough policy flexibility to ensure its economic future within the new NAFTA environment. States like Nuevo Le6n, the industrial powerhouse in northern Mexico, will surely fall into this category of North American region states. The only issue is the degree to which they will pursue this role. In fact, according to some preliminary estimates (Diaz Cayeros and Martinez 1999), the degree Figure 2.6. Regional GDP Versus Poverty Index FGT(2,2) Per Capita GDP (pesos) 0.0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 0.05 0.10 VS. SIN CHIH * COL *NL AGA *COA 0.15 A F * -JAL MOR N*NAY * QRO *QR 0.20 4,T TLAX 0.25 Jcv*VER * DGO CAM HG03W YUC GCSLP 0.30 4CRO *ZAC 0.35 0.40 0CHS OAX 0.45 0.50 Source: Authors' calculations from INEGI data. 138 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO of regional integration of some border states is extremely high: foreign trade (imports plus exports) in Baja California, Chihuahua, and Tamaulipas already represents more than 100 percent of state GDP (171, 137, and 134 percent, respectively). Our view is that these border states, among others, will eventually press for more powers-taxation, expenditure flexibility, and the like. But border and rich states are not the only ones that will press for greater regional integration. Poor states like Oaxaca and Zacatecas already have highly developed networks of migration (both temporary and permanent), while worker remittances constitute a crucial component of the local economy. These pressures will complicate, if not exacerbate, the north-south divide and perhaps cut across unexpected lines. Unless the Fiscal Pact is substantially revised, some of these northern states may con- sider with increasing seriousness the option of withdrawing from the Pact and reintroducing state taxes. The relevant point is that the transfer system (along with the rest of fis- cal federalism) must be able to accommodate the preferences of these northern states while ensuring the social and political cohesion of all states within the federation. This is a tall order, but one that the transfer system can, in principle at least, deliver. Aspects of these inevitable north-south ten- sions will be an integral part of the simulations in Chapter 5, showing alter- native scenarios of decentralization and future transfers. 3 Assignment of Spending Responsibilities and Service Delivery Enrique Cabrero Mendoza and Jorge Martinez-Vazquez THE MEXICAN CONSTITUTION REFERS to the division of some responsibili- ties-for example, it proclaims that education is the concurrent responsi- bility of the federal and state governments-but leaves explicit assign- ments to sectoral laws, such as those governing education and health. Many of the assignments in Mexico follow the generally accepted rules and principles of expenditure assignment (see table 3.1). The federal govern- ment provides public services with a benefit sphere that reaches the entire nation, such as national defense. State governments provide services with an intermediate benefit sphere, such as state roads and secondary hospi- tals, and municipal governments provide services with local benefit areas, such as sanitation and street lighting. Two features dominate the present assignment of responsibilities. First, all three levels, especially the federal and state governments, have con- current obligations for important services such as education, health, and social assistance. Second, few responsibilities are assigned at the munici- pal level, especially in education and health, where significant benefits could be obtained through decentralization to the lowest decisionmaking unit. The assignments often distinguish three elements of service delivery- regulation or normative design, financing, and implementation or actual delivery. The federal government tends to play a more significant role in normative design and financing, while states and, to a lesser extent, local governments, tend to take the lead for implementation and delivery but not financing. In Mexico, expenditure responsibilities have been decentralized on a sec- tor-by-sector basis and by unilateral decisions at the federal level. There is no all-encompassing scheme in which state governments engage in concerted agreements with the federal government to take over more responsibilities. 139 Table 3.1. Mexico's Expenditure Assignrment Expenditurefunction Federal government State governments Municipal Governments Defense * 100 percent Foreign Affairs and * 100 percent Economic Relations Labor Policies * 100 percent Monetary and * 100 percent Financial Policy Post and * Government and private providers Telecommunications Education * Setting policies and norms (SEP); financing through * Financing, implementation, * Minimal role, school transfers (Ramo 33) maintenance, and equipment maintenance, and some 0 * High schools and colleges (concurrent) (concurrent) school construction * Federal technological institutes of higher education * High schools and state universities (concurrent) * Evaluation and audit of subnational performance * Administration of programs and * Labor relations and wage-setting self-evaluation * School construction supervision * Half of the technical schools * All education in the Federal District * School construction (concurrent) * Approximately half of the technical schools * Adult education programs * Most textbook production * Most teacher training Health * Setting policies and norms (Social Security * Primary care for the rural popula- Administration); financing through transfers (Ramo 33) tion and urban poor * Evaluation and audit of subnational performance * Partly responsible for financing * Secondary and tertiary hospitals * Administration of programs and * Labor relations and wage determination self-evaluation * Most capital infrastructure decisions * Epidemiology and preventive care * Reproductive health (table continued on next page) Table 3.1 continued Expenditurefunction Federal government State governments Municipal governments Roads * Federal highway construction and maintenance * State feeder roads (construction * Local streets * Financing of rural road development and maintenance) * Implemientation of rural road development * Maintenance of secondary federal roads (with federal funds) Police and Internal * Federal transfer to the states to strengthen state police * Spedal police (concurrent) * Local public order and safety Security * Federal and border police * State public order and safety * Spedal police (concurrent) * Coordination of state and municipal public safety v. Social Assistance and * Funding through Ramo 33 and Ramo 26 * Implementation of school-lunch * Implementing of social Social Security programs infrastructure programs * Food assistance to the poor * Other programs in coordination with SEDESOL Culture and Libraries * Public libraries Parks and Recreation * Biosphere reserves * National parks (concurrent) * Local parks * National monuments * National parks (concurrent) Public Transportation * Most railways and airport operations have been * Some airports * Local transportation and transit privatized * Seaport operations being privatized (table continued on next page) Table 3.1 continued Expenditurefunction Federal government State governments Municipal governments Environmental * National standards * States can adopt their own * Land use permits Protection * Approval by INE, National Water Commission, the standards Ministry of Health, and the Ministry of Industry Water, Sewerage, and * Water supply and sewage * Garbage collection Sanitation (concurrent) * Water supply and sewage; many water systems have been privatized (but municipalities retain debt liability) (concurrent) Housing * National programs for housing development Some states have housing agencies Price Subsidies * Market intervention programs (mostly phased out) Agriculture and * Funding for state programs in irrigation, water supply, * Rural development Irrigation and hydroelectric exploration * Extension services * National irrigation programs and funding research * Drilling * Rural development, rural roads, forestry * Some research * Funding for research Other Infrastructure * Financing through Ramo 33 "social infrastructure" for * "State infrastructure" * Cemeteries the poor * Slaughterhouses * Public markets Tourism * National programs (concurrent) * State programs (concurrent) Industrial Policy * Concurrent * Concurrent Source: Information provided by the Mexican authorities, the Constitution, General Education Law, General Health Law, and Amieva-Huerta (1997). SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 143 The sector-by-sector approach enables subnational governments to par- ticipate in planning and negotiating each step of decentralization, as has occurred in the decentralization of health services. Adequate planning has not always occurred, however, as the decentralization of education illustrates. Current Assignments To get an overview of the current assignment of expenditure responsibil- ities, this section examines the distribution of expenditures at different lev- els of government, using the data available. While refinements in the data might refine the picture, the information here illustrates well the trends, strengths, and weakness of the Mexican system, providing a departure point for how to improve the policies, and of the publicly available data. Expenditure Shares and Composition of Expenditures Table 3.2 presents the share of each level of government in total expendi- tures during 1989-96; more recent data are not available. These data show that Mexico has undergone decentralized spending over the past ten years.1 Although the federal government still claims the lion's share of consolidated expenditures, its share decreased from around 84 percent in 1989 to around 75 percent in 1996. (The transfers, aportaciones, effectively earmarked for teachers' salaries, are still attributed here to the federal level, where the sub- stantive decisions were made, even though the payments were made at the state level.) The trend was not monotonic. The share of federal expenditures Table 3.2. Expenditure by Level of Government in Mexico, 1989-96 (Percent) Total Programmables State and federal Year Federal district Municipalities Federal States Municipalities 1989 83.9 13.7 2.5 86.2 9.8 4.0 1990 83.2 13.7 3.0 85.5 10.0 4.5 1991 80.4 15.9 3.7 84.3 10.9 4.8 1992 76.8 19.4 3.8 82.9 12.3 4.8 1993 74.1 21.7 4.1 82.5 12.4 5.2 1994 72.5 23.5 4.0 81.7 13.5 4.8 1995 73.9 22.5 3.5 81.1 14.4 4.5 1996 74.8 21.8 3.4 81.2 14.5 4.3 Source: Authors' calculation based on INEGI 1998. 144 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO dropped to less than 73 percent in 1994. At the subnational level, the expen- diture shares of both state and municipal governments increased over time, but in an unsteady fashion. State governments in particular increased their presence in total expenditures. The states' share was close to 22 per- cent in 1996, up from 14 percent in 1989. By comparison, municipalities' share was 3.4 percent in 1996, up from 2.5 percent in 1989. Municipalities represented 4 percent of total expenditures in 1993. The distribution of programmable expenditures tells a similar story (table 3.2). Programmable expenditures exclude expenditures for debt ser- vice and financial restructuring as well as revenue sharing and transfers. For 1996 the shares of the federal and municipal governments were high- er and the share of state governments was lower for programmable than for total expenditures. However, the share of municipal governments still was only 4.3 percent, while the share of state governments was 14.5 percent. There is little information on the type of expenditure, in either functional or economic classifications, by level of government. Table 3.3 presents the share of each level of government in total education expenditures as a per- centage of gross domestic product (GDP) during 1990-97. The table also shows expenditures in the private sector as a percentage of GDP. Given the central importance of education, decentralization of government expen- ditures still has a long way to go in Mexico. In 1997, states' share in public education expenditures was only 8.5 percent, or 0.4 percent of GDP. Municipalities' share was around 0.2 percent, or 0.01 percent of GDP. Table 3.3. National Education Expenditure as a Percentage of GDP, by Source of Funds, 1990-97 Year 1990 1991 1992 1993 1994 1995 1996 1997t Federal (SEP) 2.5 2.9 3.2 3.7 4.0 3.8 3.7 3.8 Federal (other) 0.5 0.5 0.6 0.7 0.6 0.5 0.5 0.5 Federal (subtotal of SEP and other) 3.0 3.4 3.8 4.3 4.6 4.2 4.2 4.3 State 0.7 0.6 0.6 0.6 0.6 0.5 0.4 0.4 Municipalitv 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 Public (federal, state, and municipal) 3.71 4.01 4.41 4.91 5.21 4.71 4.61 4.71 Private 0.3 0.2 0.3 0.3 0.3 0.2 0.2 0.2 Total 4.03 4.21 4.7 5.25 5.49 4.93 4.8 4.91 Federal education/ federal budget 19.1 21.7 23.9 26.4 26.1 26.7 26.4 24.6 tEstimates based on the document "General Criteria for Economic Policy." Source: Statistical annex of the Presidential Report 1997, reproduced from World Bank (1998a). SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 145 During 1981-96, the share of resources at the state level going to current expenditures increased, and the share going to capital expenditures decreased (table 3.4). The same trend was true for municipalities. For both states and municipalities, with many ups and downs, the share of capital expenditures in programmable expenditures moved from one-third or more to around one-fourth or less. The increasing relative importance of current expenditures and decreasing relative importance of capital expen- ditures seems to have continued in 1997 and 1998. State Expenditures and Financing During 1989-96 states received more resources overall but became more dependent on federal sources of revenue (data are from Instituto Nacional de Estadistica Geograffa e Informdtica, INEGI). Federal revenue sharing (par- ticipaciones) was without doubt the most important component of state income, reaching approximately 47 percent on average for the period. States' own revenues represented an average of 14 percent for the period. Table 3.4. Distribution of Programmable Expenditure by Level of Subnational Government in Mexico, 1980-96 (Percent) State Municipal Administrative Public works and Administrative Public works and expenses development expenses development Year (current) (capital) (current) (capital) 1980 30.9 69.1 71.6 28.4 1981 76.2 23.8 66.6 33.4 1982 70.9 29.1 68.6 31.4 1983 65.5 34.5 65.5 34.5 1984 57.4 42.6 60.2 39.8 1985 55.8 44.2 62.4 37.6 1986 63.0 37.0 67.8 32.2 1987 60.9 39.1 69.7 30.3 1988 59.2 40.8 68.9 31.1 1989 61.9 38.1 69.3 30.7 1990 62.6 37.4 67.3 32.7 1991 61.8 38.2 68.7 31.3 1992 65.6 34.4 69.5 30.5 1993 70.4 29.6 69.9 30.1 1994 69.5 30.5 70.4 29.6 1995 72.3 27.7 73.9 26.1 1996 81.3 18.7 74.7 25.3 Note: Excludes the Federal District. Source: Author's calculations based on INEGI (various years). 146 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO While states' own revenues increased in real terms through 1994, they dropped drastically in 1995 during Mexico's fiscal crisis. The significant increase in transfers (Ramo 26) in 1995 and 1996, and Ramo 33 during 1998-99, may have reduced tax effort by the states (figure 3.1). Moreover, the growth of current expenditure crowded out capital invest- ment. The increases in overall state income led directly to the enlargement of current expenditure. As shown in figure 3.2, the divergence of current and capital investment expenditures started in 1991 and accelerated in 1993. In 1993 and 1996, current expenditures clearly displaced capital investment expenditures. Municipal Expenditures and Financing During 1989-97 municipal income grew in real terms. However, the 1994 crisis caused a real decline in both own revenue and revenue sharing (fig- ure 3.1). Nevertheless, in 1996 the trend was reestablished, both in own rev- enue and in revenue sharing. Since 1995 there has been a significant change, Figure 3.1. State Revenue from Own Sources, Earmarked Transfers, and Participations (Constant Prices, 1993 = 100) Miles de 1994 pesos (millions of 1994 pesos) 45 40* .. 35 .----- 25 20 15 10 5 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Own source revenues --- * Participaciones Federales A Transfers Sources: Based on INEGI's database and CPI from the Bank of Mexico. SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 147 Figure 3.2. Composition of State Expenditures, 1989-96 (Constant Prices, 1993 = 100) Miles de 1994 pesos (millions of 1994 pesos) 90 80 70 60 50 40 30 20 - A'- A- . A. - -A------ O I I I 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 -U*- Net current expenditure --- A--' Capital expenditures Sources: Based on data from INEGI and INPC, Banco de Mexico. with the appearance of the decentralized resources for municipalities in Ramo 26 and its successor in Ramo 33, the Fondo de Aportaciones para la Infraestructura Social Municipal. The question is how this new source of rev- enue will affect municipal fiscal effort. It is quite possible that fiscal effort fell in 1998, the year in which municipalities began to receive the direct flow coming from Ramo 33 resources. Unfortunately, there are no detailed nationwide data for 1997 and 1998. The structure of revenues differs considerably by type of municipality. In 1993 the big urban municipalities (2 percent of all municipalities) count- ed on 40 percent or less from revenue sharing, the middle urban munici- palities (13 percent of the total) counted on between 40 and 70 percent, while the middle and small rural municipalities (85 percent of the total) counted on between 70 and 95 percent. Because of the direct transfers coming from Ramo 26 for 1995-97 and Ramo 33 for 1998, municipalities across-the- board became more dependent on federal resources. In municipal budgets, increases in real income have been accompanied by steady increases in the share of resources going into current expendi- tures. These increases may be explained by accumulated shortages in cer- 148 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO tain inputs, such as personnel, maintenance of infrastructure, or adminis- trative expenditure. There is also the possibility that municipalities have become more wasteful simply because they have expanded the local bureaucracy unnecessarily. Capital investment expenditures declined vis- a-vis current expenditures from 1989 to 1996 (table 3.4). This trend could sharpen in the future due to the presence of direct transfers from Ramo 33 that are earmarked for capital investment. Because of Ramo 33 funds, municipalities will be able to spend more for current expenditures by free- ing up funds formerly used for capital investment. However, the level of substitution will be limited by the size of Ramo 33 funds. Given that Ramo 33 is literally multiplying the funds available to municipalities and that much of the funds for municipalities in Ramo 33 are restricted to different forms of capital investment, the relative share of capital expenditures in municipal budgets may be on the rise. In the case of education, the financial responsibilities decentralized to municipalities correspond strictly to the maintenance of school structures and the provision of basic equipment to schools. Municipal expenditures on public education have been very modest since the decentralization of education started in 1992. Municipal expenditures on education have been between 0.01 and 0.03 percent of GDP, while consolidated government expenditures on education have been between 4.41 and 5.21 percent of GDP. This is true despite the significant need to maintain school structures. A sur- vey conducted in 1995 found that primary schools in 55 percent of the coun- try's municipalities were in urgent need of repair and maintenance, as were junior high schools in 41 percent of municipalities (data obtained from Secretaria de Gobernaci6n, CEDEMUN, and INEGI 1995). Despite the flow of resources from Ramo 26, municipal governments were not able to assume full financial responsibility for maintaining school facilities. It remains to be seen whether transfers under Ramo 33 will change this situation. Since 1989, municipal governments have received more responsibilities and more funds for social programs and social infrastructure. First, these funds came from Ramo 26, where the federal government had considerable discretion, and in 1998 from the formula-driven Ramo 33. The total flow of resources from Ramo 26 was equivalent to 31 percent of gross municipal spending in 1989, 49 percent in 1992, and 45 percent in 1995 (table 3.5). In some small municipalities, resources coming from Ramo 33 tripled total municipal spending in 1998. In some medium and large municipalities, they covered only 6 to 10 percent of total municipal expenditures. In essence, they altered spending by small municipalities more than they did spend- ing by medium and large ones. There is little information on how the new Ramo 33 funds are being used. However, an informal survey conducted by the Secretariat de Desarrollo Social (SEDESOL) in 156 municipalities provides some data from 1997 that is help- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 149 Table 3.5. Ramo 26 as a Percentage of Municipal Gross Expenditure, 1989-95 Year Share 1989 30.61 1990 40.18 1991 44.47 1992 48.56 1993 46.07 1994 45.55 1995 44.57 Source: Zedillo (various years). ful for understanding the subsequent changes in municipal expenditures induced by Ramo 33. In 1997, small rural municipalities spent only 38 per- cent of social infrastructure funds in the municipal seat of government, which indicates a better distribution of the benefits of the program. However, during 1998, they spent a higher percentage of the funds in the seat of government. For 1997, FISM funds were spent on drinking water projects (15 percent), roads (13 percent), education (12 percent), and paving and urbanization (16 percent). For 1998, more investment was concentrated in paving and urban- ization (37 percent), relegating to second and third place drinking water projects (9 percent) and roads (8 percent). For 1997,48 percent of the projects realized with FISM funds were exe- cuted by community committees, 35 percent directly by city councils, 12 percent in a combined way, and only 5 percent through contractors. For 1998, 35 percent were executed by community committees, 30 percent by city councils, and 30 percent by contractors. Thus in the short life of the Ramo 33, it appears that decisions became more centralized, distribution of benefits worsened within municipalities, and potentially less productive use was made of the funds. The time period is too short, however, and the survey too informal, to permit firm conclusions. Horizontal Disparities Expenditures per capita across states for 1995-97 show significant hori- zontal disparities (table 3.6). These disparities seem to decrease over time, but they remain high by international standards.2 Disparities in expendi- tures per capita are much more pronounced for capital investment than for total or current expenditures per capita. For 1995, Sonora spent 19 times more per capita than Guanajuato. For 1997, the state with the highest spending per capita, Baja California Sur, spent over five times more than 150 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 3.6. Horizontal Disparities in Expenditures Per Capita, 1997 Net expenditure Current expense Investment Total Per capita Total Per capita Total Per capita State (millions) (thousands) (millions) (thousands) (millions) (thousands) Aguascalientes 2,501 2.71 2,221 2.41 279 0.30 Baja California 6,563 2.81 6,304 2.70 258 0.11 Baja California Sur 1,845 4.51 1,786 4.36 58 0.14 Campeche 2,956 4.29 2,326 3.38 628 0.91 Chiapas 10,577 2.70 9,716 2.48 861 0.22 Chihuahua 7,856 2.66 7,009 2.38 845 0.29 Coahuila 6,702 2.97 5,618 2.49 1,083 0.48 Colima 1,749 3.41 1,661 3.24 87 0.17 Distrito Federal 39,057 4.57 28,216 3.30 10,840 1.27 Durango 4,025 2.75 3,869 2.64 155 0.11 Guanajuato 5,880 1.26 5,299 1.14 581 0.12 Guerrero 8,570 2.77 7,636 2.47 934 0.30 Hidalgo 5,842 2.64 4,743 2.14 1,099 0.50 Jalisco 13,763 2.17 12,232 1.93 1,530 0.24 Mexico 21,645 1.73 21,002 1.68 643 0.05 Michoacan 8,879 2.20 8,211 2.03 668 0.17 Morelos 3,808 2.45 3,416 2.20 392 0.25 Nayarit 2,940 3.17 2,413 2.60 527 0.57 Nuevo Le6n 9,214 2.45 8,991 2.39 223 0.06 Oaxaca 8,267 2.39 8,044 2.33 223 0.06 Puebla 10,282 2.07 9,459 1.91 824 0.17 Quer6taro 3,872 2.89 3,154 2.35 717 0.53 Quintana Roo 2,605 3.23 2,301 2.85 304 0.38 San Luis PotosI 4,987 2.16 4,744 2.06 243 0.11 Sinaloa 6,011 2.39 5,607 2.23 404 0.16 Sonora 6,251 2.84 5,775 2.62 476 0.22 Tabasco 7,396 3.97 6,328 3.40 1,069 0.57 Tamaulipas 8,422 3.18 5,797 2.19 2,625 0.99 Tlaxcala 2,598 2.77 2,518 2.69 80 0.09 Veracruz 18,685 2.60 16,213 2.26 2,472 0.34 Yucatan 4,088 2.49 3,823 2.33 265 0.16 Zacatecas 3,413 2.48 3,307 2.40 106 0.08 Minimum 1.26 1.14 0.05 Maximum 4.57 4.36 1.27 Mean 2.80 2.49 0.32 Standard deviation 0.73 0.59 0.29 Coefficient of variation 0.26 0.24 0.92 Source: INEGI 1998. SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 151 the state with the lowest, Guerrero. Moreover, the coefficient of variation in total expenditures per capita dropped from 0.45 for 1995 to 0.36 for 1997. The horizontal disparities in capital expenditures per capita for 1997 show a range of forty-sixfold between the lowest-spending state, Guanajuato, and the highest-spending state, Hidalgo. The coefficient of variation for capital expenditures per capita decreased only slightly from 0.61 in 1995 to 0.58 in 1997. Not unexpectedly, total net expenditures per capita at the state level vary with the average economic resources available to state residents as mea- sured by gross regional product per capita. In 1995, for each additional $Mexl,000 in per capita gross regional product, per capita expenditures increased around $Mexl80 at the state level.3 There is little systematic information on expenditures per capita for functional areas. Significant disparities exist across states in expenditures per student in education and expenditures per capita in health. In the case of education, expenditure per student for 1995 was $Mex939 in Estado de Mexico, $Mexl,132 in Jalisco, $Mexl,198 in Guanajuato, $Mex2,569 in Nayarit, $Mex2,836 in Campeche, and $Mex3,535 in Baja California Sur. Poorer states such as Chiapas, Oaxaca, or San Luis Potosi had fewer funds to invest in primary schools or primary health clinics than other states. These horizontal disparities may be much more significant across munic- ipalities within states than across states. About 95 percent of the inequali- ty in the "Carrera Magisterial" (grade school teacher) test scores is account- ed for by within-state differences (World Bank 1998a). The state seats of government seem to receive more resources than other municipalities, especially the rural municipalities. In other words, the level of equalization achieved by federal programs designed for the states seems to be diluted within the states. Problems with Expenditure Assignments The current assignment of expenditure responsibilities still has three sig- nificant problems-unclear assignments, a dichotomy between investment construction and maintenance responsibility, and unfunded mandates. Unclear Assignment Without doubt, the most important problem is that the assignment of expenditure responsibilities in several important areas continues to be ill- defined. These areas include water resources, higher education, roads, and general infrastructure investment. Some responsibilities remain to be devolved pending the design of appropriate legal and regulatory frame- 152 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO works. The devolution of responsibilities on the revenue side of the budget is tied to changes in revenue assignments and transfers. This is, for instance, the case for agriculture, tourism, and the environment. Many services, including health, education, roads, and rural develop- ment and agriculture, are provided in a mode of shared responsibility rather than independent assignment to a particular level of government. In theory, there is nothing wrong with shared responsibilities. The prob- lem arises when the responsibility of each level of government is poorly delineated. This is less of a problem in health services (but not insignificant) and much more of a problem in education, roads, and rural development and agriculture. The lack of clarity in assignments is responsible for the inefficient pro- vision and frequent underprovision of some services, because either the state or federal level of government attempts to free ride on the other level. Part of the confusion arises from ambiguity about Mexico's type of feder- alist system. Should the federal government work only through the states, or should it also work directly with municipalities, even though the con- stitution prohibits direct transfers? In a pure federalist system, the federal government works exclusively through the states, and the assignment of expenditures between the state and the municipalities is determined by each state. Confusion about responsibilities also arises from the funding mecha- nisms used. The use of earmarked transfers reduces the ability of subna- tional governments to allocate resources as they choose and, in many cases, allows federal agencies to micromanage the activities of subnational gov- ernments. This works to the detriment of both subnational autonomy and production efficiency; it also limits the ability of subnational governments to respond to taxpayer preferences. Dichotomy of Responsibilitiesfor Construction and Maintenance A persistent problem with the assignment of expenditure responsibilities in Mexico is the separation between the assignment of responsibility for maintenance and operation of infrastructure facilities and the assignment of responsibility for capital investment. For example, in the case of basic education, municipalities maintain school buildings, while federal and state governments carry out the vast majority of capital investments. This dichotomy reduces the expenditures for both maintenance and capital infrastructure, because each level of government can blame the other for not doing its part, and each level expects that the other will ultimately replace or renovate and maintain the property. As a result, neither main- tenance nor new investment is adequate, and most of the physical infra- structure is decrepit and poorly maintained. SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 153 Unfunded Mandates Unfunded mandates from the federal government impose expenditures on subnational governments without compensation and seem to be on the rise in Mexico. Without doubt, the most significant mandate on subnational governments is the setting of wages for education and health, which occurs at the federal level through negotiations between the national unions and the Ministries of Education and Health. The negative impact on state finances is especially pronounced in the case of education. There are 1.5 mil- lion employees in the state education sector compared with 130,000 employ- ees in the state health sector. Other examples include decisions by the Federal Electoral Institute and the Ministry of the Interior that impose extra expense on state electoral institutes. Decentralization in Specific Sectors Decentralization has followed different patterns in the various sectors: education, health, water and sanitation, other infrastructure, agriculture, the environment, roads, social programs, and police and public safety. The first two are the largest and are examined in detail below; the others are briefly reviewed. Education Sector Decentralization in education is key to the government's overall decen- tralization policy (World Bank 1998a). Mexico's education system has an unusually complex structure and history (box 3.1). The division of respon- sibilities and the educational system in general are regulated in Article 3 of the Constitution and in the General Education Law. The process of decentralization in basic education started in 1992 with the National Agreement to Modernize Basic Education. This agreement was signed by the federal government, the state governments, and the National Union of Workers in Education. The critics of the National Agreement have argued that the process resembles administrative deconcentration rather than true decentralization. This criticism has significant merit. Under the National Agreement, federal authorities (the Ministry of Public Education) remain in charge of all normative and policymaking func- tions for setting standards, developing curricula and teaching programs, training teachers, producing most textbooks, and evaluating and moni- toring performance. They also provide the states with financial resources to ensure proper coverage and quality of the educational system, and all transfers for education are earmarked for specific purposes. The assign- ment of responsibilities to the states is general and somewhat vague. Specifically, the National Agreement charges the states with (a) delivering 154 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Box 3.1. Mexico's Education System The education system in Mexico consists of mandatory basic education (which includes primary education (from ages 6 to 12) and lower secondary education (from ages 12 to 16). The latter comprises a general and a vocational/technical track. Some of the secondary education is imparted through telesecundaria, a distance-learning approach for remote areas. Prior to basic education, there is a relatively small preschool program. Beyond basic education, there is upper- secondary or bachillerato, with a duration of about three years. With bachiller- ato, students again can choose between general education and technical schools. Finally, there is higher or tertiary education, imparted by teacher-training col- leges, federal technological institutes, state universities, and autonomous universities. The financing of basic education is the combined responsibility of the fed- eral and state governments. Municipalities also contribute to school mainte- nance. For high school and college education, states have agreements vvith the federal government to share these expenditures. The federal government finances all basic education in the Federal District, adult education, and runs approximately half of all technical schools. These two activities are scheduled for decentralization but there is no concrete timetable. The role of the private sector varies by level of education. The government or public sector is the predominant provider of basic education and bachiller- ato. Public school enrollment shares in 1994 were 94 percent in primary, 93 per- cent in lower secondary, and 78 percent in upper secondary. In higher educa- tion the private sector plays a much more significant role, accounting for 46 percent of all enrollments in 1994. Source: World Bank 1998a. basic education services in their territories, (b) guaranteeing the working rights and benefits of the transferred workers, (c) integrating the federal and state systems under one system, (d) allocating increasing financial resources in real terms to basic education, (e) designing a state evaluation system, (f) proposing regional contents for teaching programs, and (g) creating super- visory bodies. The distribution between the federal and state levels of actual decisions on expenditures has changed little since decentralization began. The most apparent result of decentralization is that some of the deconcentrated organisms of the Ministry of Public Education have been replaced by state agencies. In particular, states now perform the payroll function. The current difficulties with the education system go back many years but were exacerbated by drastic budget cuts during the 1980s. During 1982-90, federal spending on primary education decreased in real terms at an annual average rate of 1.2 percent, while enrollment increased 5 percent annually. Over this period, actual spending per pupil in real terms SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 155 decreased 47.4 percent. As a result, few resources were available for main- tenance of school buildings, new construction, and basic teaching materi- als, and real wages for teachers declined 40 percent. Federal spending on education recovered sharply after 1991. As a share of GDP, real federal spending on education is now back to pre-1980s lev- els. Education expenditures continue to be concentrated at the federal level. Using average figures for 1990-97, 89 percent of education expendi- tures were funded at the federal level, 11 percent were funded at the state level, and less than 1 percent were funded at the municipal level. Of the fed- eral government's share, 33 percent was spent directly by federal govern- ment agencies and 56 percent was transferred to the states. About 95 to 99 percent of federal spending that went to the states as transfers was ear- marked for teachers' salaries, and the rest was earmarked for investment. Funds are transferred to the states for the construction of schools (CAPFCE), compensatory programs (CONAFE), adult education (INEA), and special compensatory programs (PARE and PRODEI). Data by state on the relative importance of these funds are not available. All these funds are distributed to the states through annual agreements between the federal government and individual states, and they vary by state from one year to the next. At least part of this flow is being decentralized to the states, par- ticularly CAPFCE. Other sources of education funding include SEDESOL (PROGRESA, "escuela digna") and Desarrollo Integral de la Familia (DIF, school breakfasts and scholarships). The National Agreement gave states control of operations, but not of the entire education process. States can hire and fire teachers, but they cannot do so on a large scale, nor do they control wages and benefits. States coor- dinate school construction with funds from Ramo 33 for both states and municipalities and, very rarely, with state and local funds, but at least as of the time of writing, federal authorities have not completely transferred to the states the function of construction. In practice, the Ministry of Public Education continues to have a say in the planning and programming of teaching activities and a presence in the budgeting process. In reality, states have limited capacity to operate the system and limited scope to adjust curricula and programs for regional content.4 The lack of clarity between state and federal areas of responsibility has led to confusion and inefficiency in the delivery of services at the state level. The National Agreement has contributed to the confusion by failing to cre- ate an institutional mechanism for conducting a national dialogue between federal and state authorities about the process of decentralization and by failing to clarify the assignment of responsibilities and to adapt the decen- tralized system in a way that makes expenditure more efficient. In contrast, such an institutional mechanism has been effectively at work in the health sector and in police and internal security. 156 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Municipalities and schools have had no significant responsibility or authority in the decentralization process. Municipalities have responsibil- ity for maintenance and some school construction. The latter is nominal because municipalities have to count on state approval to staff any new schools. Although municipalities in many states have expressed their desire to assume more responsibilities for basic education, few states have expressed any intent or plan to decentralize further responsibilities to the municipalities. In some states, municipalities have gained flexibility for school schedules and school lunch programs. Previously, the school lunch program was administered entirely by the federal authorities. Parent asso- ciations in some states can also decide whether to charge students for text- books. International experience shows that the delivery of educational ser- vices becomes more efficient when decentralization reaches the local level and, even more significantly, when it reaches the school level (Gershberg and Winkler, 2000). The National Agreement of 1992 brought additional problems. Its impact on the states was uneven because some states had already developed their education systems and others had not. Its rapid deployment also caused logistical problems. Every state that did not have a special agency for basic education had to create one immediately. These new agencies came into direct competition with the ministries of education and culture in some states. The National Agreement did not abolish the duality between the fed- eral and state systems of education prevailing in many states. Although the Agreement effectively unified salaries in both systems (state and federal), social benefits were not unified.5 This forced states that had their own education system in 1992 to carry dual accounting systems. The National Agreement stated that wages would be set separately in state and federal subsystems, with states negotiating with state unions and federal author- ities negotiating with the federal union. In practice, wages in the federal sub- system, which are negotiated in the Federal District by the Ministry of Public Education and the national union, set the wage level that all states have to follow in their negotiations with state unions.6 State subsystems have little choice but to follow the decisions reached at the federal level. The National Agreement also had a significant impact on the distribu- tion of federal resources across states. The current distribution is quite uneven. Before 1992, some states had extensive (state) education systems financed from state budgets, others had less extensive (state) systems, and quite a few others had no (state) education system at all. All states also had the federal system of basic education managed by the Ministry of Public Education. In 1992, in effect, the federal government took a picture of the different ratios of federal to state teachers and, based on those figures, froze its obligations to support teacher salaries and benefits in the federal sys- tem. As a result, in some states the federal government pays the vast major- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 157 ity of teacher salaries, while in other states the majority of teacher salaries are paid by the state government. This approach not only was inherently punitive of states that had made an effort to develop education, but also ignored the fact that some states have much higher population growth than others and that Mexico's population is increasingly mobile. The system of basic education appears to show significant disparities in expenditures per student. These disparities, especially when they go against poorer states, are worrisome because access to basic education is likely to be the most effective means of equalizing income in the long term. Unfortunately, the quality of education is probably worse in poorer states.7 Under the National Agreement, states have inadequate incentives to fund education. Because of these perceived incentives, the states may not be fully reporting their education budgets. As frequently argued at the fed- eral level, the states may be contributing more to education than is being reported, and the share of expenditures going to wages may be decreasing as a whole. However, these facts are not openly reported to the federal authorities because states fear that increases in their own funding will lead to decreases in federal funding. Although the federal authorities are not openly reducing their state spending at this time, historically this has been the pattern. States spending more of their own resources on education have received fewer resources for education from the federal authorities. The National Agreement supposedly assigns states the responsibility for increasing their education expenditures. This does not seem to have occurred. In a recent study that uses data from the early part of the decen- tralization process (1993 and 1994) for a group of seven states, the ratio of state-funded education expenditures to federal expenditures dropped in all seven states following the start of decentralization (table 3.7; see Merino 1998). At the national level, the share of state-funded expenditure in consolidated Table 3.7. State Expenses as a Percentage of Federal Education Expenses, 1989-94 State 1989-91 1993-94 Difference Baja California 45.4 44.5 -1.0 Federal District 10.7 8.0 -2.7 Hidalgo 4.1 3.6 -0.05 Mexico State 44.7 39.3 -5.4 Nuevo Le6n 47.6 40.2 -7.4 Sinaloa 32.8 32.4 -0.4 Tamaulipas 24.4 18.1 -6.3 158 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO expenditures fell from 13.2 percent in 1992 to 9.4 percent in 1995, and to 8.5 percent for 1996 and 1997. Clearly, the states are not matching federal expenditures on education. The challenge is to undo the problems associated with the current dis- tribution of federal resources for education. The historical pattern of fund- ing provides flexibility in design for the future. Until recently, Mexico allo- cated resources after case-by-case negotiations, with pragmatic adjustments at different junctures and circumstances. This ad hoc process has been used as the distribution benchmark for the decentralization of federal resources since 1992. This distribution benchmark has also been incorpo- rated in the Fiscal Coordination Law. These facts make it difficult to move toward more transparent, fair, and efficient formulas for allocating feder- al education funds. The most appealing option for reform is to transition over time to a "per student" funding formula, which might be adjusted for special costs of service delivery. Other formulas, such as those incorporat- ing some measure of performance or existing inequalities, have been given some consideration. As shown in box 3.2, whatever formula is chosen, there will be significant changes in the distribution of resources vis-a-vis the cur- rent status quo. In the short run, the most urgent need is to improve the transparency, equity, and efficiency of the allocation of federal education-related resources to the states. The current allocation has created incentives to accumulate teachers to the detriment of other key inputs. The absence of resources to allocate at the states' discretion prevents them from introducing their own potentially innovative projects for improving the education system. The current earmarked programs have generated inefficient adminis- trative structures and obstructed the comparison of efficiency and perfor- mance among states. For example, the relation between teaching staff (class teachers) and administrative personnel (teachers dedicated to administra- tion or supervision) varies from a distribution of 65 percent teaching staff and 35 percent administrative staff in some states to 85 percent teaching staff and 15 percent administrative staff in others. The efficiency of public expenditures is especially low in basic education, where 95 to 99 percent of all resources are used to pay teacher salaries. Little is spent on classroom materials, supplies, or maintenance. Efficiency and performance could be greatly improved if more funds were dedicated to inputs of production other than salaries. Getting these additional funds may be difficult because of the commitment and incentives to hire teachers. Nevertheless, the additional funds might be obtained by reducing the number of teachers at the margin through attrition (where enrollment is low or declining), by encouraging the private provision of post-secondary education, and by pushing for higher cost recovery in the public sector universities.8 SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 159 Box 3.2. Impact of Alternative Distribution Formulas for Education Funds The dominant criterion for the current allocation of resources to the states is the number of teachers employed in each state. This evidently generates perverse incentives, including a disproportional growth of personnel and the neglect of other inputs of production for the delivery of educational services. But even the use of this criterion does not produce a good balance across states in terms of students per teacher. For example in a group of six entities, for 1997 the proportions ranged from 17.7 to 32.8 students per teacher at the primary level, and from 13.9 to 33.8 students per teacher at the junior high level. Using other criteria would bring significant changes in the allocation of resources. Distribution by school enrollments. It is clear that equality per student has not been the funding criterion used by the federal authorities to distribute funds to the states. Using this criterion, the corresponding distribution shows that some states would have been receiving a much larger percentage of funds than they have been getting. This is the case for states such as Baja California Sur (62 percent more funds), Durango (+44 percent), and Baja California (+39 percent). At the same time other states would have been receiving a lower percentage of funds. This is the case of Guanajuato (19 percent less funds), Puebla (-16 percent), Queretaro (-15 percent), and Michoacan (-10 percent). Distribution by education efficiency. Allocation of federal funds by the level of the performance of the states' education system would promote a better use of resources. Of course, the use of this criterion would require adjustments for the fiscal capacity of the states and for inputs other than public expenditures, such as parents' level of education and income. The simplest benchmark for the distribution according to this criterion would be a national student perfor- mance evaluation system. Mexico has not yet developed such a system. Some states have experimented with standardized statewide tests (for example, Aguascalientes, Chihuahua, Nuevo Leon, and Sonora). In the absence of nation- al test score data, resource allocation by entity can be simulated using data on the states' "terminal efficiency," that is, the proportion of students that finish their basic studies. This criterion is far from optimal, since states could relax stan- dards to obtain more resources. With this criterion, the states of Estado de Mexico, Michoacan, Nuevo Leon, Puebla, and Guanajuato would receive more resources than they actually are now allocated. Baja California Sur, Campeche, Chiapas, Guerrero, Oaxaca, and Tamaulipas would receive significantly less. Distribution as a compensatory criterion. Compensatory distribution, on the other hand, would allocate more resources to those states for which education lags and/or deficiencies are larger. The objective would be to provide these states with the necessary funds to allow them to perform at national average levels. With this criterion, the states of Chiapas, Oaxaca, Estado de Mexico, Yucatan, and Veracruz would receive more resources than at present Source: World Bank 1998a, p. 49-55. 160 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Health Sector The recent process of decentralization in the health sector started with the adoption in 1995 of the Program to Reform the Health Sector, 1995-2000. The program sought to address inequalities in access to health care and the low quality and inefficiency of health services in the federal-state system admin- istered by the Ministry of Health. This system covers the rural and poor urban population. In Mexico, health care services are also provided by the social security system (Mexican Institute for Social Security [IMSS] and other insti- tutions) and by the private sector (see box 3.3). From 1983 to 1987, an attempt was made to decentralize health services, and 14 states were transferred some responsibilities for health services previously under the federal system. This system included people in rural areas who were attended to by IMSS- COPLAMAR, which has continued to work as IMSS-Solidaridad did ear- lier. During the 1980s decentralization was in fact mostly administrative deconcentration, not decentralization, since those 14 states continued to be subject to strict regulatory and budgetary controls by the Ministry of Health.9 The current decentralization of the federal-state health care sys- tem is deeper and wider than that of the 1980s. The health care system pro- vided by the IMSS is also being decentralized, and this process is in many ways more radical and more rational than that of the federal-state system. The responsibilities assigned to the federal government and to the state governments are as follows. Federal-level responsibilities: i Ensure that national health policy responds to the national and sub- national priorities and that it strengthens the national health system and the state systems * Define and review the legal norms of the health care system and pro- pose any needed changes in the legislation and implementation rules * Evaluate the delivery of health services with the goal of amending deviations, reviewing objectives, reorienting activities, and improv- ing the efficient use of resources * Establish more effective mechanisms for coordinating with other health providers and among different levels of government • Maintain through the Ministry of Health the legal framework, coor- dination, planning, and evaluation * Deal exclusively through the Ministry of Health with labor relations pertaining to all workers in the non-IMSS part of the system. State-level responsibilities: * Organize and operate, among others, primary health services, child- mother services, family planning, and mental health SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 161 Box 3.3. Mexico's Health System The health system in Mexico has several subsystems, each servicing different population groups. First, there are several social security schemes, the largest being the Mexican Institute of Social Security (IMSS), which covers the employ- ees and the self-employed in the formal sector of the economy. Other social secu- rity schemes include those for government workers (ISSSTE), the employees of PEMEX, members of the army and air force (SEDENA), and the navy (MARI- NA). Social security schemes covered approximately 48 million people in 1997- half of the population. Second, the services provided by the Secretaria de Salud (SSA) and the state govemments apply to the population not covered by social security (or "poblaci6n abierta"). This is a heterogeneous group of people from rural areas and some poor urban areas. Health services for this population are provided directly through government health facilities or indirectly through the IMSS-Solidaridad, which is a program financed by the government and actual- ly delivered by the social security institute (IMSS). In 1997, the federal-state health program provided health services to approximately 34 million people. Both the social security systems and the federal-state program for the "poblaci6n abierta" have their own organizations of doctors, clinics, and secondary and tertiary hos- pitals. Third, the private sector provides services of different quality to all income groups. The three systems operate with little or no coordination. Although access to health care is a constitutional right for all Mexicans, in 1994 approximately 10 million people had very little or no access to health care (OECD 1998). The main social security system (IMSS) is also undergoing a process of administrative decentralization. The regional deconcentration of the IMSS start- ed in 1985 with the creation of seven regional offices, each one in charge of 70 to 80 hospitals, and by giving them more flexibility, although not autonomy, in decisionmaking. More recently, the IMSS has devolved more functions and responsibilities to the seven regional directorates and has decentralized deci- sions further to Medical Area Units. There are 139 Medical Area Units, an aver- age of four per state, and each Unit typically consists of a secondary hospital and several clinics. These units have a good degree of authority over budget exe- cution. Also significant for the federal-state system, the IMSS has designed and implemented a resource allocation system based on a risk-adjusted capitation formula. This is the resource allocation used to determine the amount of trans- fers in exchange for the provision of a standard package of care for a given pop- ulation. The adjustment factor takes into account the greater costs implied by the specific demographic, socioeconomic, and epidemiological characteristics of the area. * Overview and control nutrition issues * Prevent and control environmental health threats * Oversee occupational health and basic sanitation * Prevent and control contagious diseases * Implement social assistance programs 162 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO * Maintain the current level of financial participation and attempt to increase these contributions over time. Spending by state governments on the health system was insignificant 15 years ago. More recently, state contributions to health expenditures appear to be increasing. In 1994 these contributions ranged from 5 to 15 per- cent of total state expenditures on health. There are few state-by-state data on the level and distribution of expenditures. For three states (San Luis Potosi, Aguascalientes, and Guanajuato), contributions on average almost doubled from 2 percent in 1993 to 4.4 percent in 1997. These three states show wide disparities in services. One state delivers health services to twice the population of another state with half the budget. Expenditure per patient in these three states ranged from $Mex37 to $Mexl,469 in the same year. Revenues from cost-recovery fees averaged around 10 percent, and on average 90 percent of total resources were used in recurrent expendi- tures and 10 percent in capital investment. The ongoing decentralization in health services has many positive fea- tures. Even the limited experience gained from the decentralization exper- iment in the 1980s produced useful lessons for decentralization. In the case of health, state administrations appear to have been better prepared institutionally, administratively, and technically to assume their new responsibilities. To coordinate all aspects of the reform, the National Health Council (Consejo Nacional de Salud, CNS) was created in 1995 with rep- resentation from all states and federal authorities; it has been meeting on a monthly basis. The council not only has a role in coordination but also in negotiation, dialogue, and conflict resolution. It diagnoses issues, con- ducts surveys, and prepares recommendations. This institution is perhaps the most important element accounting for the relative success of decen- tralization in the health sector. Unlike decentralization in education, more time and thought bave been given to decentralization in the health sector. During 1996 each state eval- uated the existing problems, set objectives for reorganization to increase coverage and efficiency, and issued specific proposals for reform. By 1997 all but one state and the Federal District had signed an agreement with the federal authorities for decentralization. In addition, the Ministry of Health created groups of experts to help the states strengthen their administrative capabilities. The reform created new agencies in the states, called public decentralization units (organismos publicos de descentralizaci6n, OPDs). During the decentralization, all federal health employees were transferred to the OPDs as state employees, and all state personnel were brought up to the pay levels of federal employees. The federal government paid the financial costs of this policy. These OPDs are part of the states' health bud- get and have considerable autonomy for allocating current resources in pri- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 163 mary and secondary care.10 The OPDs report to the states' health ministry on the allocation of resources and results; in turn, the health ministry reports to the state's congress and to the federal Ministry of Health. The Ministry of Health has introduced 21 indicators to measure performance at the state level and to supervise programs through federal delegations in the states. The ongoing decentralization in health services also presents important problems. One is the lack of clarity on how much autonomy state author- ities have in managing resources. The current legal framework gives states the freedom and flexibility to allocate resources, but only after they com- ply with minimum federal standards. There is still confusion on what these minimum standards are and how they should be interpreted. A sim- ilar source of ambiguity is the repeated reference to finding alternative or additional sources of financing for the health system while giving states no direct incentives to come up with additional resources. At least until very recently, the distribution of health funds from Ramo 33 created perverse incentives to streamline cost and increase efficiency. The actual formula used to distribute health transfers under Ramo 33 is not known. The dominant criterion appears to be the existing physical infra- structure in the states (hospitals, clinics, storage, and other facilities).11 Other types of information are also used. The resulting allocations achieved over time have no easy explanation. Using physical infrastructure in the for- mula for determining the distribution of health transfers creates perverse incentives to maintain obsolete infrastructure and keep excess capacity. A 1995 survey found that 23 percent of the hospitals in the federal-state sys- tem had average occupancy rates of 50 percent or less, which seems waste- ful in the Mexican context of many unfilled medical needs. Other problems at the current stage of decentralization include: * Inequitable distribution of the incidence of disease, with children and adult males in rural areas being the worst affected * Significant disparities across the states in the per capita allocation of federal health funds12 * The generally poor quality of the health services provided by the states and the Ministry of Health * Inefficient allocation of resources, with too few health professionals other than doctors * Difficult access to health facilities in rural areas13 * Funding formulas that are not based on a per capita, risk-adjusted basis14 * Insufficient operational autonomy of hospitals * No clear regulations for contracting out hospital services such as laundry or cleaning 164 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO * Little cost recovery * Inadequate budget transparency * Failure to control hospital costs adequately because there is no sepa- ration of the double function as buyer and provider of health care * Important investment decisions made at the federal level * Transfers that consist of earmarked funds with very detailed man- dates, leaving little room for discretion at the subnational level other than management of personnel15 * Perverse incentives for the states to contribute their own funds because higher matching rates from the states can result in reduced federal funding * The determination of wages at the federal level, imposing a federal mandate on the states since they need to pay the salary raises deter- mined at the federal level * Complex regulations that are issued at the federal level, impeding autonomous decisions at the subnational level. WATER AND SANITATION. There has been significant decentralization in water services. The National Water Program (1991-94) proposed to trans- fer responsibilities from the federal government to the jurisdictions that were operating the water and irrigation systems. These federal responsi- bilities were off-loaded taking into consideration the ability to pay and administrative capabilities of the jurisdictions.16 The goal of the program, not fully achieved yet, is to have fully autonomous systems in adminis- tration and financing. Back in 1989 the National Water Commission (Comisi6n Nacional del Agua, CNA) provided water to 14 urban svstems. By now the vast majority of these operations have been decentralized. Concurrent with this decentralization, a majority of the states have introduced legislation facilitating the decentralized operation of water and sewage services. The private sector has also invested in water and sewer projects in Aguascalientes, Cancun, Mexico City, and some of the northern states. Legal reforms in the states have strengthened the role of local water authorities. Currently 118 out of 135 cities with 50,000 or more inhabitants have autonomous water authorities. Municipal governments set tariffs for about one-third of these local authorities. For the rest, tariff changes must be approved by the state congress. The Federal District does not have an autonomous water authority; some tariffs are set by the national Congress and others by the municipal government which provides the service. Even though local water authorities have been able to raise overall cost recov- ery, most systems are far from self-sufficient. Performance and willingness to collect fees differ considerably from state to state. There are still signif- icant problems with cost recovery. In most cases, there is a lack of meter- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 165 ing, and water fees are flat fees, so consumer waste is high. The adminis- trative capabilities of water authorities are often weak, operational costs are high, and leakage and waste in the distribution system are significant. In the case of small companies that have not been able to survive, the respon- sibility for provision has returned to the state governments. The states also have assumed responsibility for water services for the poorest munic- ipalities, although coverage in rural areas is still very low. However, deficits are no longer financed by the federal government, and accountability and transparency in budget operations have improved. OTHER INFRASTRUCTURE. The first step in decentralization in other infra- structure was taken with the Solidarity Program managed by SEDESOL in the Salinas administration. But even in this period, all funds were federal, and states and municipalities mainly executed federal programs with funds earmarked for water, sewerage, school construction, and roads. With Ramo 33, there has been a step forward in the decentralization of these funds. Even though the funds are earmarked for a list of expenditures, sub- national governments have more discretion in how they are used. The efficiency gains from decentralization will be reaped only after subna- tional governments have even more discretion over these funds. However, further discretion makes sense only if subnational governments have the administrative and institutional capacity, especially at the local level. AGRICULTURE. Since the early 1990s many responsibilities previously per- formed by the Ministry of Agriculture have been decentralized, ranging from extension services to research. As in other areas, the response of the states has varied. In poorer states, some of the functions previously carried out by the ministry have not been taken over by state governments. There are now 32 state agricultural councils that include both govemment author- ities and producers, and their agenda is to guide agricultural development in the states. ENVIRONMENT. The 1993 Standards for Environmental Protection Law sup- posedly delegates a significant role to the states, which can adopt their own legislation and establish processes to enforce federal standards, but actu- al decentralization remains incomplete. For example, in some years, the fed- eral authorities have not reached coordination agreements with some of the states. In order to operate, many industries must obtain a permit from the local authorities and also from the National Water Authority, the Ministry of Health, and the Ministry of Industry. The Institute of Environmental Protection in the Federal District is still the only agency that can issue per- mits for hazardous wastes, imposing an additional burden on businesses and discouraging compliance. A new pilot program in 2000 provides 166 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO matching funds to states that sign performance agreements to improve their capacity for environmental protection. ROADS. The Ministry of Communication and Transportation initiated a process of decentralization to the states in 1996 through a series of agree- ments involving the construction and maintenance of rural roads and the maintenance of secondary federal roads (World Bank 1998b). Despite its poor record on the maintenance of national roads, the federal government has been reluctant to devolve management of even part of the federal road network to the states. The budgetary difficulties of the federal government in providing adequate maintenance of the primary road network was the principal impetus for the private toll road scheme, which was introduced in 1989.17 SOCIAL PROGRAM. There has been some degree of decentralization, better called deconcentration, in the national programs for social assistance (DIF). These programs now operate through state offices in all the states and in local units in more than 3,000 municipalities. States and municipalities mostly implement federal programs, but also have their own smaller pro- grams. POLICE AND PUBLIC SAFETY. The decentralization process in this area has been laid out in the General Law of Coordination of the National System of Public Security. There is a national council with membership of the federal gov- ernment and all the states including the Federal District. This council has been meeting regularly to coordinate joint financing of projects such as the construction of jails, purchase of modern equipment, and development of an information network. Institutions of Coordination Formal expenditure assignments can never be explicit enough to cover all contingencies. Similarly, conflicts can arise from diverse interpretations of the law. That is why, regardless of the vehicle and precision of formal assignments, a mechanism for dialogue and coordination between differ- ent levels of government is always needed. Mexico has no government- wide institution of this kind. The Ministry of Finance and Public Credit (Secretaria de Hacienda y Credito Publico, SHCP) has periodic meetings with the finance officers of the states to discuss the fiscal coordination agree- ments and revenue-related issues. SHCP is setting up a Decentralization Committee, chaired on a rotating basis by the three Subsecretaries, to coor- dinate decentralization within SHCP. A few sectors also have such coordi- nation mechanisms, as in health with the National Health Council and in SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 167 police and public safety with the National Council of Public Security. Other sectors, such as education, lack this type of mechanism but need it. A general feature of Mexico's intergovernmental relationships is the use of individual agreements between the federal authorities and each of the states in areas of concurrent responsibility. Typically these agreements are not made public or even ratified by the state legislatures. This system dates back many years to the time when the president effectively appointed all the governors. Although explicit and more transparent agreements would help to coordinate concurrent activities between federal and state author- ities, they are not substitutes for a national forum, such as the National Health Council, which brings all states and the federal government togeth- er in open discussion and dialogue. At the state-municipal level, COPLADE traditionally established prior- ities for all investments of the state and municipal governments. State con- gresses have to approve the budgets for all the municipalities and for the state government. Formal responsibilities include formulating and updat- ing the state's development plan, coordinating the federal, state, and local plans, and executing and evaluating state development plans. Practically all federal investment funds are channeled through this institution.18 In states where COPLADE has functioned well, it has been organized by sec- tors, and each sector may have several subcommittees. Different depart- ments of the state and its municipalities have representation in the COPLADE. In some states, the COPLADE has tried to respond to the pro- posals made by municipalities and by state government departments, but the governor clearly has controlled the ultimate decisions. One main objec- tive of the COPLADE-to incorporate municipalities in the decisionmak- ing process of federal and state capital investment programs destined for the municipalities themselves-has been only partially fulfilled. In reality, COPLADE has worked more like a state planning institution, and the role of municipal authorities has been quite limited. Coordination and conflict resolution between state authorities and municipalities should follow the model of state councils for different areas of concurrent responsibilities and have membership of state authorities and all municipalities in the state. This may be more difficult to implement in states with highly fragmented municipalities, such as Oaxaca. In these cases, the state councils could function with representation of associations of municipalities. More fundamentally, Mexico's institutional infrastructure for decen- tralization of expenditures is dominated by one key factor: the lack of autonomy at the subnational level. This seems to arise from four sources, some of which were mentioned earlier. First, a significant part of the fed- eral transfers (aportaciones) are earmarked. This means that state govern- ments have no discretion over spending outside the designated type of expenditure. This is the case with all transfers under Ramo 33 and with 168 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO other decentralization funds and programs for education, health, agricul- ture, and so forth. Because of earmarking, decentralization has brought state officials not more autonomy, only much more work in accounting for funds. Second, discretion or expenditure autonomy for the states lies with revenue sharing and own revenues. To begin with, 20 percent of all rev- enue sharing goes directly to the municipalities. Many of the federal funds require matching arrangements by the subnational governments, leaving them few discretionary funds with which to pursue their own policies and expenditure priorities. Third, federal legislation imposes on subnational governments a detailed list of legal norms and conditions. These constrain the states' ability to deviate from federal policy, to develop and support their own policies, and to introduce innovations in their own programs. Fourth, the federal government has imposed unfunded mandates on the state governments. The most significant of these is the determination of wages for teachers and health care professionals at the national level. A significant twist in municipal government autonomy has taken place with the recent introduction of Ramo 33 transfers to the municipalities. Many municipalities have seen their resources increase three- to fivefold. In an effort to prevent mismanagement, many state governments have tried to control how municipalities administer these resources. In a well-known case, Puebla established state-level criteria for the distribution and control of Ramo 33 transfers, a decision supported by the Supreme Court. Some other states have gradually introduced similar procedures. Administrative Capacity and Democratic Governance Decentralization will not increase efficiency if subnational governments do not have adequate administrative capacity or institutional development or if the institutions of democratic governance are absent or weak. In some cases, as with the Solidarity Program under the previous administration, municipal government institutions were weakened by the creation of par- allel structures. A significant feature of Mexico's fiscal federalism is the uneven distrib- ution of institutional development and administrative capacity across states and municipalities. The disparities clearly are more pronounced among municipalities than among states, with urban and larger municipalities typically having more capacity than rural and smaller municipalities. The management of state finances is influenced by the legislative frame- work of the state and by the type of laws regulating municipal finances. Three laws are important for regulating state financial management: (a) the Fiscal Coordination Law (which specifies the criteria for distributing resources among municipalities); (b) the Public Budget Process Law (which sets the process and policies for budget formulation, execution, and con- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 169 trol); and (c) the Public Debt Law (which sets the criteria, procedures, and limits for the use of debt financing). In 1997, seven states had none of these three laws, six states had only one, 15 had two, and only three states had all three, but all states have them now. The qualifications and skills of state officials directly affect the efficient use of budgetary resources. In the education sector, the professional train- ing of state functionaries appears to explain the differences in performance across states. According to a study conducted in 1997, states that had the best management also had the best-trained officials (Cabrero, Flamand, and Vega 1998). States whose officials had been in the sector the longest, however, had the worst performance. This may suggest that time in ser- vice often reduces innovation and intensifies the use of traditional approaches. At the municipal level, administrative and institutional capacity is more precarious. An official survey of municipalities in 1995 found the following deficiencies (Secretaria de Gobernaci6n, CEDEMUN, and INEGI 1995): * Even though the great majority of municipalities had a treasurer, only 39 percent had a specialized agency responsible for supervising and planning the budget process. * About 60 percent of the country's municipalities did not have a computer system in 1995; accounting was done by hand in most municipalities. * Regarding the existence of an internal legal framework, 63 percent of municipalities did not have basic regulations or manuals for admin- istrative procedures. * In roughly 80 percent of the municipalities there was no statutory framework for expenditure planning. In these municipalities, the budget document was improvised and very elemental. * In more than 70 percent of the municipalities there were neither basic rules nor manuals for the collection and actualization of the proper- ty tax. In fact, the state administered the property tax for the vast majority of these municipalities. * In 1994, 72 percent of the municipalities collected less than half of their budgeted revenues, 27 percent collected 50 to 85 percent, and less than 1 percent collected 86 to 100 percent of the expected revenues. * Almost half of the municipalities spent more than half of their bud- get on wages and salaries. This reflects the narrow margins for expen- diture decisions that are common in a large part of the country. * Finally, citizens participated in government programs in just 51 per- cent of the municipalities, almost always in public works. Although at least some states have made improvements, a comprehen- sive reassessment would be worthwhile. Without a major effort to develop 170 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO institutional capacity of the poorer, smaller, and rural municipalities, it does not make sense to push decentralization at this level. The federal govern- ment has committed to make funds available for institutional development by states and municipalities. Although many municipalities are ready to play a more significant role in the delivery of services, many more clearly are not. The general reforms promoted by local governments since 1983 and reforms to Article 115 of the Constitution have reached the limit of their efficiency because they do not recognize explicitly the diversity of Mexico's municipalities. Doing so will require both the federal and state governments to use asymmetric approaches to decentralization at the municipal level. The administrative and institutional limitations of the vast majority of municipalities are aggra- vated in cases where the entire staff of the municipality changes every three years with new elections. This is more damaging, of course, where the pool of skilled professionals is limited. The negative impact of this institution- al feature of Mexican public governance cannot be exaggerated. There may be no point in training and increasing the qualifications of municipal officials if the training process has to start again in three years or less. The majority of countries have dealt with the changes brought about by periodic democratic elections by separating elected officials or politicians (mayor, vice mayor, and local council) from technical staff (budget and tax officials, accountants, secretaries, and so forth). Often, technical staff are cer- tified and made part of a professional career-oriented national or state body. In some countries, there is a local civil service. Allowing for the per- manency and professionalization of technical staff at the municipal level is a fundamental reform that state governments need to consider. Interpreting and dealing with differences in institutional develooment and administrative capacity also require full acknowledgment that incen- tives are needed to make subnational governments develop. The lack of incentives can arise from the centralization of responsibilities and resources and from the lack of an appropriate institutional framework, including inadequate compensation for state and local officials or the perceptions that discourage honest and capable citizens from seeking public office. Toward a More Efficient Expenditure Decentralization The first priority is to assign expenditure responsibilities formally both at the federal-state level and the state-local level. This will help to clarify the existing confusion about who is responsible for what in education, health, the environment, water, and other infrastructure. Although many countries formally assign expenditure responsibilities in their constitutions, this avenue is not recommended for Mexico. The con- stitution is difficult to change, and it cannot provide the level of detail need- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 171 ed. The proper vehicle for a formal and explicit statement of expenditure responsibilities is the law, such as the much-discussed (but not yet formally drafted) Treasury Law. If this law were not to progress for some time, a sec- ond-best vehicle would be federal sectoral laws, such as the General Education Law or the General Health Law. At the subnational level, the fur- ther division and clarification of responsibilities between states and their municipalities could be formalized in the Budget Process Law for munici- pal and state governments. Where concurrent sectoral responsibilities exist, responsibilities for regulation, financing, and delivery or implementation should be allocated explicitly among the three levels of government. At the subnational level, state governments need to clarify the assign- ment of responsibilities between the state authorities (the Ministry of Education or the Ministry of Health) and municipalities and the units of implementation (schools or hospitals). The largest benefits from decen- tralization are typically achieved when decentralization reaches local agen- cies, empowering those directly involved in the delivery of services and ensuring accountability to local residents. The most appropriate vehicle for clarifying the division of responsibilities between state and municipal gov- ernments is the Budget Process Law of each state. In terms of specific sector priorities, the assignment of expenditure responsibilities could be better clarified in the areas of water resources and drainage, higher education, school construction, road construction and maintenance, and other infrastructure. The federal government could promote further decentralization from the states to municipalities, but only to municipalities with proven adminis- trative capacity. The best approach is to adopt an opting-in system, under which municipalities that have adequate institutional and administrative capacity can receive further responsibilities and resources to carry them out. It is inefficient to assign responsibility to different levels of government for capital expenditures and for maintenance, such as is currently the case with basic education and roads. In the longer run, it is more desirable to have a single level of government responsible for planning, financing, exe- cuting, and maintaining capital infrastructure. If subnational governments are deemed competent to take over the operation of services, they should also be made responsible for capital investments in this service area. This recommendation is cast in the long run because shifting capital expendi- ture responsibilities to state and local governments will require a corre- sponding adjustment in the assignment of tax instruments, revenue shar- ing, transfers, and long-term borrowing. The clarification of expenditure responsibilities will also demand addressing the problem of unfunded mandates. The most important exist- ing mandate for state government is the national determination between federal agencies and national unions of the salaries to be paid to education 172 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO and health employees. This extremely complex issue will require time and tact to solve. However, true decentralization of education and health services will not happen as long as wages are mandated from the center. The chal- lenge will be to gain the consent of the unions to decentralize the determi- nation of wages and other labor issues to the state level. Perhaps this will happen if and when a number of state governments are able and willing to negotiate higher wages and benefits than those negotiated at the center. One final issue will need to be clarified in the short run. Many state pub- lic employee pension funds are bankrupt or nearly bankrupt. The states believe that this situation developed because of mandates from federal leg- islation going back many years. Federal and state authorities need to dis- cuss this issue and work toward a compromise for funding existing liabil- ities arising from past decisions. The agreement would eventually need to establish clear responsibility for any future liabilities within the states. This may require modifying the mandates contained in current federal legislation. Institutions to Clarify Expenditure Assignments Although the National Health Council has not worked as well as desired, it has clearly shown the benefits of a nationwide institution for encourag- ing dialogue and resolving conflicts among levels of government. Along similar lines, a National Education Council could be created with the man- date to mediate problems and differences in interpretation between the states and the federal government, make adjustments in the assignment of responsibilities, coordinate educational policies, and spread knowledge on the efficient management of expenditures. This could be accomplished in the short run by reforming the Fiscal Coordination Law to require the introduction of a National Education Council. The current periodic meetings between SHCP and finance representa- tives from the states in the area of budget revenues could be expanded, or held separately, in the area of expenditures. Although all states should be represented in these institutions, it will be impossible to do the same for all municipal authorities. It is important, however, that municipal gov- ernments be represented in the dialogue. This can be addressed by allow- ing representatives of the national associations of municipalities or appoint- ed representatives of municipalities to participate in these forums. At the state level, these institutions of dialogue and coordination could be repli- cated for expenditure responsibilities between state and municipal officials. More Subnational Autonomy and Incentivesfor Increasing Efficiency INCREASE SPENDING AUTONOMY AT THE STATE AND LOCAL LEVELS. To address the lack of expenditure autonomy at the subnational level, the federal gov- SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 173 emnment needs to reduce the degree of earmarking in federal transfers and increase unrestrained or lump-sum transfers of funds. In those sectors where national priorities need to influence the allocation of funds by state authorities, such as health and education, transfers could be earmarked only by sector-with the condition that some minimum level of service per stu- dent or inhabitant be achieved. To increase the autonomy of subnational authorities, the federal author- ities could increasingly use instruments of ex post performance and accountability, in place of ex ante control and regulations focused on input usage or "production function" decisions. Control should be transferred to the greatest extent possible to state legislatures, municipal councils, and ex post audit institutions, such as the Comptroller of the Treasury. The regu- latory role of federal authorities should be limited to basic national pro- grams that involve federal funds, and regulations could continue to move toward performance-based criteria. GIVE INCENTIVES FOR INCREASED EXPENDITURE EFFICIENCY AT THE STATE AND MUNICIPAL LEVELs. Federal guidance, education, and the selective use of con- ditionality could help to spread the good practices and results across states. At the same time, the federal government needs to provide the proper incentives to state and local governments. In particular, the federal gov- ernment needs to remove as soon as possible the perverse incentives pro- vided in the Fiscal Coordination Law for the distribution of funds to edu- cation and health. These funds should not be distributed on the basis of physical capacity (registered number of school or health facilities) and number of employees (teachers and health workers). Instead, a simple for- mula could be used based on objective criteria, which cannot be manipu- lated by state authorities. One way to improve efficiency would be to distribute funds on a per stu- dent basis (for education) and a per inhabitant basis (for health). These sim- ple criteria should move state governments to seek economies in the deliv- ery of services by consolidating facilities and in the use of staff. The per capita criterion may be adjusted, via a formula, to take into account different costs of provision due to differences in price levels or in the current stock of infrastructure. However, these adjustments should be kept to a minimum to exploit the advantages of a capitation system. The transition to the new system should be incremental, given the differences in funding that desir- able new formulas will generate. In the short run, this calls for reform of the Fiscal Coordination Law to introduce over time a per student capitation for- mula, perhaps modified by educational outcomes and existing disparities. Both the federal and state governments could promote the association of municipalities. For example, approved associations of municipalities could receive additional funding for investment and maintenance of facilities. 174 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Moreover, government agencies could contract with the private sector for the full or partial delivery of public services. The experience with full privatization of services has been mixed in Mexico, but there are success stories that could be replicated throughout the country. The federal gov- ernment could play a significant role in spreading and publicizing the positive experiences and in encouraging their replication. Similarly, state governments could encourage municipalities to make greater use of contracting with the private sector. The federal government could take the lead by supporting, through Banobras, the creation of joint ventures between the public and private sectors for building infrastructure and delivering services. The most important check on the wastefulness of the public sector-or the most important incentive for an efficient use of public resources- should be the accountability of public officials to their taxpayers and elec- torate. To this end, the federal and state governments could promote elec- toral reforms that allow individuals to cast votes for members of the town council rather than for a party list, as is now the case. State governments need to provide municipalities more discretion and resources and to increase accountability by implementing units, such as schools and hospitals. State governments also need to provide incentives for better performance and effective support mechanisms to develop tech- nical and administrative capabilities. Improvement of Administrative Capacity and Democratic Governance (Medium and Long Term) and Use of Asymmetric Assignments in the Interim (Short Term) Weak administrative and institutional capacities constitute the most impor- tant impediment to increasing the efficiency of subnational public expen- ditures. Because of significant differences in administrative capacity and institutional development among states, it may be desirable to devolve more expenditure responsibilities only after each state government has achieved objective and transparent criteria that reflect a minimum level of administrative and institutional capacity. There is a need to promote the professional training of middle- and high-level officials in the finance departments of subnational governments. The federal government could make this a priority in its dealings with state governments and local legis- latures. One possibility is for the federal government to offer funds sup- porting the modernization of finance departments under the condition that subnational governments establish professional service systems at least in the area of financial management. SPENDING RESPONSIBILITIES AND SERVICE DELIVERY 175 The federal government could require states to update or develop a leg- islative framework for financial management. At the very least, this frame- work should include a fiscal coordination law for dealing with the munic- ipalities, a budget, an accounting and public expenditure law, and a public debt law. Federal policy could offer incentives for subnational governments to develop institutional and administrative capacity and to ensure that the asymmetric treatment of subnational governments does not create differ- ent classes of subnational governments, in particular an underclass of dependent governments under the paternalistic oversight of federal author- ities. The federal government could help laggard states improve their administrative and institutional capacity. States that have not promoted or developed adequate democratic governance institutions in their munici- palities-also stated in terms of objective and transparent criteria (such as municipal elections with separate executive and legislative powers, a local budget process law, and ex post audit)-should not be devolved functions or transferred funds destined for municipalities. State governments could also be encouraged to use an asymmetric approach to municipal governments depending on their administrative capacity and record of democratic governance. Without a major effort to develop institutional capacity of the poorer, smaller, and rural municipal- ities, decentralization of responsibilities should not proceed. State gov- ernments could be encouraged to help municipalities improve their admin- istrative and institutional capacities. Decentralization would proceed for those municipalities that opt to receive the training and demonstrate their administrative capacity. Strengthening the administrative capacity of municipalities will require several tracks. First, the states should provide clear incentives to munici- palities for upgrading their capacity. This may be done through a qualifi- cation system that devolves more responsibilities and resources to munic- ipalities that improve their administrative capacity and institutions. Second, states need to find the technical means to do all the training, legal advice, and so forth. INDETEC is the type of institution that could be retooled for this type of training, but its operations and current status should be recon- sidered. SEDESOL also gives technical assistance to municipalities in infra- structure, and Banobras has assisted municipal governments with evalu- ation studies for capital investment projects. Third, and most important, it is crucial to remove current institutional barriers to sustained improvement of these capabilities. In particular, municipal technical staff should not be removed every three years when elected officials are replaced. The states could be encouraged to professionalize municipal staff and to create a career ladder for municipal bureaucrats. 176 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Part of the problem is that state governments have not had the right incentives to strengthen their municipalities. It may be desirable for the fed- eral government to create a matching grant program through which state governments compete, winning earmarked resources for strengthening their own administrations, with the condition that a share of these funds go to a state program for strengthening municipal administrations. 4 Tax Assignment Alberto Diaz-Cayeros and Charles E. McLure, Jr. TAX REVENUES OF THE MEXICAN FEDERAL GOVERNMENT are unusually small as a percentage of GDP (11 percent in 1999, not including oil royalties) and of total federal revenues (76 percent). At the same time, federal transfers to states are an unusually large share of federal outlays (over half of out- lays other than debt service) and of the revenue of states (over 80 percent for states other than the Federal District), which raise little revenue on their own. The Mexican public needs more resources for major economic infra- structure and human resources-areas in which the states carry a consid- erable part of the responsibility. The conjunction of these conditions today implies that a major source of increased revenue for the public sector should be more subnational taxation, which could be offset partially by decreases in aggregate transfers from the federal level. The tax section of Chapter 1 provided some principles and a method for thinking about the major potential sources of revenues for state and munic- ipal governments. This chapter applies these principles to the situation in Mexico, evaluates the possibility of assigning particular taxes to state or municipal levels of government, and provides rough estimates of the rev- enue effects of doing so. Potential Sources of State Tax Revenues This section looks at potential sources for increased state revenues, either by increasing use of existing taxes on payroll and automobile registration, introducing new taxes at the state level in the form of surcharges on fed- eral taxes such as the income tax and VAT, or transferring excise taxes from the federal to the state level. It also reviews two potential taxes that the states should not use-corporate income tax and retail sales tax. 177 178 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Payroll Tax Twenty-one of the 32 states of Mexico rely on payroll taxes. These taxes are levied at rates that vary from 0.5 percent to 4 percent. Rates cluster in the range of 1 to 2 percent of payrolls (18 states); the modal rate is 2 percent (imposed by 11 states), and only one state levies a rate in excess of 2 per- cent (4 percent in Baja California Sur, which is due to a 2 percent surcharge to support higher education). Several states have multiple rates-lower rates for specific activities or incomes. Payroll taxes produce almost half of state tax revenues (45.6 percent, but this percentage jumps to 63.4 percent if one excludes the Federal District, the tax authority of which differs from that of the regular states because it includes the property tax). This consti- tutes 9.2 percent of own revenues of state governments-but, because of the preponderance of revenue sharing, only 2.7 percent of total net rev- enues. The percentage of states' own tax revenues among those states that do levy this tax varies from a low of 0.75 percent in Chiapas to a high of 10.74 percent in Nuevo Le6n. Conceptually, the payroll tax is not an attractive source of state revenues. The primary problem is that it is an origin-based tax on one of the key costs of production. As a result, it may distort the location of economic activity if levied at different rates in different states, unless rate differentials reflect differences in the level of public service provided to employers by the var- ious states. State officials are naturally reluctant to impose payroll taxes that are far out of line with those levied by states with which they compete for business, fearing loss of jobs to other states. A second objection is the presence of cities on or near state boundaries. Since payroll taxes are commonly collected at the place of employment, instead of the place of residence, revenues from taxes paid by commuters flow to the "wrong" state; that is, to the state where people work, instead of where they live. The problem is not merely that states where employees live have less revenue than they should and those where employees work have too much, relative to a residence-based system. That problem could be alleviated, and perhaps largely eliminated, by the structure of grants. A more pressing problem is that the exporting of payroll taxes to nonresident employees artificially lowers the political price of taxation in the state of employment. Similarly, to provide a given amount of revenue, the state where employees live must levy higher taxes than if it could levy a resi- dence-based income tax. Commuting across state boundaries seems to be a major problem pri- marily in the Mexico City metropolitan area, which spans two states (Estado de Mexico and the Federal District), but it could also be an issue in some northern cities in the future. There is no straightforward way to quantify commuting in the Mexico City metropolitan area. Since there are no per- sonal income estimates at the state level it is hard to know how much of TAX ASSIGNMENT 179 the tax base is being shifted from one jurisdiction to another. Moreover, the high prevalence of informal employment in Mexico City makes such esti- mation difficult. Some results from income distribution surveys carried out in 1996 for the Mexico City metropolitan area were released by 1999, but the aggregate results are not helpful because individual-level data analy- sis is necessary. Such analysis is a major task that should be undertaken by the local authorities, perhaps with federal support. Despite these concerns, the payroll tax might play a greater role in the finance of the states. It appears to enjoy a reasonable level of acceptance, and it is one of the few potential sources of state revenues that is adminis- tratively feasible. The way the payroll taxes are currently implemented is far from opti- mal. Each state is free to define not only the rate, but also the subject of tax- ation (which person) and the tax base (which part of their pay). While the definitions of taxpayers and tax bases seem to be generally similar, they are not identical to each other or to the payroll taxes levied by the federal gov- ernment for social security and housing. (State definitions differ, for exam- ple, with regard to the treatment of fringe benefits such as commuting allowances and the value of housing and meals provided for employees.) Moreover, each state administers its own payroll tax, following its own pro- cedures, which are independent of those used for the federal payroll tax. Duplication of administration is wasteful for both taxpayers and tax admin- istrators. Differences in both administrative procedures and tax bases cre- ate unnecessary complexity and costs (compliance and audit) for multistate firms. There is some cross-checking of information between state and fed- eral authorities, but this tends to be inefficient-and certainly less efficient than if the state and federal taxes were levied on the same base and collected by the same agency. The most efficient way to implement state payroll taxes would be as a surcharge on the payroll tax levied by the federal government. The base for the federal and state taxes should be identical, except perhaps for differ- ent floors and ceilings on taxable payrolls.1 Federal administration of state surcharges would appear, in theory, to be most efficient. States may, how- ever, challenge this proposition for any number of reasons. If states insist on retaining their own administrations, calculation of liability for state payroll taxes should nonetheless begin with the base for the federal tax. This would require federal collection of data on tax base by state for each employer, something that is not now done. State administrative procedures could be made uniform to minimize compliance costs. The modifications explored above, while important, are not funda- mental-though they may be politically difficult to enact. They would improve the system by simplifying administration and compliance and would thus economize on the costs of administration and compliance. 180 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO These improvements might lead to some increase in revenues, but these are likely to be inconsequential. Rate increases could provide increased revenues. Table 4.1 provides sev- eral estimates of the revenue potential of the payroll tax. The first column indicates present payroll tax revenues, by state. The second column pro- vides an estimate of the revenues that could be gained by the 21 states that now levy payroll tax at rates below 3 percent by increasing rates to 3 per- cent. It involves simply scaling up existing revenues by the ratio of 3 per- cent to the present tax rate. (For example, for states now levying a tax rate of 2 percent, revenues are increased by 50 percent.) The estimate suggests that revenues would increase by 160 percent, an estimation that assumes a rate elasticity of 1, while separate econometric estimates by the authors suggest that the payroll tax revenue in Mexico is rather elastic at 1.5. The third column presents the results based on econometric estimates of an equation that explains payroll tax revenues as a function of rates, employment in the formal sector, state GDP, and revenue sharing. Raising the rate to 3 percent would represent an increase in total net tax revenues2 of around 8 percent in Campeche, Colima, Morelos, and San Luis Potosi; the increase in total net revenues would be less in other states, but it is important to note that in the 11 states that do have a payroll tax, an increase of their rates to 3 percent would mean at least doubling their revenue from this source. The bottom line is that substantial revenue could be derived from the payroll tax in the states that do not now have a payroll tax, and that even in those states that do have a payroll tax, if tax subjects and bases were more similar while compliance was taken to the national average, col- lection would improve. Since the payroll tax constitutes two-thirds of state tax revenues, this increase would be important. Individual Income Tax In theory, the individual income tax is an attractive source of revenue for the states of Mexico (or even for localities). Unlike the payroll tax, it could, in principle, be levied on the basis of residence, instead of the location of employment. Besides channeling revenues to the state (or locality) most likely to provide services to the employee's household, a residence-based tax is less likely to distort the location of economic activity. In addition, there is an intangible impetus toward use of the individual income tax by the Mexican states bordering the United States as they experience economic development induced by their northern neighbor: the desire to imitate the income tax systems of the U.S. states and the Canadian provinces. Surcharges on the individual income tax are not without problems. First, residence-based taxation requires that most taxpayers file income tax declarations, an objective that is not likely to be realized-or even desir- able-soon in Mexico.3 Second, because the income tax threshold is quite TAX ASSIGNMENT 181 Table 4.1. Increase in Revenues from Increase in Payroll Tax (1996 Thousands of Pesos) Proportional increase Econometric estimate of State Current collection* with 3 percent rate increasefor 3 percent rate Aguascalientes 2,788 n.a. 51,402 Baja California 104,873 251,695 36,267 Baja California Sur 7,993 11,990 23,609 Campeche 29,071 43,607 86,962 Coahuila 60,854 182,562 138,500 Colima 102 n.a. 45,801 Chiapas 26,501 90,861 0 Chihuahua 104 208 201,469 Distrito Federal 1,929,602 2,894,403 34,118 Durango 21,540 n.a. 51,908 Guanajuato 0 n.a. 211,274 Guerrero 46,276 69,414 63,837 Hidalgo 22,773 68,319 73,892 Jalisco 305,182 457,773 131,679 M6xico 488,127 732,191 149,611 Michoacan 0 n.a. 115,530 Morelos 279 n.a. 111,063 Nayarit 12,435 18,653 25,725 Nuevo Le6n 374,368 561,552 137,476 Oaxaca 0 n.a. 61,647 Puebla 47,541 142,623 129,878 Queretaro 0 n.a. 91,057 Quintana Roo 45,985 68,978 62,062 San Luis Potosi 1,013 n.a. 112,628 Sinaloa 52,613 105,226 76,717 Sonora 118,880 178,320 29,069 Tabasco 38,175 114,525 0 Tamaulipas 182,137 273,206 0 Tlaxcala 11,322 33,966 12,697 Veracruz 10,033 n.a. 223,727 Yucatdn 41,860 62,790 33,982 Zacatecas 0 n.a. 33,428 Total 3,982,427 6,362,858 2,557,018 n.a. = not applicable. * Includes some minor personal income taxes levied in states without a payroll tax. Zero in the third column indicates an econometric estimate below the 1996 actual. A blank in the second column indicates no true payroll tax in 1996. Source: Author's calculations from INEGI data. 182 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO high and administration of the income tax is, by all accounts, quite weak, only a small percentage of the economically active population currently pays the individual income tax. The economically active population is cal- culated at 35.8 million, but only about 14 million people pay the payroll tax used to finance social security, and only 5.6 million pay the individual income tax, mostly withheld at source. The regional differences in com- pliance and individual filing are striking: 29 percent of the economically active population in the Federal District file returns, but in Oaxaca and Chiapas only 9 percent do. Thus, unlike the payroll tax, the individual income tax is not currently a "mass tax" that might provide significant addi- tional revenues to pay for state services. If the individual income tax is to serve as a benefit-related state tax, it should be paid by a high percentage of the population.4 Given the distribution of income in the states, which closely mirrors that of the country, for most states only the highest decile would be susceptible to pay income taxes, at least under the current fed- eral provision that exempts incomes under three times the minimum wage.5 Third, state individual income taxes would be characterized by sub- stantial horizontal fiscal disparity. The disparity resulting from differences in per capita income across states would be aggravated by a tax threshold and graduated rates. Under present conditions, allowing the states to levy surcharges on the individual income tax does not seem to achieve much that could not be achieved as well by the state payroll taxes.6 Even so, the option could be considered, particularly if the subjects and base of the payroll tax are made more uniform, so that payroll and income taxes could be integrated. There could also be a flat-rate state surcharge and a graduated-rate fed- eral tax; both would be imposed on the base defined by the central gov- ernment. It might be desirable to have a lower tax threshold for the state surcharges than for the federal tax, given the differences in objectives of the two taxes: benefit-related taxation in the case of the state surcharge and pro- gressive taxation in the case of the federal tax.7 This would complicate com- pliance and administration somewhat, because the two thresholds would need to be reflected in withholding tables, as well as rate tables used in com- pleting tax returns.8 Even more than in the case of payroll taxes, federal administration of the possible state or local income tax surcharges seems appropriate. Excises Excise taxes, for example, those on alcoholic beverages and tobacco prod- ucts, would be ideal candidates for assignment to the states.9 They would probably be politically acceptable as a means of financing state expendi- tures, and they could be linked to health care services provided by the states. They are reasonably assigned to the states under the principle of sub- TAX ASSIGNMENT 183 sidiarity in taxation, and because they are relatively visible, they would help ensure accountability of state officials. State excises should be levied by (or on behalf of) the states where con- sumption occurs, not the states where production or importation occurs. States where consumption occurs are more likely than producing states to incur public costs incidental to the consumption of excisable products. Moreover, taxation by producing states is likely to result in inappropriate exporting of tax burden to residents of other states. Even so, for adminis- trative reasons, many excises should be collected at the point of production or importation. The ordinary way to implement state excises on alcoholic beverages and tobacco products is to attach distinct tax stamps to excisable products destined for various states. Taxes on motor fuels are commonly collected at the retail level by using sealed pumps. The primary administrative problem might be smuggling from states with low excises to states with high excises. Smuggling is likely to be prob- lematic for products with high ratios of value to weight and volume, such as alcoholic beverages and tobacco products.l This would tend to limit the differentials between state excise rates that could be sustained. It might be appropriate for the federal government to impose floors, below which state excise rates cannot go, in order to prevent a "race to the bottom" caused by competition among states seeking to attract sales of products they know are intended for smuggling to other states-and to protect local sell- ers from the onslaught of products smuggled from other states, where they are subject to lower taxes.11 Such floors on excise rates may also be needed for a political reason. The industries producing alcoholic beverages and tobacco products are suffi- ciently powerful that it is politically difficult to maintain the level of fed- eral excises. This problem would be exacerbated by the assignment of excises on these products to the states, which would be even less able than the federal government to withstand the political influence of the indus- tries. Moreover, the industries could play off states against each other, thus encouraging a race to the bottom. While a federal floor is inconsistent with the basic reason for assigning taxes to subnational governments-provid- ing states with control over marginal sources of revenues through control of tax rates-it may be needed to avoid virtual elimination of excises on certain products, by concentrating political action in Mexico City. Of course, states should have freedom to go above the federally mandated minimum rates. To maximize the latitude of the states to use excises, the central gov- ernment should stop using them, effectively transferring them to the states. In 1996 federal excises on alcoholic beverages and tobacco products yield- ed total revenues equal to almost 11 percent of net state revenues (tobacco products 4.6 percent, beer 4.3 percent, and other alcoholic beverages 2.0 per- cent). This slightly overestimates what states could receive if they were to 184 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO directly collect these taxes, since they would receive slightly less from rev- enue sharing. Taxes on motor fuels are the most important excise. If they were fully transferred to states, they could represent 24 percent of net state revenues. It was not possible to get data on estimated consumption of these products by state, although there are regional breakdowns of the estimated poten- tial revenues, by state in the case of beer, and by region for alcohol and tobacco. This information was obtained from the beer producer industry and a national health survey on addictions. With this data, a rough estimate is made of how much state excises in these products could yield (table 4.2). Better estimates should be generated when the government makes the actual proposals for change. The greatest problem is that the available fis- cal statistics provide information about how much revenue is collected at the point of production, not at consumption. The state-by-state income dis- tribution surveys conducted in 1996, discussed earlier, provide consump- tion information, but they cover only a third of the states. States and the fed- eral government could work together to produce better consumption information. Automobile Taxes Although the registration tax on existing automobiles (tenencia) is formal- ly a federal tax, revenues are actually assigned to the states, which is rea- sonable. The present tax could thus be assigned formally to the states. In theory, states could choose different rates for the tax, depending on the desire for public services. In fact, the possibility of illegally registering in low-tax states would probably constrain these tax rates to be similar, if not identical. It would be appropriate for the federal government (or the states acting in concert) to impose a national minimum, to prevent ruinous tax competition for the registration of vehicles. Automobile taxes cannot, how- ever, provide a major source of revenue if they are not coupled with the excise on gas. The tax on new cars, ISAN, is also formally federal but its rev- enues are actually assigned to the states. It represents only around 1 per- cent of state net revenues, while tenencia reaches 8 percent. The tax base for these taxes is evenly distributed across states, since the per capita num- ber of cars, except for Mexico City, is rather similar. One would not expect great increases in the yield of these taxes since compliance with them is already near the practical maximum, except for the problem of smuggled cars from the U.S. in several border states, most notably Chihuahua. Corporate Income Tax The corporate income tax is not a particularly good way to finance subna- tional governments, for reasons explained below. It should be used only as a last resort, if necessary, to reduce vertical fiscal imbalance. TAX ASSIGNMENT 185 Table 4.2. Increase in Own Revenues from Reassignment of Major Excises by State All excises on beer and All excises on 50 percent of excises alcoholic beverages tobacco products on motorfuels Increase in Percentage Increase in Percentage Increase in Percentage own of net own of net own of net State revenues revenues revenues revenues revenues revenues Aguascalientes 62,104 7.4 118,378 14.0 96,151 11.4 Baja California 268,524 11.2 294,617 12.3 237,467 9.9 Baja California Sur 48,875 10.6 52,703 11.4 42,480 9.2 Campeche 57,450 5.8 26,629 2.7 71,626 7.3 Coahuila 150,210 7.4 235,076 11.6 240,365 11.8 Colima 50,854 8.6 66,575 11.2 54,075 9.1 Chiapas 316,621 9.0 153,920 4.4 414,020 11.8 Chihuahua 288,058 9.6 303,749 10.1 310,583 10.3 DF 599,553 2.8 834,777 3.9 930,997 4.3 Durango 93,607 8.3 154,910 13.7 158,396 14.0 Guanajuato 337,785 11.0 325,461 10.6 495,935 16.1 Guerrero 246,074 13.0 233,044 12.3 329,147 17.4 Hidalgo 146,831 4.9 154,844 5.2 235,951 7.9 Jalisco 471,239 9.8 827,764 17.2 672,341 14.0 Mexico 752,238 8.0 855,871 9.1 1,304,169 13.9 Michoac6n 309,865 13.2 306,801 13.1 433,320 18.5 Morelos 111,005 8.2 105,695 7.8 161,057 11.9 Nayarit 83,802 11.8 122,216 17.2 99,269 14.0 Nuevo Le6n 330,789 9.5 322,536 9.2 394,631 11.3 Oaxaca 265,144 7.6 262,375 7.5 370,574 10.6 Puebla 317,059 10.0 344,556 10.9 525,032 16.6 Quer6taro 84,987 6.9 91,930 7.4 140,082 11.3 Quintana Roo 96,537 10.2 29,677 3.1 79,828 8.4 San Luis Potosi 136,183 9.6 201,092 14.2 246,042 17.4 Sinaloa 197,797 9.2 332,547 15.5 268,040 12.5 Sonora 214,692 8.6 287,233 11.6 231,516 9.3 Tabasco 142,222 3.9 72,704 2.0 195,562 5.3 Tamaulipas 239,164 10.6 228,986 10.2 280,171 12.5 Tiaxcala 56,588 6.7 64,487 7.7 98,264 11.7 Veracruz 504,373 8.3 542,923 8.9 766,815 12.6 Yucathn 116,693 8.3 64,212 4.6 172,719 12.3 Zacatecas 97,968 9.8 184,153 18.5 149,576 15.0 Source: Authors' estimates. CONCEPTUAL ARGUMENTS. There is little reason to believe that corporate income taxes closely reflect the cost of services provided to business, espe- cially if the taxes are levied only on corporations. There is no reason to believe that only profitable firms consume public services or that con- sumption of public services is closely related to profits. 186 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO ADMINISTRATIVE ARGUMENTS. It is inherently difficult to determine the geo- graphic source of the income of corporations operating in more than one state (see McLure 1984).12 Firms may manipulate transfer prices to shift income to the states where tax rates are lowest. Ordinarily it is even con- ceptually impossible to divide the income of multijurisdictional corpora- tions accurately, because of the economic interdependence among activi- ties in various jurisdictions. For this reason, it is common to use formulas to apportion the tax base among subnational jurisdictions. The "appor- tionment factors" that are commonly used in the United State are payroll, property, and sales; the Canadian provinces use payroll and sales. ECONOMIC ARGUMENTS. Being origin-based taxes, corporate income taxes tend to distort the location of economic activity, be exported to residents of other states, and accentuate horizontal fiscal disparities. LOCATIONAL. EFFECTS. Using a formula to apportion income is equivalent to taxing whatever is in the apportionment formula, at a rate that depends on the nationwide profitability of the firm, relative to the apportionment fac- tors (see McLure 1980). This being the case, it may be more appropriate simply to tax the apportionment factors (for example, payroll, property, and sales) directly. In either event, the tax will have economic effects, including distortion of the location of economic activity, that resemble those of taxes on the apportionment factors. Because payroll and proper- ty (and sales attributed to the state of origin) are origin-based factors, their inclusion in the apportionment formula tends to discourage location of production in the taxing state. By comparison, if the sales factor reflects sales at destination, its inclusion in the formula should have little effect on location.13 TAX EXPORTING. The incidence of a corporate income tax that is based on apportionment of profits among the states is likely to resemble that of a tax on the factors in the apportionment formula (for example, payroll, prop- erty, and sales; see McLure 1981). The portion of the tax that is related to property is likely to be exported in part to nonresident owners of corpo- rations. HORIZONTAL FISCAL DISPARITIES. As noted in Chapter 1, origin-based taxes tend to produce horizontal fiscal disparities. In the case of a tax based on formula apportionment, the forces creating horizontal disparities are prob- ably strongest for the part of the tax that is related to property and weak- est for the part related to sales (if at destination). Figures on GDP per capi- ta by state provide some evidence on the potential for horizontal disparities that would be created by a state corporate income tax. While the Federal TAX ASSIGNMENT 187 District, Nuevo Le6n, and the oil-producing states of Campeche and Tabasco have per capita GDP (in terms of purchasing power parity) com- parable to those of the less wealthy European countries, the figures for Chiapas and Oaxaca are closer to those in Africa. DESIGNING SURCHARGES. The only sensible way to implement a state tax on corporate income would be as a surcharge on the tax of the central gov- ernment. Because of the need for uniformity in tax bases, apportionment formulas, and administrative practices, it would be a mistake to allow the states to define their own tax bases, choose their own apportionment for- mulas, or administer their own taxes. The central government would apply the relevant apportionment formula to the federal tax base to determine the income subject to the tax of the various states and then apply the state tax rates to determine liability for state taxes. It has been impossible to provide reliable estimates of how a state corporate income tax surcharge would be distributed across states. First, most of the federal corporate income tax is reported in Mexico City, where the company headquarters are often located. Second, the corporate income tax is fully integrated with the 2 per- cent asset tax, which although spread more evenly across the country, is offset by the possibility companies have of deducting the asset tax from the corporate income tax in their returns. Hence a high figure in a particular state in the asset tax could simply signal that firms in that state are in a downturn of economic activity, not reporting profits and hence income tax. The third issue is reliability of information. The numbers for corporate income tax provided by INEGI do not match those of the Cuenta de la Hacienda PuIblica Federal, which is supposedly the official source for this information. Retail Sales Tax The federal government of Mexico imposes a value added tax (VAT), and has recently proposed to allow the states to impose a retail sales tax (RST). If implemented, this proposal may create difficulties of compliance and administration-difficulties that have dissuaded most countries from attempting this particular combination of assignment of taxing powers. This subsection describes these difficulties and international experience in this area. Mexico could also consider adopting a system of dual national/sub- national VATs. The next two subsections describe problems that have com- monly been attributed to state VATs and consider briefly dual federal/state VATs. A particular dual VAT system that could work in Mexico is described in greater detail below and in McLure (1999). The proposal to combine state RSTs with the federal VAT raises two con- cerns; they relate to the achievement of consumption-based taxation and the treatment of interstate sales to achieve destination-based taxation. 188 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO TAXING CONSUMPTION. A consumption-based tax applies only to sales to households; it does not burden sales to business.14 In an ideal system, VAT is collected on virtually all sales by registered traders, except those for export, without regard to whether the sale is to a household or unregistered trader or to another registered trader. Taxation of business purchases is eliminated by allowing registered traders to deduct (take "input credit" for) taxes paid on purchases from tax due on sales. Thus the only tax that is not eliminated by input credits is that paid by households and unregistered traders. An ideal RST would directly achieve the objective of exempting sales to businesses; it would apply only to sales to households (and unregistered traders) and exempts all sales to registered traders.15 The administrative problem is how to achieve this objective without opening the door to eva- sion by households claiming to make business purchases. It would be sen- sible to condition exemption of sales to business on presentation of a busi- ness taxpayer identification number (TIN) to the seller, so that, in principle, only purchases that would be eligible for input credits under the VAT would be exempt under the RST.16 The recent Mexican VAT proposal did not contain such a provision; it simply states that sales to businesses should be exempt. The mechanics of the RST and the VAT operate in different ways. Under the RST the vendor collects tax only on sales to households and must keep track of exempt sales made to business purchasers and document the pur- chaser's eligibility for exemption. By comparison, under the VAT, since all sales are taxable, the vendor does not need to make the distinction between taxable and exempt sales, but is saddled with the task of accounting for the tax paid on purchases, in order to claim input credits. Both of these systems involve considerable administrative and compliance costs, especially for small businesses. Imposing both systems on businesses, as under the pro- posal to combine state RSTs with the federal VAT, would be quite onerous. Because the two systems operate differently, implementing both would be costly for tax administration, especially since the proposed legislation for the state RST indicates that the tax would be administered by state admin- istrations.17 Both the state and federal tax administrations would need to determine whether purchases by registered traders are for a legitimate busi- ness purpose, since only those are eligible for exemption (RST) or input credit (VAT).18 In theory, state officials could use information obtained in administration of the federal VAT (or vice versa); in fact, the requisite sharing of information probably would not occur. Moreover, in the absence of substantial cooperation between federal and state administrations, ven- dors would be forced to provide essentially but not exactly the same infor- mation to both. 19 TAX ASSIGNMENT 189 Customarily sales taxes exempt a long list of products sold to house- holds. This is true of the Mexican VAT, and it can be expected that states would also provide exemptions, because the proposal does not specify a uniform base for the RST. Exemptions substantially complicate compliance and administration, because taxpayers and tax administrators need to dis- tinguish between sales of taxable and exempt products. The situation would be even worse if the exemptions provided by the state RSTs and the federal VAT were not substantially identical. Vendors would need to dis- tinguish between four possible combinations of tax treatment, in addition to the distinction between sales to registered traders and other sales: tax- able under both federal and state taxes, exempt under both, and taxable under one, but not the other. To minimize this type of complexity, the bases of the two bases should be identical (except for the difference in treatment of sales to registered traders and to others), so that a given sale to house- holds is either taxed or exempt under both taxes. This implies, of course, identity of the bases of the state RSTs. As noted in Chapter 1, the lack of state sovereignty over the definition of tax bases is a small price to pay for autonomy over tax rates, and thus the level of public spending in the state. DESTINATION-BASED TAXATION OF INTERSTATE TRADE. To the extent that sales to households and unregistered traders pass through locally registered mer- chants, it is relatively easy to achieve destination-based taxation of inter- state trade under the RST, providing the state of origin does not tax inter- state sales to registered traders.20 Some sales to households and unregistered traders, called "remote sales" in the European Union, do not pass through locally registered merchants; mail-order sales of tangible products and electronic commerce in digital content fall into this catego- ry.21 This "disintermediated" interstate commerce poses potential prob- lems, because states of destination are likely to be unable to tax it, for rea- sons mentioned below.22 If states of origin tax interstate sales made directly to households and unregistered traders, revenues from the RST go to the "wrong" state, the state of origin, instead of the state of destination. While this may not be important in the aggregate, it could produce substantial amounts of tax exporting in some cases, for example, by the Federal District and certain other states where large cities are located and, most prominently, by the bor- der state of Tamaulipas, where most of the RST on imports would be paid. It seems more likely that states of origin would exempt all interstate sales, for competitive reasons. If states of destination could not tax sales to house- holds and unregistered traders, there would be gaps in the tax base, and local merchants would be placed at a competitive disadvantage, relative to vendors in other states. Gaps in the tax base violate vertical as well as hor- 190 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO izontal equity, because higher-income households are more likely to buy from out-of-state vendors. It would be difficult to construct a workable system of destination- based taxation of remote sales under an RST. To implement such a system with precision, it would be necessary for each vendor to (a) calculate tax on each remote sale using the tax rate of the state of destination, (b) record taxable sales and tax due, by state, and (c) file a tax return and pay tax in each state to which it makes sales. Nothing comparable would be required under the federal VAT-or under the dual state/federal VAT described below. Similarly, each of the 32 states would, in principle, need to audit ven- dors engaged in remote commerce that are located in each of the other states.23 This system seems too complicated for serious consideration, even if tax were due only from vendors making sales to a particular state in excess of a de minimis amount, since the alternative state/federal VAT described below is available. An alternative would be to subject remote sales to households and unregistered traders to a uniform tax and divide the resulting revenue among the states in proportion to estimated consumption in the various states.24 This is what would occur under the dual federal/state VAT. While this would address the problem of interstate trade, it would not address the first issue discussed above, the fundamental incompatibility in the way the VAT and the RST operate. Value Added Tax: Traditional Concerns The conventional wisdom has long been that subnational governments can- not use the value added tax effectively, especially as a stand-alone tax that they themselves administer.25 As with the RST, the basic problems involve avoidance of multiple (cascading) taxation of business purchases and trade crossing borders between taxing jurisdictions within the nation,26 especially the proper (destination basis) treatment of sales to households and unreg- istered traders.27 Evidence and analysis in the late 1990s, however, have shown ways that state governments can effectively impose a surcharge on the VAT of the national government. The next subsection explains briefly the operation of two "dual" systems of destination-based federal/state VATs; one that is especially attractive for Mexico relies on imposition of a "1compensating VAT" on trade between states; the state component of this system could be imposed either by the central government or by a con- sortium of the states. The rest of this subsection indicates why a stand-alone state VAT is prob- lematic, if not infeasible, whether imposed on the origin or destination basis. Poddar (1990, p. 105) sets out four criteria as a basis to judge state VAT sys- tems: (a) lack of interference with the location of economic activity, which is inherent in the destination principle; (b) lack of internal border controls; (c) state autonomy over tax rates; and (d) simplicity. TAX ASSIGNMENT 191 ORIGIN-BASED TAXATION. Unless imposed at the same rate by all states, ori- gin-based taxation distorts the location of economic activity and is difficult to administer, because all interstate transactions must be valued, to prevent artificial attribution of value added to low-tax jurisdictions.28 But limiting all states to a single rate undercuts one of the primary reasons for assign- ing taxing powers to subnational govemments, allowing subnational choice of tax rates. Moreover, there is a risk that triangular trade will produce undesirable patterns of fiscal flows, as has happened in Brazil.29 DESTINATION-BASED TAXATION. By comparison, destination-based taxation by states acting alone has several possible undesirable features, depend- ing on how it is implemented: internal fiscal frontiers that interfere with trade between the states, the risk that revenues will be lost (if cross-border sales to registered businesses are zero-rated and diverted to households and unregistered traders without payment of tax), onerous burdens of com- pliance and administration, and complicated clearinghouse arrangements between states of origin and destination. These potential problems are described more fully. Fiscal frontiers would impede the operation of a single market within a country. Destination-based taxation is achieved for international trade through the use of "border tax adjustments." That is, imports are subject to the VAT and exports are zero-rated (and tax collected on previous stages of the production-distribution process is refunded to exporting firms, if it exceeds their liability for the VAT on domestic sales). Implementation of this system requires fiscal frontiers, in order to collect the VAT on imports and verify that reported exportation actually occurs. This is ordinarily not a problem for international trade (except in the context of a customs union or common market), since goods commonly stop at the border for other rea- sons, including collection of customs duties and inspections for health and safety. But it is a problem within a country, where free internal trade is an important goal. Zero-rating/deferred payment provides a conceptually attractive approach for dealing with interstate trade; the European Union employs it as a "transitional" scheme. Zero-rating of sales to registered vendors in other states would eliminate tax in the state of origin, and use of deferred payment by the state of destination would avoid the need for fiscal fron- tiers between states. (Since interstate sales would not be taxed, registered vendors who import from another state would have no VAT to claim as a credit against tax on sales. Tax is deferred from the time of importation to the time tax is paid on sales.) As in the case of the RST, sales to households and unregistered traders in other states could be taxed at the rate in the state of destination (with revenue submitted to that state) or, to simplify mat- ters, at the rate in the state of origin (which would receive the revenue). The primary risk is that goods that have been zero-rated as sales to a registered 192 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO trader in another state will be diverted to use by households or unregistered traders, perhaps even in the state of origin. Taxation of all sales by the vendor at the tax rate of the state of destina- tion (with submission of revenue to that state) also achieves the conceptu- ally correct result, again without border controls.30 Moreover, it avoids the need to treat cross-border sales to registered traders and to households and nonregistered traders differently. The primary drawback of this approach is complexity, which is similar to that of attempting destination-based tax- ation under the RST, but even greater, because it would apply to all inter- state sales, not just those to households and unregistered traders. Vendors making interstate sales would potentially need to account for sales to 32 states and file returns in as many states. This alternative does not seem wor- thy of serious consideration. Taxation by the vendor at the rate of the jurisdiction of origin, with a tax credit clearinghouse, seems also to be inordinately complicated. (Moreover, sales to households and unregistered traders would be taxed at the "wrong" rate and tax revenue would go to the "wrong" state, if such sales were accorded the same treatment.) Registered purchasers would be allowed credits for all tax on inputs, but credits would need to be identified by states collecting the input tax-and thus identified by vendors claiming the cred- its, so that destination states granting credits could be reimbursed by states of origin that had originally collected tax for which credit is granted. Since states of destination are to be reimbursed for the credits they grant for taxes paid to states of origin, they have little incentive to verify the validity of credits claimed for tax paid on interstate sales. This option also does not seem worthy of further consideration. The Compensating VAT on Interstate Trade It appears that the above analyses, which have usually implicitly assumed taxation by a single level of government, may overstate the difficulty of implementing a destination-based subnational VAT in the context of a national VAT. If a subnational VAT is possible, the state VAT could be a major source of own revenues for the states of Mexico and could increase their autonomy to raise own revenue at the margin, via state control of VAT surcharge rates. The remainder of this subsection discusses two alternative ways to implement a state VAT as part of a dual federal/state system. ZERO-RATING/DEFERRED PAYMENT IN A DUAL SYSTEM. The dual system used in Canada by the federal government and the province of Quebec ("the Quebec system") relies on zero-rating and deferred payment for the taxa- tion of trade between provinces. Bird and Gendron (1998) assert that the control provided by the presence of the federal tax is enough to prevent TAX ASSIGNMENT 193 unacceptable abuse of the provincial tax.31 (It would be difficult to evade tax on an interstate sale to a household or an unregistered trader by claim- ing that the transaction should be zero-rated upon export. Only interstate sales to registered traders would be eligible for zero-rating.) This argument may be less convincing in Mexico, where tax administration is weaker.32 Some interstate purchases by households and unregistered traders mas- querading as purchases by registered traders might escape tax. Primary reliance on state administration of the state VAT, combined with federal administration of the federal VAT, would accentuate the problem, due to the risk that communication and coordination among state administrations and between them and the federal administration would not be as good as within a centralized administration. DUAL FEDERAL/STATE VAT, WITH COMPENSATING VAT ON INTERSTATE TRADE. A compensating VAT (CVAT) on interstate sales to registered traders would avoid most of the problems associated with stand-alone VATs and with Quebec's system of "unprotected" zero-rating/deferred payment.33 Under the CVAT scheme, interstate sales to registered traders would be zero-rated by the state of origin and subject to deferred payment of VAT by registered businesses in the state of destination, as in the Quebec sys- tem. The risk that zero-rated sales would be diverted to households or to unregistered traders would be avoided (or reduced) by the collection of a federal "compensating VAT" on interstate sales to registered traders. (It is assumed here that the compensating VAT would be collected by the fed- eral government in the context of the federal VAT.) Registered taxpayers would be allowed to claim credits for the compensating VAT, as well as the ordinary federal VAT, in computing federal tax liability; in principle, the two taxes need not be differentiated in calculating liability to the federal government.34 Assuming that the same compensating VAT rate would be applied to interstate sales to households and unregistered traders as to interstate sales to registered traders, vendors would need to deal with only three rates (plus zero, for international exports), only two of which would be relevant for any one domestic sale, as in any system involving taxation by two lev- els of government: the ordinary federal rate on all sales, the rate of the local state on sales within the state, and the compensating VAT rate on sales to other states. Vendors would not need to deal with the tax rates of any other states or to engage in interstate clearing of tax credits. The dual federal/state VAT system has several administrative advan- tages over the combination of a federal VAT and state RSTs discussed ear- lier. Since both state and federal taxes are value added taxes, compliance and administration would be simpler. There would be no need to distinguish 194 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO between taxable sales to households and unregistered traders, nor to exempt sales to registered traders, as under the RST. Under the VAT/CVAT system, vendors would need to distinguish between local and interstate sales (regardless of the nature of the buyer), but that seems to be easier than the multiple distinctions required under the RST. Sharing of VAT Revenue as a Transitional Measure As an interim measure, the federal government might share revenues from a joint federal/state VAT with the states on the basis of a formula reflect- ing the destination of taxable sales. Tax sharing would subsequently be replaced with uniform-rate state taxes on value added. Finally, states would be given latitude to choose their own tax rates. While it is very dif- ficult to estimate what a state VAT would collect, very preliminary calcu- lations suggest that the yield could be significant. Taking into account only domestic transactions susceptible to VAT, and keeping only the national oil company, PEMEX, susceptible to Federal VAT, if the federal government retained a value added tax rate of 12 percent, and during the transition state governments kept a 3 percent rate, state net revenue could increase by 10.9 percent. Such an increase would need to be offset by a decrease in revenue sharing, since federal VAT collection would be reduced, and hence sharable revenue. To this one would need to add an estimated 9.1 percent increase in state net revenues coming from the CVAT collect- ed for imports. Potential Sources of Municipal Revenues Municipal governments are slightly less dependent than the states on rev- enue sharing and transfers. In the aggregate municipal governments receive 64.3 percent of total net revenues from the federal government and the states, 17.6 percent from taxes, and the rest from fees. Of course, this pat- tern differs markedly across and within states. While municipalities in Baja California, Chihuahua, the State of Mexico, Nuevo Le6n, and Quintana Roo finance about a quarter of their expenditure with tax revenue, in Campeche, Chiapas, and Oaxaca this is close to 5 percent, and Morelos exhibits a dismal 1.3 percent. Thus, it would be advantageous to provide municipal governments with access to additional sources of revenue. PROPERTY TAX. The property tax (predial) is the most important source of rev- enue for municipalities, besides the transfers received from the federal and state governments (participaciones). The property tax comprises 13 percent of municipalities' total net revenue, but 74.2 percent of their tax revenues. These numbers exclude the Federal District, which does not have munici- TAX ASSIGNMENT 195 palities and hence collects predial directly. The collection of property tax is best in Mexico City, where compliance has been improved and cadastres are updated more frequently. In fact, predial collection in the Federal District is almost as much as what is collected in all municipalities together. The two most important issues for property tax collection are the updating of the tax base and the problems of compliance. While initia- tives are under way to help municipalities on both counts, successful experiences in specific municipalities could be better disseminated across the country. ExcIsEs ON PUBLIC UTILITIES. Excises on public utilities have the advantage of being easily enforced. Moreover, demand for some public utilities is quite inelastic with respect to price, so that taxation does not create serious dis- tortions. In some cases (for example, electricity) demand is also income elas- tic, but in others (water) it is income inelastic; taxation is progressive in the former case, but regressive in the latter. The proper structure of a tax on electric power deserves attention. First, like other indirect taxes, such a tax should not be applied to business con- sumption of power. Taxing power used by business introduces an element of cascading into the system. Second, if power is subject to a special munic- ipal excise, it should probably be exempt from the VAT, to prevent exces- sive disincentives to use power.35 For electricity, there is a provision for rev- enue sharing to municipalities out of the federal excises, but the formulas are complex and not transparent. Part of the problem with municipalities charging excises on electricity is that utilities are public, and will remain so unless the constitution is changed. A number of municipalities have greatly improved their systems and practices for water charges over the last few years, but as with the property tax, successful experiences are scattered across the country and need systematic dissemination. Constraints in the Constitution According to Article 73, Section XXIX, of the Mexican Constitution, the national Congress has the power to levy taxes 1. On foreign commerce 2. On the utilization and exploitation of natural resources 3. On institutions of credit and insurance companies 4. On public services under concession or operated directly by the Federation 5. Special taxes on a. Electric power 196 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO b. Production and consumption of processed tobacco36 c. Gasoline and other products derived from petroleum d. Matches e. Maguey and its fermented products f. Forestry exploitation g. Production and consumption of beer. It appears that this constitutional grant of powers to the Congress is interpreted to grant to the federal government the exclusive right to levy the taxes listed above. If so, it imposes severe limitations on the ability to achieve the tax assignments described above as appropriate. In particular, it would prohibit state excises on tobacco products, beer, and some other alcoholic beverages (but apparently not most other alcoholic beverages), and motor fuels and municipal taxes on electric power. Thus it appears that it would be necessary to change the constitution to enable the states and municipalities to levy excises on these products, as proposed above. The constitution would pose no barrier to the assignment of excises on other products (for example, soft drinks) to the states. By comparison, there appears to be no constitutional prohibition on state entrance into the sales tax area.37 A CVAT could be problematic if the system is run as a state tax arrangement, since the constitution explicitly forbids states from imposing sales tax at differential rates depending on where products come from. However, if the CVAT is a federal tax, there would be no perceived discrimination in the tax treatment of states to dif- ferent products according to their origin. It would, however, be necessary to nullify or modify the revenue-sharing conventions and the provisions of the value added tax law that prevent the states from levying a VAT as long as they participate in the revenue-sharing program. A final note on the politics of changing the constitution is in order. While ordinary laws are enacted by vote of the House of Deputies, amend- ments to the constitution must be approved first by the Senate and then by the House of Deputies. Summary: Long- and Short-Run Reassignments of Taxes Table 4.3 summarizes the actual assignment of taxes, proposed assignments, reasons for change, and caveats. The most important change proposed at the state level would be the introduction of a dual system of state and fed- eral VAT. This change should not be undertaken, however, without care- ful planning and preparation. Canada is the only nation in the world that has a two-tier VAT that functions satisfactorily, and no nation employs the proposed system of a compensating VAT. (Brazil's two-tier system does not TAX ASSIGNMENT 197 function satisfactorily now, but reforms of the VAT type are under serious consideration.) The next most important change would be to make revenues from des- tination-based excises available to the states, which would require chang- ing the constitution. These two fundamental changes could be highly ben- eficial but cannot be made quickly, for political and technical reasons. Less significant changes involving state taxes are (a) rationalization of the payroll taxes by making definitions and administrative procedures uniform and identical to those for the federal social security taxes, and (b) formally transferring taxes on registration of motor vehicles to the state level. From a technical point of view, these changes could be made relatively quickly. Of course, except for the fears that the origin-based payroll tax will dampen economic activity, increases in the payroll tax could provide more revenues for most states. At the municipal level greater reliance on the property tax may be pos- sible, but not without a modern, up-to-date cadastral survey. This, of course, takes time to implement. Beyond that, taxation of electric power may provide revenues for some municipalities, and fairly quickly. Finally, it is important to increase charges for public services where they are priced below marginal cost, which is typical. Short-term changes: * An increase in the payroll tax rate to 3 percent in those states where the current rate is below 3 percent; * A shift of all revenues from registration of automobiles to the states. Long-term changes: * A shift of revenues from 3 percentage points of the VAT from the fed- eral government to state governments, employing the dual feder- al/state VAT; * A shift of all revenues from excises on alcoholic beverages, tobacco products, and 50 percent of revenues from excises on motor fuels from the federal government to the governments of states where con- sumption occurs. Table 4.4 provides preliminary estimates of the potential revenue avail- able to the states from reassignment of taxes along the lines outlined above. It is based on the assumptions given in the table headings. Table 4.3 Current and Proposed Assignments in Mexico Tax Current assignment Proposed assignment Rationalefor proposal Caveats/comments Production Excises Federal (constitutional) Eliminate Subsidiarity (reserve for states) Consumption State (federal administration, on desti- Benefits of specific or generalized New system; interstate Excises nation basis) benefits; subsidiarity smuggling Vehicle taxes Mixed federal/state State Subsidiarity Convert purchase tax to annual tax Value Added Tax Federal/20 percent to revenue- Federal/state surcharges (Perhaps Specific or generalized benefits; sub- Requires innovative tech- sharing pool destination-based tax sharing in transi- sidiarity; broad coverage; less hori- nique tion) zontal disparities Individual Income Federal/20 percent to revenue- Federal/revenue-sharing pool Narrow coverage; horizontal dis- State surcharge required for sharing pool parities revenue Go Corporation Federal/20 percent to revenue- Federal/revenue-sharing pool Tax exporting; distortion of loca- Income Tax sharing pool tion; difficult to determine geo- graphic source of income; horizon- tal disparities Payroll Tax State State, but perhaps with federal admin- Administration with federal payroll I .ack of trust; less subnation- istration tax is more efficient al autonomy Natural Resources: Ordinary Federal/ 20 percent to revenue-shar- Federal/revenue-sharing pool Avoid horizontal disparity; ing pool provide own state revenues Extraordinary Federal Federal Avoid instability of state revenue revenues Property Tax Municipal Municipal Subsidiarity Need to improve adminis- tration TAX ASSIGNMENT 199 Table 4.4. Increase in Own Revenues from Reassignment of Various Taxes, by State (Thousands of Pesos 1996) Taxes potentially reassigned Taxes potentially reassigned in the short run in the long run Increase in payroll State State VAT of Destination-based State tax to 3 percent automobile taxes 3 percent excises Aguascalientes 51,402 68,113 162,437 372,783 Baja California 36,267 120,373 401,177 1,038,075 Baja California Sur 23,609 21,302 71,765 186,537 Campeche 86,962 27,721 121,006 227,332 Coahuila 138,500 190,812 406,073 866,016 Colima 45,801 31,123 91,354 225,580 Chiapas 0 80,239 699,445 1,298,581 Chihuahua 201,469 183,401 524,700 1,212,974 DF 34,118 2,208,338 1,572,827 3,296,324 Durango 51,908 61,449 267,594 565,309 Guanajuato 211,274 231,126 837,833 1,655,116 Guerrero 63,837 73,307 556,061 1,137,411 Hidalgo 73,892 67,005 398,615 773,577 Jalisco 131,679 526,976 1,135,853 2,643,686 M6xico 149,611 390,673 2,203,265 4,216,448 Michoacdn 115,530 159,547 732,051 1,483,305 Morelos 111,063 82,041 272,089 538,813 Nayarit 25,725 32,695 167,704 404,556 Nuevo Le6n 137,476 566,770 666,690 1,442,586 Oaxaca 61,647 49,248 626,048 1,268,668 Puebla 129,878 222,552 886,989 1,711,678 Quer6taro 91,057 97,512 236,655 457,080 Quintana Roo 62,062 72,329 134,861 285,869 San Luis Potosi 112,628 111,721 415,664 829,360 Sinaloa 76,717 134,296 452,827 1,066,424 Sonora 29,069 148,445 391,123 964,958 Tabasco 0 83,573 330,383 606,051 Tamaulipas 0 180,702 473,321 1,028,492 Tlaxcala 12,697 23,757 166,008 317,604 Veracruz 223,727 236,421 1,295,457 2,580,926 Yucatan 33,982 87,666 291,791 526,342 Zacatecas 33,428 33,716 252,694 581,274 Source: Authors' calculations from INEGI data. 5 Transfers and the Nature of the Mexican Federation Thomas Courchene and Alberto Diaz-Cayeros THIS CHAPTER ANALYZES THE CURRENT PRACTICE of transfers in Mexican fis- cal federalism and then considers alternative approaches to transfer design and implementation. The first section focuses on those principles that ought to inform the system of intergovernmental transfers in any federa- tion, Mexico included. The second section examines the constitution and the programs of the Zedillo administration for evidence on the nature of the social contract underlying Mexican federalism. Then the analysis directs attention to the description, quantification, and assessment of the current set of intergovernmental transfers, including unconditional participaciones, the mostly conditional aportaciones, and the infrastructure transfers. For each of these categories and their components, per capita data are presented both in absolute value by state and in relation to the state's ranking in terms of a poverty index. To these transfer data the following section adds data relat- ing to own-source revenues which, then, allows an assessment of the degree of vertical and horizontal imbalance in the Mexican federation. What emerges is a relatively high degree of horizontal balance, provided one excludes a few of the obvious outliers. The final section turns attention to some alternative scenarios for the evolution of Mexican fiscal federalism. Principles for Intergovernmental Transfers The analysis of Mexico's theory and practice for fiscal federalism, described in the rest of the chapter, indicates that its system of intergovernmental transfers ought to satisfy eight principles. First, intergovernmental grants should complement the associated expenditure and tax allocations in ways that make the overall system res- onate well with the needs and underlying social values of the Mexican fed- 200 TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 201 eration and society. For example, the importance of assuring relatively equal access to certain national public goods like education implies that the fed- eration will have grants earmarked for those purposes. Second, intergovernmental grants should deliver the conception and degree of equity embedded in the Mexican "political contract." This means that the equalizing or horizontal transfers to states should be tailored to deliver this degree of subnational equity. As an important aside, equity does not relate to the grants in isolation: it relates to the sum of the transfers and all other state revenues. That is, the states' overall fiscal capacity should be brought up to the equity standard. Third, the transfer arrangements should be predictable over time. The set of transfers should be designed within a multiyear framework to allow the states a corresponding multiyear planning and budgeting horizon. Although absolute predictability cannot be guaranteed, a useful initial compromise might be a federal-state agreement over three or five years (preferably overlapping the presidential sexenio and the congressional three-year terms) with provisions requiring adequate notice for any changes. The fourth principle is a variant of the last: formulas should determine transfers as much as possible. This implies that the procedures of calcula- tion should be open and transparent, such that they can be reproduced by third parties. To facilitate this, Mexico could consider the establishment of a formal or informal agency representing both the federal government and the states. Australia's Commonwealth Grants Commission is the pre- eminent exemplar of a formal, quasi-independent agency, while Canada's technical working group of federal and provincial officials represents the informal variant. An important role for formula-based transfers is to de- politicize, at least to a degree, the overall transfer regime. These formulas will presumably be altered over time as conditions merit. But at any point in time the degree of objectivity and transparency of formula-based trans- fers tends to defuse what is, in effect, a zero-sum redistributional game. The fifth principle relates to the nature of the conditions that are attached to grants associated with expenditure areas. Consistent with meeting the expenditure goals, the conditions should respect state priorities. For exam- ple, different states may be able to satisfy these conditions in different ways-ways that relate to the social, cultural, and economic needs of their respective constituencies. This argues not for a one-size-fits-all approach to conditionality, but rather for conditionality defined in terms of "equiv- alencies." In the fiscal federalism literature this has come to be associated with the term "competitive federalism" or "horizontal competition," which means allowing states sufficient flexibility to design their own bundles of goods and services consistent with the agreed upon conditions or, prefer- ably, equivalencies. This should tilt the federation in the direction of 202 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO dynamic efficiency, since demonstrably superior approaches to policy design in some states will be copied by other states. The focus here is on conditional grants: we take it as axiomatic that unconditional grants will be spent in ways that reflect state priorities. The sixth principle relates to the objective of tax decentralization. Whatever the approach finally taken to equalization (or to the amelioration of horizontal imbalances), the incentives in this system should ensure that states that increase their tax effort or increase the effectiveness of their col- lection should be able to retain a meaningful share of the increased revenue. Even in developed federations one can still find equalization programs that embody 100 percent (that is, confiscatory) taxation with respect to subna- tional government revenue increases-transfers fall fully in line with rev- enue increases. The relevant formulas should prevent this and should retain some incentive for local tax effort. The seventh principle also anticipates greater tax decentralization. The overall grant system should be designed in a way that anticipates and accommodates the desire of states to enhance tax revenue. For example, an expansion of states' access to tax revenues should automatically trigger the appropriate change in the flow of transfers; that is, the system should not have to be designed anew every time there is further tax decentraliza- tion. Any such institution of adjustment should respect principles three and six-assuring predictability and refraining from penalizing local tax effort. Finally, but hardly exhaustively, any new grant design should take the ini- tial conditions into account. Part of the rationale in this chapter is that the cur- rent set of transfers presumably embodies the implicit social contract. This must inform the evolution of transfers. Furthermore, to be politically feasi- ble, any novel transfer arrangement must be accompanied by adequate and acceptable arrangements for transition from the old to the new system. Mexican Constitution and Fiscal Federalism The constitution is the framework document for all federations. Among other things, it specifies the division of powers among and between the var- ious levels of government, including the assignment of responsibility for expenditure and taxation. In most federations, the constitution also speaks to the nature of the "federal social pact" and, therefore, provides some guid- ance on the nature and role of federal transfers. For example, in the Canadian federation, Section 36 addresses equality of opportunity for all citizens and also provides for equalization payments to provinces to "ensure that provincial governments have sufficient revenues to provide reasonable, comparable public goods at reasonably comparable tax rates." As an important aside, in 1999 the federal government and the Canadian provinces signed a social pact (Framework Agreement on the Social Union) TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 203 designed to establish and promote the internal social and economic union. The German Basic Law embraces the concept of "uniformity of living con- ditions" for all Germans and actually enshrines the details of revenue shar- ing and the interlander revenue-sharing pool. Australia's federal social pact is reflected in the guiding principle of the Commonwealth Grants Commission-"each state should be given the capacity to provide the average standard of state-type public services, assuming it does so at an average level of operational efficiency and makes an average effort to raise revenues from its own sources." In other words, the constitution usually informs the design of intergovernmental transfers. The Mexican Constitution, however, provides little guidance on the appropriate role, operational principles, and design of intergovernmental transfers. The Mexican Constitution proclaims, in Article 40, that Mexico is a representative, democratic, federal republic, composed of free and sovereign states in regard to their internal regimes. Article 124 assigns to the states the residual authority or power over all areas not explicitly assigned to the federal government. However, most of the other provisions then proceed to limit the authority of the states. In fact, the Mexican Federal Pact, as expressed in various provisions of the Constitution, reflects a cer- tain distrust of the states. Moreover, except with respect to education, there is very little in the Constitution about the desired equality of condi- tions among constitutional units. Nor does the Constitution speak to the manner in which the federation should benefit citizens: that is, most of the references to federalism relate to the division of powers, not to the living condition of citizens within the federation. Much of this distrust of states is rooted in an historical process of con- solidation of national authority where, consonant with Riker's (1964) two principles of federalism, the federal arrangement was considered the only way to keep the country together in the face of the threat of the northern neighbor and the secessionist tendencies of some states (most notably Texas, but also Yucatan). At the same time, a highly centralized fiscal and political system was established once the country was pacified at the end the otherwise unstable nineteenth century. The Constitution limits state authority by granting exclusive tax rights to both the federal and the munic- ipal governments. The granting of rights to the federal government is the result of a long process, culminating in the 1940s, when exclusive author- ity was granted to the federal government over several important taxes, although legislation to provide for explicit exclusive tax assignment was never approved by Congress (see Diaz-Cayeros 1997). The assignment of exclusive rights and responsibilities to municipalities is a more recent development, finding its expression in the 1983 reform of Article 115. The federal Constitution explicitly mentions revenue sharing twice. The first mention is in Article 73, which assigns excises (IEPS) exclusively to the 204 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO federal government and then specifies that a secondary law will determine the proportion that states will receive from those revenues. And in the case of electricity revenues, the Constitution states that the state congresses will establish a share for municipalities. The second mention of transfers is in Article 115, which gives states exclusive authority over the land tax. Here, it also provides that states should receive federal revenue shares according to conditions approved by the local congresses. These principles are important because they imply that revenue sharing is integral to the Constitution, but only for excises and for municipal governments. Revenue sharing arising from federal taxation in general is not considered explicit- ly, although it is the most important source of financing for subnational gov- ernments. Transfers to municipal governments are left for local congress- es to decide, but a provision does imply that states must share revenues with their municipalities. In terms of the overall fiscal authority, although Article 31, part 4, does not impose limits on tax assignment, it does provide that Mexicans must contribute to public expenditures of the federation, states, and municipal- ities on a proportional and equitable basis. Another article (Article 73, part 7), however, has been interpreted in practice to mean that the federal gov- ernment has no limitation whatsoever on the taxes it can establish (the arti- cle states that the federal government can establish the taxes necessary to cover its budget). States, in contrast, have very specific limitations on sev- eral tax bases. The most important state limitation is related to the high-pro- file fiscal debate of the nineteenth century, which dealt with taxes on inter- nal trade (the so-called alcabala). Such state taxes are now explicitly forbidden in Articles 116 and 117, although the provisions of those articles have often been violated de facto. The second major limitation is related to natural resources, which are now taxed exclusively under federal jurisdiction (Article 27 and Article 73). The third limitation is more recent: exclusive authority for municipalities over several areas, including property taxes, was established with the 1983 reform of Article 115. Centralization of fiscal authority throughout the twentieth century was enhanced by increasing the authority of the federal Congress in areas as diverse as control over natural resources, regulation of financial institutions, and the exclusive right to impose taxes on foreign trade and the produc- tion of specific goods (tobacco, alcohol, beer, electricity, matches, fuel, for- est products). All of these reforms were carried out by amending Article 73, concerning the attributes of Congress, which has been modified 36 times. Via Article 73, the federal government is allowed to impose taxes as needed to cover the budget; to prevent states from restricting interstate trade; to legislate with respect to oil, mining, cinemas, betting, financial ser- vices, electric and nuclear power, and labor; to coordinate states and munic- TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 205 ipalities on issues of public safety; to distribute and unify the education function "conveniently" among federation, states, and municipalities; to establish taxes on foreign trade, natural resources, credit and insurance institutions, public services granted to private providers, and sales of elec- tricity, tobacco, gas, matches, alcohol, forestry products, and beer; and to enact laws on matters of concurrence between levels of government relat- ing to population and environment. Thus except for the provision granting the federal government respon- sibility for unifying education and distributing the education function among levels of government, there is little in the Constitution relating to the nature of transfers or what they ought to accomplish. We must look elsewhere for an understanding of what federal transfers are meant to achieve. The Zedillo administration (1994-2000) has made the "new federalism" one of its major programs in government. The Program for a New Federalism (Programa para un Nuevo Federalismo, 1995-2000) explicitly states what the administration considers the most important changes needed to reinvigorate the federal arrangement. The program promises to decentralize resources in health, education, and the construc- tion of physical infrastructure, much of which Ramo 33 has achieved. The program does not explicitly address the criteria to determine federal trans- fers for those functions. It does, however, seek to strengthen state autono- my and to redistribute "authority, functions, responsibility, and resources" from the federal government to the subnational governments (Programa 1997, p. 7). It also suggests that fiscal coordination should incorporate rev- enue, expenditure, and debt. But when it comes to stating the lines of action, there are only proposals to "carry out studies" and "propose new models of expenditure that should redistribute and balance both vertical- ly and horizontally the assignment of resources for social spending and fed- eral public investment" (p. 17). On the unconditional transfers granted through revenue sharing, the program does advocate revising the formu- las for the distribution of revenue, but it fails to mention anything more spe- cific than that. Indeed, the program is rather vague, calling for "including in the formulas, representative variables of the phenomena that are effec- tively sought to be measured" (p. 17). The Program for a New Federalism gives some details on the devolu- tion of expenditure and signals the future evolution of the federation, but not on the nature and characteristics of the accompanying transfer system. So, we look elsewhere for this vital information. In particular, we focus on the de facto overall distribution of the actual transfers and own revenues across states, which reveal the preferences in the social pact underlying the Mexican federation. The degree of overall horizontal imbalance across states is much less than one might expect from the interplay of more than 20 transfer programs and own-source revenues. 206 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Current Intergovernmental Transfer Programs Transfers in Mexico accounted for 39 percent of the total federal govern- ment budget in 1999. As figure 5.1 shows, this is a remarkable share, since it is larger than what is spent in all the ministries of the central sector (30 percent); the allocations to pay for the public debt (22 percent, including financial supports for banks and debtors as well as bills pending from pre- vious periods); and support for the social security institutions and pension funds (9 percent). This 39 percent of the federal budget is made up by bud- getary ramos or items-Ramo 23 for the president's discretionary fund, which is used for salary increases and natural disasters; Ramo 25 for sup- port of education and salary increases in the Federal District; Ramo 26 for regional development; Ramo 28 for revenue sharing; and Ramo 33, start- ed in 1998, for transfers (aportaciones) for education, health, social infra- structure, and other uses. Ramos 23 and 26 are being phased out, so 28 and 33 have become the main transfers. This is not all the money that the federal government spends in states and municipalities. For example, there are matching grants for public investment projects with the states (which are substantially smaller than in previous years, but still remain important), and there are funds for social assistance such as PROGRESA, PROCAMPO, and other social wel- fare initiatives. Nevertheless, we focus here on the 39 percent share corre- sponding to funds that are transferred to states and municipalities, both Figure 5.1. Composition of the Federal Government Budget 9%// H] Central Sector 30%,/l) f0$ f $: g Q g Xf S f0:' E f j 0' tff; ff $ ffi Social Security * Debt E Transfers Souirce: LeV de Egreso 1999. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 207 conditional and unconditional. We begin with the unconditional transfers arising from revenue sharing (Ramo 28). Unconditional Revenue Sharing (Participaciones) The most important element in the Mexican federal transfer system was, until December 1997, the revenue sharing (participaciones) determined by the National System of Fiscal Coordination (Sistema Nacional de Coordinaci6n Fiscal) from 1980 onward. Laws relating to fiscal coordina- tion existed before 1980, but the current system took shape once the value added tax was established at the federal level, substituting for a federal sales tax (Impuesto Sobre Ingresos Mercantiles, ISIM) that had existed for decades in most states. ISIM was not universally embraced by all states until the early 1970s, when the overall rate was increased from 3 to 4 percent, 2.2 percent of which was assigned to the states. (Accounts of the history of rev- enue sharing can be found in Bonifaz Chapoy 1992; Martinez Almazan 1980; Martinez Cabafnas 1985; Retchkiman and Gil Valdivia 1981; and Diaz Cayeros 1995.) When the current system was established in 1980, the states surrendered not only their share of the federal sales tax, but also the mul- tiple excise taxes on consumption and production that had been accumu- lating over the years. The 1980 agreement established two contracts between states and the federal government. The first was the convenio de adhesi6n, which established that states would voluntarily join the National System of Fiscal Coordination and hence restrain their own taxing authority. In return, the states would receive unconditional revenue-sharing transfers. The second was the convenio de colaboraci6n administrativa, which provided guidelines for joint state surveillance and monitoring of federal tax com- pliance, providing incentives to states for detecting tax evasion. Originally three funds comprised the revenue-sharing system: the Fondo General de Participaciones and the Fondo Financiero Complementario, which were distributed to states, and the Fondo de Fomento Municipal, which was transferred to municipalities via state governments, but according to a federal allocation formula. The states were also required to transfer 20 per- cent of these federal revenue shares to municipalities, in accordance with the states' own formulas or allocation criteria. The overall amount for the states and municipalities was determined as a percentage (now equal to 20 percent) of the Recaudaci6n Federal Participable (RIP) or "assignable taxes." The main components of these assignable or shared taxes are revenues col- lected from the federal income tax, the value added tax, and the ordinary fees from oil. This meant that, except for the so-called "extraordinary" fees on oil and some other items, most of the domestic federal taxes were sub- ject to sharing. In addition to these funds, several specific federal taxes are shared on a case-by-case basis. The two most important for states are the property tax 208 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO on automobiles (tenencia) and the tax on new cars (ISAN), which are both fully shared-that is, 100 percent of the revenues are transferred to the states. Both tenencia and ISAN are transferred on a derivation basis, so their collection in each state is a good indicator of the underlying tax base in the respective state. This is not true for value added and income taxes, which are typically reported regionally in terms of the companies' headquarters. To this we should add excises that are shared according to complex for- mulas depending on the product or service involved. Additionally, there are annexes to the convenios de adhesi6n that establish additional percent- ages when the states share other federal nontax sources of revenue. The components of these unconditional revenue-sharing transfers appear as the first six rows of table 5.1. In addition to indicating their total values for 1999, the table also shows selected features of these transfers, including the method of allocation across provinces and whether it is for- mula based. Far and away the largest transfer is the Fondo General de Participaciones, which appears in row 1. Although the allocation formula for participa- ciones, or that of its constituent components, has varied over the years, since 1993 (when these constituent funds were consolidated in the participation fund), the allocation has been as follows: * 45.17 percent is distributed to the states on an equal per capita basis. * 45.17 percent is allocated on a historical basis (referred to as an "iner- tial" basis), starting with the states' own revenues just before the sys- tem started in 1980 and modified gradually by relative state tax effort. In the figures that follow, the effect of the historical allocation is evi- dent for the states of Tabasco and Campeche and reflects their high pre-1980 revenues as major oil-producing states. * 9.66 percent is allocated in a way that "compensates" (or that is inverse relative to) the previous two allocations. The original allocation principle gave primary weight to the derivation principle-the so-called principio resarcitorio-which in effect stipulated that no state should receive less from participaciones than they were col- lecting before joining the revenue-sharing system. This has been tempered over time by the shift toward an equal per capita basis for a significant por- tion of the overall allocation (although, as noted, the historical or inertial component still reflects some of this earlier allocation principle). Diaz Cayeros (1995) demonstrates a gradual convergence across states in per capita revenue shares; while Saucedo Sanchez (1997) shows a steady decrease in the population-weighted Gini coefficient for the distribution of state revenue shares over the years (the opposite result holds for a Gini coef- ficient weighted by state GDP, which has become more unequal). The tendency toward capitation in the largest fund has led total revenue shares under Ramo 28 to be distributed fairly equally across states. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 209 Table 5.1. Federal Funds Transferred to States and Municipalities (In Ascending Order of Federal Accountability) Percent Type of Nature of Faind Ramo Amount RFP expenditure Transfer allocation Formula General de 28 117,801 20 States must transfer at Unconditional Equal Yes Participaciones (FGP) least 20 percent to municipalities de Fomento Municipal 28 5,559 1 Mostly current Unconditional Derivation Yes (FFM) Tenencia 28 7,984 n.a. Works as own Unconditional Derivation 100 '¼. revenue in practice Automoviles nuevos 28 3,255 n.a. Works as own Unconditional Derivation 100 % (ISAN) revenue in practice Impuestos especiales 28 2,600 n.a. Mostly current Unconditional Derivation Yes Other revenue sharing 28 1,437 n.a. Mostly current Unconditional Derivation No de Aportaciones para la 33 86,481 n.a. Current (teacher Conditional Equal No** Educaci6n Basica y payrolls in basic Normal (FAEB) education) de Aportaciones para 33 14,466 n.a. Current (doctors and Conditional Equal No.** Servicios de Salud nurses payrolls) (FASS) de Aportaciones para la 33 12,245 2.037 Capital investment in Conditional Compensatory Yes Infraestructura Social the municipalities Municipal (FAISM) (public works) de Aportaciones para la 33 1.689 0.281 Capital investment, Conditional Compensatory Yes Infraestructura Social states can assign Estatal (FAISE) according to own criteria to municipalities de Aportaciones para el 33 13,098 2.5 Current, although Unconditional Equal Yes Fortalecimiento de los recommendation to muncipios (FAFM) spend on specific items: public debt and safety de Aportaciones 33 4,537 0.814 Current and capital Conditional Equal NO** Multiples (FAM) (school breakfasts and school construction) de Aportaciones para la 33/11* 1,251 n.a. Current (technical Conditional Equal No- Educaci6n Tecnoldgica institutes) de Adultos (FAETA) de Aportaciones para la 33/11* 1,062 n.a. Current (adult literacy Conditional Equal No' Educaci6n de Adultos program) (FAEA) de Aportaciones para la 33 4,700 n.a. Current, though some Conditional Derivation Yes Seguridad Publica de los capital equipment Estados v del Distrito Federal (FASP) (table continued oni next page) 210 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 5.1 continued Percent Type of Nature of Fund Ramo Amount RFP expenditure Transfer allocation Formula de Aportaciones para los 25 10,767 n.a. Current (teacher Conditional n.a. No servicios de educaci6n payroll) basica en el Distrito Federal Previsiones salariales 25 10,935 n.a. Current (teacher pay Conditional Equal No* por FAEB raises) Previsiones salariales 12 1,696 n.a. Current (doctor and Conditional Equal No*' por FASSA nurse raises) Previsiones salariales 23 1,966 n.a. Current Conditional n.a. No Desastres Naturales 23 4,000 n.a Current Conditional n.a. No * These funds come from Ramo 11 "Health" in the budget. Item 33 reports an appropriation of 0 in the fund, but article 17 establishes where the money is to come from. "The Ley de Coordinaci6n Fiscal does provide some vague guidelines, but they are so general that they do not become a binding formula. *** A formula exists, but only for a compensating part of the fund, which constitutes 1 percent of it. n.a. = not applicable. Source: Based on CIDAC (1998), using data from the Presupuesto de Egresos de la Federacidn para 1999 and the Ley de Coordinaci6n Fiscal and its reforms for 1999. Figure 5.2 presents per capita data for the total revenue shares received by each state (including municipal funds). The graph ranks states according to their degree of poverty-the Foster-Greer-Thorbecke index with a pover- ty line set at two minimum wages-so that the leftmost state is the poor- est and the rightmost is the richest. Returning to figure 5.2, the per capita totals across states are the result of the interplay of various components of transfers, although, as noted, the participation fund plays a dominant role. Some of the other funds, such as the municipal promotion transfer, are extremely unequal across states. What tilts figure 5.2 in the general direction of richer states receiving more revenues is the influence of tenencia and ISAN, both of which are fully trans- ferred to states on a derivation principle. Finally, the two prominent out- liers, Tabasco and Campeche, benefit substantially from the "inertial" com- ponent of the allocation formula (which in turn reflects their historically high, oil-related revenues, which were grandfathered into the calculations). Overall, then, figure 5.2 reveals a slight positive correlation-on aver- age, richer states receive larger per capita revenue sharing than poorer states. This is what one would expect, given the emphasis on derivation in the various transfer components. Conditional Transfers: Ramo 33 (Aportaciones) Conditional transfers have existed in Mexico at least since the creation of Ramo 26 in the late 1970s, which was designed to provide matching grants TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 211 Figure 5.2. Total Per Capita Revenue Shares by State, 1999 Per capita pesos 4,000 3,500 IPCoeff. Var. 3 0.3288 3,000 2,500 2,000 1,500 1,000 500 0~ ~~~C N ~ 'O 0......... ....... ; U U i s . > U ' ' U U - St U C States Source: Authors' calculations from INEGI data. for states and municipalities to finance public works. The set-up of condi- tional transfers at that time reflected mainly the priorities in federal invest- ment projects and not any set of rules for transfers to subnational govern- ments. Funds in Ramo 26 became extremely important during the Salinas administration (1988-94), as the cornerstone of the poverty-alleviation strategy of the Solidaridad (Solidarity) Program. But even within this pro- grammatic structure, solidarity funds remained highly discretionary and were often criticized for being managed electorally. (For the most promi- nent critiques of Solidarity, see Dresser 1991; Weldon and Molinar 1994; Bailey 1994; and Bruhn 1996; for some debates with these critics see Mogoll6n 1999; and Pineda and Gomez 1999.) These funds were trans- formed during the next administration into the so-called Municipal Social Infrastructure Fund, with their allocation determined by a complex poverty formula. This transfer was later incorporated into one of the funds in Ramo 33. Most of the Ramo 33 transfers relate to expenditures originally under- taken by the federal government that were converted to conditional trans- fers. The most important of these new transfers does not relate directly to 212 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO poverty alleviation and social infrastructure, but rather to the payroll of teachers, which was decentralized to the states through an agreement reached in 1993 (Latapi and Ulloa 1998; Merino 1998). Those funds were originally placed in Ramo 25 but were later grouped together with other transfers and placed in Ramo 33, which was created at the end of 1997 for the 1998 federal budget. The complex transfer of existing expenditures into Ramo 33 is captured in figure 5.3, based on a schema developed by Claudia Marcias Angeles and Jorge Rafael Manzano and reproduced from Guerrero (1998). Figure 5.3. Where Ramo 33 Came From Ramo 33 (1999) 1997 Federal Budget (MEc |Ramo 25. Basic Education lTransfers l ~~~~FASS |Ramo 12. Health L rServices Transfers u q ~~~~~~~~~FAIS Ramo 26. Povertvll Alleviation g on A X" "" ', 'j, i C t E . States Source: Authors' calculations from INEGI data. 214 FISCAL DECENITRALIZATION: LESSONS FROM MEXICO bined federal and state spending appears in figure 5.5. The lower (and light- shaded) area is the FAEB transfer, with the additional state expenditures displayed in the upper (dark-shaded) area. We made some key assump- tions in generating these data on state teacher salaries or basic education spending. First, current expenditures on teacher salaries by state are only available for 1996 from the Ministry of Finance and Public Credit (SCHP), so those figures were extrapolated with inflation rates of 15.72 percent for 1997, 18.61 percent for 1998, and 13 percent for 1999. Second, data were unavailable for four states (Baja California Sur, Guanajuato, Tabasco, and Tamaulipas) and the Federal District, so this exercise might underestimate overall spending in these states, because it assumes that they spend noth- ing of their own revenue on teachers. These state expenditures of their own funds on basic education are very unequal, since some states make substantial contributions to education, such as Baja California ($Mex669 per capita), Nuevo Le6n ($Mex529 pesos), and Coahuila ($Mex403 pesos), while at least six states (Aguascalientes, Figure 5.5. Education Transfers Complement State Expenditure Pesos per capita 2500 2000 1900 1000 0 Y~~~~~~ CEC J) Z C U u H Federal education transfers * State spending of own resources on education States ordered by rank in poverty index Source: Authors' calcualtions from INEGI data. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 215 Colima, Moreles, Oaxaca, Quintana Roo, and Yucatan) spend nothing. The coefficient of variation of state expenditure for teacher salaries is 1.0827 for all states, including those not reporting any expenditures, and 0.6638 for states that do report some expenditure. When one adds both federal and state expenditures (as is done in figure 5.5), the variation across states is reduced. Excluding the outliers (Baja California Sur on the top and Guanajuato on the bottom, both of which may be underestimated because we have no data for state teacher salaries), the minimum total expenditure on basic education is found in Yucatan with $Mex999 per capita, and the maximum is found in Baja California with $Mexl,739 per capita, producing a coefficient of variation of 0.1790, lower than the 0.2394 coefficient of variation for the federal FAEB transfers and well below the coefficient of variation for state spending on teacher salaries. This is an intriguing result. While the overall impact is to generate more equality across states in basic education spending, the FAEB transfer is very problematical since the allocation formula effectively discriminates against those states that have opted to use their own revenues to finance aspects of basic education. Chapter 3 recommends basing the allocation for FAEB on criteria that are more equitable across states, for example, per student capitation. To do this, however, the overall transfer system itself must also become rationalized. Prior to an overview of all transfers and own-source revenues, one other transfer program merits discussion-federal public investment. Federal Public Investment During the late 1980s and the early 1990s, about one-third of the federal gov- ernment's consolidated programmable budget (that is, the total budget of the federal government, excluding participaciones and interest payments on the federal debt) was allocated territorially to the states as federal public investment (Inversi6n Publica Federal, IPF). These funds were earmarked for specific projects considered important by the federal government, aiming to fulfill a developmentalist vision of the role of the state. Resources from IPF were two times higher than federal revenue sharing transferred to local governments (participaciones). Considering such relative magnitude, it is no wonder that IPF was a crucial financial flow to the states and has played a major role in debates on regional development in Mexico. The importance of IPF in Mexico has declined dramatically in recent years, how- ever, to the point that in 1997 it was only 11 percent of the programmable budget, equal to only 32 percent of revenue sharing, precisely at a time when other transfers to states were steadily increasing. Federal funds used to finance almost all local public goods in Mexico, although this is no longer the case. The money was allocated to subnational jurisdictions through a wide variety of federal agencies, programs, and 216 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO bureaucracies, with the collaboration of local governments. Federal spend- ing in Ramo 26-the poverty-alleviation program (PRONASOL)-com- manded much attention in recent years not just because a lot of money was involved, but also because a whole bureaucratic apparatus was put togeth- er to make the program an instrument of the federal executive; it was just one of the parts of IPF. Funds were sometimes directly exercised by the local governments, but under very strict federal guidelines and oversight. That is, in Mexico federal funds are controlled by the federal bureaucracies, although they are spent, especially since the 1970s, jointly with the lower levels of government, according to development agreements. Table 5.2 summarizes the main characteristics of the shares of IPF received by each state during 1960-93-the mean, standard deviation, coefficient of variation, and maximum and minimum. The most prominent feature of IPF is the high degree of discretion with which the federal gov- ernment could increase or decrease the share of a specific state. In a way, these are really regional development grants by the federal government and so should be left out of the analysis of the transfer system. This completes our description and brief assessment of the status quo with respect to the Mexican transfer system. The following section inte- grates state and municipal own revenues into the analysis, assessing the overall vertical and horizontal imbalances in the Mexican federation. Vertical and Horizontal Balance To provide a comprehensive overview of Mexican fiscal federalism, we need to bring state and municipal own-source revenues into the picture. This is the role of the first subsection. In the second, we present the aggre- gate revenue sources of state and municipal governments. This allows us to assess the degree of vertical fiscal balance in the federation. Moreover, by rearranging the transfers and revenues into their conditional and uncon- ditional components, we obtain a measure of subnational fiscal autonomy. The final subsection focuses on aggregate state and municipal revenues and the degree of horizontal fiscal balance that exists in the Mexican federation. State and Municipal Own-Source Revenues Mexican states have little access to dedicated sources of revenue. No source is exclusively assigned to them constitutionally, and for those where they have concurrent jurisdiction, they have typically chosen to delegate author- ity to the federal government. The taxes and revenues that the states do col- lect fall into the following areas, as elaborated in Chapter 4: payroll tax, var- ious fees, tenencia, and ISAN. The inclusion of tenencia and ISAN as own-source revenues of states merits further comment. Although, admin- istratively, they fall into participaciones (and were included in the earlier dis- TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 217 Table 5.2. Summary Statistics of State Shares of IPF, 1960-93 Standard Coefficient Maximnum/ State Mean deviation of variation Maximum Minimum minimum Tlaxcala 0.42 0.06 0.01 0.49 0.34 1.44 San Luis Potosi 1.35 0.13 0.01 1.48 1.15 1.29 Morelos 0.73 0.12 0.02 0.90 0.57 1.58 Zacatecas 0.73 0.12 0.02 0.87 0.58 1.50 Oaxaca 2.41 0.20 0.02 2.62 2.07 1.27 Jalisco 2.69 0.29 0.03 3.11 2.32 1.34 Aguascalientes 0.57 0.15 0.04 0.79 0.39 2.03 Durango 1.19 0.22 0.04 1.57 0.99 1.59 Baja California Sur 0.84 0.21 0.05 0.98 0.43 2.28 Quintana Roo 0.73 0.21 0.06 0.98 0.42 2.33 Queretaro 0.88 0.22 0.06 1.29 0.66 1.95 Colima 0.89 0.22 0.06 1.21 0.61 1.98 Mexico 3.95 0.49 0.06 4.60 3.33 1.38 Yucatdn 1.10 0.28 0.07 1.56 0.80 1.95 Guerrero 2.36 0.48 0.10 3.11 1.84 1.69 Nuevo Le6n 2.62 0.58 0.13 3.49 1.72 2.03 Baja California 2.28 0.58 0.15 3.26 1.50 2.17 Sonora 2.41 0.67 0.18 3.46 1.52 2.28 Coahuila 3.42 0.83 0.20 4.57 2.23 2.05 Nayarit 0.82 0.44 0.24 1.66 0.45 3.69 Puebla 1.90 0.74 0.29 3.36 1.32 2.55 Hidalgo 2.07 0.79 0.30 3.39 1.15 2.95 Guanajuato 2.25 0.82 0.30 3.60 1.46 2.47 Chihuahua 2.63 1.04 0.41 4.10 1.75 2.34 Veracruz 10.37 2.23 0.48 14.01 8.05 1.74 Chiapas 2.85 1.19 0.49 4.82 1.85 2.61 Sinaloa 3.11 1.27 0.52 5.20 2.03 2.56 Tabasco 4.78 1.72 0.62 7.77 2.94 2.64 Michoacan 3.77 1.69 0.76 6.87 1.95 3.52 Tamaulipas 5.11 2.25 0.99 8.16 2.20 3.71 Distrito Federal 26.34 6.35 1.53 37.27 21.18 1.76 Campeche 2.44 2.26 2.09 6.15 0.60 10.25 Source: Diaz-Cayeros 1997, table 6.2. cussion of revenue sharing), they are fully transferred to the states on a derivation principle. Hence, they are, for all intents and purposes, state own- source revenues. Thus, for the analysis here, they are included in state own- source revenues and are deducted from revenue-sharing transfers. Under the constitution, municipalities have exclusive authority (under Article 115) to provide drinking water and sewage, public lighting, mar- kets, graveyards, slaughterhouses, streets, parks and gai dens, and public 218 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO safety. To fulfill these obligations, they are assigned control of the proper- ty tax. The municipalities also obtain revenues from various fees on the pro- vision of goods and services (including water user fees in some munici- palities). Obtaining data for own-source state and municipal revenues in 1999 pesos (in order to make these comparable with the transfer data) requires several critical assumptions and extrapolations. First, the latest state and municipal revenue data (except for tenencia and ISAN) are for 1996. Lacking more recent data, we adjust these 1996 revenues for inflation to convert them to 1999 estimates. This is clearly unsatisfactory, but it is a further exam- ple of the data constraints facing researchers in this area. With this caveat in mind, figures 5.6 and 5.7 present aggregate per capi- ta own-source revenues for states and municipalities, respectively. The first observation is that we are talking about quite small amounts of money. Second, own-source state revenues tend to be higher in richer than in poor- er states. This trend is even more apparent for municipal own-source rev- enues, although these revenues are even smaller. Third, the spikes in fig- ure 5.5 for Quintana Roo and Baja California Sur reflect their importance Figure 5.6. State Own Revenue Pesos per capita 3,000 2,500 El tenencia * ISAN * Own revenue Coeff. Var=0.6819 0.6430 1.0926 2,000 1,500 1,000 500 0 .r < e. °4 C E, --4 .,0. °' a.. 4 9ee mXX: C)~~~~~~~ 22CE~ 0< U~u States Source: Authors' calculations fron INEGI data. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 219 Figure 5.7. Municipal Own Revenue Pesos per capita 400 Coeff. Var. = 0.8024 350- 300 250 200 - 150- 100..Iii I 0 Id I I nI II I mI 1 I I I I I I I_ U eH XZ_ 0 o N C FE-~~~~~~~~~ N C' U States Souirce: Authors' calculations from INEGI data. as centers for tourism. The Federal District is omitted in figure 5.6 because, while usually classified as a state, it has no municipalities. Aggregate Stubnational Revenuiies and Vertical Fiscal linbalance Total state own-revenues (the penultimate line in table 5.3) are estimated to be $Mex34.4 billion, with the comparable municipal own revenues totaling $Mex9.0 billion. Table 5.3 provides a capsule overview of state and munic- ipal access to revenues. Focusing on the final three rows of the table, aggre- gate transfers account for 87.6 percent of combined state-municipal rev- enues, with own-source revenues accounting for 12.4 percent. This is one measure of the overall vertical balance in the Mexican federation. This is a narrow conception of own-source revenues, however, in that it includes only those revenues that the states and municipalities themselves collect. One can make a case (as surely the states would) that the more appropriate definition of own revenues would include the unconditional participaciones. Participaciones account for 40.0 percent of aggregate state- municipal revenues from all sources. Under this conception, own-source revenues would account for 52.4 percent of total revenues (40.0 percent plus 12.4 percent). 220 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 5.3. Total Transfers to Subnational Governments in Mexico, 1999 (Ranked in Ascending Order of Federal Conditionality) State Municipal Total State Municipal Total (percent) (percent) (percent) Total Aportaciones 124,952 25,342 150,295 43.0 41.3 42.7 FAEBI 97,248 n.a. 97,248 33.5 0.0 27.6 FASSA 14,466 n.a. 14,466 5.0 0.0 4.1 FIS 1,688 12,244 13,933 0.6 19.9 4.0 FORTAMU'N n.a. 13,097 13,097 0.0 21.3 3.7 FAIE2 2,467 n.a. 2,467 0.8 0.0 0.7 FAAS 2,069 n.a. 2,069 0.7 0.0 0.6 FASP 4,700 n.a. 4,700 1.6 0.0 1.3 FAETvEA3 2,313 n.a 2,313 0.8 0.0 0.7 Total salarv Previsions 14,597 n.a. 14,597 5.0 0.0 4.2 Education4 10,935 n.a. 10,935 3.8 0.0 3.1 Health5 1,696 n.a. 1,696 0.6 0.0 0.5 General6 1,966 n.a. 1,966 0.7 0.0 0.6 Revenue sharing 113,756 27,055 140,811 39.2 44.1 40.0 Excises 2,600 n.a. 2,600 0.9 0.0 0.7 Incentives 1,437 n.a. 1,437 0.5 0.0 0.4 General revenue Sharing (FGP) 97,090 20,711 117,801 33.4 33.7 33.5 Municipal Promo- tion (FFM) n.a. 5,558 5,558 0.0 9.1 1.6 New cars 3,254 n.a. 3,254 1.1 0.0 0.9 Car property 7,984 n.a. 7,984 2.8 0.0 2.3 Contingency Reserve 1,389 n.a. 1,389 0.5 0.0 0.4 Other7 n.a. 785 785 0.0 1.3 0.2 Total other 10,299 n.a. 10,299 3.5 0.0 2.9 Natural disasters8 3,640 n.a. 3,640 1.3 0.0 1.0 Sistema Nacional de Seguridad P6blica 2,581 n.a. 2,581 0.9 0.0 0.7 Poverty regions 4,077 n.a. 4,077 1.4 0.0 1.2 Total transfers 255,887 52,397 308,285 88.1 85.3 87.6 Own revenue9 34,432 9,015 43,448 11.9 14.7 12.4 Total SNG resources 290,320 61,413 351,733 100.0 100.0 100.0 n.a. = not applicable. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 221 This conception of the vertical fiscal balance corresponds closely to sub- national autonomy defined as the proportion of total revenues that are unconditional (that can be spent as and where the states wish). These unconditional grants would include own-source revenues and participa- ciones and a few other revenue sources (the FAFM component of Ramo 33 that is unconditional; see table 5.1). Figures 5.8 and 5.9 present the relevant data for states and municipalities, respectively. Taking into account the vari- ability of these data across states, it is evident that for both states and municipalities the absolute amount of unconditional revenues is higher for richer states, and municipalities in richer states also have a higher pro- portion of total funds that are unconditional. The outliers in terms of the state data in figure 5.8 are Tabasco and Campeche (because of the alloca- tion of participaciones) and the Federal District, for reasons noted earlier. The outliers in figure 5.9 result either from own-source municipal revenues or from states' pass-through to municipalities of roughly 20 percent of par- ticipaciones (which again makes municipalities in Tabasco and Campeche outliers). Table 5.4 attempts to capture some of the dynamics relating to decen- tralization. The table is organized so that entries at the bottom are the most Figure 5.8. State Funds Pesos per capita 6,000 0Own Revenue Unconditional * Conditional 5,000 4,O000l Coeff. Var. = 0.2783 * 3,000 1 ||| 2,000 IIw f- 1,000 Source: and authors' c a lu Ml tiu UO GI) '0~~~~CU'0.CC C :3 s. ' 4 - 'U oC~~ ~ .C'C Z z -~ -~ z ~~~~~~~~~~~~~~~~0 States Source: SHCP and authors' calculations. 222 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 5.9. Municipal Funds Pesos per capita 1,400 N Own revenue E Unconditional U Conditional 1,200 , I | ~~~~~~Coeff. Var. =0.24296l 1,000 800 600 400 0 0 C N ,-, C S tZbi~~~~~~~~ r.Utn-cN >O U o C u C C~~~~~~~~~~~ States Soltrce: SHCP and authors' calculations. centralized (transfers), with the most decentralized (tax revenues) appear- ing at the top. The data indicate that there was some "decentralization" or increase in autonomy of state revenues during 1997-99. For example, high- ly discretionary federal transfers (row 1) fell from 10 percent of total state- municipal revenues in 1997 to 5 percent in 1999. Federal spending on behalf of the states (row 2) fell from 36.6 percent in 1997 to 1.9 percent in 1999. These were transferred to conditional aportaciones (embedded in Ramo 33), which rose from 3.2 percent in 1997 to 38.2 percent in 1999. To be sure, this shows the ongoing process of decentralization and devo- lution in Mexican federalism. However, advocates of a more decentralist future would surely point out that none of this increase has been in the cat- egories of unconditional transfer or own-source revenue, which would signal a more meaningful decentralization. As noted, these shares re- mained stable. Aggregate Subnational Revenues and Horizontal Fiscal Balance Our empirical overview of the Mexican transfer system concludes by look- ing at allocations across states of aggregate state or local revenues from all TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 223 Table 5.4. Transfers and Own Revenue as Percentage of Total Subnational Resources 1997 1998 1999 1997 1998 1999 (percent) (percent) (percent) Own revenue (taxes and fees) 32,390.0 38,450.1 43,448.7 12.7 13.5 12.1 Revenue sharing by origin (piggyback/ state rate) 6,140.6 6,712.6 7,984.0 2.4 2.4 2.2 Revenue sharing by origin (federal set rate) 4,071.7 5,333.2 8,078.2 1.6 1.9 2.2 Common pool revenue sharing 84,832.5 97,529.8 124,749.1 33.4 34.1 34.7 Unconditional Aportaciones 0 6,732.1 13,097.6 0.0 2.4 3.6 Earmarked Aportaciones 8,187.0 100,384.1 137,197.6 3.2 35.1 38.2 Matching funds 92,934.6 5,342.7 6,659.3 36.6 1.9 1.9 Negotiated and extraordinary transfers 25,528.9 25,135.6 18,237.0 10.0 8.8 5.1 Total 254,085.3 285,620.2 359,451.5 100.0 100.0 100.0 Source: Ley de Egresos, various years, and authors' calculations. sources. Figure 5.10 shows a fairly equal allocation across states after tak- ing into account the reasons for the outliers on the upside. The rationales for these outliers are similar to those noted earlier. Moving from left to right in figure 5.10, Tabasco and Campeche owe their position to the "inertial" feature of the participaciones formula. Quintana Roo is high because of a strong local tourist economy (Cancuin); Colima and Baja California Sur ben- efit from being the smallest states (in population) and, therefore, favored by the educational allotment under Ramo 33. Finally, the Federal District tops the outliers on the basis of strong local finances (that is, an effective property tax system). Dismissing these outliers, the overall picture presented in figure 5.10 is one of considerable horizontal balance. Yet this is the result of the interplay of more than 20 seemingly unrelated individual transfer programs plus own-source revenues. We view this as being more than coincidental and see the data as revealing a preference in the direction of a federal social pact. Although not explicit in the Mexican Constitution, the results in figure 5.10 signal an implicit social and political commitment to pursue a meaningful degree of horizontal balance in the Mexican federation. 224 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 5.10. Total Local Funds by State (Revenues and Transfers) Pesos per capita 7,000 Coeff. Var. = 0.2391 6,000 5,000 4,000 3,000 2,000 1,000 0 _ _ c '< ~ ~ N 0 C C ...* CCCC 2'z- u E States Souirce: SHCP and authors' calculations. The Evolution of Mexican Federalism This section speculates on some alternate futures for Mexican fiscal feder- alism and, within this, the alternative futures for the system of intergov- ernmental transfers. The Mexican federation is evolving on many dimensions, led by a remarkable devolution of administrative spending authority. The transfer of financial resources for this spending devolution has taken the form of myriad conditional-grant programs falling largely under the Ramo 33 umbrella. It is possible that these recent expenditure and transfer initiatives could endure. More likely, however, they are creative transitional arrange- ments toward a more decentralized federal system. To understand what this more decentralized future might look like, we assess the contradictions and forces for change within the existing system. The Inequality Challenge Inequality across citizens and states alike presents an enormous challenge to the Mexican federation. The north-south illiteracy divide (as reflected in figure 2.3) is striking, and the North American integration pressures and TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 225 opportunities under NAFTA may exacerbate the degree of inequality. For this reason the societal preference for approximate equality revealed in the pattern across states of the total per capita revenues deserves special emphasis. The data in figure 5.10 show considerable variability across states, but these deviations are not systematically related to how "rich" the states are. Approximate horizontal balance across states appears to be an underlying value of Mexican federalism. Since the intergovernmental transfer system is, in effect, a zero-sum game, the challenge is to generate transfer programs that can withstand political pressure. Part of the solution is to ensure that these transfers are formula based, which, in turn, will tend to depoliticize them. The experi- ence of the Municipal Social Development Fund in 1997, later transformed into a component of Ramo 33, should be helpful, since it demonstrates the possibility of establishing a formula that, although still too complex, can accommodate poverty criteria while being politically palatable. A viable scenario for the future will need an overall transfer system that addresses the enduring inequality issue in fair and acceptable ways. The NAFTA Challenge Mexican states, especially the northern tier, are being progressively drawn into NAFTA's orbit. Effectively, they are becoming North American region- states, along the lines of Ontario (as suggested in Courchene and Telmer 1998), and they will behave as such. They will use their available powers (and surely press for more powers) to make their states attractive locations for Mexican, North American, and foreign investment. Foreign policy will still be made in Mexico City, but the important wheeling and dealing on detailed trade issues will, to an increasing degree, rest with the states and governors. This is happening in Canada, and it seems inevitable that it will also intensify in Mexico. This means the states will press for additional economic and fiscal pow- ers in pursuing their regional-international interface within NAFTA. Many of these states are pressing for more freedom on the tax front. Legally any state can unilaterally withdraw from the voluntary tax concordat and rein- stitute their own taxes and collection systems. This gives individual states a lot of negotiating leverage and contrasts to a situation like Argentina's, where any change in the fiscal transfer and tax system requires unanimous approval of all states and the federal government. Greater tax autonomy of these regions appears to be inevitable. Given the inequality in tax bases across states (and particularly the north-south dimension), this could create fiscal havoc within the fiscal-federalism framework if it is not done carefully. The challenge is twofold: first, to accommodate the growing desires of selected states to acquire greater tax autonomy and, second, to design the 226 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO overall transfer system to satisfy both the desire for more tax autonomy on the part of higher-revenue states and the equality imperative. As produc- tivity and wages rise in these NAFTA states, they will make the case that their cost of public services is rising as well. (Canadian evidence suggests that roughly 60 percent of all provincial expenditures on public services relate to wages.) The southem states can and will make a convincing case that their expenditure needs are greater. Balancing these legitimate concerns will be priorities in the model outlined. Alternatives for the Future First, accountability and transparency at the state level are improving and will continue to improve. There will be tensions nonetheless, particularly regarding the "two-stage" procedure characterizing municipal transfers. Funds are first given to the state and then are allocated by the state to the municipalities, sometimes according to federal formulas, sometimes accord- ing to the state's own criteria. To be sure, accountability and transparency will vary across states, but over the longer term the states will have the ana- lytical and managerial capacity to accommodate increased decentralization. Second, the system of intergovernmental transfers should be formula based and should be set in a multiyear framework. Enhancing state-level analytical and managerial capacity will not mean much if they are stuck in the context of highly variable and unpredictable transfer regimes. Third, as the system decentralizes, it will need some federal-state insti- tutional machinery to secure the benefits of decentralization. This institu- tional evolution could take the form of a formal organization (along the lines of Australia's Commonwealth Grants Commission) or a more informal vari- ant (along the lines of Canada's federal-provincial "executive federalism" arrangements). In any event, the following are among the likely roles for enhanced federal-state cooperation and coordination: CONDITIONALITY ISSUES. Many of the transfers will probably have condi- tionality attached to them. This will be appropriate in the transition and to some lesser extent in any steady state. The issue is who sets these condi- tions. Initially, the answer probably has to be the federal government. Over time, however, it will become increasingly important to bring the states more formally into the determination of these conditions or principles. INTERNAL TRADE ISSUES. As the federation decentralizes, state actions may fragment the internal Mexican economic union, either intentionally or inadvertently. Hence, a critical ingredient of any decentralization process is to preserve and, indeed, to promote the economic union. This harks back to McKinnon's principles of market-preserving federalism. TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 227 DISPUTE RESOLUTION. Initially, perhaps the federal government will serve as umpire. As decentralization proceeds, however, pressures will develop for a federal-state umpire-eventually a system of federal-state dispute-res- olution mechanisms, complete with voting rules and sanctions. CITIZENS AND DEMOCRACY. Too often fiscal federalism arrangements involve only governments. But federalism is presumably a system that also ought to benefit citizens. One way to accomplish this would be with a nonjusti- ciable citizens charter as an integral component of fiscal federalism. This might include, as a minimum, provisions granting citizens the rights for equality of access to some national public goods, such as education and health. Or it could be more proactive, allowing citizens to access dispute resolution procedures to redress perceived mistreatment under the new fis- cal regime. To some extent, this may exist under the Mexican law of amparo, but the citizens charter would spell things out more specifically for feder- alism issues. A Decenitralist View of Mexican Fiscal Federalism With the previous section as backdrop, we now attempt to spell out the defining features of a longer-term model that, in our view, is consistent with the existing values and norms and the emerging pressures being exerted on the Mexican federation. We give only the key features of the model, although we are specific about the forces driving the model. The basic approach is general and amenable to alteration. The model has six general principles: tax decentralization, fiscal equity, conditional citizenship transfers, a federal-state fiscal covenant, a federal- state commission, and transitional guarantees. We deal with each in turn. TAX DECENTRALIZATION. The first principle is to devolve more taxation authority to the states. Chapter 4 assesses the likely candidates for state tax- ation, including personal income taxation (perhaps an integration of pay- roll and personal income taxation), commodity sales taxation, and state sur- charges on the value added tax. State access to some of these tax sources would require a constitutional amendment. In other areas, the federal gov- ernment need not formally devolve authority. Rather, it might provide tax room for the states to apply surcharges on federal taxes, with collection still under federal (or national) supervision. Our goal here is not to recommend constitutional changes, but rather to note that selected states will progres- sively become more active players in the taxation area and that, somehow, the federation must accommodate this. The problem arising from this recommendation, if implemented in iso- lation, is that the well-endowed states will have access to more per capita revenues than the poorer states. Thus accommodating one pressure in the 228 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO federation would create another. How do we address this emerging chal- lenge to horizontal balance? EQUALIZATION AND FISCAL EQUITY. The second principle or component of the model is that the federation would mount an equalization program to offset the horizontal imbalances arising from increased state taxation autonomy. The specific proposal follows Canadian lines. The federal government would provide unconditional transfers to ensure that all states are brought up to the agreed-upon equalization standard. For presentation purposes, this equalization standard is defined as what an average state's per capita revenues would be if the all-state tax rates were applied to the commonly defined state tax bases (see Courchene 1998 for more detail). This implies that all low-fiscal-capacity states will receive equalization payments to ensure that their own-source revenues are brought up to the national aver- age of all states. Three implications follow. First, these equalization payments must be unconditional, since the tax revenues of the richer provinces are likewise unconditional. Second, the existence of an equalization program will (or should) mean that all states are in favor of tax decentralization-richer states because they can pocket their per capita revenues in excess of the nation- al average and poorer states because they receive unconditional equaliza- tion payments that bring their per capita revenues up to this national stan- dard. Third, this can be a revenue-neutral exercise from the federal government's standpoint. The enhanced taxation autonomy of the states as well as the equalization transfers can come out of reductions in the existing participaciones. Although this low-fiscal-capacity compensation is termed an equaliza- tion payment, in reality it is more like a social transfer, since it allows low- tax-capacity states to be put on some socially determined equal footing with high-tax-capacity states. Moreover, this equalization standard is, in prin- ciple, quite flexible-for instance, it could be defined in terms of the top 10 states rather than the all-state average. The key point is that much of the federal "cohesion" issue for inequality can be "solved" within the context of tax decentralization. THE CONDITIONAL CITIZENSHIP TRANSFERS. With much of the social cohesion of the federation addressed in the context of the tax-autonomy-cum-equal- ization context, the stage is now set for a new approach to conditional trans- fers, namely that they should be equal per capita across all states. Their con- ditionality is predicated on the assumption that they are designed to deliver citizen-related public goods (health, education) equally to all Mexicans. It would be possible to have this aggregate transfer defined in terms of spe- cific public goods, along the lines of the Ramo 33 funds. Although strict con- TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 229 ditionality might apply at the outset, the spirit of this proposal is eventu- ally to define conditions as a set of agreed-upon principles relating to min- imum acceptable standards for these expenditures. This would allow states some flexibility in how they most efficiently achieve these standards and principles. A FEDERAL-STATE FISCAL COVENANT. The fourth principle of the proposed model is the creation of a Mexican federal-state fiscal covenant. In turn, this covenant would include: * A code offiscal conduct, defined to ensure that states do not use their increased tax autonomy to fragment the Mexican economic union. More positively, the role of the covenant would be to preserve and promote the national fiscal union. * A social union agreement, which would ensure that Mexican citizens moving from state A to state B do not suddenly find themselves deprived of social services. This may not be an issue today, but it could emerge as the federation decentralizes. There are many ways to accommodate this, but it must be accommodated. * A citizens rights charter, which would be a political document, not a legal document. The charter would ensure that the evolution of Mexican federalism takes adequate account of related citizen con- cerns. Part of this would presumably be reflected in the principles that would attach to the conditional-citizenship grants. Part also would be reflected in a proposition that citizens have the right to move and work anywhere and everywhere in the federation (with the qualifi- cations accepted on a mutual-recognition basis, as in the recent Canadian social union agreement). If the authorities wanted to go fur- ther still, they could allow citizens to invoke dispute-resolution pro- cedures. A FEDERAL-STATE COMMISSION ON MEXICAN FISCAL FEDERALISM. The fifth building block is a federal-state monitoring and coordinating agency. The development of such an agency is required on three grounds-technical, policy, and politics. The newly created Decentralization Committee in the Ministry of Finance is, with its Technical Secretariat, an important step in this direction. On the technical front, and critically important if the Canadian federa- tion is a guide, there is much to be done. For example, running an equal- ization program is technically demanding-handling increasing amounts of data, appropriately defining tax bases, determining the timing of the pay- ment stream for equalization transfers, and the like. On the policy side, a federal-state monitoring and coordinating agency would be an ideal forum 230 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO for working out, cooperatively, the set of conditions and principles that would inform the spending of the equal per capita conditional transfer. On the political front, the challenge would be to develop agreed upon dispute- resolution and dispute-avoidance procedures relating to the many issues that will surely arise. A variety of alternative institutional structures could fill this role. TRANSITIONAL GUARANTEES. This final principle concerns a transitional sol- idarity program for the states, and perhaps for the federal government as well. There would be winners and losers in a direct transition to any model, although under reasonable assumptions the steady state could be a win- win solution for all. Appropriate transitional mechanisms are therefore needed to make the change generally desirable and politically acceptable. This scenario is our conception of one potential future for Mexican fis- cal federalism. We believe that it addresses in an appropriate fashion the various forces at play in Mexico today, but it is only one of many pathways along which the system might evolve. Given the substantial effort involved in documenting the status quo with respect to the transfer system, we now present some empirical simulations relating to assumptions underlying this model. Simulating Transfers with Tax Decentralization We present two simulations relating to the evolution of the transfer system. Consistent with the recommendations elsewhere in this book, these simu- lations seek to assess the degree to which Mexican fiscal federalism can accommodate increased tax devolution to subnational governments. We focus on the combined state-municipal level, as in figure 5.10. By construc- tion, these simulations are revenue neutral from the vantage point of the federal government. The methodology underpinning the simulations essen- tially marries the features of the status quo with characteristics of the sce- nario extrapolated in the previous section. MEDIUM-TERM SIMULATION. The medium term simulation, with parameters drawn from tax decentralization recommendations in Chapter 4, is designed as follows. First, we assume a relatively modest increase in devolution in tax assignment. Specifically, the payroll tax is adopted by all states at a 3 per- cent rate; a proportion of excises on beer, tobacco, and alcohol is transferred to state governments as own revenues; 30 percent of the (now federal) value added tax paid by individuals becomes a state value added tax; and 50 per- cent of the overall personal income tax rate is passed through to the states. These tax reassignments lead to an increase in state own-source revenues that is roughly double the status quo. The financing of this tax devolution is assumed to come out of the existing revenue-sharing grants (Ramo 28). TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 231 Second, we overlay an equalization program on these own-source rev- enues, which brings the per capita value of own-source revenues for low- fiscal-capacity states to the national average. Funds required for this equal- ization transfer also come from the Ramo 28 participaciones. Third, since the sum of these two components does not exhaust the cur- rent aggregate level of participaciones, we allow the remainder of participa- ciones to be distributed in accordance with shares in the existing formula. This implies, for instance, that the revenue-sharing "spikes" for Tabasco and Campeche still exist, but at attenuated levels. Fourth, and finally, we allocate the existing value of the conditional trans- fers (Ramo 33) to the states, but with a key difference: this transfer is now redesigned to be equal per capita across states. Figure 5.11 presents the results of the modest tax decentralization sce- nario. The lowest part of the figure relates to existing own-source revenues; the next portion is related to the additional own-source revenues transferred to the states as a result of the simulation. The combined value of these own- source revenues is then equalized. As a result of equalization, all states now have access to the all-state average of own-source revenues, that is, the value Figure 5.11. Medium-Term Simulation (Modest Tax Reassignment Proposal, Revenue-Sharing Formulas Unchanged) Pesos per capita 7,000 E own revenue * additional revenue Cl equalization 6,000 E revenue shares 3 social transfer 5,000 4,000 3,000 2,000 X X i 1,000 0 C) : G 12 . 0. - E U >~~ ~ ~ cJ U States Source: Authors' calculations. 232 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO of equalization is equal to the difference between the equalization standard (just under $1,000 per capita) and the sum of the states' own revenues. As is to be expected, equalization payments flow to the poorer provinces (the left side of the figure). Next, the remainder of the participaciones are allo- cated to the states in terms of their current shares of these transfers (hence the spikes for Tabasco and Campeche). Finally, the equal per capita con- ditional transfer is then superimposed on all states. In comparison with figure 5.10-the status quo-three features stand out. First, there would be much less variation across states. This is a result of the equalization transfer, on the one hand, and per capita equality of Ramo 33, on the other. Second, most poorer states would lose a little compared to the status quo, even though they would get a larger share of the (smaller) pie of transfers. Third, total revenues per capita would tend to be slightly high- er for larger states, compared to the current system which penalizes them. This is a revenue-neutral simulation, so there are losers as well as win- ners. These gains and losses appear in table 5.5. Compensating the states that lose would cost $Mexl9 billion per year, or 15 percent of the Ramo 28 transfer. This estimate of the cost of compensating losers is the upper bound, because the costs would be less if both winners and losers were gradually brought up or down to their equilibrium levels. And assuming that the overall revenue pie is growing in real terms and even more in nom- inal terms, all states could be guaranteed their existing levels until the new system begins to generate equivalent overall revenues. If the new system gave states greater authority and incentive to increase their own revenues, then the revenue pie would grow faster and bring all the states more quick- ly to a situation with higher revenue than before. LONGER-TERM SIMULATION. The longer-term simulation embraces much more tax decentralization. It is designed (a) to reflect the emerging reality if and when states do take on greater tax authority, and (b) to test the degree to which the general model outlined earlier can handle a tax-decentralized Mexican federation. The problem here is that we need a way of estimating what a substantial increase in derivation-based tax decentralization would look like across states. This model has only three components-a large own- source revenue component, an equalization component, and an equal per capita conditional transfer. First, we assume that the entire Ramo 28 revenue-sharing fund will be transferred into tax room for the states. Although this could take the form of a variety of state taxes, we assume that the allocation across states will follow the average tenencia and ISAN allocations, which are based on derivation. Second, as in the previous simulation, we now equalize these own- source revenues (that is, bring low-fiscal-capacity provinces up to the all- TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 233 state average of per capita own-source revenues). Since the tax devolution (under the first point) "exhausts" Ramo 28, the funds for equalization must come from Ramo 33. Third, the remaining funds in Ramo 33 will be allocated across states in equal per capita terms. These two simulations assume a very passive reaction on the part of the states, which are not assumed to increase their tax rates on their existing taxes. Moreover, they are assumed to maintain the initial assigned tax rates on any new tax bases they acquire. Were they to increase these tax rates, their revenues could be higher than reported. Any such state tax-rate- enhanced increase in revenues should not be offset by a reduction in over- all participaciones. Figure 5.12 presents the simulation results for the scenario. The first important feature of this simulation is that own revenues (assumed to be according to current allocations for tenencia and ISAN) rise sharply for rich- er states, as one would expect. Figure 5.12. Long-Term Simulation (Revenue Sharing Transformed into Derivation-Principle Own Revenue) Percent 80 60 6i h 40 20 00 U0 . . C C C ~~~'m 'O E) ' ;~ ~~7 'm i ° e mEtS :=<82° oOSa o J S 2 4 '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Z~~~~~~~~~~~~~~ aC CT 0 2u States U Enhanced revenue U Equalization grant Cl Social transfer Source: Authors' calculations. 234 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 5.5. Fiscal Cost of Compensation Compensa- Gain! Compensa- Per Gain! tion Compensa- Per loss as tion Compensa- capita loss as grant tion capita percent grant tion gain! percent of per grant gain! of status per grant loss status quo capita ($) loss quo capita ($) Chiapas -498 -14.0 498 1,954,587,633 -594 -16.8 594 2,332,560,241 Oaxaca -158 -5.2 158 545,593,836 -105 -3.4 105 362,865,537 Zacatecas -407 -11.8 407 561,655,896 -498 -14.4 498 685,916,935 Guerrero -610 -17.3 610 1,886,370,780 -581 -16.4 581 1,797,544,942 Puebla 409 16.4 n.a. n.a. 455 18.2 n.a. n.a. San Luis Potosi -161 -5.2 161 370,695,816 -157 -5.1 157 361,934,273 Hidalgo -331 -10.1 331 733,423,229 -343 -10.4 343 758,823,574 Yucatan -253 -7.7 253 414,717,312 -322 -9.8 322 527,984,604 Veracruz 135 4.8 n.a. n.a. 115 4.0 n.a. n.a. Tabasco -1,546 -27.1 1,546 2,881,316,687 -2,753 -48.3 2,753 5,130,807,449 Campeche -1,717 -31.3 1,717 1,183,654,212 -2,528 -46.1 2,528 1,742,040,044 Durango -587 -16.2 587 858,788,283 -673 -18.6 673 984,689,013 Michoacan 56 2.0 n.a. n.a. 120 4.2 n.a. n.a. Tlaxcala -451 -12.6 451 422,760,670 -627 -17.5 627 586,992,779 Guanajuato 280 10.6 n.a. n.a. 315 11.9 n.a. n.a. Queretaro -368 -10.3 368 493,780,762 -637 -17.8 637 854,571,467 Navarit -828 -20.8 828 766,590,170 -1,032 -25.9 1,032 956,083,680 Quintana Roo -670 -15.3 670 540,852,517 -495 -11.3 495 399,832,049 Tamaulipas -486 -13.7 486 1,288,460,782 -219 -6.2 219 578,997,802 Jalisco 216 7.5 n.a. n.a. 608 21.2 n.a. n.a. Mexico 421 16.5 n.a. n.a. 396 15.5 n.a. n.a. Morelos -196 -6.1 196 304,948,531 -278 -8.6 278 432,081,645 Aguascalientes -606 -16.0 606 560,052,569 -831 -22.0 831 768,158,068 Chihuahua 19 0.6 n.a. n.a. 590 18.1 n.a. n.a. Coahuila -89 -2.5 89 201,711,595 16 0.5 n.a. n.a. Sinaloa -139 -4.3 139 349,170,404 137 4.3 n.a. n.a. Nuevo Le6n 728 20.5 n.a. n.a. 181 5.1 n.a. n.a. Colima -1,174 -26.0 1,174 601,290,276 -1,113 -24.6 1,113 569,919,272 Baja California Sur -1,865 -33.0 1,865 763,443,224 -1,573 -27.9 1,573 643,678,428 Sonora -424 -11.3 424 933,104,902 702 18.6 n.a. n.a. Baja California -220 -6.0 220 512,619,529 929 25.3 n.a. n.a. DF 598 10.3 n.a. n.a. 14 0.3 n.a. n.a. Total fiscal cost of compensation 19,129,589,616 20,475,481,801 Fiscal cost as percentage of revenue sharing (Ramo 28) 15.0 16.1 n.a. = not applicable. Note: States rank by descending rate of poverty. Source: Authors' calculations. The second feature is that, as a result of the pattern of own-source rev- enue, equalization now plays a much more critical role, that is, the equal- ization payments for Oaxaca and Tlaxcala appear, from figure 5.12, to be as large as their equal per capita social transfer. As a result, the overall pat- tern of per capita revenue (not affected by the addition of the third tier, the equal per capita conditional grants) is one of equal per capita revenues for TRANSFERS AND THE NATURE OF THE MEXICAN FEDERATION 235 roughly half of the states, and then a rising level of overall per capita rev- enues for the richer states. Our rationalization for this is that the equalization payment compensates for weak fiscal capacity and to a degree for expenditure needs, while lev- els across the rest of the states address the increasing "costs" of providing public services as their wages rise, and also provide an incentive for levy- ing taxes. Moreover, it is relatively easy to defend these transfers politically. Some version of equalization exists in virtually all federal systems (except the United States) and, with equalization in place, the per capita equality of the conditional transfer would presumably be viewed as equitable. But these distributional implications could be altered. Assuming that the own-source revenue pattern stays constant, the most obvious way to increase per capita revenues of the poorer states would be to raise the equalization standard. In figure 5.12, this standard is set at the all-state per capita average. Were it set higher, say at 110 percent of the average, the per capita value of the conditional grant would be reduced, and all of this would be transferred to the poorer states via equalization. Beyond this, one could introduce some additional "expenditure needs" components to modify the allocations of the conditional grant. And so on. The obvious problem with the results presented in figure 5.12 is that some states lose a lot-especially Tabasco and Campeche-since the old par- ticipaciones revenue-sharing formula is gone and has been replaced by derivation-based tax decentralization. (But this result will arise under any scenario that contemplates enhanced subnational taxing authority but excludes the oil sector from the base). The costs of full compensation to the losers (table 5.5) under the simulation in figure 5.12 are estimated at 16 per- cent of the existing Ramo 28 transfers. SUMMARY. These simulations illustrate the effects of introducing a combi- nation of reforms-rationalizing the allocation of Ramo 33, introducing the notion of an equalizing transfer, and having greater subnational taxing autonomy. The debate about transfers needs to take the form of compar- ing alternate simulations of new systems, rather than complaints about indi- vidual programs or threats to disrupt the pacto fiscal or the political process if special grants of some sort are not forthcoming. Given Mexico's more than 20 separate transfer programs, simulation is an ideal way to experi- ment with alternative transfer mechanisms and, by implication, with alter- native futures for Mexican fiscal federalism. Our intention is not to defend these particular simulations. Rather, the goal is to drive home the point that in the not-too-distant future the Mexican federation will likely become more decentralized on the tax side and more challenged by cross-state inequality. These simulations attempt to design and test a transfer system that accommodates this reality. It is hoped that 236 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO this will encourage others to simulate different and novel transfer mecha- nisms that accomplish similar objectives. If experience in federal systems elsewhere is a guide, it is not enough to develop a system of transfers that, at any point in time, delivers these objectives-the transfer systems them- selves must resonate with citizens and states alike and be able to adjust to changing circumstances. We believe our approach to transfers will pass this political test. As noted earlier the proposed model and the simulations represent only a fraction of the possibilities, but the more determining factor is that any significant increase in the states' tax autonomy will lead to a wide diver- gence in per capita own revenues. Were this to become reality, it might be politically very difficult to redesign the 20-odd existing transfers in ways that could accommodate this own-revenue increase and at the same time deliver an appropriate degree of horizontal balance. Therefore an equal- ization program may become a valuable instrument in the fiscal federal- ism arsenal. In any event, the evolution of the Mexican federation and Mexican fis- cal federalism will be determined not by the dictates of any given model but, rather, by the political, social, and economic values and norms of Mexican society. Viewed from this perspective, our study has aimed to give a framework for the variety of ways that creative designs for intergovern- mental transfers can achieve the society's objectives. 6 Subnational Borrowing and Debt Management Marcelo Giugale, Fausto Hernandez Trillo, and Joao C. Oliveira As a result of political opening and administrative decentralization, sub- national governments in Mexico are demanding fiscal autonomy in order to improve their access to financial markets. But because the move toward fiscal decentralization is so recent, the allocation of fiscal responsibilities and borrowing autonomy among levels of government (federal, state, and municipal) is still being developed. State governments in Mexico have as yet limited flexibility in fiscal decisions, which limits their capacity to bor- row for public investments. After several past attempts at reform, a new innovative regulatory system for subnational borrowing is being imple- mented which combines market discipline with rules. Subnational Debt The growth of subnational spending for public services and capital invest- ments (in health, education, and basic infrastructure) does not yet pose a major threat to Mexico's macroeconomic stability. This distinguishes Mexico's experience from that of other large Latin American countries, notably Argentina and Brazil (Dillinger and Webb 1999; Burki, Perry, and Dillinger 1999). Nevertheless, Mexico's subnational government debt grew from $Mex27 million in 1994 to $Mex71.6 million in 1998 (see table 6.1 which includes debt of both direct and indirect administrations, but not contingent debt). The 1994-95 financial crisis, and the ensuing increase in interest rates, expand- ed the states' debt stock in real as well as nominal terms, but the bailout package put in place by the federal government in 1996-97 reduced it con- siderably (figure 6.1). In 1997 subnational government debt represented 25 percent of the debt owed or guaranteed by the Ministry of Finance and 237 238 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Public Credit (Secretaria de Hacienda y Credito Publico, SHCP), 10 percent of total public debt (including Banco de Mexico debt), and only about 2 per- cent of national GDP. This compares favorably with Argentina, where subnational debt is 6 or 7 percent of GDP, and with Brazil, where it approaches 20 percent (Dillinger and Webb 1999). Table 6.1. Mexico: Total Debt, 1994-98 (Millions of Pesos) State 1994 1995 1996 1997 1998 Mexico 4,843 8,643 13,396 16,609 185,74 Nuevo Leon 2,348 6,427 5,463 6,706 7,470 Jalisco 2,811 3,371 3,876 4,006 4,418 Sonora 3,150 4,869 6,085 3,672 3,990 Sinaloa 873 1,337 1,677 1,931 2,212 Chihuahua 921 1,215 1,538 1,689 1,593 Baja Cal. Norte 999 960 1,214 1,380 1,528 Guerrero 515 858 983 1,168 1,255 Queretaro 1,282 1,090 1,016 1,061 1,163 Quintana Roo 450 643 740 842 1,009 Chiapas 1,024 992 1,088 961 931 Durango 552 462 606 713 806 San Luis Potosi 345 426 543 599 708 Coahuila 515 926 1,116 593 666 Tabasco 518 343 411 431 598 Guanajuato 405 411 464 517 569 Puebla 156 321 308 351 478 Baja Cal. Sur 304 296 350 450 452 Morelos 144 232 244 365 399 Yucatan 305 288 320 372 290 Tamaulipas 368 531 363 315 279 Oaxaca 260 147 192 202 261 Michoacan 249 256 251 216 251 Campeche 499 460 518 419 228 Aguascalientes 364 307 339 287 227 Colima 191 263 291 237 192 Zacatecas 123 380 468 235 136 Navarit 222 187 178 115 104 Veracruz 348 379 262 78 52 Hidalgo 22 14 16 12 10 Tlaxcala 136 52 0 0 0 Subtotal 25,255 37,099 44,329 46,545 50,864 Fed. District 1,703 2,772 8,322 11,958 20,763 Total 26,958 39,872 52,652 58,503 71,627 Source: SHCP data. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 239 Figure 6.1. Mexico: Subnational Governments' Total Outstanding Debt (1996 Prices) Millions of pesos 54,000 - 53,000 i 52,000 / 51,000 l / 50,000 49,000 48,000 - 47,000 - 46,000 45,000 I l 1994 1995 1996 1997 1998 Years Source: SHCP and authors' calculations. Although subnational government debt continues to benefit from the effects of the 1995-96 bailout package, it recently began to increase con- siderably. Subnational debt increased 5.6 percent in real terms in 1998, main- ly as a result of primary fiscal deficits of the Federal District (development and contractual debt with the commercial banks for the metro), the State of Mexico (development and contractual debt with commercial banks for the Atlacomulco Highway), the State of Guanajuato (bonds issued to pay for toll roads), and municipalities of Merida (bonds) and Garza Garcia (com- mercial and contractual debt with commercial banks for the construction of a tunnel linking this city to Monterrey). Although it is expanding, subnational indebtedness does not yet threat- en Mexico's macroeconomic management, because its share in the portfo- lio of the financial system is still relatively small. Two factors explain the relatively small size of aggregate subnational debt. First, subnational gov- ernments' access to capital markets is limited by their lack of borrowing capacity.1 Second, the frequent implicit and explicit bailouts by the feder- al government softened subnational governments' budget constraints before their fiscal shortfall became a debt; that is, the federal government absorbed their potential debts. The first factor, a consequence of the archi- tecture of Mexican fiscal federalism, indicates that reforms are needed to 240 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO accord subnational governments adequate access to capital markets. The second factor was a consequence of ad hoc interventions by the federal gov- ernment through ex post, extraordinary, and discretionary transfers. This second factor, in general politically motivated, indicated that the inter- governmental relationship in Mexico still embodied many channels that lead to moral hazard incentives. The federal government has acted in 1999- 2000 to correct both factors by establishing a new market-based regulato- ry framework for subnational borrowing. This framework is explained in detail later on. The subnational debt in Mexico is concentrated in a few states (see fig- ure 6.2). During 1994-98, out of the nation's 32 states, six (Federal District, Jalisco, State of Mexico, Nuevo Le6n, Sinaloa, and Sonora) were responsi- ble for four-fifths of the subnational debt outstanding (table 6.2). Among Figure 6.2. States' Real Debt, 1994-98 At 1996 prices (pesos) 25,000 20,000 15,000 10,000 5'000 A - , 0- 1994 1995 1996 1997 1998 Years - Jalisco - - -k- - Nuevo Leon --( Sonora - Mexico X Sinaloa - Fed. District OE Others Source: Authors' calculations fron SHCP data. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 241 these six entities, the Federal District, State of Mexico, and Nuevo Le6n con- tinued to expand their debt after the federal debt relief operation; the oth- ers, as a group, reduced their real stock of debt and their share of total debt. The ratio of debt to revenue reflects an important aspect of financial vul- nerability, but its relevance depends on the degree of fiscal autonomy of the states.2 Therefore, the concept of revenue, for this particular purpose, should exclude all received transfers tied to nondebt expenditures. Table 6.2 shows the states according to three types of debt ratios: debt to total rev- enue, debt to disposable revenue, and debt to own revenue. Great disper- sion is observed in the degree of indebtedness among states in any of these concepts, more so in the case of own revenue. Debt to disposable revenue, which includes only own revenue and unconditional transfers (Ramo 28), indicates that the degree of indebtedness varies from a maximum of 4.1 (Sonora) to a minimum of 0.02 (Hidalgo). The Federal District is ranked among the least indebted states, because of its relatively large capacity to collect own revenue. The eight most indebted states are Sonora, Nuevo Le6n, State of Mexico, Queretaro, Quintana Roo, Baja California Sur, Jalisco, and Sinaloa, all with debt ratios greater than 1. The solvency and capacity to pay for debt in the medium term require analysis for the first three states, in particular. Excessive indebtedness may have equity implications. In Mexico there is a clear, positive correlation between state indebtedness and state per capi- ta GDP. Therefore, debt bailouts are likely to be regressive, since financial relief tends to go to richer-than-average states. Fiscal Imbalances and Debt Accumulation The stock of debt and the degree of indebtedness alone do not reveal the financial weakness of the Mexican states. In fact, the relatively "small" size of the outstanding debt of subnational governments in Mexico does not cor- respond to the capitalization of their past "large" fiscal deficits. The reason is that a substantial portion of their fiscal deficits has been repeatedly relieved by the federal government through extraordinary, discretionary transfers (to cover unanticipated wage increases, investment expansion, and so forth) and other forms of bailouts (the 1995 ad hoc transfers for debt reduction and rescheduling). Since 1989 the aggregate fiscal deficit of states has always been larger than the sum of increases in indebtedness and changes in liquid assets of the states. The systematic federal government interventions to fill the finan- cial gaps reveals the soft side of states' budget constraints. Figure 6.3 shows the evolution of the states' primary surplus or deficit and its financing. The states' fiscal stance deteriorated precipitously until 1993 (when the aggregate primary deficit reached 0.4 percent of national GDP). Since 1994 the situa- tion has improved, and the statistics even show a primary surplus in 1995. 242 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Table 6.2. Mexico: Total Debt, 1994-97 (Debt/Revenue Ratio in Descending Order of Total Debt) Total revenue Disposable revenue State-own revenue 1994 1997 1994 1997 1994 1997 Mexico 1.0 0.9 1.0 1.4 5.9 9.6 Nuevo Le6n 0.8 0.8 1.0 1.2 2.9 4.8 Sonora 1.1 0.6 1.6 1.1 4.4 6.6 Baja California Sur 0.5 0.4 1.2 0.8 16.7 12.4 Queretaro 0.9 0.3 1.6 0.7 6.9 4.9 Quintana Roo 0.6 0.4 1.1 0.7 5.2 2.9 Sinaloa 0.4 0.4 0.7 0.7 5.5 4.5 Aguascalientes 0.4 0.2 0.7 0.6 3.8 3.2 Jalisco 0.8 0.3 0.8 0.6 3.9 3.6 Baja California Norte 0.6 0.4 0.7 0.5 3.1 2.3 Chihuahua 0.3 0.3 0.6 0.5 2.0 1.6 Durango 0.8 0.2 0.9 0.5 8.7 5.0 Guerrero 0.2 0.5 0.4 0.5 2.9 3.4 Campeche 0.5 0.2 0.9 0.3 8.1 1.6 Colima 0.3 0.2 0.6 0.3 5.6 4.2 San Luis Potosi 0.3 0.3 0.4 0.3 5.2 6.5 Chiapas 0.2 0.1 0.6 0.2 3.6 0.9 Coahuila 0.2 0.1 0.4 0.2 2.1 1.4 Morelos 0.1 0.1 0.2 0.2 0.6 1.2 Yucatan 0.3 0.2 0.4 0.2 2.8 1.5 Zacatecas 0.2 0.1 0.2 0.2 2.0 2.7 Guanajuato 0.1 0.1 0.2 0.1 2.0 0.6 Michoacan 0.1 0.0 0.2 0.1 1.8 0.6 Nayarit 0.2 0.0 0.5 0.1 2.9 0.8 Oaxaca 0.1 0.0 0.3 0.1 5.7 1.1 Puebla 0.0 0.1 0.1 0.1 0.5 1.0 Tamaulipas n.a. 0.1 n.a. 0.1 n.a. 0.7 Hidalgo 0.0 0.0 0.0 0.0 0.3 0.1 Tlaxcala 0.2 0.0 0.3 0.0 2.9 0.0 Veracruz 0.1 0.0 0.1 0.0 1.3 0.2 State average 0.4 0.3 0.7 0.6 3.7 3.4 District federal 0.1 0.4 0.1 0.4 0.2 0.6 Overall average 0.4 0.3 0.5 0.5 1.6 1.8 Source: SHCP and authors' calculations. Data for Tabasco are not available. However, a closer look into the data reveal that during 1995-97 the states as a group were not generating a primary surplus before the extraordinary transfers, which should properly be counted as a financing item (below the line) not a revenue component; and that the primary deficits continued dete- riorating even after 1995, because the debt restructuring did not lead, in SUBNATIONAL BORROWING AND DEBT MANAGEMENT 243 most cases, to any effective fiscal adjustment in the states' budgetflows. The financial deal involved basically a debt stock relief, and it did not resolve the structural fiscal imbalances. As a consequence, the fiscal stance of some Mexican states was not sustainable at least up to 1997, and they need seri- ous fiscal adjustment. The distance between real primary balance and real primary balance excluding extraordinary transfers basically shows the size of the bailout that benefited the states after 1995. Figures 6.4 through 6.9 depict the fiscal stance of the most indebted states and the Federal District. On the one hand, the states of Mexico and Sonora have been experiencing some fiscal improvements in the past two years, although they are still operating with insufficient primary surplus to sustain their high level of current indebt- edness. On the other hand, the other states did not adjust up to 1997 and continued to increase their primary deficits up to that year. The states most- ly improved their direct fiscal positions in 1998-99 as the extraordinary fed- eral transfers phased out. These statistics do not reveal some important con- tingent liabilities for subnational governments; such as guarantees to public enterprises and pay-as-you-go pension and health schemes for state employees. Still incomplete estimates reveal that the size of outstanding contingent debt is daunting. Figure 6.3. Mexico: Aggregate Subnational Governments' Fiscal Deficit, 1989-97 1997 prices (pesos) 4,000,000 2,000,000 0 1990 \991. 1992 1993 1994 .1995 1996 1997 -2,000,000 1989 -4,000,000 -6,000,000 -8,000,000 -10,000,000 -12,000,000 Year *4. BAL. PRIMARIO REAL -41- BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. 244 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 6.4. Jalisco: Fiscal Deficit, 1989-97 1997 prices (pesos) 2,000,000 1,500,000 1,000,000 500,000 0 1989 1990 1991 1992 1993 1994 1995 1994. 1997 -500,000 -1,000,000 - -1,500,000 Year -_- BAL. PRIMARIO REAL 4 BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. Figure 6.5. State of Mexico: Fiscal Deficit, 1989-97 1997 prices (pesos) 2,000,000 1,000,000 0 1989 1992 1993 1994 9 1996 1997 -1,000,000 1990 199 -2,000,000 -3,000,000 -4,000,000 -5,000,000 -6,000,000 Year -_- BAL. PRIMARIO REAL . BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 245 Figure 6.6. Nuevo Leon: Fiscal Deficit, 1989-97 1997 prices (pesos) 600,000 400,000 200,000 1 / 0 1989 199O3 1191 1992 1993 1994 4935 194 1997 -200,000 -400,000 -600,000 -,- - -. -800,000 * ,' -1,000,000 4I -1,200,000 Year _- BAL. PRIMARIO REAL - . - BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. Figure 6.7. Sonora: Fiscal Deficit, 1989-97 1997 prices (pesos) 700,000 600,000 500,000 400,000 300,000 200,000 100,000 0 1/ . -100,000 1989 1990 1991 1993\ 199/ .7.1995 1996 1997 -200,000 . -300,000 Year S BAL. PRIMARIO REAL * BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. 246 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Figure 6.8. Federal District: Fiscal Deficit, 1989-97 1997 prices (pesos) 3,000,000 2,000,000 1,000,000 0 1989 1990 1991 1 _ 55 1996 1V7 -1,000,000 -2,000,000- 1 -3,000,000 -4,000,000 -5,000,000 -6,000,000 Year a BAL. PRIMARIO REAL * BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. Figure 6.9. Sinaloa: Fiscal Deficit, 1989-97 1997 prices (pesos) 800,000 600,000 400,000 200,000 0 ~~~~~~1992 W 1990 19-91-7 s 1993 1994 195 1996 IV7 -200,000 -400,000 -600,000 -800,000 Year _ BAL. PRIMARIO REAL * BAL. PRIM. REAL (Excluding extraordinary transfers) Source: Authors' calculations from SHCP data. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 247 Subnational Borrowing: The Experience Up to 2000 Subnational governments can borrow mainly from development banks (38 percent in 1997) and commercial banks (62 percent). Other sources are available but rarely used by the states. The State of Guanajuato used bonds to finance construction of a road in 1997. The bond is guaranteed by the fees charged for use of the road. A fund (fideicomiso) was created to collect all the road revenues. The total amount of the issue was $Mex84 million. The capital of Yucatan (Merida) has also issued a municipal bond. These are the only recent subnational experiences with bonds. The municipality of Garza Garcia in Nuevo Le6n, which is contiguous to Monterrey, needed a tunnel to connect two municipalities separated by a mountain. The government of Garza Garcia, in coordination with the state government, put this con- struction to a referendum proposing that a property tax surcharge in Garza Garcia be instituted for seven years to finance the construction. People in the municipality passed the referendum, and the municipality borrowed the resources from a commercial bank (Bancomer) using the property tax surcharge as a guarantee. The Institutional and Legal Design The institutional and legal design of the National Fiscal Coordination Law is important because it contains the incentives for both creditors and bor- rowers; some of those incentives are not desirable. Subnational government borrowing is partly regulated by the national Constitution. The Federal Congress has the power to establish the bases on which the executive branch may arrange loans and take responsibility for public debt. All subnational governments must respect the criteria con- tained in Article 117, Section 8, for states; and Article 115, Section 6, for municipalities. The Constitution states that subnational governments can borrow only in Mexican pesos and only from Mexicans, and they can bor- row only for productive investments. As a result, Banobras-a federal government development bank-and other financial institutions have found a way to lend in pesos with funds obtained in foreign currencies from international financial institutions, taking the exchange risk. The details for guaranteeing credits are contained in Article 9 of the National Fiscal Coordination Law, created in 1980, which states that sub- national governments can borrow from commercial and development banks to finance investment projects only after receiving authorization of the local congress. This law also states that around 20 percent of federal tax revenue must be transferred to state and local governments (that is, Mexico has a revenue-sharing system). In fact, the main source of revenue for sub- national governments comes from federal transfers (participaciones fed- erales), which, on average, account for about 85 percent of their total income. 248 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Before a reform implemented in January 1997, after the tequila bailout, Article 9 allowed states and the Federal District to use their federal trans- fers as collateral. In case of arrears or a threat of default, the federal gov- ernment would deduct debt service payments (on registered debt) from rev- enue sharing before the funds were transferred to states each month. This arrangement began in the 1980s when the banking system was nationalized. Each year the state government would propose the debt level, and the state congress would approve the ceiling. This included the debt of munic- ipalities. Municipalities, in principle, could incur debt, but the state had to guarantee it. That is why the state congress had to approve municipal debt. The institutional arrangement before the 1995 crisis was simple. For par- ticipaciones to be used as collateral, states only needed to register the new debt contract with the SHCP, after receiving authorization from the local congress. This debt was backed by participaciones. SHCP could, in princi- ple, deny the new debt and thus control the indebtedness of subnational governments, but this rarely happened. This legislation had two implications for the behavior of suppliers and debtors. First, banks had incentives to make loans to subnational borrow- ers, because the credit risk was virtually nil, being guaranteed by the fed- eral government. Second, states had incentives to borrow because the fed- eral government would always bail them out. The latter is explained as follows. Provided that federal transfers were the main source of revenues for the states and that current expenditures represented an average of nearly 80 percent of total expenditures, most of the disposable income of states and the Federal District was tied. This means that, if their revenue-sharing transfers were diverted to service debt, they would not be able to operate; this in turn brought high political costs at both the local and federal levels. Consequently, the federal government saw no alternative but to bail out the defaulting state by making the payments without deducting the amount from participaciones. This was typical when a commercial bank exercised a cross-default acceleration clause that would completely use up the state's participaciones or leave it without enough to pay for essential services. These two points could explain in part the overborrowing in subnational credit markets and the lack of an explicit local regulation for borrowing or the lack of an obligation to present or publish financial statements. This, in principle, made it very difficult for lending institutions to evaluate a pro- ject. Until 1999 lending institutions rarely conducted an evaluation because the credit was risk-free. When credit is not rationed and markets are competitive, economic the- ory suggests that spreads should be very small, reflecting only adminis- SUBNATIONAL BORROWING AND DEBT MANAGEMENT 249 trative costs. This is not the case in Mexico, where spreads are as high as 10 percentage points (see figures 6.4 through 6.9). In the absence of the pos- sibility of a bailout, economic theory further predicts an equilibrium with credit rationing, where banks would be forced to evaluate the risk of a pro- ject, which in turn would force states to disclose information, both of which would promote market discipline. To induce market discipline in subnational borrowing, Article 9 of the National Fiscal Coordination Law was reformed in 1997 to confer new obligations on state and local governments. Subnational governments could still use debt to finance their investment projects, and many still use their federal transfers as collateral. However, banks could not ask the Treasury Department to discount the corresponding amount from a default- ing state's federal transfers. They had to arrange the collateral according to state debt laws; that is, both parties had to create a repayment mecha- nism. In other words, subnational governments were responsible for repay- ing their contracted debts when federal transfers were used as collateral. In addition, they were obliged to publish their level of debt, and in turn, banks had to evaluate the risk of a project. The modification had two purposes: to force states to exercise financial discipline and to force banks to analyze project risk when making a loan. These changes, in principle, were intended to induce discipline in subna- tional credit markets in three wavs: (a) agents would respond to changes in interest rates; (b) states and local governments would define mechanisms under which borrowing is optimum and would be forced to present their financial statements when soliciting a credit; and (c) the possibility of a bailout would be reduced significantly because the federal government would be kept out of the market. Did the Modification of the Law Induce Market Discipline? Article 9 of the National Fiscal Coordination Law was modified to induce market discipline (a necessary but not sufficient condition for avoiding sub- national bailouts). Its impact cannot be evaluated, however, because the changes it sought to impose were effectively circumvented. After modifying Article 9 in January 1997, subnational governments had difficulty obtaining credit, especially from commercial banks. For this rea- son, in 1997 the federal government and the states designed a temporary scheme by which states would give the federal government a mandate to deduct debt service from participaciones. Under the mandate, banks were not forced to take losses, and they did not have incentives to evaluate the risk of the credit because they could obtain the participaciones independently of project evaluation. These actions, which were contrary to the spirit of the modified Article 9 (originally intended to instill discipline in the credit 250 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO markets) also perpetuated the moral hazard problems. As long as the fed- eral government stayed involved in the bank-subnational transaction, it was perceived as a guarantor in subnational credit markets. Thus, in spite of the well-meaning reforms of January 1999, several weaknesses remained in the regulatory system for subnational borrowing. These weaknesses, which were addressed in a subsequent wave of reforms in early 2000 (see next section), included: * Commercial banks still did not have an incentive to evaluate project risk because they were not effectively required to take any losses. This comes from the mandate given to SHCP. Most debt was issued by commercial banks. This debt was cheaper, but commercial banks did not evaluate the risk of the project and thus issued credit much faster than development banks, which did evaluate risk. * 'States and municipalities until at least 1998 still saw SHCP as the provider of last resort. This changed to some extent in 1999, when Ramo 23 virtually disappeared as a source of discretionary transfers. Ramo 23 was intended as a discretionary account of the president. Recently, a new fund was created for natural disasters, and another was created for wage increases. Ramo 23 could be used to help out states, and it was used from 1995 to 1998. In 1999 all resources were transferred to Ramo 33, and no resources are left to help out states although the case of Nuevo Leon in August 1999 shows how an emergency need for public works, combined with the threat of with- drawal from the Pacto Fiscal, could obtain special federal expenditures for what the state would ordinarily have financed. * Matching transfers, such as for state universities, reduced the flexi- bility for managing debt. Such matching grants have the practical effect of earmarking some of the states' own revenue, cutting into what was counted as disposable revenue in our earlier calculations. * Vertical imbalance persisted. As long as alternative sources of revenue were not developed, states would not be creditworthy. Most of their revenues were already tied. * Domestic banks were virtually the only sources of debt financing for subnational governments. Alternative sources are needed to better access domestic as well as foreign savings. * Before 2000, one rating agency gave debt ratings that considered past, not prospective, debt issues, and were on a Mexican scale (not a global scale). These ratings overlooked contingent liabilities and did not consider the perverse effects of fiscal imbalances, the matching transfer system, the moral hazard problem of past bailouts, and so forth. Debt had to be rated more seriously, as the new system will require. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 251 A New Reform Initiative The Overall Casefor Subnational Borrowing Regulation Fiscal decentralization in Mexico has been a political decision, and it seems irreversible. Mexican authorities are justifiably concerned with the risks involved in its scope, implementation, sequencing, and speed. Authorities are conscious about and well motivated by the political, efficiency, and equi- ty benefits resulting from decentralization, but they are also aware of the possible tradeoffs between increased autonomy in expenditure and revenue decisionmaking and responsible macroeconomic management. The potential of subnational governments to destabilize the country's macroeconomic situation is well known, especially in a federation like Mexico, where states are constitutionally sovereign in their territorial domain. Because states are free to increase outlays, even under a balanced budget they may affect macroeconomic equilibrium, since public expen- diture multipliers tend to be larger than revenue multipliers.3 Moreover, decentralized decisions tend to amplify the procyclical effect of fiscal pol- icy and, in the absence of appropriate policies, to increase public debt. Subnational governments tend to increase expenditures during periods of economic expansion but are more reluctant to reduce expenditures during recessions. This reflects soft budget constraints in which subnational gov- ernments operate with a deficit and increase their indebtedness during recessions. Although the federal government has frequently used discre- tionary grant transfers (the so-called transferencias extraordinarias) to rescue states in financial trouble, the states of Coahuila, Guerrero, Mexico, Morelos, Nuevo Le6n, Puebla, Quintana Roo, Sinaloa, Sonora, Tamaulipas, Zacatecas, and the Federal District increased their indebtedness substan- tially during the 1995 recession. Soft budget constraints and increasing indebtedness of subnational gov- ernments may have deleterious macroeconomic effects in the short run, because of their direct impact on monetary expansion, inflation, interest rates, and balance of payments. In the medium and long term, excessive subnational indebtedness may crowd out private investments and reduce economic growth, and may have a perverse intergenerational equity effect, especially if the social rate of return on public spending is low and subna- tional governments cannot internalize all the benefits. Therefore, a prior condition to guarantee successful and sustainable decentralization in Mexico is to ensure that decentralization, on the one hand, improves the social rate of return on public expenditures and, on the other, does not jeopardize short-term macroeconomic stability. Any poli- cy strategy option should include incentives to assure that (a) a hard bud- get constraint is always in place, (b) public investments generate the high- est possible social rate of return, and (c) public borrowers show enough 252 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO capacity to pay back their loans (borrowers must be creditworthy). The incentives will be more effective if subnational governments are involved in the process of shaping the overall fiscal envelope and are accountable for their share of fiscal effects. Management of Subnational Government Debt: The Initial Options Who should evaluate the capacity of subnational governments to pay, in order to avoid excessive indebtedness, and how to do the evaluation? Many systems of subnational government debt management are possible for the Mexican government.4 The most general and common systems are (a) financial market discipline; (b) strict case-by-case control by the feder- al government; and (c) the establishment of explicit, general rules by a national forum. Sometimes a combination of these systems is applied, depending on the particular market conditions. RELIANCE ON MARKET DISCIPLINE. Market discipline is the most desirable code of behavior and set of benchmarks to follow. However, to work prop- erly and effectively, market discipline has to be strict. This has rarely been achieved even in federations with well-developed financial markets. Many governments have decided not to rely solely on market discipline. Similarly, the following market failures prevail in Mexico, which means that market discipline alone may prove ineffective: * Restrictions on thefinancial market. Market discipline is only effective if the financial market is free and open. The financial market is not entirely free or open in Mexico. Restricted access to foreign capital markets limits the available options, and compulsory allocation of resources (including those to official financial agencies and parastatals and the placement of government bonds) amplifies the borrowing capacity of the public sector and creates a suboptimal financial sec- tor portfolio. * Lack of transparency. Adequate dissemination and availability of infor- mation and full transparency on debt outstanding and capacity to pay are essential to market discipline. In Mexico, the effort to obtain reli- able financial information, especially from subnational governments, is not a trivial endeavor. Not all states and municipalities follow a standardized plan of accounting, keep clear and uniform registers of their assets and liabilities, or publish and disseminate information on debt and capacity to pay on a satisfactory, systematic, and reliable basis. Moreover, extra-budget/contingent liabilities, hidden either under indirect indebtedness through parastatals or under soaring pen- sion fund obligations, demand considerably more transparency in Mexico. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 253 * Moral hazard. In the presence of moral hazard incentives, market dis- cipline cannot serve as an effective check on subnational govern- ments' excessive indebtedness. As mentioned earlier, in Mexico, the federal government has often intervened to rescue subnational gov- ernments in financial difficulties. These frequent bailouts (either by means of ad hoc extraordinary grant transfers or across-the-board debt rescheduling) have fed expectations of future rescue operations and encouraged moral hazard behavior on the part of both borrow- ers and lenders. As an example, the automatic federal guarantee and liquidation of debt (through the sequestration of federal transfers), still offered by Article 9 (via the so-called mandates from the states), encouraged lenders to disregard risk evaluation, and subnational governments to incur irresponsible indebtedness. * Insensitivity to market signals. Market signals (interest rates and the pos- sibility of market exclusion) can discipline borrowers to seek finan- cial policies that are consistent with solvency. But for market disci- pline to be effective, the borrower must be sensitive to market signs. Increases in the interest rate should stop or at least make the borrower review its decision to borrow. However, governors and municipal presidents have rarely given adequate weight to market signals when determining their expenditure. In Canada, market discipline alone is relied on to control subnational government indebtedness, and private rating companies evaluate public sector creditworthiness in a competitive environment. However, even in Canada, with its well-developed financial market, market discipline has been unable to check the excessive indebtedness of subnational govern- ments. In the mid-1990s subnational debt reached 23 percent of GDP, prompting Canadian provinces to adopt fiscal adjustment programs to restrain themselves. But this only happened after the exclusion point had practically been reached, entailing high social costs. Brazil, and to a certain extent Argentina, which do not have the neces- sary market conditions in place, have also taken some sort of market dis- cipline approach since the late 1980s, with disastrous consequences. In Brazil, subnational debt jumped from 1 percent of GDP in the early 1970s to 20 percent in the mid-1990s, and in the past ten years the federal gov- ernment intervened three times with large bailout operations to rescue the creditors of states, and twice to rescue the creditors of municipalities. In Mexico, given present market conditions, relying solely on market dis- cipline is not recommended, at least for the time being. Therefore, adequate regulation for checking excessive subnational government indebtedness is needed. In order to minimize distortions and encourage development of market practices, the necessary regulation should mimic desirable market 254 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO discipline to the extent possible. But to what extent should the federal government or, alternatively, a national forum secretariat directly control subnational indebtedness in Mexico? DIRECT ADMINISTRATIVE CONTROL. At the other extreme from the market is the enforcement of direct administrative control from the center to check excessive subnational government indebtedness. The direct control approach, however, has been used more frequently by unitary countries and less so by federations. States in Mexico also enjoy ample autonomy, and the direct control system may not be an adequate approach. Direct controls (a) require the federal government's approval for each credit operation proposed by subnational government or (b) prohibit sub- national governments from accessing capital markets directly. The approval of each credit operation inevitably requires an evaluation of the financial terms and conditions under which each operation is contracted. India is an outstanding example of central government approval on a case-by-case basis-the central government is a major creditor and guarantor of sub- national governments, and the Constitution requires such intervention. During the 1980s, Australia prohibited subnational governments from accessing capital markets and centralized all loans through the Australian commonwealth government (the Loan Council), which then on-lent to the subnational governments. The direct control system in Australia was not effective, and now subnational governments are free to access capital mar- kets directly. The Loan Council functions were restructured, and excessive indebtedness is now checked ex ante through a cooperative approach at the national level and is monitored closely ex post. In Mexico the problems with a direct control system are clear: (a) cen- tralizing all credit operations would involve the federal government direct- ly in micromanaging each credit operation of every subnational govern- ment (the opposite of fiscal decentralization), entailing an unnecessary increase in federal bureaucracy and undesirable inefficiencies in the finan- cial system; and (b) such prohibitions would be incompatible with the Mexican Constitution, which allows states free access to the domestic cap- ital market, restricted only by approval of the state congress and by state indebtedness laws. There are, nevertheless, strong arguments in favor of some prohibi- tions, or at least severe limitations, on some subnational government cred- it operations, in order to prevent systemic damage. These are operations concerned with (a) central bank financing, (b) noninvestment expenditure financing (the golden rule), (c) short-term loans for liquidity assistance, and (d) external financing. These kinds of subnational financing operations are banned in Mexico. Central Bank independence in Mexico rules out any kind of direct government financing. Similar rules could regulate other offi- SUBNATIONAL BORROWING AND DEBT MANAGEMENT 255 cial financial institutions (for example, Banobras and Nacional Financiera). In addition, the Mexico Constitution prohibits expenditure financing for purposes other than investment (the golden rule). Yet, in the case of Mexico, the golden rule is not a sufficiently restrictive criterion, since (a) a clear def- inition of what constitutes investment is not yet in place, and (b) public sav- ings need to be generated to help finance public investments. Given the initial stage of decentralization in Mexico, the other two pro- hibitions (against short-term loans and external loans) will remain for the time being, and the government may want to revisit them in due time. Also because of the golden rule, subnational governments cannot yet use short- term loans for liquidity assistance purposes and are losing the opportuni- ty to smooth out their cash flow through the year in order to optimize their financial inflows with outflows. The use of this kind of short-term loan by subnational governments is quite common and appropriate (for example, in the United States, Brazil, and other countries), as long as these govern- ments can be forced to liquidate their debt during the same fiscal year. The Mexican Constitution prohibits subnational governments from accessing foreign capital markets. Some compelling arguments support this prohibition. First, because changes in external debt have direct impacts on macroeconomic variables (exchange rate, foreign exchange reserves, and money supply), the federal government wants to keep full control over short-run stabilization policies to prevent federal policies from being either neutralized by or detrimental to subnational governments. Second, allow- ing the states and federal government to approach international capital mar- kets in a fragmented fashion would be less efficient and would result in less favorable debt terms than a coordinated approach from the center. Third, there is always the possibility of a contagion effect in which the default of any one state would affect the creditworthiness and risk rating of other states and the federal govemment. Fourth, official international lenders usu- ally require a federal government guarantee. The arguments in favor of prohibiting external debt by subnational gov- ernments seem inescapable for the time being. Yet, as market conditions evolve and fiscal decentralization expands, the prohibition should be reex- amined, since significant opportunities may be lost by not accessing the global capital market. A cooperative approach, with greater involvement, commitment, and accountability of subnational governments and improved coordination by the federal government, will make the arguments for pro- hibition less compelling and create conditions for a less radical approach based on more effective rules. RULES-BASED APPROACH. There are strong reasons to support an adequate rules-based approach to curb subnational government access to capital mar- kets. Rules can only be effective if they can be substantiated in a simple, 256 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO transparent, and across-the-board set of legally binding instruments (the Constitution or ordinary laws). In general, these rules should comprise quantitative limits and procedural norms that respect or imitate, to the extent possible, the market practice of good financial discipline and cred- itworthiness indicators. Being constantly submitted for review, some of these rules should be established as preventative measures, others should only be implemented according to need. A rules-based approach has been adopted to different degrees in Brazil, Korea, Japan, Spain, and the United States, and other countries. Mexico can benefit from the extensive interna- tional experience in this area. The most common and essential lessons learned include * Limit the borrower's maximum debt service ratio. Subnational govern- ments should not be allowed to incur further debt if their debt service ratio (flow of interest due and amortization divided bv flow of dis- posable revenue) exceeds a certain limit, say 12 percent. A debt ser- vice commitment above this limit will likely jeopardize the delivery of normal public services. * Limit the borrower's maximum level of total indebtedness. Subnational gov- ernments should not be allowed to incur further debt if their total indebtedness indicator (ratio of outstanding debt, including indirect and contingent liabilities, to disposable annual revenue) exceeds a cer- tain limit, say 0.8. This indicator of indebtedness will capture the debt burden of loans and credits that are still benefiting from a grace peri- od. * Limit banks' portfolio exposure to the public sector. Banks' portfolio expo- sure to the public sector should be constrained by a certain maximum limit. This limit should be enforced on total bank-by-bank assets to subnational governments as a group and to each public sector entity individually. Stricter norms and supervision should be applied on the official credit institutions (for example, Banobras, Nacional Financiera). * Enforce strict bank reserve requirements. Besides the regular reserve requirements imposed on banks by the monetary authority, special provisions could be enforced on their operations with subnational governments. A special regulatory and supervisory framework could be in place to preempt problems with subnational governments that are experiencing financial difficulties. * Pass and regulate a public entity bankruptcy law. In preparation for mar- ket discipline, the government could introduce a municipal and state bankruptcy law defining debt workout procedures in the case of default. SUBNATIONAL BORROWING AND DEBT MANAGEMENT 257 * Pass and regulate a law offiscal accountability. This law would seek to stabilize public accounts over time (by limiting reiterated, excessive deficits and imprudent buildup of public debt), increase efficiency of public administration, achieve fiscal transparency (by clearly speci- fying an accounting code and full disclosure), and ensure the account- ability of public officials, placing the risk on lenders rather than the federal government. * Encourage dissemination of risk rating of subnational governments. To improve transparency and encourage the financial system to operate as close as possible to market discipline, the practice of creditwor- thiness analysis should be encouraged. In the United States and Canada, this practice is very common, and private risk-rating com- panies play a central role in helping subnational governments tap the capital markets, and in helping lenders gauge risks and limit subna- tional government's excessive indebtedness. Because of market fail- ures, such practices are not well established in developing countries. Rules-based systems of checking excessive subnational government indebtedness have the great advantage of being transparent and impartial- qualities that minimize political bargains and discretionality. Possible dis- advantages are that excess inflexibility can damage the system and that some states will try to circumvent the rules. Although these disadvantages may operate in the short run, in the medium and long run, the rules can be changed and adjusted to new circumstances and necessities. As for short- run rigidities, the purpose of the rules is precisely to harden the budget con- straints. Mexico's New Regimefor Subnational Borrowing: 2000 and Beyond Faced with the challenge of both creating market-based mechanisms that ex ante would prevent excessive subnational borrowing and conveying a credible signal that the federal government would not, ex post, bail out par- ties involved in such borrowing, in late 1999 the Mexican federal govern- ment introduced (with full effect as of April 2000) a new regulatory frame- work for debt management by the states and municipalities. This framework has six components which innovatively combine market dis- cipline with rules. The first two changes reinforce the government's com- mitment not to have bailouts. The next three (combined with the long-stand- ing constitutional prohibition of foreign borrowing by states) limit ex ante the states' borrowing, in order to reduce the chance for states to create a debt crisis that would make the rules too costly for the federal government to enforce. The sixth change, pertaining to the government development banks, serves both purposes. The six components are: 258 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO 1. A renunciation, and ensuing removal from the federal budget for 2000, of the Executive's power for discretionary transfer. Naturally, this pol- icy proved uncontroversial with the opposition-dominated Congress. 2. The abolition of the mandatos. This left the states and their creditors to make their ownfideicomiso arrangements for the collateralization of debt with participaciones or other revenue flows, assuming the legal risks involved and without recourse to the federation. 3. The elimination of the so-called regimen de excepci6n for borrower concentration limits. Now the regular rules apply also to subnation- al debt, limiting the extent of financial-sector damage that one single state can cause and signaling that state debt must be evaluated on a basis similar to other debt. 4. The establishment of a link between the capital risk weighting of bank loans to subnational governments and those governments' cred- it ratings. In particular, two, current, published, global-scale, local-cur- rency credit ratings performed by internationally reputable credit rat- ing agencies are to be used by bank regulators to assign capital risk weightings (between 20 and 115 percent) to loans given to states and municipalities. The rules for assigning these weights are fully speci- fied by regulation. This innovative scheme, which is in line with the Basle Committee's recommendations of June 1999, is based on the distance between the rating obtained by the subnational borrower in question and the rat- ing of the local-currency debt of the federal government. To control for agency shopping, two ratings are called for by the regulation and, in case of large discrepancies (more than two grades of distance), the capital weighting of the worse rating applies. The purpose of these regulations is, of course, to make the pricing of bank loans a function of the underlying risk of the subnational bor- rower, especially in the new framework of no federal intervention. Financially weaker states are likely to be priced or rationed out of the market (and become conditional clients of the development banks; see below), while stronger states would see the price of loans fall. 5. Registration of subnational loans with the federal government was made conditional upon the borrowing state or municipality being cur- rent on its publication of debt and associated fiscal statistics from the preceding year's final accounts, and on its debt service obligations toward the government's development banks. At the same time, and to make that registration appealing, unregistered loans are automat- ically risk-weighted by the regulators at 150 percent. This additional incentive to transparency has the purpose of ensuring that private con- tracting between the subnational borrowing and the credit rating agencies does not lead to the withholding of a minimum of quanti- SUBNATIONAL BORROWING AND DEBT MANAGEMENT 259 tative information on the borrower finances. Ultimately the disci- pline on these governments will come most effectively if voters and opposition parties have access to full information about the fiscal behavior of the administration in office. 6. Finally, and as a matter of corporate policy, federally owned devel- opment banks are to make new loans to states and municipalities (and their organiismos descenitralizados) only when the loan in question qual- ifies for registration and its capital risk-weighting is less than 100 per- cent. This policy, coupled witlh the conditions for registration men- tioned above, makes the development banks part of the rigor of the new regulatory scheme, rather than potential loopholes to it. Lending to weaker subnationals, however, is not forbidden. It is, after all, these clients whom the development banks have a mandate to assist. Instead, subnational loans with risk weightings of more than 100 percent are allowed if the loan package contains a technical assistance component funded by an international development bank. This latter arrangement conveys the signal that, when the loan is particularly risky (and correspondingly expensive), its origination and supervision are subject to a neutral, independent party. Assuring Success: Key Inplementationi Issuies Mexico's new regulatory arrangement for subnational borrowing is a novel and promising start in putting local finances on a sustainable path. It is like- ly to generate momentum for further reform, notably in the distribution of tax and expenditure responsibilities among the three levels of government (currently embedded in the Pacto Fiscal). Because of the weak association between their access to bank credit and their creditworthiness, states and municipalities previously had no incentive to pay the political cost of rais- ing local taxes or rationalizing local expenditures. With that association now firmly in place, subnational government will now take a more proactive and conservative approach to their finances. This will benefit the country as a whole. But, for those benefits to occur, several elements will need to converge, most of which are within the con- trol of the federal authorities. Among those that federation can control are the continuing application of the new rules to the development banks and the continuing commitment to reniounce to discretionary budget powers. These factors would become all the more important if, in July 2000, a sin- gle party both won the presidential election and gained control of Congress. Also within the policymakers' control will be the quality and enforce- ment of rules for the definition of bank capital, especially the new and sounder rules enacted in September 1999. The implementation of these rules will be critical if higher capital risk weightings for riskier borrowers are to be translated into more expensive loans, rather than fictitious capital allo- 260 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO cations (like deferred taxes) without true opportunity cost to the bank owners. Similarly, while capital risk-weightings will be a function of rating dis- tances (from the federation), it will be up to the regulators to require that the ratings are performed and posted on a global scale. This will increase the reputational exposure of the credit rating agencies involved, and assure more thorough initial and follow-up analyses. Finally, and beyond the policymakers' control, the level of competition in the market for nonlending bank services to subnational government will play a role in how differential capital risk-weightings are passed through and reflected on lending interest rates. If states and municipalities have plenty of alternative providers of the agency services that they contract from banks (for example, payroll payments and cash management), the latter may be reluctant in their loan price calculations to account fully for the high- er capital allocations that subnational loans may require in the new regu- latory framework. Notes Overview 1. In the late 1990s some states have shown interest in complementing the Fiscal Pact by piggybacking state taxes on the federal income tax or value added tax. 2. A ramo is a section of the budget. 3. However, the primary purpose of these transfers between 1995 and 1998 was to bail out overindebted states. 4. For the Federal District, the federal Congress approves both the annual bud- get and borrowing. 5. The agreements are often used, in effect, to conunit local governments to administer the property tax. Formation and updating of the fiscal cadastre are often-but not always-the responsibility of the state government. 6. Comit6s de Planeaci6n para el Desarrollo del Estado (COPLADEs) and Comites de Planeaci6n para el Desarrollo Municipal (COPLADEMUNs) should serve the purposes of intersectoral and public-private coordination at the state and municipal levels. However, these are predominantly spending mechanisms that are not usually concerned with fiscal responsibility or with balancing the revenue and spending columns of the fiscal equation. As a result, they are not effective for mobilizing additional local and regional resources. 7. Sector fragmentation of the budgets also discourages intersectoral coordina- tion within one level of government (not to mention intergovernmental and inter- sectoral coordination) for allocating resources in a specified region or subregion. 8. A cadastre is an official record of property holdings and values. 9. Tax efficiency is measured as the ratio of tax revenue to the cost of tax admin- istration. 10. If the federal government were simply to cede power to the states to levy a given tax, the "typical" state rate would be the rate previously levied by the fed- eral government. 11. Excessive indebtedness of states may have equity implications. In Mexico, there is a clear positive correlation between state indebtedness and state per capi- ta GDP. Therefore, debt bailouts tend to be highly regressive because the poorer states end up paying for the financial relief of richer states. 12.This calculation was for 16 states plus the Federal District, which employ about half of the total state employees. 13. This link between credit ratings and regulatory standards is in line with a recent consultative paper issued by the Basel Committee on Banking Supervision ("A New Capital Adequacy Framework," Basel Committee on Banking Supervision, Basel, June 1999). More generally, linking regulation to credit ratings is a common 261 262 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO practice in the financial sector of many countries (including Mexico), for example, in regulating corporate bond issues or insurance companies' investments (see Annex VI of IMF [19991 International Capital Markets-Developments, Prospects and Key Policy Issues, World Economic and Financial Surveys, at www.imf.org/external/pubs/ft/icm/1999/index.htm). 14. Banobras will be able to lend to a state with more risky credit ratings if the loans have an institutional development component funded by the Inter-American Development Bank, the World Bank, or other international development agencies because this may help improve the state's credit rating (especially among low-rated states). 15. By limiting the discretionary power of the federal authorities, the prevalence of rules will also directly support ongoing congressional efforts to grant more independence to subnational governments, including recent amendments to Article 115 of the Constitution. Perspectiva General 1. De un tiempo a esta parte, los estados han mostrado mas interes en comple- mentar el Pacto Fiscal cargando los impuestos estatales al impuesto sobre la renta federal o al impuesto al valor agregado. 2. Un ramo es una secci6n del presupuesto. 3. Sin embargo, el prop6sito fundamental de esas transferencias entre 1995 y 1998 fue sacar de apuros a los estados que se habian endeudado de manera exagerada. 4. Para el Distrito Federal, el Congreso federal autoriza el presupuesto y el endeudamiento anual. 5. En efecto, a menudo se recurre a convenios para obligar a los gobiernos locales a que administren los impuestos prediales. La formaci6n y actualizaci6n del catastro fiscal son a menudo-pero no siempre-responsabilidad del gobiemo estatal. 6. Los Comit6s de Planeaci6n para el Desarrollo del Estado y los Comites de Planeaci6n para el Desarrollo Municipal deben servir para la coordinaci6n puibli- ca-privada e intersectorial en los niveles estatal y municipal (COPLADEs y COPLADEMUNs, respectivamente). Sin embargo, se trata de mecanismos pre- domninantemente de egresos que, por lo general, no se ocupan de la responsabilidad fiscal rii de equilibrar las colurnnas de gastos e ingresos de la ecuaci6n fiscal. Por ende, no son eficaces para transferir recursos adicionales de caracter local y regional. 7. La fragmentaci6n sectorial de los presupuestos tambi6n entorpece la coordi- naci6n intersectorial dentro de un nivel de gobierno para asignar recursos en una regi6n o subregi6n especifica (por no mencionar la coordinaci6n intergubernamental e intersectorial). 8. El catastro es el registro oficial de la tenencia y valores de bienes raices. 9. La eficiencia fiscal se mide como la proporci6n del ingreso fiscal con respec- to al costo de la administraci6n fiscal. 10. Si el gobierno federal simplemente fuera a delegar poder a los estados para gravar un determinado impuesto, la tasa estatal "tipica" seria la tasa previamente fijada por el gobierno federal. 11. El endeudamiento excesivo de los estados puede tener repercusiones de equidad. En M6xico, existe una evidente correlaci6n positiva entre el endeu- NOTES 263 damiento estatal y el producto interno bruto per capita de los estados. Por con- siguiente, los rescates de deuda tienden a ser ligeramente regresivos porque los esta- dos mas pobres terminan pagando la ayuda financiera de los estados mas ricos. 12. Este calculo era para 16 estados mAs el Distrito Federal, que tiene contrata- dos a cerca de la mitad de todos los trabajadores piublicos estatales. 13. Este vinculo entre calificaciones de credito y normas reguladoras va a la par con un reciente documento consultivo publicado por el Comite de Basilea para la Supervisi6n Bancaria (Un Niuevo Marco de Suficiencia de Capital, Comite de Basilea para la Supervisi6n Bancaria, Basilea, Junio 1999). En terminos mas generales, vin- cular la reglamentaci6n a las calificaciones crediticias es una practica normal en el sector financiero de muchos paises (entre ellos Mexico); por ejemplo, al regular las emisiones de bonos de empresas privadas o inversiones de compafifas asegurado- ras (consulte el Anexo VI del Fondo Monetario Internacional [1999], International Capital Markets-Developments, Prospects and Key Policy Issues, World Economic and Fiancial Surveys, en www.imf.org/external/pubs/ft/icm/ 1999/index.htm. 14. Banobras podra otorgar prestamos a un estado con calificaciones crediticias mas arriesgadas si los prestamos incluyen un factor de desarrollo institutional financiado por el Banco Interamericano de Desarrollo, el Banco Mundial o algin otro organismo internacional de fomento ya que esto puede ayudar a mejorar la cal- ificaci6n crediticia del estado en cuesti6n (sobre todo entre los que tengan una cal- ificaci6n baja). 15. Al limitar el poder discrecional de las autoridades federales, la preponde- rancia de reglas tambien apoyarA de manera directa los esfuerzos que esta real- izando el Congreso para conceder mas independencia a los gobiernos estatales, incluyendo las recientes enmiendas al Articulo 115 de la Constituci6n. Chapter 1 1. Borrowing is efficient because it solves the problem of liquidity and allows local governments to match the timing of consumption with payment for those services. Having one generation of taxpayers pay for the capital equipment and then allow posterior generations to consume the services free of charge is unfair. 2. In general it seems that an origin-based sales tax would be more appropriate than a tax on commercial property because firms providing services generally have relatively little tangible property. But public services may be closely related to the existence of tangible property, especially real property, the base of most suc- cessful property taxation. It may appear that a source-based corporate income tax would also be appropriate in this case, but there are several problems with this rea- soning. Services are presumably not provided only to corporations, and the value of such services is not likely to be closely related to profitability. 3. The result under a flat-rate surcharge on the base of the central government, as in some states of the United States, is very different from that under the Canadian system of imposing provincial surcharges on the tax liability to the national gov- ernment. Since the national tax in Canada is levied at graduated rates, the provin- cial surcharge is also progressive, with the adverse effect already mentioned. 4. In addition, the retail sales tax provides some incentive for vertical integra- 264 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO tion. This is reduced by exempting goods bought for resale or direct use in pro- duction, for fuels, and for utilities; such exemptions vary from state to state. 5. By far the most important cause of tax exporting in the United States is the deduction for state and local income and property taxes that is allowed in computing liability for federal income tax, a provision that is hard to justify. Taxes are also exported to owners of firms that cannot shift taxes to others (customers, suppliers, or labor). Exporting to out-of-state customers is not likely to be important, because of pressures from untaxed competitors. 6. For more detailed descriptions, see McLure (1995), Martinez-Vasquez, McLure, and Wallace (1995), and Asian Development Bank (forthcoming). As explained there, the lack of oblast fiscal autonomy is much worse than suggested by the description in the text. Tax sharing commonly varies across taxes and between oblasts and is set to provide the revenue needed to finance a given level of expen- ditures. Where shared taxes do not provide adequate revenues, the central gov- ernment provides subventions. Expenditure levels reflect a combination of norms, historical patterns, inflation adjustment, and negotiation. The German lander also rely heavily on shared taxes, particularly the value-added tax. In Canada the fed- eral government collects the harmonized sales tax for itself and several of the poor Maritime Provinces, dividing the provincial share on the basis of estimated con- sumption in the various provinces. 7. The pervasiveness of revenue sharing conditions the applicability of this statement, since it implies that the federal government does not keep all the rev- enues from any tax that enters the coparticipation pool. 8. Under the state value added tax proposed in Chapter 4, interstate transactions would be zero-rated by the state of origin and subject to the compensating value added tax. Purchases by registered traders would be taxed by the state of destina- tion on a deferred payment basis. It appears that the problem of assuring consis- tency of state taxation would be vastly more difficult in the absence of the com- pensating value added tax. The state of destination would have to rely on the federal govemment to see that the state of origin is not making zero-rated sales to households in its jurisdiction. Chapter 2 1. Memorias de la Primera Convencion Nacional Fiscal, 1925, p.6. Chapter 3 1. However, expenditure shares can be a misleading measure of decentraliza- tion. Subnational governments can have more resources pass through their bud- gets, but may not have more discretion over these expenditures. 2. These data are from the Ministry of Finance and Public Credit (SHCP). Data for total revenues available to the states for 1999 indicate much lower horizontal disparities across states (see Chapter 5). This may be because there was equaliza- tion in 1998 and 1999, or that the data sets are not consistent or comparable. 3. For the 23 states with complete data. The regression coefficient was statisti- cally significant at the 1 percent confidence level, and the R-square was 0.57. 4. The degree of control actually exerted by the ministry depends on the state, in particular the governor and his team. One possible reason for the lack of decen- NOTES 265 tralization is that the ministry has not downsized significantly as a result of the National Agreement and federal employees have found it hard to change their roles. 5. This dual system existed for a number of years. Currently, even though the two systems have been formally integrated, they continue to be administered sep- arately in many states. 6. The negotiations between the Ministry of Public Education and the national union are circumscribed by the budget allocation approved by Congress in the fed- eral budget for federal teachers. 7. For example, parents in Zacatecas and Oaxaca complain that the absenteeism rate of teachers is very high (World Bank 1996). 8. The equity issue could be addressed by providing special scholarships for low- income students. 9. Any differences between decentralized and other states were mostly due to differences in initial conditions and not to decentralization per se (OCDE 1998). 10. However, wages and salaries are fixed at the national level through negoti- ations between the Ministry of Health and the national union. 11. INEGI (1998, p.82) reports that the distribution of health transfers in the Fiscal Coordination Law is based on a capitation formula adjusted by mortality, pover- ty, and expenditure on health services. This is not described in the law itself. It could be that federal authorities plan to move toward this very desirable reform. 12. Despite increases in federal health expenditures and considerable investments in poorer states, until 1997 the Ministry of Health budget allocations per capita were inversely correlated with the poverty level of the state. Reportedly, the Ministry is using a new formula that takes into account demographic and epidemiological char- acteristics (OCDE 1998). 13. The average cost of transport to the nearest health center for treatment in Zacatecas is more than one month's eamings (World Bank 1996). 14. Per capita, risk-adjusted funding is more equitable and also more likely to contain costs (as opposed to payment for service rendered). 15. According to the Ministry of Health, the states manage 70 percent of all bud- get resources in health services, but 70 percent or more of these funds are for wages and salaries that are determined at the federal level. 16. Since 1990 the Irrigation District Authority has been transferred to produc- ers themselves, and cost recovery has improved for those districts still administered by the National Water Authority. 17. For the most part, toll roads have not succeeded due to a combination of fac- tors, including poor feasibility analysis, high tariffs, and the concessionary process. 18. Since 1985 the federal government has delivered funds for investment pro- grams only when channeled through the COPLADE. Chapter 4 1. This statement is not meant to imply that the base of the existing federal tax is optimal or even desirable. It may be that all payroll tax bases, including that of the federal tax, should be made to conform to the best of the state bases. The point is that all the bases should be the same. Uniformity of bases might be achieved by federal mandate or-perhaps more acceptable politically-by allowing credits for some portion of state taxes against a federal payroll tax, provided the states adopt 266 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO the federal base. The latter technique has been used to achieve uniformity of state taxes on transfers of real estate. 2. Mexican subnational governments accounts report "gross revenues" as the rel- evant measure of own revenue collection. That is not a correct measure of fiscal capacity because it includes borrowing, revenue sharing, and funds collected on behalf of other levels of government. Net tax revenues refers to revenues collected by a subnational government without borrowing, revenue sharing, and the so-called revenue "por cuenta de terceros." It is the sum of Impuestos, Derechos, Productos, and Aprovechamientos, which are the relevant tax and nontax revenues. All esti- mates in the tables are calculated as a percentage of net tax revenues. 3. If residence-based taxation were feasible, it might make more sense to reserve surcharges on the individual income tax for use by municipal governments, because of the scarcity of tax bases that are suitable for use by local governments. This is espe- cially true if, as suggested below, state surcharges on the VAT are feasible. 4. A tax limited to a small portion of the population invites "tyranny of the major- ity." Whether this is more problematic for a federal or state tax is unclear. 5. These figures are calculated from state income distribution survey results released by 1999, but the raw data, which would produce a better analysis, could not be obtained from INEGI. 6. A long-run objective might be to merge the payroll and individual income tax systems, allocating to the states the capacity to levy flat rate income tax surcharges, while leaving the progressive element at the central level. This would necessitate unification of the two tax bases and the administration of the two taxes-a step that would presumably allow some savings in costs of compliance and administration. Whether this is a sensible objective from the point of tax assignment depends in part on one's view of the nature of the services provided by the states and the preva- lence of commuting across state boundaries. If labor income is thought to be a rea- sonable proxy for the consumption of such services and if (a) most such services are provided where people work, rather than where they live, or (b) cross-bound- ary commuting is not overly important, it would be better to employ the (source- based) payroll tax to finance them, even in the long run. If total income (including nonlabor income) were thought to be a better proxy, if most services are provided where people live, or if cross-boundary commuting is important, unification of the income and payroll taxes into a single residence-based tax would be more appro- priate, if it can be achieved. 7. One might think of the flat-rate state tax as being levied at the "basic" rate on virtually all income, with the graduated rates of the federal tax being levied only on income above a certain level. 8. Depending on the structure of personal deductions and credits, deductions and credits for personal expenditures, and other features of the individual income tax, the situation could be even worse. This cannot be assessed without a thorough examination of the structure of the individual income tax, which is beyond the scope of this paper. 9. Absent the constitutional requirement that taxes on hydrocarbons should be federal taxes, taxes on motor fuels could also usefully be assigned in part to the states to help cover the costs of highways and roads. 10. In the case of motor fuels the problem is more likely to take the form primarily of (a) households residing in high-tax states filling their cars in neighboring low- tax states, and (b) interstate truckers concentrating purchases of fuel in low-tax NOTES 267 states. The former abuse is difficult to prevent, but seems unlikely to be quantita- tively significant, except in a few cases (for example, the Federal District and the State of Mexico). To prevent the latter abuse, if Mexico were to assign motor fuel excises to the states (following relaxation of the constitutional prohibition of state taxes on hydrocarbons), consideration might be given to the "base state" approach employed by all the states of the United States and some of the provinces of Canada. Under this approach, truckers pay the tax due where they purchase fuels. But they apportion road use, and thus consumption of fuel, among the states where they operate, based on mileage logs, and file returns with their "base state" (ordinarily the state or province where they have their primary place of business) showing tax liability to each state, based on estimated consumption and tax rates in the state. The taxpayer makes one payment to its base state if, on balance, the sum of resid- ual tax due all states is positive (and receives a refund if the sum of residual liabil- ities is negative). Each base state then remits to (or receives from) an interstate clear- inghouse the net balance due other states. Given Mexico's participation in the North American Free Trade Association with Canada and the United States, it seems appropriate for it to consider joining this arrangement, especially if its states are given the power to levy taxes on motor fuels. 11. The ease of smuggling from neighboring countries effectively limits the level of excises. Above a certain level, increases in excises are likely to produce less revenue, not more. 12. It is also generally inefficient to use "geographic separate accounting" for such a purpose, since such accounts generally are not needed for any other reason. 13. It is worth noting that over the past 25 years there has been a tendency for the states of the United States to shift from a formula that accords equal weight to the three factors to one that places double weight on sales (and even to formulas based entirely on sales), presumably to reduce the implicit burden on origin-based factors. 14. This discussion concentrates on sales to households and unregistered traders and those to registered traders. (It is assumed that sales to unregistered traders should be taxed, as a surrogate for taxing their sales.) It does not examine details such as the treatment of sales to and by govemments and nonprofit institutions, because these are not central to the issues at hand. 15. The approach employed in the state RSTs levied by 45 states (and the District of Columbia) in the United States is to allow registered traders to purchase certain types of products by presenting "resale exemption certificates" to their suppliers. All states exempt purchases of goods to be resold without a change of condition and most purchases of goods to be incorporated in goods for resale. Beyond that, prac- tice varies from state to state; some states exempt, inter alia, goods to be used direct- ly in production, fuels, public utilities, agricultural implements, and seeds bought by farmers. For a more complete discussion, see Due and Mikesell (1994). It has been estimated that, on average, 40 percent of the state tax base involves sales to busi- ness, with wide variation around that average; see Ring (1989 and 1999). Thus Mexico should not emulate the typical RST levied by the American states, which deviates substantially from a pure consumption-based tax. 16. The usual problem with exemption systems is that they place the vendor in the unenviable position of determining whether each sale is taxable or legally exempt, and create an incentive for the vendor to "look the other way" when a household purchase masquerades as a business purchase. Under the Mexican pro- 268 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO posal it appears that the vendor need only ascertain that the purchaser has a valid VAT TIN. 17. If, contrary to the Mexican proposal, the state RST were to be implemented by the federal tax administration, costs of administration would be substantially lower than under state administration of the RST, and costs of compliance would be somewhat lower. The federal administration would determine for purposes of both the VAT and the RST whether a particular purchase was for business purposes, and thus eligible for VAT input credits and exemption from the RST. Taxpayers would need to deal with only one tax administration. 18. Much is made of the "cross-checking" that is made possible by the VAT- the fact that credits taken by registered purchasers must be shown on invoices issued by vendors, and thus subject to tax. This is probably less important than confirm- ing that purchases are for legitimate business use. 19. If experience in other countries is any guide, taxpayers would not-as they should-be able simply to give state auditors photocopies of information provid- ed to federal auditors. 20. This discussion refers only to the RST; the next subsection discusses the anal- ogous problems under the VAT. Exemption of interstate sales to registered busi- nesses by the state of origin would occur naturally under the RST, provided the state's tax administration respects the right of the out-of-state business purchaser to make exempt purchases. 21. Note that this discussion pertains directly only to "remote selling," sales of tangible products sent to the customer across state lines. It does not concern "cross- border shopping," in which the customer buys a product "across the counter" in a state where she or he does not live and takes it home for use there. It seems virtu- ally inevitable that the latter transactions will bear tax at the rate prevailing in the state where the sale is made (the state of origin), and not that of the state of desti- nation, except in rare cases, for example, where goods must be registered in the state to be used there, as in the case of automobiles (and perhaps boats and planes). This places revenues in the "wrong" state. People may engage in cross-border shopping in order to reduce taxes when two adjacent jurisdictions have markedly different tax rates. This abuse is inevitably difficult to prevent. It is, of course, important only where metropolitan areas straddle the boundaries of subnational jurisdictions, as in the case of Mexico City. Interstate sales of digital content over the Internet ("elec- tronic commerce") could be handled in the same way as interstate sales of tangi- ble products. The most difficult problems involve sales originating outside the coun- try. Sales occurring within the country could, in principle if not in actuality, be monitored. It is difficult to tax a transaction in digital content between an unknown customer and a foreign seller (perhaps located in a tax haven), especially if payment is made in untraceable money. See McLure (1997 and 1999). 22.This may seem not to be an important problem in Mexico at the present time. It is certainly an important problem in the United States and Canada, and will grow in importance as the economy of Mexico develops. Part of the problem in both of Mexico's northem neighbors can be traced to constitutional restrictions on the tax- ing powers of the states and provinces. In the United States these restrictions were imposed on the states by the U.S. Supreme Court because of the overwhelming diversity of state and local tax laws; McLure (1999) describes the situation in the United States. There is no reason that Mexico need be bound by similar artificial restrictions. NOTES 269 23. In theory, each of the 32 states could rely on the tax administrations of the other 31 states to serve this purpose. Indeed, perhaps the "base state" approach described above in footnote 9 could be employed for the sales tax. The latter is one alternative being considered by the National Tax Association's Telecommunications and Electronic Commerce Tax Project, described in McLure (1999). One of the Project's concerns is that state tax administrators might not be diligent in auditing their own taxpayers on behalf of other states. This concem seems equally telling in Mexico. In any event, it would be feasible only if the tax base of the various states were essentially identical. With federal administration of the state RSTs, returns for all states would be filed with the federal administration. This would not relieve the most burdensome aspects of compliance, the need to distinguish the state of des- tination of remote sales. 24. Some technique such as this might be necessary in the case of electronic com- merce in digital content, since the vendor may not know the location of the buyer. See McLure (1999). 25. See, for example, Boadway (1997) and (McLure) 1980a. Indeed, even shar- ing of revenues from a national tax has been thought to involve similar problems. See McLure (1995). 26. Much of this discussion is based on an excellent paper by Poddar (1990). It is assumed that intemational trade would be taxed under the destination princi- ple, as is the virtually universal practice. Thus, exports would be zero-rated, and the full value of imported products would be subject to tax. 27. As above, it is assumed that unregistered traders are to be treated like house- holds, since they are "outside the system," and thus not eligible for credits for tax paid on purchased inputs. 28. Even with uniform rates the value of goods crossing internal borders may be contentious. Values attached to such trade are a matter of indifference to tax- payers, but subnational governments would not be indifferent, because the valu- ation at the border would affect their tax bases. Thus taxpayers could be caught between tax authorities in the two jurisdictions. 29. The more affluent states of the south tax imports and the poorer states of the northeast rebate tax on exports. See Longo (1982). To ameliorate the resulting inequities, lower rates are applied to interstate sales than to intrastate sales, seri- ously complicating the system. 30. Poddar (1990, pp. 110-11) calls this a joint national-state VAT when imposed in the context of a two-tiered national/subnational VAT. 31. At the risk of excessive proliferation of labels, one might refer to the Quebec system as employing "unprotected zero-rating/deferred payment" and the CVAT system proposed below as involving "protected zero-rating/deferred payment." 32. It is worth noting that in this case the province administers both the federal and provincial VATs. The VAT in Mexico was administered by the states for some years prior to 1991, but the clearest effect of such measure was a drop in revenue, attributable probably to compliance problems. Thus,. the federal component of any dual system employed in Mexico-and perhaps the state component-would pre- sumably be administered by the federal tax authorities. In some countries there is a further issue: whether subnational governments trust the central govermrnent to deliver the revenue collected on their behalf. 33. In recent years there have been several proposals in Argentina and Brazil for a dual state/central VAT. In Argentina Libontti and Salinardi (1994) have proposed 270 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO state use of "unprotected" zero-rating and deferred payment. Drawing on a pro- posal originally made for Brazil by Varsano (1995), Gonzalez Cano (1996) and Fenochietto have proposed an "WA compartido" or shared VAT that would com- bine zero rating/deferred payment with an additional "compensating" VAT on interstate sales to registered traders along the lines proposed below. See McLure (1998). 34. It should be noted that international trade poses no conceptual problem in this case. State tax is collected on imports and credit is allowed for the tax paid on imports. Similarly, exports are zero-rated for the state tax, as well as the federal tax. 35. This statement assumes that the price of power approximates the marginal long-run cost of producing it. If this is not true, marginal cost pricing should be intro- duced; taxation may be one way to do this. 36. In addition, Section IX of Article 117 provides that "the States may not in any case... .levy duties on the production, storage, or sale of tobacco in a manner distinct from or with rates greater than those authorized by the Congress of the Union." 37. A common sense interpretation of the following article should not pose any problem: "Article 118. Nor shall the States, without the consent of the Congress of the Union.. establish ship tonnage dues or any other port charges, or levy imposts or taxes on imports or exports." Sales taxes and excises levied on a destination basis include consumption of imported goods in their bases. While they may be collect- ed at the point of importation, they are not taxes on imports. Chapter 6 1. Total lending from commercial banks to subnational governments in early- 1998 was only about 4 percent of their total portfolio. 2. It is not possible to compute the debt service ratio, because figures for inter- est due and amortization are not available. The available statistics onthe fiscal/finan- cial flows only show cash payments. 3. The impact of increased subnational government expenditure is bigger when they are financed with deficits, as is often the case in Mexico. 4. For a survey and a discussion of the relevant international experience see Ter- Minassian and Craig (1997) and Lane (1993). Bibliography Aguilar Villanueva, Luis F. 1996. "El federalismo mexicano: funcionamiento y tareas pendientes." In Hernandez Chavez, Alicia, ed., ,Hacia un Nuevo Federalismo? Mexico: Fondo de Cultura Econ6mica. Alzati, Fausto. 1998. "The Political Economy of Growth in Mexico." Ph.D. Dissertation. Cambridge, MA: Harvard University. Amieva-Huerta, Juan. 1997. "Mexico." In Ter-Minassian, ed., Fiscal Federalism in Theory and Practice. Washington, D.C.: International Monetary Fund. Arellano Cadena, Rogelio, ed. 1996. Mexico Hacia un Nuevo Federalismo Fiscal Mexico: Fondo de Cultura Econ6mica. Gobierno del Estado de Puebla. Asian Development Bank. Forthcoming. Fiscal Transition in Kazakhstan. Manila: Asian Development Bank. Bailey, John. 1994. "Centralism and Political Change in Mexico; The Case of National Solidarity." In Wayne Cornelius, Ann Craig and Jonathan Fox, eds., Transforming State Society Relations in Mexico: The National Solidarity Strategy. San Diego: Center for U.S.-Mexican Studies. Barro, Robert, and Xavier Sala-i-Martin. 1991. "Convergence Across States and Regions." Brookings Papers on Economic Activity No. 1. Washington, D.C. . 1995. Economic Growth. New York: McGraw-Hill. Basel Committee on Banking Supervision. 1999. "A New Capital Adequacy Framework." Basel, Switzerland. June. Bird, Richard M., and Pierre Pascal Gendron. 1998. "Dual VATs and Cross- Border Trade: Two Problems, One Solution?" International Tax and Public Finance 5: 429-42. Boadway, Robin. 1997. "Reforming the Fiscal Arrangements in Argentina: Lessons from Industrialized Federations." March 4. Processed. . 1998. "The Economics of Equalization: An Overview." In Robin Boadway and Paul Hobson, eds., Equalization: Its Contribution to Canada's Economic and Fiscal Process. Kingston, Ontario: Queen's University, John Deutsch Institute for the Study of Economic Policy. 271 272 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Boadway, Robin, and Paul Hobson. 1993. Intergovernmental Fiscal Relations in Canada. Toronto: Canadian Tax Foundation. Bonifaz Chapoy, Beatriz. 1992. Finanzas Nacionales y Finanzas Estatales. Mexico: UNAM. Bruhn, Katherine. 1996. Social Spending and Political Support: The Lessons of the National Solidarity Program in Mexico World Politics. Burki, Javed, William Dillinger, Guillermo Perry, and others, 1999. Beyond the Center: Decentralizing the State in Latin America. Washington, D.C.: World Bank. Cabrero, Enrique. 1995. La nueva gesti6n municipal en Mexico, Ed. Miguel Angel Porrua-CIDE, M6xico. . Coordinador. 1996. Los dilemas de la modernizacion municipal, Ed. Miguel Angel Porrua-CIDE, Mexico Cabrero, E., L. Flamand, C. Santizo, and A. Vega. 1998. "Claroscuros del nuevo federalismo mexicano. Estrategias en la decentralizaci6n fed- eral, y capacidades en la gesti6n local." Gestion y P6litica Publica (CIDE) 6(2). CIDAC. 1998. "F6rmulas para la asignaci6n de recursos del presupuesto de egresos de la federaci6n a estados y municipios." Presentation to the Planning and Budget Committee, Mexican Chamber of Deputies. Mexico. Processed. Commonwealth Grants Commission. 1999. Report on General Revenue Grant Relativities: 1995 Update. Canberra: Australian Government Publishing Services. Courchene, Thomas J. 1984. Equalization Payments: Past Present and Future. Toronto: Ontario Economic Council. . 1998. "Renegotiating Equalization: National Polity, Federal State, International Economy." C. D. Howe Commentary No. 113. Toronto: C. D. Howe Institute. . 1999. "Subnational Budgetary and Stabilization Policies in Canada and Australia." In James Poterba and Jurgen von Hagen, eds., Fiscal Institutions and Fiscal Performance: A National Bureau of Economic Research Conference Report. Chicago: University of Chicago Press, 301-48. Courchene, Thomas, and Colin R. Telmer. 1998. From Heartland to North American Region State: The Social, Fiscal and Federal Evolution of Ontario. Toronto: University of Toronto, Faculty of Management. . 1998. "Revenue Redistribution and Federal-Provincial Fiscal Relations in Canada." In Robin Boadway and Paul Hobson, eds., Equalization: Its Contribution to Canada's Economic and Fiscal Process. Kingston, Ontario: Queen's University, John Deutsch Institute for the Study of Economic Policy. Diaz Cayeros, Alberto. 1995. Desarrollo Econ6mico e Inequidad Regional: Hacia un Nuevo Pacto Federal en Mexico. Mexico: M. A. Porruia. BIBLIOGRAPHY 273 ______. 1997. "Asignaci6n Politica de Recursos dentro del Federalismo Mexicano: Incentivos y Limitaciones." Perfiles Latinoamericanos 10 (January-June): 35-73. ___ . 1997. "Political Responses to Regional Inequality: Taxation and Distribution in Mexico." Ph.D. Dissertation, Duke University. Diaz Cayeros, Alberto. Forthcoming. "Diversidad regional, jurisdicciones concurrentes y desarrollo." In Jorge Javier Romero and Maria Amparo Casar, eds., Federalismo Mexicao o. Mexico: M. A. Porrua. Diaz Cayeros, Alberto, and Beatriz Magaloni. 1998. "Autoridad pre- supuestal del poder legislativo en Mexico: una primera aproxi- maci6n." Politica y Gobierno V(2). Diaz-Cayeros, and Martinez Uriarte. 1999. "Tensiones del Federalismo: Crecimiento y Pobreza en las Regiones." Presentation to the Campaign Team of Presidential Candidate Vicente Fox. Mexico: CIDAC. Processed. Dillinger, William, and Steven B. Webb. 1999. "Fiscal Management in Federal Democracies: Argentina and Brazil." Policy Research Working Paper 2121. Washington, D.C.: World Bank. Dillinger, William, Guillermo Perry, and Steven B. Webb. 2000. "Macroecnomic Management in Decentralized Democracies: The Quest for Hard Budget Constraints in Latin America." In Javed Burki and Guillermo Perry et al., eds., "Decentralization and Accountability of the Public Sector in Latin America." Proceedings of Anntual World Bank Conference on Developnmenit in Latin Amnerica. 1999. Washington, D.C.: World Bank. Dresser, Denise. 1991. "Neopopulist Solutions to Neoliberal Problems. Mexico's National Solidarity Program." La Jolla: Center for U.S.-Mexican Studies. Due, John F., and John L. Mikesell. 1994. Sales Taxation: State antd Local Structure and Administration. Second Edition. Washington, D.C.: Urban Institute Press. Duncan, Harley T., and Charles E. McLure, Jr. 1997. "Tax Administration in the United States of America: A Decentralized System." Bulletin for International Fiscal DoculmentatioI 51(2): 74-85, February. Also forth- coming in Spanish translation in Hacienda Publica Espaniol. Fenochietto, Ricardo. (No date). "El IVA Compartido: Una Herramienta Uitil para el Reemplazo del Impuesto sobre los Ingresos Brutos y la Descentralizaci6n de Tributos." Processed. Garman, Christopher, Stephan Haggard, and Eliza Willis. 1999. "The Politics of Decentralization in Latin America." Latin American Research Review 34(1). Gershberg, Alec, and Donald Winkler. 2000. "Education Decentralization in Latin America: The Effects on the Quality of Schooling." In Javed Burki and Guillermo Perry, et al., (eds.), "Decentralization and 274 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Accountability of the Public Sector in Latin America." Proceedings of the Annual World Bank Conference on Development in Latin America and the Caribbean 1999. Washington, D.C.: World Bank. Gonzalez Cano, Hugo. 1996. "La Reforma Tributaria de Brasil y Posible Aplicacion del Nuevo IVA Federal y Estadual para el Reemplazo del Impuesto a los Ingresos Brutos." Boletin de la DGI 513: 1391-97, September. Gonzalez Casanova, Pablo. 1965. La democracia en Mexico. Mexico, D.F.: Ediciones ERA. Guerrero, Juan Pablo, et. al. 1998. "La Descentralizaci6n de los Fondos de Aportaciones Federales (Ramo 33): Analisis Preliminar de sus Efectos y Riesgos." M6xico: CIDE. Hobson, Paul. 1998. "Is there Too Much Revenue Redistribution through Canada's Fiscal Equalization Program?" In Robin Boadway and Paul Hobson, eds, Equalization: Its Contribtution to Canada's Economic and Fiscal Progress. Kingston, Ontario: Queen's University, The John Deutsch Institute for the Study of Economic Policy, 157-174. International Monetary Fund (IMF). 1999. "International Capital Markets- Developments, Prospects and Key Policy Issues, Annex VI." World Economic and Financial Surveys. www.imf.org/extemal/pubs/ft/icm/ 1999/index.htm. Instituto Nacional de Estadistica Geografia e Informatica (INEGI). 1998. "El ingreso y el gasto piiblico en Mexico: Diversas ediciones y poder ejecutivo federal." In Cuarto informe de gobierno. Mexico, D.F. . 1996. Encuesta Nacional de Ingresos y Gastos de los Hogares. Mexico: INEGI. . 1989. Encuesta Nacional de Ingresos y Gastos de los Hogares. Mexico: INEGI. _______. 1999. Encuesta del Estado de Campeche de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado de Coahuila de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado de Guanajuato de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado de Hidalgo de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado de Jalisco de Ingresos y Gastos de los Hogares 1996. M6xico: INEGI. . 1999. Encuesta del Estado de Oaxaca de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado de Tabaco de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. . 1999. Encuesta del Estado del Distrito Federal de Ingresos y Gastos de los Hogares 1996. Mexico: INEGI. REFERENCES 275 ______. 1999. Encuiesta del Estado del Estado de Mexico de Ingresos y Gastos de los Hogares 1996. M6xico: INEGI. . 1999. Encuesta Ingreso Gasto de los Hogares. Mexico: INEGI. Keen, Michael, and Stephen Smith. 1996. "The Future of the Value Added Tax in the European Union." Economic Policy 23: 373-420, October. Lane, Timothy. 1993. "Market Discipline." International Monetary Fund Staff Paper. Washington, D.C. Latapi, Pablo, and Manuel Ulloa. 1998. "Inconsistencias de la F6rmula de Distribuci6n de los Recursos Federales para Educacion y Propuesta de Alternativas." In Comnisi6n de Desarrollo Social Instrumentos de Distribuci6n de los Recursos del Ranio 33. Mexico: Camara de Diputados. Levy, Santiago. 1998. "Analisis Metodol6gico de la Distribuci6n de los Recursos en el Ramo 33." In Cornmisi6n de Desarrollo Social Instrumentos de Distribuci6n de los Recuirsos del Ramo 33. Mexico: Camara de Diputados. Libonatti, Oscar, and Mario Salinardi. 1994. IVA Provincial: Una Propuesta para Sn Implementaci6n. Buenos Aires. Longo, Carlos. 1982. "Restricted Origin Principle under Triangular Trade Flows: Implications for Trade and Tax Revenues." Journal of Developmnent Economtiics 10: 103-12. Martinez Almazan, Raul. 1980. "Las Relaciones Fiscales y Financieras Intergubernamentales en M6xico." M6xico: INAP. Martinez Cabafnas, Gustavo. 1985. "La Administraci6n Estatal y Municipal en M6xico." M6xico: INAP. Martinez-Vasquez, Jorge, Charles E. McLure, Jr., and Sally Wallace. 1995. "Subnational Fiscal Decentralization in Ukraine." In Richard M. Bird, Robert D. Ebel, and Christine I. Wallich, eds., Decentralization of the Socialist State: Intergovernmental Finance in Transition Economies. Washington, D.C.: World Bank, pp. 281-319. McKinnon, Ronald I. 1997. "Monetary Regimes, Government Borrowing Constraints, and Market-Preserving Federalism: Implications for EMU." In Thomas J. Courchene, ed., The Nation State in a Global/lnformation Era: Policy Challenges 5. The Bell Canada Papers on Economic and Public Policy. Kingston, Ontario: John Deutsch Institute for the Study of Economic Policy, 101-142. McKinnon, Ronald I. 1998. "Monetary Regimes, Government Borrowing Constraints, and Market-Preserving Federalism: Implications for the EMU." In Thomas J Courchene, ed., The Nation State in a Global/Information Era: Policy Challenges. Queen's University: The John Deutsch Institute for the Study of Economic Policy, 101-142. McLure, Jr., Charles E. 1980a. "State and Federal Relations in the Taxation of Value Added." Jornial of Corporation Law 6: 127-39. - 1980b. "The State Corporate Income Tax: Lambs in Wolves' Clothing." In Henry J. Aaron and Michael J. Boskin, eds., The 276 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Economics of Taxation. Washington, D.C.: The Brookings Institution, 327-346. ______. "The Elusive Incidence of the Corporate Income Tax: The State Case." Public Finance Quarterly 9(4) :395-413, October. . "Defining a Unitary Business: An Economist's View." In Charles E. McLure, Jr., ed., The State Corporation Income Tax: Issues in Worldwide Unitary Combination. Stanford, CA: Hoover Institution Press, 89-124. _____. "The Sharing of Taxes on Natural Resources and the Future of the Russian Federation." In Christine Wallich, ed., Russia and the Challenge of Fiscal Federalism. Washington, D.C.: World Bank, 181-217. ______. "Revenue Assignment and Intergovernmental Fiscal Relations in Russia." In Edward Lazear, ed., Economic Reform in Eastern Europe and Russia: Realities of Reform. Palo Alto, CA: Hoover Institution Press, 199-246. . 1997. "Taxation of Electronic Commerce: Economic Objectives, Technological Constraints, and Tax Law." Tax Law Review 52 (Spring): 269-423. ____. 1998. "Protecting Dual VATs from Evasion on Cross-Border Trade: An Addendum to Bird and Gendron." Processed. (Forthcoming.) _____. 1999. "Electronic Commerce and the State Retail Sales Tax: A Challenge to American Federalism." International Tax and Public Finance 6(2): 193-224, May. McLure, Charles E., Jr., Christine Wallich, and Jennie I. Litvack. 1995. "Special Issues in Russian Federal Finance: Ethnic Separatism and Natural Resources." In Richard M. Bird, Robert D. Ebel and Christine 1. Wallich, eds., Decentralization of the Socialist State: Intergovernmental Finance in Transition Economies. Washington, D.C.: World Bank, 379-404. Merino. M. Coordinador. 1994 En busca de la democracia municipal, Ed. Colmex, Mexico Merino, Gustavo. 1998. "Las transferencias federales para la educaci6n en Mexico: Una evaluaci6n de sus criterios de equidad y eficiencia." Gesti6n y Politica Puiblica (CIDE) 7(2). Mieszkowski, Peter, and Eric Toder. 1983. "Taxation of Energy Resources." In Charles E. McLure, Jr. and Peter Mieszkowski, eds., Fiscal Federalism and the Taxation of Natural Resources. Cambridge: Lexington Books, 5-91. Mogoll6n, Olivia. 1999. "Posibles influencias electorales y partidistas en la distribuci6 de recursos del Fondo de Desarrollo Social Municipal." Washington, D.C.: The World Bank. Processed. Molinar Horcasitas, Juan, and Jeff Weldon. 1994. "Electoral Determinants and Consequences of National Solidarity." In Wayne Cornelius, Ann Craig and Jonathan Fox, eds., Transforming State Society Relations in REFERENCES 277 Mexico: The National Solidarity Strategy. San Diego: Center for U.S.-Mexican Studies. Musgrave, Richard A. 1983. "Who Should Tax, Where, and What?" In Charles E. McLure, Jr., ed., Tax Assignment in Federal Countries. Canberra: Centre for Research on Federal Financial Relations, 2-19. Navarrete, Juan. 1995. "Convergencia: Un Estudio para los Estados de la Repuiblica Mexicana." Documento de Trabajo del CIDE, Divisi6n de Economia, No. 42. Mexico: CIDE. Navarrete, Juan. 1995. "Convergencia: Un Estudio para los Estados de la Repiublica Mexicana?" Documento de Trabajo del CIDE, Divisi6n de Economia No.42. Mexico: CIDE. Oates, Wallace E. 1972. Fiscal Federalism. New York: Harcourt Brace Jovanovich, Inc. _____. 1983. "Tax Effectiveness and Tax Equity in Federal Countries: Commentary." In Charles McLure, ed., Tax Assignment in Federal Countries. Canberra: The Australian National University, Federalism Research Centre. . 1994. "Federalism and Government Finance." In John M. Quigley and Eugene Smolensky, eds., Modern Public Finance. Cambridge, MA: Harvard University Press, 126-51. Organizaci6n para la Cooperaci6n y el Desarrollo Econ6mico (OCDE). 1998. "Capitulo especial: Reforma del sistema de salud." Estudios econ6micos de la OCDE. Paris. Pani, Alberto. 1926. "Convocatoria." Primera Convenci6n Nacional Fiscal (1926) Memoria. Mexico: Secretaria de Hacienda y Credito PuIblico. Pineda, Sandra, and Marcela G6mez. 1999. "El Reparto Municipal del Pronasol: Criterios de Asignaci6n en Aguascalientes y Michoacan." B.A. Thesis, ITAM, Mexico. Poddar, Satya. 1990. "Value-Added Tax at the State Level." In Malcolm Gillis, Carl S. Shoup and Gerardo P. Sicat, eds., Value Added Taxation in Developing Countries. Washington, D.C.: World Bank, 104-12. "Programa para un Nuevo Federalismo 1995-2000." 1997. M6xico: Diario Oficial de la Federacion, 6 August. http://www.cedemun.gob.mx/ nuevofed.doc. Retchkiman, Benjamin, and Gerardo Gil Valdivia. 1981. "El Federalismo y la Coordinaci6n Fiscal." Mexico: UNAM. Riker, William. 1964. Federalism: Origin, Operation, Significance. Boston: Little Brown. Ring, Raymond J., Jr. 1999. "Consumers' Share and Producers' Share of the General Sales Tax." National Tax Journal 52(1) (March, 1999): 79-90. Saucedo Sanchez, Alberto. 1998. "Hacia el Federalismo Fiscal. La Reforma del Sistema de Asignaci6n de Participaciones Federales a los Estados." Mexico: INAP. 278 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Scott, Robert Edwin. 1959. Mexican Government in Transition. Urbana: University of Illinois Press. Secretaria de Gobernaci6n, CEDEMUN, and INEGI. 1995. Censo nacional de desarrollo municipal. Mexico, D.F. Secretaria de Hacienda y Credito Publico (SHCP). 1973. Primer Informe Sobre las Relaciones Fiscales Entre la Federacion y los Estados. Mazatlan: SHCP. Stepan, Alfred. 1997. "Toward a Comparative Theory of Federalism." All Souls College, Oxford University. Processed. Ter-Minassian, Teresa, and Jon Craig. 1997. "Control of Subnational Government Borrowing." In Teresa Ter-Minassian, ed., Fiscal Federalism in Theory and Practice: A Collection of Essays. Washington, D.C.: International Monetary Fund. Trejo, Guillermo, and Claudio Jones. 1998. "Poverty and Income Distribution." In Susan Kaufman and Luis Rubio, eds., Mexico Under Zedillo. Boulder: Lynne Rienner. Varsano, Ricardo. 1995. "A Tributac,o do Comercio Interestadual: ICMS versus ICMS Partilhado." Texto para Discussao No. 382. Brasilia: Instituto de Pesquisa Econ6mica Aplicada, Setembro. Walsh, Cliff. 1996. "Making a Mess of Tax Allocation: Australia, as a Case Study." In Paul Boothe, ed., Reforming Fiscal Federalism for Global Competition: A Canada-Australia Comparison. Calgary: Western Centre for Studies in Economic Policy. Watson, W. G. 1986. "An Estimate of the Welfare Gains from Fiscal Equalization." Canadian Journal of Economics 19: 298-308. Weingast, Barry. 1995. "The Economic Role of Political Institutions: Market- Preserving Federalism and Economic Development." Journal of Law and Economics and Organization 11: 1-31. Wilkie, James. 1978. "La Revoluci6n Mexicana. Gasto Federal y Cambio Social." M6xico: Fondo de Cultura Econ6mica. Willis, Eliza, Christopher Garman, and Stephan Haggard. 1999. "The Politics of Decentralization in Latin America." Latin American Research Review 34(1). World Bank. 1996. "Mexico: Poverty Reduction. The Unfinished Agenda." Report 15692ME. Washington, D.C., December. . 1998a. "Mexico: Advancing Educational Equity in the Context of Decentralization." Phase I draft report. Washington, D.C., November. _____. 1998b. "The Transport Sector in Mexico: An Evaluation." Report 18153. Washington, D.C., June 30. Zedillo, Ernesto. Various years. Informe de gobierno. Finanzas pziblicas estatales y municipales. Mexico, D.F. Ziccardi. A. Coordinador. 1995. La tarea de gobernar, Ed. Miguel Angel Porrua-IISUNAM, Mexico Author Biographies ENRIQUE CABRERO MENDOZA, a Mexican national, obtained his master's degree in public administration from the Centro de Investigaci6n y Docencia Econ6micas (CIDE) in Mexico City and subsequently undertook his doctoral studies in public policy at the Hautes Etudes Commerciales (HEC) in France. For the past ten years Cabrero Mendoza has been a pro- fessor and researcher at CIDE, where he is also director of the public administration program and founder of the joumal Gesti6n y Politica Pablica. He is a member of the National System of Researchers (SNI) in Mexico. In addition to a large number of articles and chapters, Cabrero Mendoza is author or editor of several major works, the most recent of which include La nueva gestion municipal en Mexico (1995); Los dilemas de la modernizaci6n municipal (1996); Las polfticas descentralizadoras en Mexico, 1983-1993 (1998); and Gerencia puTblica municipal (1999), all published jointly by Editorial Miguel Angel Porrua and CIDE. THOMAS COURCHENE, a Canadian national, was educated at the University of Saskatchewan (Honours BA, 1962) and Princeton University (Ph.D., 1967). From 1965 to 1988 he was a Professor of Economics at the University of Western Ontario. Dr. Courchene spent the fall term of 1986 as a visiting Professor at Ecole nationale d'administration publique (Montreal). For the academic year 1987/ 88, he occupied the John P. Robarts Chair in Canadian Studies at York University. In 1988, he accepted the Directorship of Queen's new School of Policy Studies (1988-92). Currently, Courchene is the Jarislowsky-Deutsch Professor of Economic and Financial Policy at Queen's, is a member of the Department of Economics, the School of Policy Studies and the Faculty of Law, and is the Director of the John Deutsch Institute for the Study of Economic Policy. Dr. Courchene was Chair of the Ontario Economic Council of Canada from 1982 to 1985, has been a Senior Fellow of the C.D. Howe Institute since 1980, was a former member of the 279 280 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO Economic Council of Canada, is a Fellow of the Royal Society of Canada (elected 1981) and is a Past President (1991-92) of the Canadian Economics Association. He received an Honourary Doctorate of Laws from the University of Western Ontario in 1997. In April of 1999, Thomas Courchene was invested as an Officer in the Order of Canada. Recently, Thomas Courchene was awarded the 1999 Molson Prize from the Canada Council for the Arts. ALBERTO DiAZ-CAYEROS, a Mexican national, holds a Ph.D. and an M.A. in Political Science from Duke University, and a B.A. in Economics from the Instituto Tecnologico Autonomo de Mexico (ITAM). After working for many years at CIDAC, Centro de Investigacion para el Desarrollo, A.C. a think tank in Mexico City, he is now Assistant Professor in the Department of Political Science at the University of California, Los Angeles. While at CIDAC, he worked on policy projects related to technology, housing, the North American Free Trade Agreement, savings, federalism and munici- pal government. During the 1996-97 academic year he was a Fellow at Harvard University writing his dissertation, which won in 1998 the American Political Science Association Award for the best dissertation in the field of Political Economy. He has taught in Mexico at ITAM, in both the economics and the political science departments, at Duke and Stanford Universities in the United States, and at Humboldt Universitat in Berlin. He has published in the areas of federalism, regional development, and political economy of budgets. During the 1999-2000 academic year he was a Visiting Assistant Professor at Stanford University. MARCELO GIUGALE, an Argentine/Italian national, holds a Ph.D. and an M.S. in Economics from the London School of Economics, and a B.A. in Economics from Universidad Catolica Argentina. After a spell in academia, he joined the World Bank in 1989 as an economist in its financial research department. From 1990 to 1994, he was a Senior Economist in the Middle East Operations Vice-presidency, supervising Egypt's structural adjustment program and leading the Bank's reconstruction work in post-war Lebanon. From 1994 to 1998, Mr. Giugale was a Principal Economist in the Eastern Europe and Central Asia Region, responsible for the Bank's lending and analytical economic work in Lithuania and Kazakhstan. In September 1998, he became the Lead Economist for the Mexico Department, the posi- tion he currently holds. He held teaching positions at the London School of Economics and the American University in Cairo, and has published in the areas of applied econometrics, finance, business economics and eco- nomic development. Over the last six months, he led the World Bank team that delivered the recently approved US$600 million Decentralization Adjustment Loan for Mexico. AUTHOR BIOGRAPHIES 281 FAUSTO HERNANDEZ TRILLO, a Mexican national, holds a Ph.D. and an MA in Economics from The Ohio State University, and a B.A. in Economics from the National University of Mexico. He has written several academic arti- cles and two books in the area of Financial Emerging Markets. He held teaching positions at The Ohio State University, Universidad de las Americas-Puebla and CIDE. He has been a consultant for the Inter American Development Bank and the World Bank. Recently, he obtained the first place in the Daniel Cosio Villegas Prize of Economics. Currently, he is the dean of the Economics College at CIDE. JORGE MARTINEZ-VAZQUEZ, a Spanish national, holds a Ph.D. in Economics from Washington University in St. Louis, Missouri where he was a Fullbright Scholar. He also holds Lcdo degrees in Law and Economics from the University of Barcelona, and an Lcdo in Political Science from the University of Madrid. Dr. Martinez-Vazquez is now Professor of Economics and Director of the International Studies Program at the Andrew Young School of Policy Studies at Georgia State University in Atlanta. Dr. Martinez-Vazquez has worked on-site in thirty different countries on fis- cal decentralization, tax policy, tax administration, social safety net, fiscal management and budgeting, and other economic development issues. He has been an advisor to many foreign governments and worked as a con- sultant for the World Bank, USAID, IMF, ADB, IDB and the UN. Dr. Martinez-Vazquez has published widely on public finance and fiscal reform issues, and is the author of several books, numerous articles in refereed jour- nals, working papers and book chapters. Over the last two years, he has been the director of the Russia Fiscal Reform Project, a multimillion dollar and multi-year effort funded by USAID to advise Russia's federal gov- ernment and six regions in the areas of intergovernmental fiscal relations, tax policy, and tax administration reform. CHARLES E. McLURE, JR. is a Senior Fellow at the Hoover Institution at Stanford University. Prior to joining the Hoover Institution he was Vice President of the National Bureau of Economic Research (1977-81) and Cline Professor of economics at Rice University (1965-77). As Deputy Assistant Secretary of the Treasury for Tax Analysis from 1983 to 1985, McLure was responsible for developing the Treasury Department's pro- posals to President Ronald Reagan that became the basis of the Tax Reform Act of 1986, the most comprehensive reform of the income tax since its intro- duction in 1913. He was also Staff Director of the Working Group on Worldwide Unitary Taxation appointed by Treasury Secretary Donald Regan at Reagan's request. McLure has served as a Senior Economist on the staff of the President's Council of Economic Advisers, as a consultant 282 FISCAL DECENTRALIZATION: LESSONS FROM MEXICO to various agencies of the U.S. government, and an adviser to several inter- national organizations, including the World Bank, the United Nations, the International Monetary Fund, and the InterAmerican Development Bank. He holds degrees in economics from the University of Kansas (B.A., 1962) and Princeton (M.A., 1964; Ph.D., 1966). VINH NGUYEN, a United States national, is the Chief of Staff in the Mexico Country Management Unit, helping coordinate and facilitate the various sectoral activities under the department's work program. Since joining the World Bank in 1991, he has spent his time equally between direct opera- tional activities (mostly in the Human Development sector) and fiduciary and administrative areas such as quality enhancement, portfolio manage- ment, budget, and office administration. Except for a brief turn in the Europe and Central Asia region, he has been involved mostly with Mexico, Central America and parts of the Caribbean. He has an MBA from George Washington University and a BA from Amherst College. JOAO C. OLIVEIRA, a Brazilian national, holds a Ph.D. in Economics from University of Cambridge, England, an M.S. in Economics from The London School of Economics, England, and a B.A. in Economics from Universidade de Sao Paulo, Brazil. Intermittently, since 1972, he has taught economics at University of Sao Paulo, University of Campinas, and University of Brasilia (Brazil). He worked as policy researcher and adviser for the Brazilian Government since 1974, and was Deputy-Secretary of the Brazilian National Treasury from early 1996 to early 1998. He has published in the areas of agri- cultural policy, portfolio management, economic development, public finance and debt management, and subnational government finance. Mr. Oliveira joined the World Bank in 1989 as an economist at the World Bank Institute, and then moved to operations as a senior economist for Angola. During 1994 and 1995 he worked as a senior economist responsible for Mozambique at the IMF. Presently Mr. Oliveira is a senior economist at the ECSPE (Fiscal Decentralization) in the Eastern Europe and Central Asia Region. FERNANDO ROJAS, Colombian national, lawyer, holds LL.M./ITP from Harvard Law School and M.P.A from Harvard University. He was an independent consultant on fiscal, financial and institutional decentraliza- tion issues for most Latin American countries from 1976 through 1996. Mr. Rojas has taught fiscal policy and state reform at several Colombian uni- versities. He has held visiting professor and researcher positions at Stanford University, the University of Wisconsin, Harvard University, York University (Canada) the Institute of Social Studies and the University of Amsterdam (The Netherlands). Mr. Rojas joined the World Bank in 1998. AUTHOR BIOGRAPHIES 283 STEVEN B. WEBB, a United States national, holds a Ph.D. in Economics and an M.A. in History from the University of Chicago, and a B.A. in Economics from Yale University. He taught economics and economic history at the University of Michigan 1978-86 and then was a visiting senior economist at the U.S. Department of State. He published research primarily on German commercial and industrial policy in the 19th century and the hyperinfla- tion in the 1920s. He joined the World Bank in 1988 as an economist in the Macroeconomics and Growth division of the Research Department. There he produced operational reports and published research on the political economy of policy reform and central bank independence. As a senior coun- try economist he moved to the Caribbean country operations division in 1992. With the reorganization of 1997, he moved to the Poverty Reduction and Economic Management department of the Latin America region, where he has specialized in fiscal decentralization issues, for both research and operations. (Em THE WORLD BANK 1818 H Street, N.W Washington, D.C. 20433, U.S.A TELEPHONE: (202) 477-1234 FACSIMILE: (202) 477-6391 TELEX: MCI 64145 WORLDBANK MCI 248423 WORLDBANK INTERNET: www.worldbank.org E-MAIL: books@worldbank.org ISBN 0-8213-4732-2