Reassembling Social Security: A Survey of Pensions and Health Care Reforms in Latin America Published in association with the Pan-American Health Organization

Reassembling Social Security

A treasure trove that marries an analytical approach with deep, wide-ranging and long-standing understanding of country experiences. A must have for anyone interested in the tortuous path of social security reforms in Latin America and an up-to-date assessment of their outcomes. Nicholas Barr, Professor of Public Economics, London School of Economics and Political Sciences This book is an outstanding achievement. Few could have written a book on welfare reform in Latin America with the depth of knowledge, acute understanding, and easy accessibility that Mesa-Lago is able to use to illuminate on this subject. Our understanding of welfare reforms in the region is in large part due to his work. A wonderful bequest for those researching this area. Armando Barrientos, Brooks World Poverty Institute at the University of Manchester This is a unique and updated piece of academic work with a policy oriented approach to Latin American social security issues. Because pensions and heath care dominate most of Latin American social policy agenda but also have extremely relevant economic and fiscal consequences, this book is a must for everyone interested in a comprehensive view of this heterogeneous region. Therefore, the book is a valuable resource for experts on pensions, health care, social security, and Latin America. Fabio M. Bertranou, Senior Social Security Specialist, International Labour Organizacion (ILO), Santiago, Chile Without a doubt, Dr. Mesa-Lago’s study is the most comprehensive, critical and accurate account of the evolution of social security systems—including an impact evaluation of recent reforms— in the 20 Latin American countries. A must read for policy-makers and scholars interested in social security systems. Núria Homedes, MD, DrPh, Director of Global Health, University of Texas-Houston, School of Public Health, and Antonio Ugalde, PhD, Professor Emeritus, University of Texas-Austin, Department of Sociology This book is an elegant and well-researched tour-de-force of Latin American pension and health reforms and their impacts. Drawing on five decades of work on social security systems in Latin America, Carmelo Mesa-Lago bridges the gap between pensions and health care through a crosscutting analysis of vexing policy issues. This book will become an invaluable asset for all those grappling with the complexities of social policies in Latin America. Gerard M. La Forgia, Lead Health Specialist, World Bank Carmelo Mesa-Lago is one of the masters on economics of Social Security in Latin America. In the last three decades, his ideas, books, and accurate papers had influenced many generations of social economists and policy makers in every country of the Region. This book is a new masterpiece that will contribute to update the knowledge about the achievements and challenges of the last generation of reforms on social security in Latin America. André Medici, Senior Social Development Specialist, Inter American Development Bank This comprehensive volume is written by the most knowledgeable expert on Latin American pension and health care issues, a long-standing observer of social policy making on the subcontinent. His razor-sharp analysis and recommendations deserve many readers in Latin America and beyond. Dr. Katharina Mueller, Professor of Social Policy, Mannheim University of Applied Sciences

Reassembling Social Security A Survey of Pensions and Health Care Reforms in Latin America

Carmelo Mesa-Lago

1

3

Great Clarendon Street, Oxford ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Oxford University Press 2007 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2008 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn, Norfolk ISBN 978–0–19–923377–9 10 9 8 7 6 5 4 3 2 1

In view of its current relevance, I reproduce herein the dedication in my first social security book published thirty years ago: This book is devoted to the millions of workers and peasants in Latin America who suffer from lack of coverage or poor protection against social risks. It is intended as a modest contribution to the long quest for a universal, unified, standardized, and equitable social security system in the region. Social Security in Latin America: Pressure Groups, Stratification, and Inequality (University of Pittsburgh Press: 1978)

PREFACE: REASSEMBLING SOCIAL SECURITY IN LATIN AMERICA

Since the first social security programs were introduced in Germany in the 1880s, steady progress has been made to improve one of the most important instruments of social protection and welfare designed by human beings. More than half a century has elapsed since the International Labour Organization (ILO) approved the ‘minimum norm’ that established the following crucial social security principles: (a) universal coverage; (b) equal treatment; (c) solidarity; (d) comprehensiveness, sufficiency, and quality of benefits; (e) unity, state responsibility, efficiency, and social participation in the administration; and ( f ) financial sustainability. The two most important social security programs, in terms of the number of insured and beneficiaries, as well as revenues/expenditures, are old-age, disability and survivors pensions, and sickness-maternity or health care. Latin America was a continental pioneer when it introduced these programs in the Southern Cone in the 1910s and 1920s. Ultimately, these programs were implemented in all twenty countries in the region, albeit with significant differences in coverage and benefits. The conventional social security principles reigned without challenge until structural reforms commenced in the 1980s and more so in the 1990s. The structural pension reform (‘privatization’) introduced by Chile in 1981 gradually influenced other countries in the region, as well as in Central and Eastern Europe, and spurred reform debate in Western Europe and the United States. These pension reforms not only challenged the technical social security international organizations but also affected the design of policies by international financial institutions. Health care reforms, which began in the 1970s and the 1980s, had spread to all twenty countries in the region by the 1990s. The health care reforms have been less radical and more diverse than the pension reforms but both have reassembled or re-engineered social security programs transforming several key principles and setting new goals. Although pension and health care reforms have social objectives, such as extending coverage and improving equity and quality of benefits, of equal or more importance, have been the following economic aims: (a) maintaining the financial–actuarial equilibrium of the systems and

Preface vii

fiscal stability to cope with population aging in the case of pensions and increasing costs of health care; (b) establishment of the principle of equivalence or a strict relationship between contributions and benefit levels as an incentive for enrollment and payment of contributions; (c) total or partial replacement of state and/or social insurance monopolies or quasi-monopolies in pensions and health care by private insurance, financing and provision, often combined with decentralization, particularly in health care, and the separation of these functions from those of regulation and supervision that are left to the state; (d) development of insurance and provider markets as well as competition among administrators and providers; (e) granting the insured freedom of choice between competing pension administrators and health care insurance firms and providers, in pursuit of more efficiency and lower administrative costs; and ( f ) advancement of capital markets and increasing national saving promoted by pension reforms. Without a doubt, these reforms were the most important social development in Latin America in the last century. In 2004, 160 million workers in Latin America were affiliated with social insurance pensions but only 74 million were active contributors (one-third of the labor force), with 66% participating in public programs and 34% in private schemes. On the other hand, by 2001 about 136 million people were covered by social insurance health care (41% of the total population if Brazil’s population and public system are excluded) and 15 million by private insurance (59 million including Brazil) for a total of 151 million insured, leaving 325 million uninsured. Coverage by the public sector is impossible to estimate, but if those with access to Brazil’s public system were added, the total covered would reach about 62% of the population. There were many valuable resources consulted for this book, including: (a) the health care reports of the Pan American Health Organization (PAHO), including profiles of the twenty Latin American countries that describe the features and analyze the results of health reforms; (b) the country pension and health care reform studies and regional statistics from the Economic Commission for Latin America and the Caribbean (ECLAC); (c) technical documents from the ILO and the International Social Security Association (ISSA); (d) comparative world studies and statistics from the World Health Organization (WHO); and (e) regional and country reports from the World Bank and the Inter-American Development Bank (IADB). Despite this wealth of information, we still lack a comprehensive, integrated, and comparative study of the pension and health care reforms for all of Latin America that describes their features and evaluates their results. This book fills the void in the literature through a systematic comparison of pension and health care reforms in all twenty Latin American

viii Preface

countries, which categorizes reform models based on their diverse characteristics and evaluates their impact on social security principles based on standardized, recent statistics, and other data. In addition, the new goals and assumptions of the reforms are contrasted with actual results. This book is divided into four parts: Part I describes the state of social security prior to the reforms, Part II focuses on pension reforms and their effects, Part III deals with health care reforms and their effects, and Part IV provides policy recommendations in both areas. Chapter 1 summarizes the general evolution of the six conventional social security principles related to pension and health care and indicates which principles had been implemented in the region before the start of the reforms; it also describes the transformation of the conventional principles by the reforms and introduces their new goals. This chapter sets the basis for later comparisons of the social security situation before and after the reforms, as well as to test if the modified conventional principles and new goals forged by the reform have been implemented. Chapters 2 and 7 summarize the major features and key objectives of the pension and health care reforms, develop taxonomies of such reforms and identify the external influences and domestic factors involved in the reform process. The central part of this book undertakes an analytical comparison of the varied pension and health care reforms, evaluates their effects on the six conventional social security principles, and determines whether the new reforms’ goals and assumptions have been implemented and materialized. For that purpose, forty-two standardized tables systematically contrast statistics and other data for each of the twenty countries in the region, most of them with data as recent as 2005 or 2006. Four chapters are devoted to each pension reform (3 to 6) and health care reforms (8 to 11), the latter chapters are larger than those on pensions due to the complexity and diversity of health care reforms. Each chapter ends with a summary of findings on the impact of the reforms on the conventional principles and testing if new goals or assumptions have materialized. Of particular importance are the answers to the following questions: Have the reforms been successful in increasing coverage and access of the general population? Have health care reforms impeded the concentration of private providers on high- and middle-income groups with lower risks, neglecting poor- and low-income groups with higher risks? Do the new health systems offer a universal, basic package of benefits and have they enhanced quality of services? Have the reforms of both programs improved financial equity, solidarity, and gender equality? Have health reforms achieved an effective decentralization that transfers adequate authority, resources, personnel, and services to administrators closer to the participants? Has the state fulfilled its functions of regulating and

Preface ix

supervising, financing costs of the pension transition and providing social assistance to the uninsured population? To what degree has privatization been attained by both types of reform? Is there real competition in the market that effectively gives users freedom of choice and, if so, has it resulted in improved efficiency and reduced administrative costs? Do the participants have available the information and skills needed to select the best administrators and insurance providers? Do the insured participate in the administration of the systems and are their opinions taken into account to improve services? What has been the effect of the reforms on health expenditures, on the distribution of insurance monies among the three health sectors, and on out-of-pocket expenses? Have reforms accomplished a better financial equilibrium and sustainability of the systems and, in the case of pensions, also expanded capital markets and national saving? Have the new incentives controlled evasion and payment delays? Have the reforms improved health status indicators of the population? Are the private systems insulated from political and state interference? Chapter 12 offers detailed policy recommendations both of a general nature and specific to countries or issues, suggests methods to address the identified problems in the region and thereby improve pension and health care systems in the future. Finally, more than 600 bibliographic sources consulted for this book are listed in the Bibliography. Despite the author’s efforts to verify statistics and information with international, regional, and country experts, as well as to provide the most up-to-date information possible, the enormous scale of this project, and the extensive amount of material covered in this book, impeded a complete review. Therefore, it is probable that errors remain and that information for certain countries is not current. Finally, this book deals with very controversial issues, many of them charged with ideology, polarizing the field with extremes: either totally in favor or totally against the reforms. Although full objectivity is impossible, the issues are addressed in a scholarly fashion and supported with solid data, balancing the positive and negative aspects of the reforms and identifying advantages and disadvantages of public and private systems to correct flaws of both and improve all types of systems and reforms. It is my sincere hope that this book will stimulate debate, improve understanding of these reforms and, above all, contribute to better pensions and health care for the peoples of Latin America and elsewhere.

ACKNOWLEDGMENTS

It took more than four painstaking years to complete this book. Initially, I wrote separate monographs in Spanish on pension and health care reforms, under the sponsorship of the UN Economic Commission for Latin America and the Caribbean (ECLAC) in Santiago de Chile, which published both: Las Reformas de Pensiones en América Latina y su Impacto en los Principios de la Seguridad Social (2004) and Las Reformas de Salud en América Latina y el Caribe y su Impacto en los Principios de la Seguridad Social (2006). In 2006 and early 2007, I integrated the two monographs under a common framework, the data was updated, and the resulting text was considerably trimmed, edited, and translated to English. The original version of the pension part discussed public and private pensions in two separate sections that were fused in this book, whereas the original treatment of health care reforms contained separate analyses for each of the twenty countries, which were merged and compacted. These structural changes strengthened and made the comparisons more transparent. The most recent data found in this book is from 2006 even though changes in the health care and pension systems are ongoing and new data are available every day. Many people and institutions provided valuable help and financial support for this project. Andras Uthoff conceived the original idea and asked me to undertake the project, obtained financing from ECLAC and commented on the two monographs in Spanish; Daniel Titelman endorsed the project also. The Pan American Health Organization provided a grant for the translation, integration, compacting, and updating of the health care part that allowed the author to work full time on the said part; the support of Pedro Brito and Pedro Crocco is gratefully acknowledged. The Center for Latin American Studies of the University of Pittsburgh awarded two research grants in 2003–6 to finance four part-time research assistants to help gather the bibliography and data from the Internet: Gerald Hunter, Lindsey Jones, José Castro, and Javier Vazquez who also did the manuscript reformatting. Numerous and insightful comments were made by Nicholas Barr, Armando Barrientos, and Katharina Müller (the latter also provided many useful suggestions concerning policies) on the first English draft of the pension part of this book, and by Fabio Bertranou, Cecilia Acuña, Gerard La Forgia, and André Medici on the original Spanish

Acknowledgments xi

version of the health care part, as well as by three anonymous referees for OUP. Lists of questions on the pension part were submitted to and answered by officials and experts on twelve countries: Fabio Bertranou and Alfredo Conte Grand on Argentina; Alberto Bonadona on Bolivia; Helmut Schwarzer, Vinicius Pinheiro, and Rafael Ferreira on Brazil; Alberto Arenas de Mesa and Pamela Gana on Chile; Fabio Durán and Adolfo Rodríguez Herrera on Costa Rica; Omar Everleny Pérez Villanueva on Cuba; Jefrey Lizardo and Hernando Pérez Montás on the Dominican Republic; Manuel Israel Ruiz on Nicaragua; René Luciani on Panama; Eliana Carranza on Peru; and Alvaro Forteza, Heber Galli, and Ernesto Murro on Uruguay. Similar lists on the health care part were answered by officials and experts in thirteen countries: Rubén Torres and Carlos Vassallo on Argentina; Andre Medici and Sergio Piola on Brazil; Manuel Inostroza Palma on Chile; Juliana Martínez on Costa Rica; Jefrey Lizardo on the Dominican Republic; Efrem Karolys, María del Carmen Quevedo, and Francisco Penia on Ecuador; Luis José Martínez Villalba on México; Larry Valladares on Nicaragua; René Luciani and Lilian González on Panama; María Elena Ramírez de Rojas on Paraguay; Luis Manrique on Peru; José Enrique Fernández and Patricia Triunfo on Uruguay; and Marino González on Venezuela. In addition, bibliography, documents, data, and/or comments were provided on pensions and/or health care by Lilia M. Archaga de Quirós, Anida Bastidas, Fabio Bertranou, Geraldo Biasoto, Hans-Ulrich Bünger, Sergio Cesaratto, Carmen Corral de Solines, Pedro Crocco, Emilio Cueto, Fabio Durán, Iván Espinoza, Donatella Fabbri, Carlos Filgueira, Rolando Franco, Nélida Gambogi, Michael Gautrey, Rogelio Gómez, Orville Goodin, Roberto Gutiérrez, Núria Homedes, Gerard La Forgia, Marcelo Lalama, María Elena López, Nehemías López, Thomas Manz, Félix Martín, Francisco Mendoza, Eduardo Morón, Rossana Mostajo, Gustavo Nigenda, Lizette Ochoa, Ondina Olivas, Francisco Piena, Ariel Pino, Reiner Radermacher, Roberto Rodríguez Escobar, Ladina Saboz, Ana Sojo, Carmen Solorio, Anja Stuckert, Antonio Ugalde, and Rocío Zegarra. I gratefully acknowledge the valuable support of all these individuals and institutions but take full responsibility for what is said in this book. Last, but not least, during the four years it took to write this manuscript, my wife Elena gave me her continued support and freed me of many obligations so that I could dedicate more time to writing. She and our three daughters have my solemn promise that this will be my final book, whose long and excruciating process has convinced me that at 73 it is finally time to enjoy retirement.

This page intentionally left blank

CONTENTS

List of Tables and Appendices Abbreviations

PART I.

INTRODUCTION: THE STATE OF PENSION AND HEALTH CARE IN THE REGION BEFORE THE REFORMS

1 Social Security Principles, Enforcement in Latin America, and Modifications by the Reforms 1.1 Introduction 1.2 Principles, regional enforcement, and modifications by the reforms 1.2.1 Universal coverage 1.2.2 Equal treatment 1.2.3 Solidarity and income redistribution 1.2.4 Comprehensiveness and sufficiency of benefits 1.2.5 Unity, state responsibility, efficiency, and social participation in administration 1.2.6 Financial sustainability 1.2.7 Reform assumptions/goals: promotion of national saving and capital markets 1.2.8 Reform assumption: immunity to state and political interference 1.2.9 Conclusions: enforcement of principles in the region before the reforms

PART II.

xx xxiii

3 3 5 5 8 10 12 14 17 21 21 21

PENSION REFORMS AND THEIR EFFECTS

2 Pension Reforms: Taxonomy, Goals, and Actors 2.1 Private and public systems: taxonomy in Latin America 2.2 Types of reforms: structural and parametric 2.2.1 Structural reforms 2.2.2 Parametric reforms or lack of reforms 2.3 Goals and assumptions of pension reforms

27 27 28 28 31 32

xiv Contents

2.4 External influences, internal actors, and politico-economic environment 2.5 One single type of reform/model or several types/models? 3 Effects of Pension Reforms on Universal Coverage 3.1 Statistical coverage of the labor force by the contributory system 3.1.1 Private systems 3.1.2 Public systems 3.1.3 Comparison of private and public systems coverage 3.2 Legal and statistical coverage of groups difficult to affiliate 3.2.1 Urban informal sector 3.2.2 Rural sector 3.2.3 Measures to improve coverage of difficult groups 3.3 Social assistance pensions for the uninsured and impact on poverty 3.4 Coverage of the older population 3.5 Impact of the reforms on coverage 3.5.1 Coverage in private and public systems, and regional trends 3.5.2 Groups difficult to cover 3.5.3 Social assistance and poverty 3.5.4 Coverage of the elderly 4 Effects on Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 4.1 Equal treatment 4.1.1 Survival of privileged schemes and their unequal entitlement conditions 4.1.2 Gender inequality 4.2 Solidarity and income distribution 4.2.1 Solidarity versus equivalence 4.2.2 Mechanisms for and against solidarity 4.3 Comprehensiveness and sufficiency of benefits 4.3.1 Retirement ages and spans, and contribution years in all systems 4.3.2 The pension formula in public pensions 4.3.3 Adjustment of private and public pensions 4.3.4 Levels of noncontributory pensions in private and public systems 4.3.5 Are private pensions higher than public pensions?

32 34 36 36 36 39 39 40 41 45 46 47 51 52 52 53 54 55 58 58 58 59 63 63 64 67 68 70 72 73 74

Contents xv

4.4 Impact of the reforms on equal treatment, solidarity, and comprehensiveness/sufficiency 4.4.1 Equal treatment 4.4.2 Solidarity 4.4.3 Comprehensiveness and sufficiency 5 Effects on Unity, State Responsibility, Efficiency, Costs, Social Participation, and Reform Goals 5.1 Unification versus segmentation 5.2 The role of the state 5.3 Reform goals: freedom of choice and privatization 5.4 Reform goal: competition 5.5 Information 5.6 Efficiency 5.7 Administrative costs 5.8 Social participation in the administration 5.9 Impact of the reforms on unity, state responsibility, efficiency, costs, social participation, and reform goals 5.9.1 Most systems are still segmented or highly segmented 5.9.2 The state role continues to be fundamental rather than subsidiary 5.9.3 Reform goals: freedom of choice and privatization 5.9.4 Reform goal: competition does not work properly or does not exist 5.9.5 The insured lack information 5.9.6 Uneven gains in efficiency 5.9.7 Higher administrative costs in private than in public systems 5.9.8 Absence of social participation in private systems 6 Effects on Financial Sustainability and Reform Goals 6.1 Contributions 6.2 Compliance 6.3 Components of fiscal costs in private systems in the transition and beyond 6.3.1 Operational deficit of the public system 6.3.2 Recognition bond 6.3.3 Minimum pension 6.3.4 Other state guarantees 6.4 Estimates and projections of fiscal costs in private systems 6.5 Financial and actuarial equilibrium in public systems 6.6 Population aging

76 76 77 78 83 83 87 89 93 95 97 98 102 105 105 106 106 107 107 108 108 109 112 112 116 120 120 122 122 123 123 126 131

xvi Contents

6.7 Reform goals/assumptions: private systems promotion of national saving, capital markets and returns, and immunity against political interference 6.7.1 Capital accumulation 6.7.2 Increase in national saving 6.7.3 Development of the capital markets and portfolio diversification 6.7.4 Capital returns 6.7.5 Immunity against state and political interference 6.8 Impact of the reforms on financial sustainability and reform goals/assumptions 6.8.1 Private systems impose higher contributions on workers 6.8.2 Compliance deterioration 6.8.3 Projection of fiscal costs in private systems 6.8.4 Financial and actuarial equilibrium in public systems 6.8.5 Population aging 6.8.6 Reform goals/assumptions: promotion of national saving, capital markets and capital returns, and immunity against state-political interference

PART III.

132 133 134 136 140 143 145 145 145 146 146 147

147

HEALTH CARE REFORMS AND THEIR EFFECTS

7 Health Care Reforms: Taxonomy, Objectives, and Actors 7.1 Health care and pension reforms compared 7.2 Definition of health care reform 7.3 Historical summary of the reforms 7.4 Reform goals 7.5 Taxonomy of health care systems and their reforms 7.6 External influences, internal actors, and political environment 7.7 Measuring the effects of health care reforms

155 155 158 159 159 160

8 Effects of Health Care Reforms on Universal Coverage 8.1 Legal coverage and targets 8.1.1 Coverage by the three sectors 8.1.2 Fulfillment of reform coverage targets 8.2 Estimates of population coverage 8.2.1 Problems to estimate coverage 8.2.2 Analysis of the available data on coverage: currently and before the reform

167 168 168 173 173 176

161 163

177

Contents xvii

8.2.3 Rough estimate of regional coverage 8.2.4 Trends in coverage by sector 8.3 Coverage by location, urban–rural zones, and ethnic groups 8.4 Coverage by income level 8.5 Coverage of the self-employed and other groups difficult to incorporate 8.6 Access and utilization 8.7 Causes of poor coverage and nonfulfillment of targets 8.8 Impact of the reforms on coverage and fulfillment of targets 8.8.1 Legal coverage, groups difficult to incorporate and fulfillment of targets 8.8.2 Population coverage and comparisons before and after the reform 8.8.3 Geographical, ethnic and income disparities in coverage 8.8.4 Access and utilization 8.8.5 Causes for low coverage and nonfulfillment of targets 9 Effects on Equal Treatment, Solidarity, and Comprehensiveness/Sufficiency 9.1 Equal treatment 9.1.1 Standardization of benefits and subsisting disparities 9.1.2 Financial equity by income levels and geographic areas 9.1.3 Inequalities in human and physical resources and in health indicators: geographic areas, health sectors, and ethnic groups 9.1.4 Gender inequalities in access and care 9.2 Solidarity and income distribution 9.2.1 Solidarity vis-à-vis equivalence 9.2.2 Mechanisms in favor and against solidarity 9.3 Comprehensiveness and sufficiency of benefits 9.3.1 Basic package of benefits 9.3.2 Catastrophic illnesses and high-complex procedures 9.3.3 Perceived quality of benefits 9.4 Impact of the reforms on equal treatment, solidarity, and comprehensiveness/sufficiency 9.4.1 Equal treatment and equity 9.4.2 Solidarity 9.4.3 Comprehensiveness/sufficiency

179 179 180 181 184 186 189 193 193 195 196 197 198 203 203 204 206

209 211 212 213 214 217 217 224 225 227 227 229 230

xviii Contents

10 Effects on Unity, State Responsibility, Efficiency, Costs, Social Participation, and Reform New Goals 10.1 Unity or integration of the system 10.1.1 Integration and coordination in public sector and social insurance 10.1.2 Multiplicity in the private sector 10.2 Reform goal: separation of functions 10.3 Role of the state: regulation and supervision 10.4 Reform goal: decentralization 10.4.1 Degree of decentralization achieved 10.4.2 Decentralization of functions 10.4.3 Evaluation of decentralization results 10.5 Reform goal: competition 10.6 Reform goal: privatization 10.7 Reform goal: freedom of choice 10.8 Efficiency 10.8.1 General indicators of efficiency 10.8.2 Major problems of efficiency 10.8.3 Measures to improve efficiency 10.9 Administrative costs 10.10 Social participation in the administration 10.10.1 Social participation in the reform process 10.10.2 General evaluations of social participation 10.10.3 Participatory bodies, representation, functions, and effectiveness 10.11 Impact of the reforms on unity, state responsibility, efficiency, costs, social participation, and reform goals 10.11.1 Integration and coordination 10.11.2 Reform goal: separation of functions 10.11.3 State role: regulation and supervision 10.11.4 Reform goal: decentralization 10.11.5 Reform goals: competition, privatization, and freedom of choice 10.11.6 Efficiency and administrative costs 10.11.7 Social participation 11 Effects on Financial Sustainability and Efficacy 11.1 Financial sustainability 11.1.1 Health expenditure trends, distribution by sectors and per capita 11.1.2 Sources of financing 11.1.3 Compliance 11.1.4 Reform goal: shift from supply to demand subsidies

235 236 236 240 241 246 248 249 252 253 257 261 264 266 267 270 271 273 276 277 278 279

284 284 284 285 285 287 288 289 293 293 294 299 308 313

Contents xix

11.1.5 Financial and actuarial balance 11.1.6 Impacts of the reforms on financial sustainability 11.2 Efficacy 11.2.1 Unfeasibility to measure the reform impact on efficacy 11.2.2 Evolution of health indicators between 1990 and 2002

PART IV.

315 320 324 324 325

TOWARDS A BETTER SOCIAL SECURITY IN THE FUTURE

12 Policies on Pensions and Health Care 12.1 Pensions 12.1.1 Coverage 12.1.2 Equal treatment 12.1.3 Solidarity and income redistribution 12.1.4 Comprehensiveness and sufficiency of benefits 12.1.5 Unity, state responsibility, efficiency, costs, social participation, and reform goals 12.1.6 Financial sustainability 12.1.7 Reform goals: promotion of national saving, capital markets, and capital returns 12.1.8 Reform assumption: immunity against state interference 12.2 Health care 12.2.1 Coverage 12.2.2 Equal treatment 12.2.3 Solidarity and income redistribution 12.2.4 Comprehensiveness and sufficiency of benefits 12.2.5 Unity, state responsibility, efficiency, costs, social participation, and reform goals 12.2.6 Financial sustainability 12.2.7 Impact on health indicators (efficacy)

335 335 336 341 342 343 345 349 352 355 355 357 362 364 365 367 376 381

Appendices

383

Bibliography

387

Index

412

LIST OF TABLES AND APPENDICES

2.1 Models and characteristics of private and public pension systems in Latin America, 2006 3.1 Percentage of the labor force covered by private systems only, based on affiliates and active contributors, December 2005 3.2 Percentage of the labor force covered by private and public contributory pension systems, based on active contributors 3.3 Population groups difficult to cover by pensions between 2001 and 2004, and social assistance pensions in force in 2006 3.4 Legal coverage on pensions of groups difficult to incorporate, 2005–6 3.5 Percentage of the populations aged 65 and above covered by pensions in private and public systems, between 2000 and 2003 4.1 Normal ages of retirement and retirement span in private and public pension systems, c.2005 4.2 Other entitlement conditions for old-age pensions in private and public systems, 2005 5.1 Unification versus segmentation in private and public pension systems, 2005 5.2 Degree of privatization: distribution of contributors between private and public pension systems, c.2004 5.3 Freedom of choice of the insured in private pension systems, 2005–6 5.4 Competition in private pension systems: size of insured market, number of administrative firms, and concentration, December 2005 5.5 Administrative costs in private systems as percentage of wages, December 2005 5.6 Administrative costs as percentage of total taxable wages in private and public pension systems, between 2001 and 2005 5.7 Social participation in the administration of private and public pension systems, 2005–6 6.1 Contributions by employers, workers, and state in private and public pensions systems, c.2005 6.2 Affiliates in private pension systems that contributed in the past month, December 1998 to December 2005, and June 2006 (in percentages) 6.3 Fiscal costs of the structural reform in private pension systems, 2006 6.4 Domestic estimates and projections of fiscal costs in private systems, at the start of the reform or later, compared with those of the World Bank, 2005 (deficit as % of GDP) 6.5 Indicators of financial sustainability in public pension systems, between 2000 and 2004

29 37 38 42 43 50 61 71 85 90 91 93 99 101 103 113

116 121

124 127

List of Tables and Appendices 6.6 Stock of financial assets accumulated in private pension systems in 2005 and reserves in public pension systems, between 2000 and 2005 6.7 Percentage distribution of pension fund in private systems and three public systems by investment in financial instrument, December 2005 6.8 Average annual real gross rates of capital returns in private pension systems and in public systems with invested financial reserves (partly funded: CPC) 7.1 Years of start of the reform and subsequent changes, health system types, and main features of the reforms, 2005–6 8.1 Legal mandatory coverage of health social insurance by type of occupation, dependents, and pensioners, 2005–6 8.2 Health care coverage/access as percentage of the total population before the reform and in 2001–4, and as percentage of the labor force in 1994–8 and 2000–3 8.3 Total population covered by health social insurance distributed by income quintiles, between 1996 and 2003 9.1 Sufficiency of health care benefits and perceived quality, 2003–5 10.1 Degree of unity/coordination or segmentation, number of sectors and separate health-care programs, c.2005 10.2 Separation of functions in health-care systems, c.2005 10.3 Decentralization of the public health sector, between 1998 and 2005 10.4 Indicators of competition and freedom of choice in health systems, between 2000 and 2005 10.5 Indicators of degree of privatization in health-care systems, between 1998 and 2004 10.6 Selected indicators of health-care efficiency, between 1997 and 2004 10.7 Administrative costs as percentage of expenses in the three health-care sectors, between 1999 and 2005 10.8 Social participation in administration of public and social insurance sectors, between 2002 and 2005 11.1 Total health expenditures as percentage of GDP (1997–2003) and distributed by three health sectors, and health expenditure per capita in international dollars, 2003 11.2 Financing sources of health-care systems, c.2005 11.3 Contributors to social insurance in 1990 and 2002, and evasion and payment delays between 1997 and 2005 (in percentages as specified) 11.4 Financial balance in the social insurance and private health sectors, between 2000 and 2004 11.5 Selected health indicators in Latin America, 1990 and 2002 11.6 Ranking of the 20 countries based on health indicators, 1990 and 2002 (ordered from 1 best to 20 worst) Appendix 1 Total population coverage in Argentina: affiliates by health sector, 1989–2004 Appendix 2 Total population coverage in Chile: affiliates by health sector, 1984–2003

xxi

133

137

140 156 169

174 182 222 237 242 250 259 263 268 275 280

295 300

309 316 326 329 383 384

xxii List of Tables and Appendices Appendix 3 Total population coverage in Colombia: affiliates by health regime, 1993– 2002 Appendix 4 Total population coverage in Mexico: insured and uninsured, by health sector and program, 1985–2002 Appendix 5 Comparison of WHO estimates on the percentage distribution of total health-care expenditures in Latin America in 2001, based on 2004 and 2006 series

384 385

385

ABBREVIATIONS

AIOS

ARS

ASS AUGE/GES BCU BPS CCSS CI CONSAR CPC CSS CPI DILOS DRG EAP EBAIS EMP

EPS

ESE ESS EsSalud FAEC FCS

Asociación Internacional de Organismos de Supervisión de Fondos de Pensiones (international association of superintendencies of private pensions) Administradoras del Régimen Subsidiado (administrators of the subsidized health regime), Colombia, and Administradoras de Riesgos de Salud (health care administrators), Dominican Republic Administradoras de Servicios de Salud (health administrators) Guatemala Garantías Explícitas en Salud (basic health care package and other guarantees), Chile Banco Central del Uruguay (Uruguay’s central bank) Banco de Previsión Social (social insurance institute), Uruguay Caja Costarricense de Seguro Social (social insurance institute), Costa Rica Catastrophic illnesses Comisión Nacional del Sistema de Ahorro para el Retiro (superintendence of private pension funds), Mexico Collective partial capitalization (partly funded) financial pension regime Caja de Seguro Social (social insurance institute), Panama Consumer Price Index Directorio Local de Salud (local health board), Bolivia Diagnosis-Related-Groups payments Economically Active Population or labour force Equipos Básicos de Atención Integral en Salud (local basic health teams), Costa Rica Empresas de Medicina Prepaga (prepaid health enterprises), Argentina, and Empresas Médicas Provisionales (health providing enterprises), Nicaragua Empresas Promotoras de Salud (health providing enterprises), Colombia, and Entidades Prestadoras de Salud (health providing enterprises), Peru Empresas Sociales Estatales (state health providing enterprises), Colombia Empresas Solidarias de Salud (health providing enterprises), Colombia Seguro Social en Salud (social insurance health institute), Peru Fundo de Ações Estratégicas e Compensação (compensation health fund), Brazil Fondo de Compensación Solidario (solidarity health fund), Chile

xxiv Abbreviations FF FNR FNS FNS FONASA FOSYGA FSC FSR GDP HCP IADB IAMC IDSS IESS IFO IGSS IHSS ILO IMF IMSS INP INS INSS IPD IPS

ISAPRE ISS ISSA ISSS ISSSTE IVSS

Fully funded financial regime in pensions Fondo Nacional de Recursos (compensation health fund), Uruguay Fundo Nacional de Saúde (national health fund), Brazil Fondo Nacional Solidario (solidarity health fund), Bolivia Fondo Nacional de Salud (public-social insurance health program), Chile Fondo Solidario y de Garantía (solidarity health fund), Colombia Fondo Solidario de Compensación (compensation health fund), Chile Fondo Solidario de Redistribución (solidarity health fund), Argentina Gross Domestic Product Highly-complex procedures in healthcare Inter-American Development Bank Instituciones de Asistencia Médica Colectiva (collective not-for-profit private health providers), Uruguay Instituto Dominicano de Seguros Sociales (social insurance institute) Dominican Republic Instituto Ecuatoriano de Seguridad Social (social insurance institute), Ecuador International Financial Organizations Instituto Guatemalteco de Seguridad Social (social insurance institute), Guatemala Instituto Hondureño de Seguridad Social (social insurance institute), Honduras International Labor Organization International Monetary Fund Instituto Mexicano del Seguro Social (social insurance institute for private workers), Mexico Instituto de Normalización Previsional (institute of standardization of public pensions), Chile Instituto Nacional de Seguros (occupational risks institute), Costa Rica Instituto Nicaragüense de Seguridad Social (social insurance institute), Nicaragua Implicit pension debt Instituto de Previsión Social (social insurance institute), Paraguay, and Instituciones Proveedoras de Servicios (health provider institutions), Colombia Instituciones de Salud Provisional (private health providers), Chile Instituto de Seguro Social (social insurance institute), Colombia International Social Security Association Instituto Salvadoreño del Seguro Social (social insurance institute), El Salvador Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado (social insurance institute for civil servants), Mexico Instituto Venezolano de los Seguros Sociales (social insurance institute), Venezuela

Abbreviations IMAE MINSA MINSAL MINSALUD MINSAP MSDS MSPAS

MSPBS MSPP NDC NGO OISS ONA OS PAB PABA PAC PAHO PAYG PBS PCSB PMO POS POSS RGPS PSF PSS

SAFJP SAFP

xxv

Institutos de Medicina Altamente Especializada (complex medicine providers), Uruguay Ministerio de Salud (ministry of health), Costa Rica, Panamá, Peru Ministerio de Salud (ministry of health), Chile Ministerio de Salud (ministry of health), Colombia Ministerio de Salud Pública (ministry of health), Cuba Ministerio de Salud y Desarrollo Social (ministry of health), Venezuela Ministerio de Salud Pública y Asistencia Social (ministry of health), Guatemala, and Ministerio de Salud Pública y Asistencia Social (ministry of health), El Salvador Ministerio de Salud Pública y Bienestar Social (ministry of health), Paraguay Ministerio de Salud Pública y Población (ministry of health), Haiti Notional Defined Contribution System Non Government Organizations Organización Iberoamericana de Seguridad Social Office National d’Assurance Vieillesse (old-age social insurance), Haiti Obras Sociales (social insurance health care providers), Argentina Piso de Atenção Básica (basic health care package), Brazil, and Plan de Atención Básica (basic health care package), Colombia Piso de Atenção Básica Ampliado (expanded basic health care package), Brazil Planes de Atención Complementaria (supplementary health care plans), Colombia Pan American Health Organization Pay-as-you-go financial pension regime Plan Básico de Salud (basic health care package), Dominican Republic Programa de Cuidados Sanitarios Básicos (basic healthcare package) Paraguay Programa Médico Obligatorio (basic health care package), Argentina Plan Obligatorio de Salud (basic healthcare package of contributory regime), Colombia Plan Obligatorio de Salud del Régimen Subsidiado (basic health care package of the subsidized regime), Colombia Regime Geral de Previdência Social, Brazil Programa de Saúde da Família (family healthcare program), Brazil Proveedores de Servicios de Salud (health care providers), Dominican Republic, and Prestadoras de Servicios de Salud (healthcare providers), Guatemala Superintendencia de Fondos de Jubilaciones y Pensiones (superintendence of private pensions), Argentina Superintendencia de Fondos de Pensiones (superintendence of private pensions), Chile, Peru

xxvi Abbreviations SBC SBS SENASA SESPAS SFS SIAB SIAS SIBASI SILAIS SILOSS SIS SISBEN SNSS SP SPNS SPS SSA SSC SUMI SUS SPVS SSS UCS UPC WHO

Superintendencia Bancaria de Colombia (superintendence of banking and pensions). Seguro Básico de Salud (basic healthcare package), Bolivia Seguro Nacional de Salud (national healthcare insurance), Dominican Republic Secretaría de Salud Pública y Asistencia Social (ministry of health), Dominican Republic Seguro Familiar de Salud (family health insurance), Dominican Republic Sistema de Informação de Atenção Básica (health information system), Brazil Sistema Integrado de Atención a la Salud (integrated healthcare system), Guatemala Sistemas Básicos de Salud Integral (local basic health program), El Salvador Sistemas Locales Integrados de Salud (local basic health program), Nicaragua Sistemas Locales de Seguridad Social (local basic health program), Honduras Seguro Integrado de Salud (integrated health insurance), Peru Sistema de Información de Salud (health information system), Colombia Sistema Nacional de Servicios de Salud (public national health care service network), Chile Superintendencia de Pensiones (superintendence of pensions), Costa Rica, El Salvador Sistema Público Nacional de Salud (national health system), Venezuela Seguro Popular de Salud (popular health insurance), Mexico Secretaría de Salud (ministry of health), Mexico Seguro Social Campesino (peasants social insurance), Ecuador Seguro Universal Materno-Infantil (maternal-infant health insurance), Bolivia Sistema Único de Saúde (unified health system), Brazil Superintendencia de Pensiones, Valores y Seguros (superintendence of pensions, securities and insurances), Bolivia Superintendencia de Seguridad Social (superintendence of public social security), Chile Unité Communal du Santé (communal health boards), Haiti Unidad por Capitación (capitation unit), Colombia World Health Organization

PART I

INTRODUCTION: THE STATE OF PENSION AND HEALTH CARE IN THE REGION BEFORE THE REFORMS

This page intentionally left blank

1

Social Security Principles, Enforcement in Latin America, and Modifications by the Reforms

This chapter first describes the conventional principles of social security, tracing its antecedents and focusing on the approved norms by the International Labour Organization (ILO); it also assesses whether such principles were in force in Latin America in 1980, at the eve of the so-called ‘lost decade’ and the start of the pioneering structural social security reforms in Chile, and detects trends in the early 1990s when reforms started in other countries of the region. Finally the chapter summarizes the modifications in conventional social security systems by the pension and health care reforms as well as the new goals or assumptions introduced by such reforms.

1.1. Introduction More than a century has elapsed since the trilogy of social insurances against the social risks of old-age, disability, and death was introduced in Germany by Otto von Bismarck in 1883–9, a model later expanded to Europe and other industrialized countries, based on the principles of compulsory nature, contributions by employers and workers, and state regulatory role. In 1919, at the end of World War I, the ILO was founded and enthroned social insurance as a key tool for the pro-

tection of workers and their families, starting with maternity care; the first generation of ILO agreements was based on the concept of social insurance and applied to certain groups of workers. In 1935 the United States enacted the Social Security Act, the first law to use that term, replicated by New Zealand in 1938 (ILO-ISSA 2001a; Humblet and Silva 2002).1 But the modern concept of the term social security was developed by William Beveridge in his report Social Insurance and Allied Services published in 1942, which proposed a ‘social security plan’ integrating social insurances, social assistance, and voluntary supplementary insurances. His report identified six principles, including the unification of administrative responsibility, comprehensiveness of benefits, and a flat rate of contribution (Beveridge 1942). In 1944, when World War II was ending, the ILO Declaration of Philadelphia raised social security to the category of international instrument and proclaimed the need to expand its coverage. The Universal Human Rights Declaration of 1948 established that all members of society have an individual right to social security, defined by the ILO as ‘the protection that society provides to all its members, through a series of public measures against economic and social deprivation that otherwise would provoke the disappearance or drastic reduction of income

4 Reassembling Social Security due to sickness, maternity, occupational accidents and diseases, unemployment, disability, old age and death, as well as protection in the form of medical assistance and aid to families with children’ (ILO-ISSA 2001a). The International Conferences of the ILO, with ‘tripartite representation’ (workers, employers, and governments), approved several ‘Conventions’ and ‘Recommendations’ reinforcing social security principles.2 The second generation of these instruments was approved after World War II, based on the wider concept of social security and geared to all the population. The most important resolution (No. 102 of 1952) established the so-called ‘minimum norm’ of social security that integrated various previous instruments and set basic requisites related to risks, benefits, and entitlement conditions. Other important groups of conventions and recommendations were approved in the 1960s and 1980s and the most recent in 2000.3 Both types of instruments contain the conventional principles of the ILO that reigned in the world without any significant challenge until the end of the 1980s. Economic, social, and demographic changes that occurred in the last two decades of the twentieth century, combined with the influence on social security by international and regional financial organizations (World Bank, International Monetary Fund—IMF, InterAmerican Development Bank—IADB) and the globalization process have reassembled the traditional social security and created a new paradigm, through structural reforms some of which deviate from the conventional principles and have new, different goals and assumptions. The ILO Conference of 2001 approved a convention declaring that social security continues to rely on its essential principles but facing new challenges: ‘there is no single right model of social security’ in the world, ‘each society must determine how best to ensure income security and access to healthcare’, but ‘all systems should conform to certain basic principles’ (ILO 2001a: 2). A review of all international social security conventions and recommen-

dations done by the ILO in 2001–2 concluded that they are up to date and pertinent, albeit some require adaptation and all must be better diffused (Humblet and Silva 2002). Latin America introduced its social insurance programs just before other developing nations in Africa, Asia, and the Middle East; at the end of the 1970s all Latin American countries had such programs in force but with significant differences among them. In 1980, before Chile started structural reforms, the countries were classified and ranked in three groups by the author, based on the date of inception of their first social insurance programs of pensions and sickness-maternity, as well as their degree of development (measured by eleven indicators): (a) pioneer or high, (b) intermediate, and (c) latecomer or low. The characteristics of the three groups and the rank of each country within each group were as follows. The pioneer-high group (Uruguay, Argentina, Chile, Cuba, Brazil, and Costa Rica4 ) was the first to introduce social insurances in the region (in the 1920s and 1930s), by 1980 had the highest coverage and development in such programs, as well as aging populations with the highest life expectancy, but their systems suffered stratification, high costs, growing deficit, and financial and actuarial disequilibria. The intermediate group (Panama, Mexico, Peru, Colombia, Bolivia, Ecuador, and Venezuela) introduced their programs mostly in the 1940s and 1950s, influenced by the Beveridge report and the ILO conventions, and by 1980 had reached a middle level of coverage and development in their programs, which were less stratified, had lower costs, and faced a better financial situation than those in the first group, although some of them were approaching disequilibria. The latecomerlow group (Paraguay,5 Dominican Republic, Guatemala, El Salvador, Nicaragua, Honduras, and Haiti) was the last to implement their programs (in the 1960s and 1970s) and by 1980 their populations were the youngest and their life expectancy the shortest, their

Social Security Principles programs were relatively more unified and suffered fewer financial problems than the other two groups, but endured a lower coverage and were the least developed (Mesa-Lago 1989). The notable differences among these three groups had an impact on the enforcement of social security principles. The principles identified and described in this chapter are not always specified and grouped together exactly in the same fashion by the ILO. Based on several ILO documents, however, the author has classified, clustered some principles, and arranged all of them in a logical sequence, an order that does not necessarily imply ranking or priority, because the set of principles is highly interrelated. The six fundamental principles are: (a) universal coverage; (b) equal treatment; (c) solidarity and income distribution; (d) comprehensiveness and sufficiency of benefits; (e) unity, state responsibility, efficiency, and social participation in the administration; and ( f ) financial sustainability.6 The problems confronted by social security at the start of the 1990s created a fertile terrain for reforms. In addition, neoliberalism, globalization, international competition, population aging, and other factors pushed for such reforms. The important financial components of social security, both from the revenue and expenditure sides generated an interest in international financial organizations to become involved in the field. Two important reports of the World Bank influenced the reforms in the world and the region: Investing on Health (1993) and Averting the Old-Age Crisis: Policies to Protect the Old and Promote Growth (1994). Other actors involved in the reforms are analyzed in Chapters 2 and 7 of this book. The rest of this chapter systematically traces the general evolution and summarizes each principle, evaluates its enforcement in Latin America before the reforms, and identifies the modifications introduced in the conventional principle and the new goals and assumptions introduced by pension and health care reforms.

5

1.2. Principles, regional enforcement, and modifications by the reforms 1.2.1. Universal coverage 1.2.1.1. Evolution of the principle Beveridge (1942: 122) defined this principle as ‘comprehensiveness in respect both of the persons covered and their needs’ and considered it a tool to abolish poverty. The Declaration of Philadelphia of 1944 proclaimed that all members of society should be covered, hence the need to extend social security to guarantee basic income to those who need it and provide integral health care to all. An ILO recommendation in 1944 stated that social insurance should protect ‘all salaried and independent workers and their dependent family.’ Conventions 102 of 1952 and 128 and 130 of 1967 respectively determined minimum coverage for persons in monetary benefits (particularly pensions) and health care (and sicknessmaternity monetary benefits), based on three alternative percentages: (a) in pensions 50% of the total number of salaried workers, as well as their spouses and children, and in health 100% of all the salaried labor force; (b) in pensions 20% of the economically active population (EAP or labor force) that is resident, and in health 75% of the EAP, together with their spouses and children; and (c) in pensions 50% of all residents, and in health 75% of all residents, but when dealing with monetary benefits the scope is limited to all residents whose resources do not exceed a certain limit.7 To facilitate ratification, these instruments provided flexibility, allowing the exclusion from such minima of groups difficult to incorporate, such as temporary, family without payment, home and agricultural workers. Temporary exceptions are also granted to developing countries, restricting coverage to industrial enterprises with a given number of workers, although the countries must periodically report to the ILO whether the causes

6 Reassembling Social Security of said exceptions persist or not (ILO 2001a; Humblet and Silva 2002). Coverage began in organized groups, large enterprises, formal employers, and urban salaried workers, who were the easiest and fastest to identify, affiliate, collect contributions from, and keep records and individual accounts on (in pensions). The ILO proposed that coverage would be gradually extended, but it could not be done in many countries due to the predominance of small enterprises, domestic service, self-employment, temporary and home work, and subsistence agriculture. Coverage largely depends on the degree of industrialization, the size of the formal sector of the economy, and the age of the program, therefore it is considerably higher in the most developed countries and lower in the least developed. Economic crises, structural adjustment reforms, globalization, and other factors have generated higher unemployment, as well as the growth of jobs without social security protection, such as partial, occasional or temporary, self-employment, in micro enterprises and the informal sector, provoking a decline in coverage and new challenges to that principle (ILO 2000a; ILO-ISSA 2001a). In 2000 the ILO acknowledged that the extension of coverage was the major test confronted by social security systems; to meet such challenge and cope with the problems explained above, in 2001 the Conference launched a campaign to expand and improve coverage, although each country has to choose the strategy to accomplish social security for all. ‘The state has a priority role in the facilitation, promotion and extension of coverage of social security . . . to those who are not covered by existing systems’ (employees in small enterprises, self-employed, migrants, and informal workers, particularly women), taking into account the diverse needs and contributory capacities of these groups. When they cannot be covered in an immediate and compulsory manner, it is necessary to introduce voluntary coverage, micro insurance, or social assistance (ILO 2001a: 2–4).

1.2.1.2. Enforcement in Latin America Before the social security reforms in Latin America, the principle of universal coverage ruled in a few countries. Measuring coverage in each country and comparisons were difficult for the following reasons: estimates were based on affiliates in some countries and on contributors in others; some included virtually all the insured while others excluded those under separate schemes; in the few countries with social assistance pensions, reported coverage was limited to the contributory program excluding noncontributory beneficiaries; and health care coverage was considerably less precise than that of pensions because of unreliable data on the dependent family and those effectively protected in the public health sector. There were no historical comparative series published by the ILO or other international or regional organizations, therefore the following are author’s estimates for 1980. The six countries of the pioneer-high group had attained or were close to universal coverage, stretching from 67 to 96% of the total population in sickness-maternity and from 62 to 87% of the EAP in pensions.8 The intermediate group had lower coverage in both programs and only three countries reached about half of the total population and the labor force, fluctuating from 10 to 54% in sicknessmaternity and from 19 to 52% in pensions. The latecomer-low group endured the lowest coverage, ranging from 4 to 19% and from 8 to 36%, respectively (excluding Haiti). Pension coverage increased in all countries in 1970– 80 (largely due to the urbanization process and expansion of the salaried labor force), except in three; it was not possible to do a similar comparison on health care due to lack of data in 1970. Despite economic crisis and the expansion of the informal sector in 1980– 90, coverage of sickness-maternity estimated in sixteen countries grew in thirteen and declined in three, while pension coverage estimated for seventeen countries rose in eleven and was stagnant or fell in six (Mesa-Lago and Bertranou 1998).9 In all countries, the public

Social Security Principles health sector was expected to provide care for the poor and low-income uninsured, a segment considerably smaller in the pioneer-high group than in the other two groups where the majority of the population was uninsured and hence more difficult to protect. Only the six countries of the pioneer-high group had social assistance or noncontributory pensions, which increased their coverage figures, but the remaining fourteen countries lacked those programs despite wider poverty incidence (Mesa-Lago 1989). Estimates of coverage in Latin America allow an approximate application of ILO prescribed minima. The weighted regional average of coverage was 61.2% in 1980, both of the EAP in pensions and the total population in health care; excluding Brazil (that had 54% of the regional EAP and total population), the regional average decreased to 42.7% (Mesa-Lago 1989). The latter was still twice the ILO prescribed minimum of 20% of the EAP in pensions, but short of the minimum of 75% of the resident population in health care. Weighted regional averages circa 1990 (before the reforms began, except in Chile) were 63.4% of the EAP in pensions and 63.8% of the total population in health care, an increment of more than two percentage points in each over 1980. Excluding Brazil (that had 45% of the regional EAP and 35% of the regional population in 1990), coverage in Latin America was 44.4% in pensions and 51.8% in health care, an increase over 1980 and above the ILO minimum in pensions but still below the minimum coverage in health care (coverage from Mesa-Lago and Bertranou 1998; EAP and total population from ECLAC 2003c).10 These data were limited to the major contributory programs and hence excluded coverage in separate schemes and by social assistance; adding those groups coverage increased significantly in some countries, for instance in Costa Rica from 76 to 83% and in Uruguay from 69 to 86% (Mesa-Lago 1989; ISSA 2003b). Furthermore data also excluded protection by the public health sector. The ILO (2000a) estimates that ‘social security systems in the majority of developing countries

7

cover less than half of salaried employees’, hence Latin America was above developing countries in this parameter. Regional averages of coverage, however, hid significant differences among countries in 1990. Pension coverage oscillated from 73 to 87% of the EAP in the pioneer-high group, 13 to 64% in the intermediate group, and 9 to 27% in the latecomer-low group, while health care coverage of the total population in the three groups were from 74 to 96%, 16 to 58%, and 6 to 22%, respectively. Only in five countries was pension coverage lower than 20% of the EAP but in ten countries health care coverage was lower than 50% of the population (Mesa-Lago and Bertranou 1998). Factors that explain differences in coverage are the degree of industrialization and urbanization, the size of the formal or salaried sector, and the age of the system. Countries in the pioneer-high group have the highest proportions of industrialization, urbanization, and salaried labor force and the oldest systems, while the opposite is true of countries in the latecomer-low group where the majority of the EAP is informal or agricultural and is excluded from coverage. In the least developed countries (all in the latecomer-low group and part of the intermediate group) coverage began in the capital city, later expanded to urban areas, and then to large agricultural plantations geared to exports, and so forth (Mesa-Lago 1994).

1.2.1.3. Modifications by the reforms Supporters of the reforms question the financial capability of social security or public systems to implement the principle of universal coverage, arguing that they encourage individuals to depend on the state for their social protection thus harming their initiative, and giving two reasons for the low population coverage in many countries: the social security contribution is perceived as a tax instead of a saving or an advanced fee for services, and high contribution rates and their disconnection with individual risks and the amount and quality of benefits discourages

8 Reassembling Social Security affiliation and creates incentives for evasion. They also contend that the population will have a stronger incentive to affiliate to private insurance firms and providers than to public systems, hence increasing coverage because the shift from a public pension system to a private one will reestablish the broken link between contributions and benefits, stop the flow of workers to the informal sector, and infuse workers’ incentives to participate in the formal sector. In addition fiscal resources that are inappropriately allocated to social security will be targeted on the most vulnerable groups of the population thus expanding coverage (World Bank 1993, 1994; Mitchell 1998). According to World Bank officials the ‘literature on pension reform clearly states that increasing coverage is both an objective and a predicted result of implementing a multipillar system with a large privately funded component’ (Gill, Packard, and Yermo 2005: 98).11 Health care reformers claim that the constitutional principle of universal coverage is not enforced in practice and that it is better to target scarce resources on the poor through a basic package of benefits. But World Bank officials now sustain that public support for the reforms of the 1990s might have been undermined because they focused almost purely on instruments for improving efficiency (payment mechanisms, shift from supply to demand subsidies, decentralization, autonomy of units, and separation of functions) rather than on specific benefits for the target population. A new more effective approach is to grant the people the right to explicit entitlements or specific health benefits that can be demanded by users and must be guaranteed and financed by the state (Baeza and Packard 2005).

1.2.2. Equal treatment 1.2.2.1. Evolution of the principle Beveridge recommended a uniform flat rate of contribution for all insured and standardized comprehensive health services to all

members of the community, both regardless of income although adjustable according to different characteristics of the insured groups (wage earners, self-employed, employers, farmers, housewives), and those who postponed their retirement and kept contributing would have a higher pension (although health care would be the same for all). Critics alleged that such a policy would harm individual initiative and benefits would be very low, refuted by Beveridge with the arguments that all the population would have access to a subsistence benefit, that higher benefits would be financially unfeasible and that those with resources could resort to supplementary voluntary insurance. Another approach to equal treatment is ‘equity’ understood as the elimination of unfair and avoidable differences in the access to benefits. The Universal Declaration of Human Rights banned any type of discrimination based on race, gender, language, religion, politics, nationality, ownership, income, or other causes. The ILO considers that ‘social security should promote and be based on the principle of gender equality’ and ‘equal treatment is a guiding principle of social security’ (Greber 1997: 19). Few countries, however, implanted equal benefits as recommended by Beveridge and the majority established a linkage between contributions and pensions. Highincome groups were initially excluded with the rationale that they were self-protected or could pay private provision but, when they were eventually incorporated, such linkage was reinforced, although some countries set ceilings to contributions and pensions. Entitlement conditions were expected to be equal for all insured, avoiding any discrimination by occupation but certain groups of workers that already had enterprise or guild health plans, superior to that of the general social insurance program, resisted the giving away of their advantages. Civil servants frequently had separate schemes with better benefits, making it difficult to incorporate them into the general program. In most countries there were three separate health care sectors (public, social insurance, and private) with significant

Social Security Principles differences in access to and quality of services. Finally there was gender discrimination resulting from wage differences between men and women who perform the same job, and the fact that women devote a considerable time to raising children and caring for older or disabled relatives without receiving payment and not contributing to social security. Until the end of the twentieth century gender equality had not been actively promoted by any ILO instrument (Greber 1997; ILO-ISSA 2001a). The Conference convention of 2001 declared that ‘social security should promote and be based in the principle of gender equality’; to that end it is not sufficient that the law stipulates the equal treatment for both sexes under similar situations, but also that effective measures are taken to guarantee in practice such right to women. Female access to employment will support granting them social security benefits in their own right instead of through a dependency relationship with an insured male. Women should be paid for their contribution to raise children and take care of other relatives. ‘Each society should consider introducing positive discrimination in favour of women where systemic [negative] discrimination is faced’ (ILO 2001a: 3–4).

1.2.2.2. Enforcement in Latin America Latin American countries in the pioneer-high group (and some in the intermediate group) had, since the end of the nineteenth century, programs that protected civil servants and military men, a similar evolution to the rest of the world as acknowledged by the ILO (2000b: 23). Social insurances evolved in a fragmented manner through multiple schemes that gradually and exclusively covered diverse occupational groups, each with its own insured and legislation, generating a stratified system with unjustified differences in coverage, entitlement conditions, financing, and benefits, contrary to the principle of equal treatment. The best organized, most powerful groups gained earlier and virtually full coverage, more generous entitlement conditions,

9

better benefits and fiscal subsidies, while the opposite was true of the least organized and powerful groups.12 Between 1960 and the 1980s, processes of unification of schemes and standardization of entitlement conditions of divergent degree took place in all pioneer-high countries although maintaining some important inequalities. Intermediate countries’ programs were established later, learned from the problems afflicting the pioneers, were influenced by the Beveridge report, and established a general social insurance program, but some of the previously existing schemes survived and new separate schemes created. Latecomerlow countries established less stratified systems albeit with important exceptions (MesaLago 1978, 1989; see Section 1.2.5 ). Throughout Latin America the armed forces had separate pension and sickness-maternity programs and/or hospitals, except Costa Rica that disbanded them and Panama that integrated them into the general program; civil servants often had their own separate schemes, a phenomenon that also occurred in developed countries. Both groups usually enjoyed more liberal entitlement conditions and benefits than in the general program, for instance earlier retirement based on years of service regardless of age, as well as a more generous formula to estimate pensions, and higher benefits. Other occupational groups with privileged schemes were congressmen, judges, teachers, and university professors, as well as employees in banking, public utilities, petroleum, and social security institutions. Conversely, agricultural workers, selfemployed and domestic servants, when they were covered, endured stricter entitlement conditions and meager benefits. As in most of the world, social insurance pensions in the region were based on certain linkage between contribution and benefit (Mesa-Lago and Bertranou 1998). Usually there were differences in the quality of health facilities and services among various groups: the hospitals and services for the armed forces were often the best (in some countries together with those of a small private sector), followed by those of social

10

Reassembling Social Security

insurance and, lastly, by those of the ministry of health in charge of the uninsured.13 It was also frequent that the best services and personnel were found in the capital city (where they were concentrated), as well as in urban zones and the most developed regions, whereas rural areas and the least developed regions had the worst services and scarce personnel, when they had them. Latin America achieved important advances in health, but with significant variance among countries, as well as among occupational groups, regions, and areas within one country (MesaLago 1992; PAHO 1998a). Most pension programs established a retirement age for women five years younger than for men, which combined with a female longer life expectancy and lower contribution density, resulted in a lower pension. Nevertheless, many countries set an equal minimum pension for both sexes and the pension formula used unisex mortality tables, thus facilitating transfers from male to female insured (Bertranou and Arenas de Mesa 2003).

1.2.2.3. Modifications by the reforms A positive measure of pension reforms is that they usually standardize entitlement conditions and the formula to estimate the benefit in public systems as well as in the new private systems. Conversely, health care reforms usually don’t have among their objectives such standardization and in practice pursue the opposite goal: to provide diversified packages of benefits to individuals and groups, except for a universal basic package. The immense majority of reforms has neither attempted to incorporate separate pension schemes for powerful groups with superior entitlement conditions and benefits nor pursued the integration of the three health sectors and standardization of their differentiated services (except for the basic package). The elimination or reduction of gender discrimination to achieve equality in access and benefits is disregarded by the reforms, despite the fact that some elements of private systems

may accentuate such inequalities, like paying lower pension annuities for women, who live longer than men, and adverse selection in health care plans because of the higher costs of pregnancy care.

1.2.3. Solidarity and income redistribution 1.2.3.1. Evolution of the principle Under Beveridge’s viewpoint all the population must be affiliated and contribute to social security to guarantee its sustainability and there should be solidarity between generations and between the healthy and the sick, to generate a progressive redistribution. He argued that increases in production alone were insufficient to eradicate poverty and raise living standards, but an adequate distribution of the national product was required too. Social security would contribute to those goals in two ways: (a) tripartite financing through contributions from workers and employers (the selfemployed only pays his own contribution) and fiscal subsidies based on general taxes; and (b) a universal flat benefit that helps the lowerincome groups combined with social assistance for the poor totally state financed. Additional voluntary insurance was to be financed by the insured or their trade unions or mutualaid societies and not entitled to fiscal subsidies. ‘The Plan for Social Security is first and foremost a method of redistributing income, so as to put the first and most urgent needs first, so as to make the best possible use of whatever resources are available’ (Beveridge 1942: 170). There is a strong interrelation between the principles of universality, equal treatment, and solidarity, so when one of them does not work properly it affects the other two. According to the ILO, the principle of solidarity calls for all workers to join and contribute to the general system, but there are groups with separate schemes that resist such integration. From an economic viewpoint, the objective and global effect of social security

Social Security Principles should be horizontal and vertical redistribution of income.14 Horizontally, the healthy and active transfer resources to the sick, disabled, or retired, an effect only significant, however, when coverage is universal or quite extended but nil or reduced when coverage is restricted to a minority of the population. Vertically, high-income groups transfer resources to lower-income groups through contributions (high-income insured pay more, although often with a ceiling that tops this effect), benefits (the minimum pension, equal health care), and state subsidies; the latter is accentuated where the general tax system is progressive (ILO-ISSA 2001a).15 Finally there are intergenerational distribution effects, as the young transfer resources to the old, as well as gender redistribution effects, when men transfer resources to women to compensate the discrimination suffered by the latter. The Conference of 2001 ratified that social security is an important tool to promote solidarity, income redistribution, and poverty alleviation. ‘The fundamental challenge posed by the informal economy is how to integrate it into the formal economy. This is a matter of equity and social solidarity . . . Support for vulnerable groups in the informal economy should be financed by society as a whole.’ In pension systems of defined benefit based on pay-as-you-go (PAYG), the risk is collectively assumed, but in systems of individual saving accounts (fully funded), persons assume the risk. ‘While this is an option which exists, it should not weaken solidarity systems which spread risks throughout the whole of the scheme membership.’ Voluntary additional pension schemes ‘could be a valuable supplement but, in most cases, cannot replace the compulsory [basic] regimes . . . Governments should consider that all fiscal support or incentive to those [supplementary] regimes should be assigned instead to middle and lowincome workers’ (ILO 2001a: 2–4). The social insurance of sickness-maternity is based on a pool of risks and resources typical of the solidarity principle: the insured pay according to their economic capacity but health care is basically uniform for all,

11

albeit adjusted to risk of the individual and his family. There is a difference between the pension program, with a linkage between contributions and benefits, and the health care program where benefits are not related to contributions. This implies that the young, healthy, single men should pay relatively more, in relation to the benefits received, than the old, sick, married with family, and women. The intergeneration contract means that the young pay higher contributions, despite relatively low health risk in the short and middle term, so that when reaching old-age their contributions and those from the new generation can finance higher health costs (ILO-ISSA 2001b).

1.2.3.2. Enforcement in Latin America All social security programs in Latin America legally proclaimed the principle of solidarity between income groups, generations, and healthy and sick populations. Nevertheless, the positive impact of such principle was reduced in reality, notwithstanding notable differences among countries, due to four reasons: stratification (especially in the pioneer-high group), low coverage (in all countries of the latecomer-low group and half of the intermediate group), ceilings on contributions, and skewed allocation of health resources. Those insured in separate schemes did not contribute to the general program but got more benefits of better quality and often received fiscal subsidies financed by the entire population. In countries with low coverage, the employer contribution frequently was transferred to prices, while the state contribution was financed by taxes mostly paid by consumers, therefore most of the population, low income and uninsured, helped to finance the protection of the middle income minority. Privileged schemes covering higher-income affiliates received fiscal subsidies, leading to an additional regressive transfer within the insured segment. The majority of countries imposed a wage ceiling on contributions (as it was normal too in countries outside of the region), attenuated in countries that established a minimum and maximum

12

Reassembling Social Security

pension. The bulk of health care funds was assigned to the social insurance sicknessmaternity program, predominantly curative and covering a minority of the population in most countries, whereas a small part of those funds went to the ministry of health, the main provider in the public sector, which was responsible for overall prevention and care of the uninsured population that normally was the majority and had low income or was poor. The system tended to generate a regressive impact on distribution, albeit reduced or compensated where coverage was higher, the stratification was smaller, and there were social assistance programs. Only the six pioneerhigh countries had social assistance pension programs, precisely those that enjoyed the least poverty incidence, while the remaining fourteen countries lacked those programs despite their highest poverty incidence. The informal sector was generally excluded, although some countries had schemes for selfemployed and domestic servants. In Costa Rica, the state promoted the incorporation of the self-employed paying the employer’s contribution in sickness-maternity but not in pensions (Mesa-Lago 1978, 1989; Mesa-Lago and Bertranou 1998).

1.2.3.3. Modifications by the reforms and new principle of equivalence The principle of solidarity is either ignored or receives scant attention by the reforms, which criticize public systems in the region because of their failure to implement solidarity in practice, as most of them don’t cover the majority of the population and grant fiscal subsidies to middle-income insured while lacking protection to low-income and poor groups. Many reforms aim to expand health care services to the poor and low-income groups, but with one exception none have included social assistance pensions for the needy. Reformers also allege that, due to population aging and the increase in life expectancy, costs of pensions and health care will progressively increase along with the burden on younger generations, hence rejecting solidarity among

generations. Employer contributions should be eliminated or reduced and transferred to workers because they create distortions in the labor market: increase labor costs, encourage the shift from formal to informal employment, reduce the nation competitiveness abroad, and generate incentives for the substitution of labor by capital, all of which result in higher unemployment and informality (World Bank 1994). The reforms have replaced the principle of solidarity by a new principle of ‘equivalence’16 that establishes a much tighter link between contributions and benefits to promote incentives for affiliation and compliance, and ensure equilibrium between revenue and expenditure (Titelman, Uthoff, and Jiménez 2000). The new principle raises three questions: whether there would be adverse risk selection by providers imposing higher premium on insured persons of high risk than those with low risk, whether such providers would exclude chronic diseases and rescind contracts for old people, and whether the poor and low-income population would be able to afford the premiums charged (ILO-ISSA 2001b). Solidarity, therefore, must be procured outside of the private system through state benefits targeted on those in need and financed with taxes instead of contributions. The reforms do not have as an objective the redistribution of income; the principle of equivalence in private systems is expected to maintain existing inequalities and the redistribution function is considered inappropriate to private systems and left to the state.

1.2.4. Comprehensiveness and sufficiency of benefits 1.2.4.1. Evolution of the principle Social security must protect against all social risks and their benefits must be sufficient to cover an adequate minimum. The Beveridge plan ensured benefits of a subsistence level, regardless of income, although they could be improved in a voluntary manner; the plan was comprehensive as it combined three

Social Security Principles instruments: (a) social insurances (the most important) that included pensions, sicknessmaternity, occupational risks and diseases, employment promotion, and children’s allowances; (b) social assistance for the uninsured in need (submitted to a means test) granting them lower benefits than under social insurance; and (c) voluntary insurance for those with enough resources to buy either supplementary pensions or more personalized health care with better hotel services (room, food, etc.). ILO convention 102 of 1952 set minimum norms in the content and level of the following benefits: health care; paid leave in sickness and maternity; old-age, disability, and survivor’s pensions; health care and monetary benefits in occupational accidents and diseases; unemployment compensation; and family allowances. These benefits did not necessarily have to be introduced jointly and fully, but could be in a gradual fashion. The setting of minimum benefits faced serious obstacles due to notable differences in the degree of development among countries. A distinction was made between flat pensions and those whose levels depend on salary and contributions, establishing a minimum replacement rate of 40% for an average wage earner with thirty years of contribution (the rate was increased to 45% in 1967). Since 1944 the ILO has also supported the adjustment of monetary benefits to the cost of living. The 2001 Conference did not add concrete norms in this area, only stated that pension systems had to ensure adequate levels (ILO 2001a). All of the above refer to social insurance programs, which are distinguished from social assistance in several features: social insurance is financed by contributions and generates a right, whereas social assistance is financed by the state, it is not a right but granted based on available resources, and the beneficiary must be in state of need (ILO-ISSA 2001a).17

1.2.4.2. Enforcement in Latin America Before the reforms, all Latin American countries had in force social insurances of sickness-

13

maternity, pensions, and occupational accidents and diseases, hence the insured population was legally protected against the most important social risks (complying with the ILO norms, except in unemployment and family allowances), although enforcement depended in practice on the degree of coverage. Entitlement conditions in pensions were usually liberal although more in pioneer-high countries and less in latecomer-low countries. Ages of retirement were relatively low while life expectancy at the time of retirement exhibited a rising trend (in Costa Rica and Uruguay, ages of early retirement were reduced as life expectancy increased); formulas to calculate the pension used an average of the last three to ten years of wages; and in half of the countries, the basic replacement rate was above the 45% established by the ILO minimum norm. As time elapsed, in a growing number of countries (particularly the pioneerhigh) the linkage between contribution and pension level became more tenuous, a result of generous entitlement conditions, the program maturity, population aging, and insufficiency of contributions. In the immense majority of nations with historical data, the adjustment of pensions to the cost of living led to an increment of their real level until 1980, but high inflation in the 1980s reversed that tendency with a couple of exceptions (Mesa-Lago and Bertranou 1998). In the mid 1990s, the average pension as a percentage of the average income ranged from 43 to 51% in Argentina, Chile, Mexico, and Panama, but was only 19% in Ecuador and Nicaragua (ILO 2000b). Entitlement conditions to sickness-maternity in the region were relatively generous, for instance, almost all countries granted coverage to the insured, dependent spouse, and children albeit with some exceptions and restrictions. Costs of this program steadily rose due to the liberality in entitlement conditions, inflation, epidemiological transition, rising life expectancy, administrative inefficiency, and predominance of the costly curative medicine over preventive medicine, all of which had an adverse effect on the supply

14

Reassembling Social Security

and quality of health services. In the 1980s, the economic crisis and structural readjustment programs reduced public health budgets and services deteriorated further (Mesa-Lago 1992, 1994; Mesa-Lago and Bertranou 1998; ILO 2001a).

1.2.4.3. Modifications by the reforms The reforms presume that the level and quality of benefits in the private system will be better than that in the public system, because of the elimination of monopolies or quasimonopolies in insurance and provision, the introduction of competition, superior managerial efficiency, punctual payment of contributions, higher capital returns of the pension fund, and positive macroeconomic effects (see Section 1.2.7). Health care reforms advocate the establishment of a basic package of benefits for all or part of the population, independent from risk; with middle and highincome groups entitled to buy supplementary and better services through higher premium or co-payments. Few reforms advocate coverage for costly catastrophic illnesses or high-complex care. Pension reforms introduced voluntary contributions for supplementary benefits affordable by upper-middle and high-income groups.

1.2.5. Unity, state responsibility, efficiency, and social participation in administration 1.2.5.1. Evolution of the principle The Beveridge Plan stipulated the unification of administrative responsibility aiming to promote efficiency and reduce costs. A single Social Insurance Fund would collect all contributions and process and pay all benefits, while its major three programs would be managed and supervised by the Ministry of Social Security to unify and coordinate policy. Unity would not necessarily imply centralization, thus the system would be decentralized through local agencies close to the insured and knowledgeable of their needs.

Managerial unity in social security would make it cheaper than private insurance and save resources based on the following reasons: elimination of multiple programs with their own administration; unification of registration, collection, and payments; consolidations of facilities, equipment, and personnel; economies of scale; not-for-profit nature of social insurance; establishment of a single legal regime that would simplify knowledge and requests for benefits; easier access of the insured and portability of its coverage and benefits when changing jobs; and elimination of conflicts of jurisdiction among various entities. But the trend toward social security unity was obstructed by the previous inception of schemes covering various groups of powerful insured who resisted integration. On the other hand, the need to have a centralized register to help in the identification, affiliation, and change of jobs is not as important now due to the use of electronics in interconnected computer networks (ILO-ISSA 2001a). Despite the advantages of unification, when elaborating convention 102 the ILO decided— in view of the economic, political, and socialsecurity diversity among member countries— that it was not advisable to recommend a single and uniform administration, although it established the principle of state responsibility. More recently, the ILO noted that countries with universal coverage or close to it commonly have a central ministry responsible for social security policy. The ILO and WHO recommend the integration of the social insurance sickness-maternity program with the public health care system, as well as the overall coordination of policy by the ministry of health, both politically difficult to implement. The ILO acknowledges that the ideals of efficiency and low-cost administration are not always met in practice (Greber 1997; ILO-ISSA 2001a). According to the ILO interpretation, the existence within the social security system of two regimes, one public or social insurance and another private, is not by itself incompatible with the convention 102, because the latter permits to organize social security through

Social Security Principles different means, providing they respect the fundamental principles on which the structure of social security should continue to rely: state responsibility in the last instance and participation in the administration by those concerned. Regardless of the type of administration chosen, the state must be responsible for the good management of the institutions and services to ensure the protection guaranteed in the conventions. In addition, the insured must participate in the administration or play a consultative role (when management is exercised neither by a government department responsible before congress nor by an institution regulated by public authority); in some cases employers should have representation also (Humblet and Silva 2002). The participation of the insured in management is an important complement of democracy that continues to be entirely justified today and, theoretically, is a social security principle, although convention 102 does not fully guarantee such participation except in certain cases (Greber 1997). Furthermore, social participation helps the systems to reflect the needs and aspirations of those whom social security serves and to fulfill the obligations of the contributors (ILO 2000a). The Conference of 2001 declared that ‘schemes should be managed in a sound and transparent manner, with administrative costs as low as practicable and a strong role for the social partners.’18 It also ratified the effective participation and the important role of social interlocutors in the development of policies through bipartite administrative organs (workers and employers) or tripartite organs (the previous two and the government). Finally the Conference recommended a social dialogue among workers, employers, and government with the objective of generating social consensus and political will to act, particularly when debating a reform of the system (ILO 2001a: 2, 6, 29).

1.2.5.2. Enforcement in Latin America Unity in the management of social security existed in few Latin American countries:

15

Costa Rica, Cuba, and to a lesser extent in Panama. The process of stratification in the pioneer-high group induced a multiplicity of programs without central coordination and often without supervision. In Chile there were 160 programs of social insurance in 1979 (90 of old-age, seniority, and survivor pensions), in Uruguay there were more than 50 programs in 1967 (10 of pensions), and in Costa Rica 10 independent pension programs for public employees were created after the inception of the general system. The intermediate group was less stratified (Panama only had one general system), nevertheless Colombia had some 1,000 programs of social insurances, Bolivia 51 pension programs, Mexico 7, and Venezuela 13. The latecomer-low group was the most unified, for instance one single program in Nicaragua, but three public pension programs in El Salvador and Guatemala, and seven in Honduras. All this provoked higher administrative costs, as well as serious problems of control and supervision, and lack of portability. The process of unification and standardization in the 1960s, 1970s, and 1980s reduced in some countries the number of institutions and/or established a central agency in charge of the administration or at least the coordination and supervision of pensions, for example in Argentina, Brazil, Chile, and Uruguay. Usually there were three health sectors in the large majority of countries in the region: (a) the public, mainly represented by the ministry of health, which had the direction but in virtually all countries lacked power and resources to enforce it, and was entrusted with the poor or low-income uninsured population, in addition there were other public providers, like states, provinces, or departments, and municipalities (Argentina, Brazil, Colombia, Mexico); (b) social insurance that covered the formal sector with middle income and was frequently stratified, in some countries managed by trade unions (Argentina) or mutual-aid societies (Uruguay); and (c) the private sector, normally small, geared to the high-income group through diverse mechanisms, such as prepaid plans, insurance, and independent facilities

16

Reassembling Social Security

and health professionals. Such segmentation pushed costs up, left rural areas without care, and led to overlapping of social-insurance and ministry services in some urban areas, and wasted scarce resources. Despite recommendations from the Pan American Health Organization (PAHO) and ILO, only three countries achieved the integration of health services19 (Mesa-Lago and Bertranou 1989; Mesa-Lago 1994). The state role and fiscal responsibility in social security varied among countries, more in the pioneer-high group than in the latecomer-low group. In Cuba the government financed the considerable deficit of the pension program, while in Costa Rica the state transferred to social insurance the cost of social assistance pensions. In the pioneer-high group, due to the PAYG regime and the usual deficit of the pension program, fiscal subsidies became necessary and rapidly grew, but they were nonexistent in the latecomer-low group because of its better financial status and youth of both its pension program and population. Regarding health care, Cuba’s entire public national system was financed by the state, whereas Costa Rica’s social insurance virtually was the only provider and financer and the state had to transfer to it the cost of care for the poor. In a good number of countries, the ministry of health, the comptroller office, the central bank, and/or the ministry of finance had functions of direction, supervision, and control of social security. Nevertheless, contrary to the law and with few exceptions, the state role was customarily negative: forcing or putting pressure on pension programs to invest their reserves on government debt paper or to deposit such reserves in the central bank disbursing low or no interest; often not paying its legal contributions to social insurance (as an employer and third party) or the cost of assistance pensions or health care for the poor or even the expenses of covering its own employees on health care; signing payment agreements with social insurance but without adjusting the debt to inflation and with meager or negative interest rates; and often interfering in the

administration of social insurance by hiring unneeded and sometimes inept employees. In the immense majority of the region, before the reforms, the sickness-maternity social insurance functioned with direct provision, that is, it had its own facilities and personnel to care for the insured, unlike everywhere else in the world including Europe. Indirect provision existed in a few countries in which social insurance contracted with public and/or private facilities and health professionals to provide services in geographic areas where it did not have its own infrastructure. The advantages of direct provision are the control of prices and quality, as well as an easier planning of the infrastructure, while the disadvantages are the virtual monopoly, the bureaucracy, the lack of incentives to offer good-quality services, and the potential conflicts of interests. The advantages of indirect provision are the separation of interests, better incentives for quality through competition, and multiple methods of management and organization, whereas the disadvantages are the need of a legal framework, strict and enforceable regulation and supervision, and skilled administrators in social insurance capable of negotiating with providers, as well as risks of higher costs (ILO-ISSA 2001b). Administrative costs of social insurance tended to be smaller in countries with the highest coverage and vice versa, due to economies of scale and the fact that those countries were the most urbanized (except Costa Rica) thus facilitating the administration. In 1980–9 the average administrative cost took from 4 to 6% of total expenditures in Argentina, Costa Rica, Panama, and Uruguay, but fluctuated from 12 to 21% in Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, and the Dominican Republic. The principal component of administrative expenses was salary and fringe benefits; powerful trade unions had contracted salaries and other labor conditions that ranked among the best. In the sickness-maternity program, a common problem was the low hospital efficiency measured by the percentage of bed occupancy and the

Social Security Principles average days of stay (Mesa-Lago and Bertranou 1998). Virtually all countries had tripartite participation in the administration of social insurance, through representatives of workers, employers, and government in their directive councils or overseeing committees. In some countries, pensioners also had representation and in two of them (Argentina and Uruguay) they organized associations with enormous political and lobbying power. And yet, in several countries the government had the majority in the management councils or committees or controlled the appointment of the tradeunion representatives.

1.2.5.3. Modifications by the reforms and new goals of competition and freedom of choice Reformers argue that the state monopoly in social security or public systems has led to excessive bureaucracy, inefficiency, and waste of resources thus generating high administrative costs. The reforms support the separation of functions: those of insurance, administration, financing, and provision should be transferred to the private sector and the market, while those of policy, regulation, and supervision should be left to the state (also the direction of the health care system). The government should also create a marketfriendly environment, guarantee minimum benefits (pensions, basic health package) to the poor, higher-risk persons, and the insured who fail to meet the entitlement conditions, and finance the costs of the structural pension reform. Reformers strongly believe that a private system is more efficient than a public one because of competition, an important new goal of the reform. Replacing a monopolistic public or social security institution with multiple private providers that compete among themselves in the market (forprofit corporations with exclusive dedication in pensions, multiple types of insurance firms, and providers in health care) will improve administrative efficiency and reduce costs. On

17

the other hand, private insurance firms and providers seek profit and resort to advertisement to attract clients (nonexisting expenses in public systems) that add to administrative costs. Various elements are needed for competition to function properly: access to information by the insured person who also should have skills to make decisions, transparency of the system, and the proper legal framework of regulation-supervision (World Bank 1994). Another key new goal of pensions and health care reforms is the freedom of choice granted to the insured to select between the public and private systems, as well as among the most efficient and best providers; freedom of choice is essential for competition and this, in turn becomes the foundation of other goals and assumptions of the private system. The administration and supervision of pension systems is considered to be quite complex and thus respectively entrusted to corporations and technical bodies, excluding any participation by workers, pensioners, and employers. Most health care reforms promote several forms of users’ participation and social control in the public sector, usually through advisory councils or committees at the municipal or community levels, but normally exclude all social participation in private insurance firms as well as supervisory bodies.

1.2.6. Financial sustainability 1.2.6.1. Evolution of the principle The Beveridge report stated that the social security plan had to be financially viable according to the economic capacity of the country; to comply with that demand the plan did not have to be enforced all at once, but could be in stages setting priorities in its implementation and gradually advance when the needed resources became available. ILO convention 102 stipulated that ‘the state must ensure that actuarial studies and estimates are periodically done to maintain the financial equilibrium and . . . before

18

Reassembling Social Security

any modification in the benefits, contributions and taxes . . . The cost of benefits and administrative expenses must be financed in a collective manner through contributions, taxes or both . . . Methods of financing must avoid that low-income persons endure too heavy charges and also take into account the economic situation of the country and covered persons . . . Total contributions paid by the insured should not exceed 50% of total resources devoted to protection’, hence the rest must be financed by employers and/or the state (Humblet and Silva 2002: 13–4). Nonetheless the ILO acknowledges that international social security norms have not responded to changes occurred in recent times and virtually lack guidance in the search for just and viable financial methods to confront that challenge, and that existing conventions are insufficient for nations to financially support social security under the current economic situation (Greber 1997). Three financial regimes are employed to maintain the equilibrium of social security long-term schemes such as pensions: (a) fully funded capitalization, based on a ‘defined’ contribution expected to guarantee equilibrium for an undetermined period of time, this method can be collective (originally utilized in social security as the ‘general average premium’ but now out of use) or individual (used in contemporary systems of individual accounts); (b) collective partial capitalization (CPC) (usually the ‘scaled premium’) that guarantees a relatively large period of equilibrium, subject to periodic actuarial evaluations that often result in an increment of the contribution for the next period; and (c) PAYG, where income and expenditures must be balanced annually or in very short periods (typical of the oldest and most mature pension programs). The ILO supports three ‘classic principles’ in the investment of social security reserves: (a) security, to guarantee the future of the funds (convention 102 stipulates that the state should not use social security reserves for other purposes, such as to cover fiscal deficit); (b) adequate capital returns, to maximize

the fund, attenuate the rise in contributions, and maintain the purchasing power of benefits; and (c) liquidity, to have sufficient cash available when needed. A controversial objective has been that investment has a social function (health, housing, etc.), although done efficiently to ensure the fundamental goal of guaranteed benefits (OIT-AISS 2001a, 2001b). And yet many social investments were not done efficiently and led to low or negative capital returns that in the long run harmed the viability of benefits. Financial regimes to maintain equilibrium in social security short-term programs such as health care don’t need substantial reserves as pension programs usually do, although they should have a contingency reserve for unforeseen temporary events, set as two months of current expenditures, but should be higher at the start of operation of the program or where coverage is low and social risks are high. Because such reserve is not substantial, normally there is no significant investment in health care programs. The Conference of 2001 recommended to broaden the sources of funding for social security through, for example, tripartite financing . . . With the objective of sustainability, pension systems should be financially viable in the long run. Therefore it is necessary that they conduct periodic actuarial projections and introduce the needed adjustments . . . It is essential to make a full actuarial evaluation of any proposed reform before adopting new legislation (OIT 2001a: 4–5).

1.2.6.2. Enforcement in Latin America Before the reforms, the financial sustainability of social security varied among Latin American countries. Social security expenditures as a percentage of Gross Domestic Product (GDP) was the highest in the pioneer-high group and the lowest in the latecomer-low group; there were also important differences in the distribution of social security expenditures: in the pioneer-high group the majority of expenses went to pensions, while in

Social Security Principles the latecomer-low group the majority went to sickness-maternity. These differences were caused by the maturity of the programs and the aging of their populations, but general trends exhibited an increase in both the social security burden on GDP and the percentage of total expenditures allocated to pensions (ILO 2000b). In the pension program, the pioneer-high group was based on PAYG, and suffered financial-actuarial deficit covered with fiscal subsidies, the only exception being Costa Rica which had CPC and generated a financial surplus although facing actuarial disequilibrium. In the intermediate group most countries had CPC but also endured actuarial disequilibrium (Colombia, Ecuador, Panama, and Venezuela); the rest legally had CPC but in practice were under PAYG or rapidly approaching it (Bolivia, Mexico, and Peru). In the latecomer-low group four countries had CPC albeit with actuarial disequilibrium (El Salvador, Paraguay, Guatemala, and Honduras; the last two in a better situation) and the other three had virtual PAYG (Dominican Republic, Nicaragua, and probably Haiti). The sickness-maternity program was based on PAYG with contingency reserves and financial sustainability of diverse degree among countries. The principles of investment (security, adequate capital return, and liquidity) were not enforced in the majority of countries. A frequent problem was that the pension program had heavily invested in the construction of hospitals and equipment of the sicknessmaternity program, sometimes even subsidizing part of its current expenditures; as time elapsed and with the growing percentage of total social security expenditures going to pensions, such transfers became unfeasible and provoked a deficit in the pension program, with a similar effect on the sickness-maternity program. Despite having actuarial teams in all social security institutions—of diverse technical level—periodic actuarial studies were not always conducted and, when they were performed their recommendations were often not implemented.

19

In 1989 the average regional distribution of social security revenue by source was: 44% employers, 25% workers, 19% capital returns, 9% the state, and 3% other sources. Therefore the ILO minimum norm that the salaried worker should not finance more than 50% of the total contribution was enforced, because wage earners contributed an average of only 32% vis-à-vis 68% contributed by employers and the state (excluding capital returns and other sources; excluding the state the proportions were 36 and 64%, respectively). Nevertheless, the self-employed had to contribute a percentage equal to the sum of the percentages of contribution from salaried workers and employers (they lacked the latter), a significant barrier to their affiliation.20 Two-thirds of total revenue came from contributions, whose percentage over the wage bill was the highest in the pioneer-high group and the lowest in the latecomer-low group. In most countries the state legally had to share in financing the system as a third party, either with a contribution set as a percentage of the wage bill or as a sum to help in the costs of benefits or the administration, although such contribution often did not materialize. In some countries the law imposed taxes (e.g. on alcohol, tobacco, the lottery) whose proceeds accrued to social insurance, but more frequently taxes on services related to the activity of occupational groups with privileged separate schemes to help finance them. The high contributions (particularly in the pioneer-high group), combined with poor control of compliance, light sanctions badly enforced on transgressors, as well as inflation, generated incentives for employers’ evasion and payment delays, and the state was one the principal debtors to social security in most countries. In the pioneer-high group there were neither reserves nor investments, except in Costa Rica. In the rest of the countries pension reserves were normally invested in the following instruments: (a) long-term public debt securities or deposits in the central bank or a

20

Reassembling Social Security

state bank, without adjusting the principal to inflation and with a fixed interest below the market rate; (b) hospital infrastructure of the sickness-maternity program or administrative buildings none of which generated capital returns; (c) housing plans or mortgages or personal loans that pursued a social objective but brought heavy losses; (d) real estate or enterprises with low yields; and (e) hard currency deposits that occasionally the state converted into national currency later devaluated. High inflation in the 1980s and the start of the 1990s contributed to both reserve devaluation and negative interest rates. As a result, real capital returns (adjusted to inflation) in ten countries for which data were available were negative in 1981–94 and provoked a sharp drop in the value of real reserves (Mesa-Lago and Bertranou 1998). All of the above show that the principles of security, adequate capital return, and liquidity were not enforced in the region. Two causes of this problem were the absence of capital markets regulated and supervised by the state, capable of offering sufficient, diversified, and relatively secure instruments for investment of social security reserves, and state intervention to use such reserves with the purpose of financing its fiscal deficit.

1.2.6.3. Modifications by the reforms Reformers criticize public systems in Latin America because of their lack of incentives to contribute that resulted in evasion and payment delays, which in turn led to actuarial disequilibrium, as well as financial imbalances in the pioneer-high group, and increasing fiscal subsidies. In addition, the state normally failed to honor its financing contributions as third party and employer, as well as its obligation to transfer the cost of health care to the poor and social assistance pensions paid by social insurance, hence aggravating the imbalances. The lack or poor linkage between contributions and pension amount, the calculation of the benefit based on an average of the latest or highest salaries, and the decline in the real value of pensions also encouraged

noncompliance, as well as under-declaration of wages. There were even greater incentives for under-declaration of salaries in health care because a higher contribution usually did not increase the level and quality of care. Finally the investment of pension funds in public systems did not generate adequate capital returns and in some cases they were negative and depleted the reserves. It is reasoned by reformers that the population will have a stronger incentive to affiliate and contribute to private insurance firms and providers than to public systems. Private pension systems will reduce contributions rates, evasion, and market distortions, and the shift of the employer contribution to the worker will cut payroll taxes paid by enterprises and increase employers’ ability to hire and keep their employees (World Bank 1994; James 1996; Mitchell 1998). It is also presumed that a private pension system based in the principle of equivalence, defined contribution (expected to be unchanged for indefinite time), and fully funded financing will ensure financial sustainability and actuarial equilibrium. Therefore actuarial evaluations would not be required, although the reform should have projections of its fiscal costs during the transition. An important point made is that the reform actually reveals or makes explicit the hidden or implicit debt in public systems of PAYG, a debt that tends to increase with the maturity of the scheme and population aging, therefore the fiscal cost of the transition is a logical result but, contrary to the hidden debt in public systems, transition costs will eventually decline and disappear. Investment of the pension fund will be done by private managers in a more efficient and diversified manner, and hence with higher capital returns than in public systems. Private health care systems based on equivalence should not confront problems of financial sustainability because premiums are adjusted to individual or family risks and the premium is raised when risk increases. Contrary to pension reforms, health care reforms do not place emphasis on the elimination of the employer contribution (with the exception of Chile).

Social Security Principles

1.2.7. Reform assumptions/goals: promotion of national saving and capital markets Private pension systems are expected to generate a significant capital accumulation, increase national saving, develop capital and financial markets, and diversify the investment portfolio, all of which would promote growth, employment, and wages. It is also assumed that there is no need for a country to have a previous capital market because the reform will encourage it due to the high capital accumulation in the pension fund and the need to invest it in instruments such as stocks and bonds. All these assumptions were based in turn on several previously discussed postulates on the effects of private pension systems: an increase in coverage and affiliation, better compliance (less evasion, payment delays, and under-declaration of salaries), and higher capital returns. Furthermore in some Latin American countries the said assumptions were turned into key objectives of the reform, as in the preamble of the Mexican reform law and the first Salvadorian legal draft of reform. The economic character of these assumptions/goals was a novelty vis-à-vis the traditional social character of social security principles, reflecting the diverse interests of economists and international financial organizations (World Bank, IMF, and IADB) on the one hand, versus the concerns of experts and international social security organizations (ILO, ISSA) on the other hand (see Mesa-Lago 1996).

1.2.8. Reform assumption: immunity to state and political interference In contrast with the traditional ominous statepolitical meddling in public pension systems in Latin America, reformers assert that private systems are immune against state interference and political manipulation, a feat explained by the private nature of the administration and the ownership of the individual account

21

by the insured (World Bank 1994; Mitchell and Zeldes 1996). A more cautious viewpoint is that the private system does not ensure absolute protection against political risks, but relatively more safeguards than the public system (Cortázar 2003). Actually the World Bank (1994) had acknowledged, prior to most reforms, that the imperviousness or better protection of the private system depended on the premise that the state behaves properly, stopping its previous interference and properly conducting its newly assigned functions of regulation, supervision, and financing of the transition.

1.2.9. Conclusions: enforcement of principles in the region before the reforms The previous analysis of the enforcement of social security principles in Latin America at the eve of the reforms shows mixed results and significant diversity among countries: (a) The average regional coverage rose in 1970–90, despite the severe economic crisis of the 1980s, and tripled the ILO minimum norm on pensions in 1990 but was below the minimum norm on health care, however such data excluded noncontributory pensions as well as public and private sector coverage in health care. Countries in the pioneer-high group were above the ILO minimum norm in pensions and met the norm in health care but virtually none of the countries in the latecomer-low group did; only five countries had a pension coverage below the minimum norm, but ten countries were below the norm in health care. (b) The principle of equal treatment was eroded by stratified systems that proliferated mainly but not exclusively in the pioneer-high group, introducing unjustified inequalities in coverage, entitlement conditions, benefits, and financing among various occupational groups; although the said inequalities were reduced by unification and standardization processes, still privileged schemes subsisted

22

Reassembling Social Security

for the armed forces in virtually all countries, for civil servants in the majority and for other sectors in several countries. The lower ages of retirement for women than men, combined with a higher life expectancy and smaller contribution density among women, generated lower pensions, even though such inequalities were mollified with an equal minimum pension and unisex mortality tables. There were also inequalities in health care services between urban and rural zones, as well as among geographic regions. (c) The principle of solidarity was legally proclaimed in all countries, albeit diminished in reality by existing stratification, low coverage in half of the countries, contribution ceilings in most of them, and generalized skewed allocation of health care resources. Social security normally had a regressive impact on distribution, attenuated or compensated in countries with high coverage, low stratification, and inclusion of low-income groups. (d) The principle of comprehensiveness and sufficiency of benefits also legally ruled in the region but in practice was subordinated to the level of coverage. Entitlement conditions were quite liberal in the pioneer-high group but much stricter in the latecomer-low group. In most countries the adjustment of pensions exceeded the rise in the cost of living until 1980, but deteriorated thereafter because of high inflation and other causes. (e) The principle of unity in administration was found in only three countries, due to rampant stratification especially in the pioneer-high group, though the process of unification reduced the number of separate schemes and created a central administrative agency in some countries. The overlapping and lack of coordination between social insurance and the public health care sector persisted in most countries. State responsibility ranged from total to very restricted but, with few exceptions, the role of the government was negative. Administrative costs were low in countries with the highest coverage and high in those with lowest coverage. Virtually all countries had tripartite participation in management but in some of them the government had a majority in the

councils or controlled the selection of workers representatives. ( f ) The principle of financial sustainability had a diverse enforcement; nine countries had CPC albeit enduring actuarial disequilibrium of divergent magnitude, six countries had or were closed to PAYG with significant deficit, and the pioneer countries (except one) had PAYG, suffered the greatest deficit and required fiscal subsidies. The general trend was toward an increment in both the cost of social security over GDP and the share of pensions in total social security expenditures, while the share of health care declined. The region enforced the ILO minimum norm that the salaried worker should not pay more than 50% of the total contribution because only 32% was paid by workers vis-à-vis 68% by employers and the government (36% and 64% excluding the government). But the high contribution (especially in the pioneer-high group) generated perverse incentives for employer evasion and payment delays and the state was one of the principal debtors to social security. The investment of the pension reserves was generally inefficient, because of its heavy concentration in public debt securities and a few instruments that generated low capital returns; in half of the countries such returns were negative; these problems were largely caused by the lack of capital markets and state interference. Investment in the health care program was normally small or nonexistent, the pension program often invested heftily in hospital infrastructure, positive under a social viewpoint but poor financially; as the pension program matured, it could no longer transfer resources to health care and eventually both programs were overcome by deficit particularly in pioneer-high countries. The evaluation of Latin American performance in enforcement of social security principles should take into account the diverse level of development of its countries and be compared with the rest of the world. Available statistics indicate that the majority of the region was ahead of other developing countries, but below the developed countries. The ILO observes, however, that

Social Security Principles most industrialized countries enjoyed or were closed to full employment, basically formal employment, regulated labor markets, a fair distribution of income, high levels of compliance, and universal coverage. Conversely, countries of middle development (as Latin America) confronted serious problems: high open unemployment and underemployment, a significant and growing informal sector, very unequal distribution of income, and lack of a tax system capable of efficiently collecting social security contributions and financing social assistance programs. Under

23

such notable socioeconomic differences, it is obvious that social security principles could not function in the same fashion. And yet the pioneer Latin American countries introduced their first social insurance programs before the United States and Japan, and all the region had a sickness-maternity scheme (albeit of diverse importance according to their population coverage) while the United States lacked and still don’t have such a program and a considerable proportion of its population is without any type of health insurance (ILO 2000b; Roberts, Stafford, and Ashworth 2002).

Notes 1. The term ‘social security’ is used herein in its ample, contemporary meaning, embracing all contributory programs (pensions, sickness-maternity, occupational risks, unemployment, and family allowances), as well as noncontributory programs (social assistance). The term ‘social insurance’ is utilized in its historical sense and when specifically referring to the two key contributory programs of pensions and health care subject of the book. 2. ‘Conventions’ are approved in the annual ILO Conference, set minimum conditions and are submitted for ratification (albeit nonmandatory) to the national legislative bodies. ‘Recommendations’ embrace more advanced conditions than conventions and are not submitted for ratification. Both instruments are designed for universal implementation in a flexible manner to facilitate their potential ratification among 175 member countries with significant differences among them (ILO-ISSA 2001b; Humblet and Silva 2002). 3. The most important social security conventions related to this book are: 102 of 1952 (minimum norm), 162 of 1962 (equal treatment), 128 of 1967 (old-age, disability, and survivors pensions), 130 of 1969 (health care), 157 of 1982 (maintenance of rights), and 183 of 2000 (maternity care). The complete text of all instruments is at http://www.ilolex.ilo.ch:1567/english/docs/convdisp.htm 4. Based on the date of creation of its programs and young population, Costa Rica was in the intermediate group, but based on coverage, development, and costs of such programs it was in the pioneer-high group. 5. In 1980 Paraguay was ranked at the tail of the intermediate group, largely because of the date of inception of its programs, but in 2002 its coverage was among the lowest in the region, its population among the youngest, its life expectancy among the shortest, and its financial situation was relatively fair; due to these reasons the author decided to move Paraguay to the latecomer-low group. 6. In health care there is another principle, efficacy, not considered by the ILO, but that will be evaluated in Chapter 11. 7. In maternity, the minimum percentages of convention 102 are 50% of the total number of salaried women and the spouses of salaried men or 20% of the resident EAP, while convention 183 of 2000 gives all the women employed. 8. Cuba is excluded from these figures as it does not publish statistics on coverage; the author roughly estimated it in 1980, as 93% in pensions and close to 100% in health care (MesaLago 1989). 9. Health care coverage increased in Chile, Colombia, Costa Rica, Ecuador, El Salvador, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, and Venezuela, and declined in Argentina, Bolivia, and Guatemala. Pension coverage rose in Argentina, Chile (but see analysis later), Costa Rica, Ecuador, El Salvador, Honduras, Mexico, Panama,

24

Reassembling Social Security

10.

11.

12. 13.

14. 15.

16.

17.

18. 19. 20.

Paraguay, Dominican Republic, and Venezuela, stagnated in Colombia, and fell in Bolivia, Guatemala, Nicaragua, Peru, and Uruguay (Mesa-Lago and Bertranou 1998). Pension coverage refers to the insured worker and the right to survivor pensions for dependent spouse and children; health care includes dependent spouses as well as children below the legal age of adulthood with few exceptions. Conversely, the president of the association of private pension administrators (AFP) in Chile has asserted that ‘The higher or lower coverage of the labor force is not a task of the AFP’ (Errázurriz 2004). Armed forces and civil servants had total coverage, while self-employed, domestic servants, and agricultural workers were either not covered or endured the lowest coverage. Two notable exceptions before the structural reforms were Costa Rica’s social insurance that covered with equal treatment both the insured and the uninsured poor, and Cuba’s public national health system that provided equal treatment to the entire population (except the armed forces and security personnel that had their own services). The ILO (2001b: 69) warns, however, that redistribution cannot be achieved mainly through social security but by macroeconomic policies. For Greber (1997) solidarity implies that the poor receive benefits without contributions, others receive more than their contributions, and high-income insured pay contributions that benefit the first two groups. According to Gruat (1997) this principle that he calls ‘individual equity’ has a spectrum that goes from a very tight bond (private schemes of individual accounts) to an almost nonexistent linkage (flat universal pensions, minimum pensions). Within the parameters of the poverty threshold used by UNICEF, the minimum social assistance pension should be between 35 and 50% of the national average salary (cited by Gruat 1997). As Beveridge, the ILO holds that because social insurance does not pursue profit as its private counterpart, its management costs are lower (ILO-ISSA 2001b). Conversely, in the non Latin Caribbean virtually all countries had integrated national public health care, a legacy of the British model (Mesa-Lago 1992). Costa Rica’s sickness-maternity program was an exception because the state contributed the equivalent of the employer contribution on behalf of self-employed workers of low income.

PART II

PENSION REFORMS AND THEIR EFFECTS

This page intentionally left blank

2

Pension Reforms: Taxonomy, Goals, and Actors

This chapter first defines public and private pension systems, as well as structural and nonstructural (parametric) reforms. It also develops a taxonomy of pension systems in Latin America that distinguishes three private models from the public model and summarizes the characteristics of all, classifies the twenty countries into both systems and models and identifies public systems with parametric reforms and those without reform. The chapter also summarizes the goals of both types of systems and reforms (as well as assumptions of structural reforms), and briefly reviews the factors that explain the adoption of structural pension reforms in the region.

2.1. Private and public systems: taxonomy in Latin America In the last decade there was a vibrant international discussion on social insurance ‘private’ and ‘public’ pension systems (ILO 2000a). Because both terms are conceptually ambiguous and ideologically charged, a deeper analysis requires more precision. For that purpose four essential but different features have been identified in both types of systems, related to their contributions, benefits, financial regime, and administration. The public system is characterized by: (a) nondefined contribution, because rates are not fixed for all time but tend to increase in the long run due to the maturity of the pension scheme and aging of the population; (b) defined benefit, as the law determines the

formula to calculate the pension, for instance, it sets replacement rates based on average wages, as well as minimum and maximum benefits (these norms, however, are not always honored in practice due to financial difficulties); (c) financial regime of PAYG without reserves or CPC with partial reserves and pool of risks among the insured;1 and (d) public administration either by an autonomous social insurance entity or directly by the state. The private system is characterized by: (a) defined contribution, because it is supposedly set in the long run without variation (although rising life expectancy should eventually force its increase or a cut in the pension level); (b) nondefined benefit, which means that the amount of the pension is uncertain because it will depend on the sum accumulated in the individual account of the insured, which in turn will depend on the salary level, contribution density, capital return on the individual account fund, and macroeconomic factors such as economic growth, inflation, etc.; (c) fully funded (FF) financial regime based on the ownership of the individual account; and (d) private administration, although in some countries might be ‘multiple’, meaning either private, public, or a mix of the two. Another major difference between private and public systems is that ‘in the [private] system both the losses from economic downturn and the gains from development are individualized and appropriated [while] in PYGO systems, the benefits and costs are socialized and subject to redistributive rules defined by

28

Reassembling Social Security

the government’ (Pinheiro 2004: 115). In a private system ‘the individual faces all the risks’: macroeconomic, demographic, political, investment, and annuities, while in a public system risks are shared more broadly as they can be borne by the pensioner through a lower benefit or shorter retirement period, by workers and employers through higher contributions or by future pensioners or contributors if the pension fund is capable of borrowing (Barr and Diamond 2006: 36–7). Table 2.1 elaborates a taxonomy of pension systems in Latin America, classifying its twenty countries into private and public systems and showing the four divergent characteristics of both, with some exceptions (for other taxonomies see Müller 2003; Bertranou 2006). The division between private and public systems is somewhat arbitrary because private systems have important elements that are typical of public ones: they are mandatory instead of voluntary, and the state finances some benefits (minimum pensions, the transfer of insured contributions from the public to the private system), as well as ongoing pensions during the transition, and in some countries guarantees minimum capital returns and benefits in case of bankruptcy (see Chapter 6.3).

2.2. Types of reforms: structural and parametric The debate between public and private pension systems is intrinsically linked with two different types of pension reforms implemented in the world: structural or nonstructural (parametric). Structural reforms radically transform a public system replacing it totally or partially with a private system. Nonstructural or parametric reforms financially strengthen a public system in the long run, by increasing the age of retirement or the contributions or tightening the formula to calculate the benefit or a combination of those changes.

2.2.1. Structural reforms There are three general models of structural pension reforms in Latin America: substitutive, parallel, and mixed. Table 2.1 (upper segment) classifies the ten countries that enacted and implemented structural reform laws, according to the model they followed, and shows the year of starting operation (Ecuador and Nicaragua structural reform laws have not been implemented and they still had public systems at the end of 2006). Out of the ten reforms implemented, four are in the pioneer-high group (Argentina, Chile, Costa Rica, and Uruguay), four in the intermediate group (Bolivia, Colombia, Mexico, and Peru), and the remaining two in the latecomer-low group (Dominican Republic and El Salvador). The bottom segment of the table shows the ten countries that still have public systems and identifies those that have enacted parametric reforms.2 Five countries have implemented the substitutive model: Chile (the pioneer, in 1981), Bolivia and Mexico (1997), El Salvador (1998), and the Dominican Republic (it started in 2003 and was to be gradually implemented in 2005–6, but the process has been halted).3 Nicaragua’s reform law of 2000 followed the substitutive model but had not been implemented at the end of 2006 (see below). The substitutive model shuts down the public system (new affiliates are not allowed) and replaces it with a private system. Its four basic characteristics are those already identified, except Mexico and the Dominican Republic where the administration is multiple; in Mexico the benefit can be defined or not defined.4 In a study comparing structural reforms in twenty-two countries in Latin America and Central and Eastern Europe, Kazakhstan was the only one that had introduced a substitutive reform outside of Latin America (Mesa-Lago and Hohnerlein 2002). A commission of experts in Chile made recommendations to improve the substitutive system and correct its weakness instead of replacing it, ‘oriented to the needs of the

Pension Reforms

29

Table 2.1. Models and characteristics of private and public pension systems in Latin America, 2006 Model, country, and year of initiation of reform

System

With structural reforms (private) Substitutive model Private Chile (1981) Bolivia (1997) Mexico (1997) El Salvador (1998) Dominican Rep. (2003–?)a

Contributions Benefits

Financing

Management

Defined

Undefined Fully funded (FF) Privateg

Parallel model Peru (1993) Colombia (1994)

Public or Private

Undefined Defined

Defined Unfundedf Undefined FF

Public Privateg

Mixed model Argentina (1994) Uruguay (1996) Costa Rica (2001)

Public Undefined and Private Defined

Defined Unfundedf Undefined FF

Public Multipleg

Without structural reforms (public) Public Undefinede Brazilb b Cuba Ecuadorc Guatemala Haiti Honduras Nicaraguad Panamab Paraguay Venezuelab

Defined

Unfundedf

Public

Sources: Mesa-Lago (2004), updated with ISSA (2005, 2006a, 2006b) and data from the text. a Only the contributory regime is in force (since 2003), the implementation of the subsidized regime (for the poor) and the contributory-subsidized regime (for self-employed and other groups) have been postponed. b Parametric reforms recently introduced or under way. c Various articles of the structural reform law in Ecuador (2001) were declared unconstitutional in 2005 and a law with minor revisions was rejected by congress; therefore, still this country has a public system. d The reform law in Nicaragua (2000), substitutive model, was repealed in 2005; therefore, it has a public system. e Defined contribution in part of the scheme for employees in the private sector in Brazil (notional defined benefit); a defined contribution second pillar to the public mixed system in Panama is stipulated by reforms of 2005–6. f Unfunded in Argentina, Peru, and Uruguay, as well as in Brazil, Cuba, Haiti, and Venezuela, but CPC in Colombia and Costa Rica, as well as in Guatemala, Honduras, Panama, and Paraguay. g Multiple (either public, private, or mixed administrators) in Colombia, Mexico, and Dominican Republic, as well as in all mixed models; the second pillar of the new mixed model of Panama is managed by social insurance.

30

Reassembling Social Security

beneficiaries instead of the interests of the administrators’; 90% of such recommendations were incorporated into a legal draft of reform finished at the end of 2006 to be submitted to congress in 2007 (Consejo 2006; Gobierno 2006). The parallel model has been applied in two countries: Peru (1993) and Colombia (1994).5 In this model the public system is not shut down but undertakes a parametric reform (more so in Colombia than in Peru), a new private system is created and the two compete against each other. The public system has its four typical characteristics, except that in Colombia the financial regime is CPC instead of PAYG. The private system also has its four typical features, except that in Colombia the administration is multiple. No country outside of Latin America has followed this model, probably because of its complexity. In Colombia a second reform in 2002 introduced significant modifications in the parallel system. The mixed model has been implemented in three countries: Argentina (1994), Uruguay (1996), and Costa Rica (2001); it integrates a public system that is not shut down and grants a basic pension (first pillar) with a private system that provides a supplementary pension (second pillar). The public pillar has its four typical characteristics as does the private pillar, except that the administration is multiple in the three countries. This model is the most extended outside of Latin America; it is applied in at least twelve countries in Central, Eastern, and Western Europe (Mesa-Lago and Hohnerlein 2002). Ecuador’s reform law of 2001 followed the mixed model but had not been enforced by the end of 2006 (see below). In Argentina, the crisis and erosion of the private pillar in 2001 led to a debate and the enactment of a reform law in 2007 that allows changes between the mixed and public system every five years, and mandatory affiliates in the public system workers with certain characteristics (Ley 26, 222 2007). In Panama, a parametric reform law in 2005 and a compromise reached in 2006 stipulate a variant of a mixed model but with the second saving

pillar publicly managed and other differences (Ley 51 2005; Asamblea 2006). The private system in the three models has two components: one mandatory second pillar and one voluntary third pillar. The same administrator manages both accounts but usually separates them; contributions to both accounts benefit from deferred taxes; in most countries the amount accumulated in the voluntary account cannot be withdrawn until the time of retirement (a minority allows early withdrawal under certain conditions). At the time of retirement, the worker can withdraw the amount in the voluntary account or combine it with the deposit in the compulsory account to calculate the pension. Placement of Ecuador and Nicaragua in the taxonomy was difficult because their reform laws stipulated a private system (substitutive and mixed respectively), but they have not been enforced. In Ecuador, the Constitutional Court ruled twice in 2005 that several articles of the reform law of 2001 were unconstitutional and suspended its implementation; a revised law submitted to congress was rejected also in 2005, but there was a new legal draft at the end of 2006 (IESS 2006c). Nicaragua’s structural reform law of 2000 was repealed in 2005; a new law was approved by congress in May, but several of its articles were vetoed by the Executive and in October congress postponed the law until January 2007 (ISSA 2005; Ley 568 2006). These two countries still had a public system by the end of 2006 and, obviously, their private systems never generated effects, therefore both were included among public systems but, when appropriate, the legal features of their private systems (particularly Ecuador) are explained in the book. The unfolding reform in Panama is also tricky to situate in the taxonomy, because it had not been implemented by the end of 2006, furthermore it stipulates a second pillar of defined contribution (partly voluntary and partly mandatory, depending on the age of the insured) that is managed by the social insurance institute; Panama was therefore maintained among public systems but occasionally

Pension Reforms explaining its legal partial private features (see below).

2.2.2. Parametric reforms or lack of reforms The eight countries that have kept public systems are Brazil, Cuba, Guatemala, Haiti, Honduras, Panama, Paraguay, and Venezuela. In most of these countries, there has been more public discussion on the reform (Cuba being an exception) than in the majority of countries that passed structural reforms except for Colombia and Costa Rica (see MesaLago et al. 1997a, 1997b; Mesa-Lago 2000b, 2000c, 2000e; Saldaín 2003; Schwarzer 2004). Because the structural reform law of Nicaragua was repealed in 2005 and that of Ecuador declared partly unconstitutional in the same year, these two countries still have public systems in place thus increasing the total number of public systems to ten. All of them have the typical four characteristics of public systems with the exception of Panama. Two of these countries are in the pioneer-high group (Brazil and Cuba), three in the intermediate group (Ecuador, Panama, and Venezuela), and five in the latecomer-low group (Guatemala, Haiti, Honduras, Nicaragua, and Paraguay). The ranking of the twenty countries by their degree of social security development indicates that more countries in the pioneer-high group have adopted structural reforms than parametric reforms, while the number is similar in the intermediate group and smaller in the latecomer-low group. Financial and fiscal pressures that mainly afflicted the pioneerhigh group may explain such a trend, but does not elucidate why Brazil and Cuba didn’t follow it. One public system has recently implemented a parametric reform, another two are in the process of approving one, and others are considering such reform. Brazil reformed the general program that covers employees in the private sector in 1998–2001, introducing measures to fortify its finances, as well as a

31

financial regime with elements of notional defined contribution (NDC),6 and keeping the defined benefit and public administration features; a reform of the program for civil servants approved in 2003–4, included a gradual process of unification and standardization of entitlement conditions, significantly cutting privileged benefits particularly for future generations (MPAS 2002a; Ferreira 2003; Schwarzer 2004; Pinheiro 2005).7 In Panama, a parametric reform approved by tripartite consensus in 1998 was postponed despite two ILO actuarial reports that showed a worsening actuarial disequilibrium; after many public discussions, a parametric reform law was passed by Congress at the end of 2005 but public demonstrations forced its suspension by the Executive. A compromise reached in 2006 stipulates a modified public version of a mixed model that coexists with a pure public model and whose savings pillar is defined contribution but publicly managed (Ley 51 2005; Mesa-Lago 2005a; Asamblea 2006). Venezuela approved a structural reform following the mixed model in 1998 but did not implement it due to a change in government; the current administration suspended the said reform and enacted an organic law of social security in 2002 that mandates the integration of all schemes into a general system, which will continue with a nondefined contribution, defined benefit, CPC, and public administration, although a second pillar of supplementary pensions is contemplated. At the end of 2003, an Interministry Commission submitted an implementation plan to the Executive but was not approved, because of the lack of regulations, and therefore the law of 2002 had not been implemented by the end of 2006 (LOSSS 2002; República 2003; Andrade 2006; González 2007). Cuba has studied for a decade a parametric reform and the ministry of labor and social security elaborated legal drafts in 2003 but by the end of 2006 only marginal changes had been introduced (MesaLago 2003b; Sandó 2003). Guatemala, Honduras, and Paraguay have elaborated and

32

Reassembling Social Security

discussed various projects of both parametric and structural reforms of the substitutive and mixed models, but the integral reforms had not been approved by April of 2006; a minor parametric reform was implemented in Honduras in 20018 (Mesa-Lago et al. 1997a; Mesa-Lago 2000c; Balsells 2002; Durán 2003; Saldaín 2003). A parametric reform of the public pension pillar within the mixed system of Costa Rica was enacted in 2005, and similar parametric reforms are being considered in Argentina and Uruguay (CCSS 2005b; ISSA 2005). It was not possible to obtain recent information from Haiti (IADB 2001).

2.3. Goals and assumptions of pension reforms Structural and parametric reforms, as well as private and public pension systems, share some key common goals (for instance extension of coverage, ‘consumption smoothing’,9 and financial sustainability) but each places emphasis on divergent objectives. Structural reforms and private systems predominantly pursue economic-financial or market goals like cuts in fiscal subsidies, mandatory individual savings (second pillar), elimination of public monopolies through competition, and promotion of better investment results. Parametric reforms and public systems accentuate social targets such as solidarity and equity. The various international organizations involved in social security also share such distinctions, thus the World Bank and IMF underscore economic-financial goals, while the ILO and ISSA underline social goals (on the debate among these international organizations, their paradigms, and goals see MesaLago 1996). Recent changes in the stance and goals of both types of international organizations have moved them somewhat toward a convergence (ILO 2001a; Gill, Packard, and Yermo 2005). A recent study argues that pensions have multiple goals and that some experts and institutions focus on one ignor-

ing the others: ‘It is right to debate the relative weights accorded to various objectives, but policy analysis that focuses on a single objective . . . will be flawed’ (Barr and Diamond 2006: 22). Private and public pensions face a similar challenge in terms of maintaining and expanding coverage and providing poverty relief, particularly in Latin America with the growing informalization of the labor market, and increasing aging and lack of incomeprotection of the elderly; expanding coverage to the active and passive population has become therefore a common objective to all types of reforms and international organizations. Important assumptions of structural reforms are that they will increase national saving, develop capital markets, diversify the investment portfolio, and generate higher capital returns, as well as reduce administrative costs (due to competition and improved efficiency), provide better pensions and protect the private system against political and state interference. The book discusses the arguments behind these assumptions and whether they have materialized.

2.4. External influences, internal actors, and politico-economic environment The literature on the political economy of pension reforms is wide and with divergent viewpoints; this is not the subject of this book, hence it will not be dealt with in depth but will only summarize key issues in the debate. The first issue is why structural pension reforms were passed in some countries and not successful in others. In the 1980s and early 1990s, it was sustained that such types of reform were only feasible under autocratic political regimes (Chile under Pinochet, Peru under Fujimori), but the proliferation of reforms in the rest of the 1990s and the start of the new century showed that such radical transformation was also possible

Pension Reforms in democratic regimes (for instant, in Costa Rica, El Salvador, and Uruguay). The second issue deals with the internal and external factors that facilitated or impeded the structural reforms. In a previous work, the author and a pension scholar examined the diverse political economy circumstances that led to pension privatization in nine Latin American countries reaching the following conclusions: (a) the type of political regime partly determined the public–private mix in the reforms but did not fully elucidate the outcomes in all countries, hence additional explanations were searched for (b) main actors pushing for the reforms were neoliberal economists in the ministries of finance and economics, as well as international financial organizations (IFO), employers, enterprises and the financial sector; (c) opposition forces to the reforms included leftwing parties, the social security bureaucracy, powerful trade unions, associations of pensioners, and weak ministries of labor in some countries; (d) the margin of maneuver of these actors was influenced by existing institutional arrangements, as well as political factors and economic conditions; (e) legal limitations for the reform included constitutional norms that impeded privatization; ( f ) the degree of political control of the Executive over Congress, the links of trade unions with the government, and the capacity of some groups to prevent or support the reform law through mechanisms of direct democracy were also important political factors; (g) economic conditions that facilitated the reforms were pressures to promote national saving and capital markets, heavy fiscal subsidies to public pension systems, and a high degree of indebtedness with IFOs, conversely, the potentially high fiscal costs of the transition to private systems was an important blocker of the reform; (h) the reaction of domestic policymakers to external pressures from IFOs varied from alignment to occultation and rejection; and (i) some aspects of the reform design had a strategic importance, such as hiding the magnitude and length of fiscal costs during the transition, obfuscation and packaging tactics, offering ownership rights

33

over individual accounts as well as better private pensions vis-à-vis insecure and low public pensions, and deciding the right moment for the reform (Mesa-Lago and Müller 2002; for other analyses see Huber and Stephens 2000; James and Brooks 2001; Madrid 2003; Weyland 2004; Rodríguez-Oreggia 2005). A third issue is the weight of Chile in the rest of the region. Weyland convincingly argues that Chile became a model of pension reform because it surged at a time of mounting criticism on the problems faced by public systems, which seemed impossible to solve within the prevailing paradigm. Chilean reformers rejected the latter and substituted it with the bold new paradigm of privatization inserted into the framework of neoliberal policies, and promised to solve the public system problems and achieve broader longterm goals (converting the fiscal deficit into national saving that would promote capital markets, economic growth, and employment) thus mobilizing additional supporters to the reform. He also argues, less successfully, that the high status of Chile as a nation contributed to turn its policy into a regional model, while cultural and socioeconomic similarities enhanced its probability of emulation, but Chile was an international political outcast in 1980–1 because of the Pinochet dictatorship. Weyland adds that the Chilean model ‘became more attractive when the country returned to democracy, yet maintained the pension scheme’. Actually the first country to follow Chile’s substitutive model was Bolivia in 1997, despite significant socioeconomic differences with Chile. He rightfully notes that the enormous transitional costs of Chile’s full privatization, induced other countries to scale down the bold model and adopt parallel or mixed systems (in 1993–2000), easier to finance and gain political approval in democratic countries. His rationale, however, does not explain the virtual full imitation of the Chilean model in other democracies such as Dominican Republic, El Salvador, and Mexico. Diffusion was helped by ‘third parties’ (IFOs) that may ‘push a model that conforms to their normative or ideological

34

Reassembling Social Security

orientation, but does not fit the specific needs of the recipient country’, and yet what worked in one country may not work in another, it may even make things worse. Finally he asserts that the Chilean model was ‘oversold’, imitated before starting to pay benefits to a substantial number of people, hence was ‘a matter of projection or speculation not fact’ (Weyland 2004: 10–11, 23, 273). A fourth issue relates to the influence of IFOs, particularly the World Bank, on shaping the structural pension reforms in the region. In a compilation by Weyland (2004) Joan Nelson argues that the World Bank had little influence, preceding its 1994 report, on the reforms of Argentina, Colombia, and Peru that occurred prior to that report. In turn Sarah Brooks contends that the IFOs played an important role in the diffusion of private pension models in the region but was far from decisive in shaping any final reform models. Brooks based her assessment on Argentina, Brazil, Mexico, and Uruguay, but a wider spectrum of cases (Bolivia, Dominican Republic, El Salvador, and Nicaragua) could have led to a different conclusion. Finally Weyland (2004: 242–4) asserts that ‘the World Bank never pushed a single, specific model of pension reform [Chile] but offered a menu of [‘permissible’] options from which different countries could pick and choose’. In practice the Bank mainly followed a single model from 1993 until the end of that decade (see Müller 2003); but recently the Bank has taken a more open position (see below).

2.5. One single type of reform/model or several types/models? Latin America is a mosaic of pension reforms and models. Instead of a single universal model of structural reform in half of the countries, they actually have three distinct models and with considerably variety in the characteristics of countries within each model; simi-

larly, there is significant diversity in the other half of the region with public systems, several of whom have undertaken parametric reforms. Usually countries have adapted the reform to their own financial, economic, demographic, social security, and political needs and peculiarities, but some copied a foreign model that could not function due to the absence of essential elements. An important report by World Bank officials evaluating the effects of structural pension reforms in the region in the last decade, which revises some of the original assumptions of the initial Bank report in 1994, will be analyzed in this part of the book. The officials acknowledge that ‘no magic formula for success exists’, and offer options without advocating any in particular as optimal, because they ‘would vary widely by country, level of development and administrative capacity’ (Gill, Packard, and Yermo 2005: xviii, 204– 5). They also note that their report was stimulated by several events in 2002–3: the crisis of Argentina’s private system that prompted a legal draft allowing the change of the insured from the private to the public system; an attempt to include a similar right in the constitutional reform in Peru, and the approval in Bolivia of financing social assistance using funds from the private system. Additional signs of trouble in structural reforms and the need to revise their initial assumptions and correct their flaws occurred in 2005–7: Ecuador reform law declared partly unconstitutional, Nicaragua law repealed, Dominican Republic postponement in the implementation of two reform regimes, El Salvador serious obstacles to financing the reform transitional debt, Argentina law of reform of the system, Chile legal draft of reform of its private system and Uruguay debate to allow a shift of insured from the mixed to the public system. The next four chapters evaluate the effects of structural pension reforms on the conventional social security principles, systematically compare performance in private and public pension systems in each of such principles, and test whether the reform new goals have

Pension Reforms materialized. Chapter 3 deals with coverage; Chapter 4 with equal treatment, solidarity, and comprehensiveness-sufficiency; Chapter 5 with unity, state responsibility, efficiency, and social participation; and Chapter 6 with financial sustainability and the reform new goals. The analysis is based on comparative standardized statistics mostly for 2005–6, legislation and numerous technical documents from

35

the twenty countries. The structural reform in Chile is twenty-five years old, in Argentina, Colombia, and Peru about thirteen years, and in Bolivia, El Salvador, Mexico, and Uruguay close to one decade, only in Costa Rica and Dominican Republic are the reforms recent, therefore there is enough time and accumulated evidence to assess most effects even if preliminary in some of them.10

Notes 1. The international debate presents a dichotomy between PAYG and FF regimes, which ignores the existence of CPC regimes in Latin America. 2. All countries in the English-speaking Caribbean continue to have public PAYG systems of defined benefit, have not undertaken structural reforms and continue to be financially solvent because their schemes are relatively recent and they have young populations (Osborne 2004; ECLAC 2006). 3. The Dominican Republic’s private system has three regimes: contributory (implemented in 2003), subsidized for the poor (planned to start in 2004 but postponed that year and not enforced by 2006); contributory-subsidized for self-employed and other low-income groups (planned to start in 2005 but also postponed). After five years of the law enactment the system is paralyzed due to opposition from various sectors and talks about a counterreform; in 2006 the President called for a national summit to discuss the situation (Lizardo 2006). 4. In Mexico all insured at the time the reform can choose, at retirement, the best pension offered by either the closed public system (defined benefit) or the private system (undefined benefit). Such an option (unknown in the rest of Latin America and in Europe) was a political compromise to get trade union support in order to secure at least the former public pension level. 5. In August 2006, the Colombian government announced that the social insurance public system would be terminated prompting a heated debate in congress. 6. The NDC, existing in Italy, Kyrgyz, Latvia, Mongolia, Poland, and Sweden, is based on contributions credited to individual (‘notional’) accounts that simulate a FF regime; there are neither physical reserves nor money deposited in the accounts nor investment in financial markets but the state credits the accounts with a notional interest rate; at the time of retirement the state converts the ‘sum’ accumulated in the notional account into an annuity that often is adjusted to the increase in life expectancy (Barr and Diamond 2006; Holzmann and Palmer 2006). 7. Brazil is a unique case in the region because it implemented parametric reforms in the first public PAYG pillar, did not create a mandatory private FF second pillar, and developed a voluntary third pillar of supplementary pensions (Pinheiro 2005). Conversely, structural reforms emphasized the second pillar, neglected the first, and paid little attention to the third. 8. A legal draft in Honduras in 2005 followed the mixed model (Congreso 2005) but had not been approved when this book was finished. 9. A technique that defers consumption allowing persons to transfer consumption (through contributions) from their productive years to their retired years, either done collectively (public pension systems) or by individual savings (private pension systems). 10. On the world debate on the effects of structural pension reforms, see Orszag and Stiglitz (2001); Barr (2002); Mesa-Lago (2002, 2005c); Müller (2003); Gill, Packard, and Yermo (2005); and Barr and Diamond (2006).

3

Effects of Pension Reforms on Universal Coverage

Supporters of structural reforms have repeatedly asserted that the private system provides stronger incentives for affiliation and labor force coverage than the public system: individual accounts, strict correspondence between contributions and level of pensions, and better information and transparency.1 World Bank officials confirm that increasing coverage is both an objective and a predicted result of implementing a private pension system, and assert that Bank loans for structural reforms in Argentina, Mexico, and Nicaragua had as a main goal the expansion of coverage (Gill, Packard, and Yermo 2005: 98). But such an assumption is rejected by Guillermo Arthur, President of the International Federation of Pension Fund Administrators who claims that ‘coverage [is not] a responsibility of the social security system itself, but rather the result of . . . the labour market’ and the state should take care of it (IFPFA 2003: 446).

3.1. Statistical coverage of the labor force by the contributory system This section estimates coverage of the labor force (mainly salaried) by contributory pensions in private and public systems, compares both estimates, and identifies problems in them.

3.1.1. Private systems Estimates of coverage of the labor force in private systems are available based on affiliates

(all those enrolled in the system) and active contributors (affiliates who contributed in the last month). The former is considerably higher than the latter and is overestimated because a worker may have enrolled in the system, contributed only once or for a few months, and later stopped paying and still appear as an affiliate (see below). Coverage based on active contributors is more accurate although might be underestimated because an affiliate who did not contribute in the last month may pay in the following months or still be covered if he had already accumulated enough contributions to earn at least a minimum pension. To determine real coverage, the insured contribution density is needed, an information usually unavailable.2 Household surveys can also measure coverage but face problems that will be discussed later. Table 3.1 shows that the weighted average of the ten private systems at the end of 2005 was 60% of the labor force based on affiliates but dropped to 25% relying on active contributors. Chile had 116% of its labor force covered based on affiliates, a clear indication of overestimation, furthermore, that figure did not take into account the uninsured labor force and those covered by the armed forces; Chilean coverage fell to one-half (60%) based on active contributors, and in El Salvador, Mexico, and Peru such coverage was about one-third of that based on affiliates. Five factors may explain the gap between the two estimates: (a) many affiliates are unemployed and don’t contribute, for instance, 13–18% in Argentina, Colombia, Dominican Republic, and Uruguay and 8–9% in Chile and Peru in 2004 (ECLAC 2004b);

Effects of Pension Reforms on Universal Coverage

37

Table 3.1. Percentage of the labor force covered by private systems only, based on affiliates and active contributors, December 2005 Coverage (% of labor force) based on: Private systems/countries

Affiliates

Active contributorsa

Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averageb

60.8 25.9 116.5 30.4 75.2 32.4 43.8 80.4 30.3 45.2 60.2

24.5 12.3 59.6 15.4 49.1 15.3 17.3 30.2 11.0 25.2 25.1

Sources: AIOS (2006). a Affiliates who contributed in the last month, except Mexico in the last two months. b Weighted average.

(b) some affiliates are no longer in the labor force because they have emigrated or are seasonal or temporary or part-time workers or women who are raising their children; (c) evasion and payment delays are serious and growing (see Section 6.2); (d) doubling counting in the number affiliated occurs when frequent transfers of affiliates between administrators are not processed promptly; and (e) the insured minimize their contributions just to gain access to the minimum pension and maximize the state subsidy when the fund accumulated in their individual accounts is insufficient to finance such pension (Arenas de Mesa 2000; Bertranou 2001; Mesa-Lago 2001b; Uthoff 2002; Lizardo 2004). Comparisons of coverage before and after the structural reform are feasible based on the more accurate statistics on active contributors, also available for public systems. Coverage of the labor force in the ten private systems between the year before the structural reform

and 2004 is shown in Table 3.2 (first two columns, top segment) with several caveats. First, coverage is overestimated because all insured in public systems of parallel models, in public pillars of mixed models, and remaining in closed public systems in 2004 were counted in the private system (five million) when actually they were in the public, increasing private coverage particularly in Argentina, Colombia, Costa Rica, and Uruguay. Second, overall coverage is underestimated because it is limited to the principal program and excludes insured in separate schemes (see Section 5.1), if these were included, coverage would increase in countries with a significant number of insured in those schemes. Third, statistics before the reform could be based on different periods of contribution (for instance, from one to six months), while 2004 data are standardized in the last month. Table 3.2 indicates that coverage decreased in all private systems: from 50 to 24% in Argentina, 12 to 10% in Bolivia, 64 to 57% in Chile, 32 to 22%

38

Reassembling Social Security

Table 3.2. Percentage of the labor force covered by private and public contributory pension systems, based on active contributors Coveragea (%) before structural reform

Coverage (%) 2004

Coverage (%)b 2000–3 household surveys

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagec

50 12 64 32 48 30 26 37 28 73 38

24.3 10.5 57.3 22.2 46.6 14.2 20.1 28.0 14.8 58.8 26.3

34.6 9.9 58.2 n.a. 50.1 n.a. 29.7 38.5 18.9 55.3 35.8

Public systemsd Brazil Ecuador Guatemala Honduras Nicaragua Panama Paraguay Venezuela Averagec

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

45.2 19.4 20.2 18.9 16.4 53.4 8.5 20.5 39.0

45.1 21.9 19.6 n.a. 18.7 n.a. 13.9 35.1 41.0

a

Sources: First two columns from Mesa-Lago (2004), updated with BPS (2005a), CCSS (2005a), IESS (2005), IGSS (2005), INSS (2005a), Mesa-Lago (2005a), MPS (2005b), IVSS (2006), Table 5.2, and labor force from ECLAC (2005c). Last column from Rofman 2005, except weights by author. n.a. = non available. N.A. = not applicable. a In private systems: percentage of the labor force covered by the public system before the structural reform and jointly by private and public systems in 2004; in public systems: percentage of the labor force covered in 2004 (except Brazil in 2003, Honduras in 2001, and Paraguay in 2000); excludes insured persons with separate schemes: the armed forces in all countries, civil servants in some countries, and other small groups. b Contributors in all systems, programs and schemes as percentage of the labor force, based on household surveys. c Weighted: columns 2 and 3 based on total number of contributors and the total labor force (2004 for all private systems and 2000–4 for public systems); last column weighted by labor force in 2000. d Ecuador and Nicaragua are included in public systems because their private systems have not been implemented; Cuba is excluded because it does not publish statistics on coverage; the most recent rough estimate for Haiti is 1.6% in 1999 and was excluded.

Effects of Pension Reforms on Universal Coverage in Colombia, 48 to 47% in Costa Rica, 30 to 14% in the Dominican Republic, 26 to 20% in El Salvador, 37 to 28% in Mexico, 28 to 15% in Peru, and 73 to 59% in Uruguay. The weighted average based on the labor force of the ten countries fell from 38 to 26%. Relying on estimates of labor force coverage in eight private systems (excluding the Dominican Republic and Uruguay) based on household surveys in 1998–2000, World Bank officials revised the initially predicted increase in coverage by structural reforms reaching generally negative conclusions, although with nuances: After rising modestly as a result of the reforms, coverage ratios have stalled at levels of about half of the labour force in those countries where [it] is highest [Chile and Mexico]. . . . stagnant coverage ratios are indicative of scepticism of the new system despite its virtues [therefore] it may be reasonable to question the effectiveness of the current [private] systems in creating an attractive instrument for retirement savings . . . scores of people are left uncovered just as they were under the old [public] systems . . . there is no clear evidence of increased coverage in Latin American countries that implemented pension [structural] reforms . . . pension coverage remains low and inequitably shared among income groups . . . (Gill, Packard, and Yermo 2005: 5, 8, 14, 89).

The statement from World Bank officials that the most dramatic increase in coverage occurred after the Chilean reform of 1981, although stalled later, is questioned by two standardized historical statistical series published in that country, both based on total active contributors as a percentage of the labor force. The first shows a peak of 79% in 1973 (the year of the coup d’état), a fall to 64% in 1980 (before the reform) and to 29% in 1982 (one year after the reform), then a gradual increase to 58% in 1997, and stagnation thereafter (SAFPb 2002b). The second series exhibits a decline from 62% in 1975 to 48% in 1980 and to 42% in 1982, then a steady climb to 62% in 1997 but a fall to 58% in 2000 (Arenas de Mesa and Hernández 2001).3 In both series, therefore, coverage in 2000 was below the level of 1973–5.

39

3.1.2. Public systems Estimates of coverage in public systems face a similar problem to that in private systems, the lack of standardized historical series and additional nuisances: dearth of statistics for Cuba and obsolete estimates for Haiti,4 estimates previous to 2004 (2000–3) in three countries, and the inclusion of the insured in public systems/pillars in the estimates of coverage of private systems. Finally, it is impossible to compare coverage before and after the reform as in private systems because there has not been a structural reform in public systems and we lack a historical series. Coverage in the principal program in the eight public systems between 2000 and 2004 was Panama 53%, Brazil 45%, Venezuela 21%, Guatemala 20%, Ecuador and Honduras 19%, Nicaragua 16%, and Paraguay 9%. The weighted average of the eight countries was 39%, but excluding Brazil the average declined to 20% (Table 3.2, second column, bottom segment).5 Coverage in Cuba is probably very high and in Haiti very low, the addition of those data would not change the public averages because these two countries have fairly similar labor forces.

3.1.3. Comparison of private and public systems coverage An entirely accurate comparison of coverage between private and public systems is restrained by the statistical problems already explained, particularly because public systems have diverse periods in which the latest contribution is made (the last month or the last six months or even the last year), whereas such a period is standardized as the last month in private systems.6 Taking into account those caveats, Table 3.2 (second column) indicates that the weighted average coverage in the principal program in public systems (39%) was higher than in private ones (26%); excluding Brazil, however, the public average (20%) was lower. Data show similarities and differences in coverage between the two types of systems and in their ranking within the region in the

40

Reassembling Social Security

1980s. Regardless of the type of system, countries in the pioneer-high group have the highest coverage and those in the latecomer-low group the lowest coverage, with two exceptions. Coverage ranged from 45 to 59% (from highest to lowest) in Uruguay, Chile, Panama, Costa Rica, and Brazil (possibly Cuba), all in the pioneer-high group, except Panama (intermediate group) that notably improved its coverage and regional ranking in the last two decades; but coverage in Argentina, largely due to the crisis, shrank below the regional average at the level of intermediate countries. Coverage in the intermediate group stretched from 11 to 28%: Mexico, Colombia, Venezuela, Ecuador, Peru, and Bolivia; Mexico exceeded Venezuela, reversing the position of the 1980s; whereas Peru’s coverage diminished sharply to 15%, the third lowest, at the level of latecomer-low countries, and Bolivia became the region second lowest. Coverage in the latecomer-low group oscillated from 9 to 20%, the highest in Guatemala and El Salvador (tied and superior than coverage in Bolivia and Peru and close to Colombia’s), followed by Honduras, Nicaragua (also higher than Bolivia and Peru), and the Dominican Republic; Paraguay lagged behind with 9%, the lowest regional rate with the probable exception of Haiti. Although more precise estimates are needed (based on the same contribution period and year), the above comparison suggests that public systems, not only kept their ranking in the region, but some improved it. Pre-reform regional weighted averages of coverage (61% in 1980 and 63% in 1990 before reforms except for Chile) cannot be compared with coverage in 2004 because the latter are based on active contributors. The weighted regional average of 31% in 2004 (based on Table 3.2) was still above the ILO prescribed minimum of 20% of the EAP in pensions. The above estimates of labor force coverage in private and public systems are based on the insured in the principal program and exclude separate schemes. The third column of Table 3.2 (relying on household surveys in fourteen countries taken in 2000–3) includes insured in all programs based on active contributors. A comparison of estimates in the second

column (excluding separate schemes) and third column (including those schemes) at approximately the same time in fourteen countries results, as expected, in a higher coverage in the third column particularly in countries with a significant number of insured in separate schemes. The author has estimated a weighted average coverage of 41% in public systems vis-à-vis 36% in private systems, influenced by the high coverage and weight of Brazil. Seven countries exhibit a significantly higher coverage in the third column ranging from 4 to 15 percentage points: Argentina, Costa Rica, El Salvador, Mexico, Paraguay, Peru, and Venezuela; the largest increases are in Venezuela, due to its multiple separate schemes, and in Argentina and Mexico, explainable by the inclusion of federal and state-province employees and other groups.7 Surprisingly, coverage figures for Brazil are identical despite the inclusion of federal and state employees in the third column that should add about five points over the estimate in the second column (MPS 2005b). The figures based on household survey exclude five countries (Panama and Colombia among them) and are afflicted by various problems: The quality of the instruments and their representativity is not uniform . . . there are differences in geographical coverage (in some cases the surveys are exclusively urban, while in others they are nationwide), as well as differences in the way the questions are formulated and processed [therefore] one should be cautious about using them to analyze coverage [and] to make international comparisons (Rofman 2005: 12–13).8

Survey figures fit into coverage trends analyzed before among the three groups, but differ with several conclusions on the ranking of coverage in private and public systems.

3.2. Legal and statistical coverage of groups difficult to affiliate Previous coverage data are essentially related to salaried workers in the urban formal

Effects of Pension Reforms on Universal Coverage sector; most pension programs in the region have been designed mainly for those workers because they have stable employment with long-term contracts, medium and high salaries, fair to high contribution density, retire at normal ages, and are mostly men. But in Latin America, the majority of workers do not meet those conditions: (a) 47% of the urban labor force is in the informal sector of the economy as self-employed, employees of microenterprises, domestic servants, home, seasonal or part-time workers, laborers without contract, and unpaid relatives; (b) 56% of the rural labor force is self-employed or unpaid family workers and others are seasonal workers, sharecroppers, squatters, etc.; (c) indigenous peoples are a majority or a substantial segment of the population in most Central American and Andean countries, largely out of the formal labor market and concentrated in the least developed, rural, and poor regions; (d) informal and agricultural employment are unstable, have low productivity, and meager and irregular income; (e) these two sectors also endure low contribution density due to job instability, periods of unemployment or temporary work or, in the case of women, leaving jobs to raise their children; and ( f ) manual workers suffer physical deterioration that impedes them from retiring at the statutory age. Most of these workers therefore face serious obstacles for coverage, and the problem is aggravated in private systems, because even if these workers affiliate, most of them cannot accumulate enough years of contribution or save enough to qualify for a minimum pension. Finally, the law usually discriminates against most of these workers, either by granting them only optional coverage or excluding them altogether (Arenas de Mesa 2000; Bertranou 2001; Mesa-Lago 2001b; Table 3.3). This section focuses on two major groups of workers difficult to cover: (a) the urban informal sector particularly self-employed, employees of microenterprises, and domestic servants, and (b) the rural sector including various types of agricultural workers, as well as peasants. Table 3.3 (first three columns) shows the importance of these groups in the labor force, while Table 3.4 summarizes their

41

legal coverage status: compulsory, voluntary, and excluded or under restrictions or special regimes. The scarce and nonstandardized statistics on coverage of these groups, as well as measures to improve coverage, will also be discussed.

3.2.1. Urban informal sector The informal sector increased from 42.8% of employment in Latin America in 1990 to 46.5% in 2002 (66% of all newly employed are informal), as a result of cuts in public formal jobs, a lower growth of employment in large enterprises than in the labor force, the flexibilization in contracting labor, and the expansion of jobs in self-employment, microenterprises, and domestic service (ILO 2002a, 2003a; Tokman 2004). Defined by its low skills and productivity, the informal sector averaged 46.7% of the employed urban labor force in the region in 2001–4, ranging from 49 to 63% in nine countries: Bolivia, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, Paraguay, Peru, and Venezuela (Table 3.3). Excluding two countries, the rest are the least developed in the region and five belong to the latecomer-low group. Self-employment alone averaged 29.8% and ranged from 30 to 46% of the employed urban labor force in the cited nine countries, as well as in the Dominican Republic, Colombia, and Peru, the first in the latecomer-low group and the last two in the intermediate group. The lowest percentages in both the informal sector (29–43%) and self-employment (3.7–26%) are in the pioneer-high group (Argentina, Brazil, Chile, Costa Rica, Cuba, and Uruguay), and in the two most advanced in the intermediate group (Mexico and Panama). The informal sector is very difficult to cover by social insurance pensions; hence the larger such a sector is the more are the obstacles to extend coverage. A comparison of Tables 3.2 and 3.3 indicates that the less developed a country is, the higher is the size of the informal sector and lower the coverage of the labor force, and vice versa. The self-employed are a heterogeneous group with divergent skills, types of work, and income. According to statistics from Chile’s

42

Reassembling Social Security

Table 3.3. Population groups difficult to cover by pensions between 2001 and 2004, and social assistance pensions in force in 2006

Systems/countries

Informal sector (% of urban labor force)a

Self-employed (% of urban Rural self-employed, employed etc. (% of rural labor force) labor force)b

Social Poverty assistance incidence (% pensions of population) in force

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay

37.0 62.8 28.6 44.4 31.4 43.0 49.5 41.4 60.0 42.6

17.1 45.7 15.0 38.5 18.1 30.6 32.4 19.0 42.0 21.8

n.a. 86.0 32.0 54.8 26.4 55.1 40.5 35.4 80.5 n.a.

29.4e 62.4 18.7 50.6e 20.3 44.9 48.9 37.0 54.7 15.4e

X Xg X

Public systems Brazil Cuba Ecuador Guatemala Haiti Honduras Nicaragua Panama Paraguay Venezuela Regional averagesc

41.9 n.a. 52.4 54.1 n.a. 54.2 55.6 34.5 53.8 53.9 46.7

23.6 3.7d 34.2 36.0 n.a. 36.7 35.4 21.0 30.1 39.6 29.8

64.7 n.a. 60.4 62.3 n.a. 63.0 57.2 56.3 69.4 44.9 55.6

38.7 20.0f 49.0e 60.2 n.a. 77.3 69.3 34.0 61.0 48.6 41.7

X X

h

X h

X

i

h

Sources: Based on ECLAC (2005a); Mesa-Lago (2005b) author’s regional averages except for poverty incidence (averages exclude Cuba); social assistance pensions from legislation. n.a.= non available. a Percentage of employed urban labor force that is either unskilled self-employed, domestic servant, or employee in microenterprises, all with low productivity. b Percentage of rural labor force that is either self-employed or nonpaid family workers. c Nonweighted except for poverty incidence. d Percentage of the civilian total labor force. e Only urban. f Only in Havana. g An annual flat sum (Bonosol) granted to all those aged 65 and above who were 21 years or older at the end of 1995 regardless of income, not a supplementary means-tested social assistance pension. h Laws of reform stipulate social assistance pensions but they had not been implemented by 2006. i Ecuador has a small assistance program that includes old people and other poor groups, created in 1998 it was closed and then reopened in 2003; its implementation has been slow and confusing and there are no data on beneficiaries.

Effects of Pension Reforms on Universal Coverage

43

Table 3.4. Legal coverage on pensions of groups difficult to incorporate, 2005–6 Self-employed

Domestic Servants

Agricultural Workers

Systems/Countries Obligatory Voluntary No Obligatory Voluntary No Obligatory Private Systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Public Systems Brazil Cuba Ecuador Guatemala Honduras Nicaragua Panama Paraguay Venezuela

X

a

X X X

Xb a

X X X X

X

X a

c a a

X X X X X X

X X, salaried only X X, with exceptions X, salaried only X X, large farms only Special regime n.a. X

X X X X

X X X X X X Xc X X

X X X X Xe c d d

Xc Xd Xd

Special regime X, state farms and coops Special regime X, large farms only X, if more than 10 workers Xe X, if more than 3 months n.a. X

Source: Legislation; no information available for Haiti. n.a. = non available. a Reform laws make coverage obligatory but have not been implemented. b Became obligatory in February 2006. c The ongoing reform makes coverage obligatory for high-income self-employed and voluntary for those with low income; domestic servants are voluntarily covered until regulations are approved. d Laws not implemented yet made coverage compulsory. e The new social security law of 2005 (postponed until 2007) makes mandatory the incorporation of these workers.

private system, the self-employed have a considerably lower income than salaried workers; a household survey of 2003 indicates that the self-employed are concentrated in the two extremes of the income distribution: the poorest and the wealthiest quintiles, probably the unskilled and high-income professionals, respectively. Chile’s social protection survey of 2002 revealed that the affiliated self-employed have lower average schooling

years than affiliated salaried workers and less alternative sources of old-age protection (savings, financial investment, and insurance, but not housing), and that affiliates who have always been self-employed have considerably lower contribution density than those who have always worked as salaried (Bertranou and Vázquez 2006). There is no significant difference between private and public system in terms of legal

44

Reassembling Social Security

coverage. That of self-employed is obligatory only in four pioneer-high group countries: Argentina, Brazil, Uruguay, and Costa Rica (compulsory covered in the first public pillar but excluded in the second private pillar). Reform laws in Colombia, the Dominican Republic, Ecuador, Paraguay, and Venezuela also make compulsory such affiliation but these laws have not been enforced. Panama’s ongoing reforms in 2006 make coverage mandatory for high-income self-employed and voluntary for those who earn less and are the immense majority. Legal coverage is voluntary in fifteen countries: Bolivia, Chile, Colombia, Cuba, Ecuador, El Salvador, Guatemala, Haiti (probably), Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, and Venezuela; and these workers are currently excluded in the Dominican Republic (Table 3.4). Legal mandatory coverage of the selfemployed seems to increase statistical coverage (Argentina, Brazil, and Uruguay), whereas the legal voluntary alternative has resulted in very low coverage (Chile), unless proper incentives are provided (Costa Rica). Argentinean self-employed accounted for 24% of the urban employed labor force in 2001 and 30% were reported as active contributors; small low-income entrepreneurs (monotributistas) also have compulsory affiliation but we lack data on their coverage (Schulthess and Felippone 2003; Bertranou and Vázquez 2006).9 Brazilian self-employed represented 26% of the urban employed labor force in 2000 and 23% of them were covered whereas in Uruguay the respective figures were 22 and 29% in 2004 (Bertranou and Vázquez 2006; Schwarzer 2006). Voluntary affiliation has led to low coverage even in pioneer countries, as the case of Chile demonstrates after two decades of structural reform: self-employed workers account for 15% of the employed labor force and their coverage decreased from 12 to 5.4% in 1986–2004 whereas the number of uninsured increased 80% in the period; affiliation in the private system was essentially by skilled professionals and accounted for only 2.4% of total affiliates in 2005 versus 25% in the closed pub-

lic system; combining both, only 7% of the self-employed were insured (SAFPb 2005; SSS 2005; Arenas de Mesa and Mesa-Lago 2006). In 2004–5, the self-employed were only 4% of total insured in Bolivia and Colombia; in Mexico 0.1% of insured in the principal program (IMSS) were self-employed or 0.13% of the estimated informal sector (SBC 2005; Díaz 2006; SPVS 2006a, 2006b). In eight public systems coverage of the self-employed is voluntary and very few are protected, for instance, only 0.2% of total insured in Nicaragua were voluntary insured including self-employed. In Paraguay coverage is compulsory by law but not enforced and only 0.2% were covered in 2002 (Saldaín 2003; INSS 2005a). Conversely, in Costa Rica where the self-employed also had voluntary coverage in the public pillar, they accounted for 21% of the urban employed labor force but 24% of them were covered in 2001 because to join the health care program and receive a state subsidy it was obligatory to be affiliated to the pension program (Martínez and Mesa-Lago 2003). In 2003–4 domestic servants comprised from 3 to 7% of the employed urban labor force in fifteen countries and 8 to 10% in three (ECLAC 2005a). Although these figures may underestimate the actual number of these workers in some countries (because they lack a contract or employers don’t report them), still their proportion is considerably lower than that of the self-employed. Legal coverage of domestic servants is also better than that among self-employed: it is obligatory in twelve countries, nine in private systems and three in public systems, and is made mandatory by laws in Nicaragua, Panama, Paraguay, and Venezuela, albeit not enforced yet. Coverage is voluntary in Cuba and El Salvador, and these workers are currently excluded in Honduras and Nicaragua. Even where coverage is compulsory, due to the special nature of this type of work, enforcement of the law is difficult, particularly if the worker lacks a contract. Data on coverage of domestic servants is very scarce; in Brazil affiliation is obligatory and only 27% of them were covered in 1999, but coverage rose to 100% among domestic servants who had a contract

Effects of Pension Reforms on Universal Coverage (Pinheiro 2001). In Costa Rica coverage is also mandatory and 39% of domestic servants were contributors to the general program in 2004, compared with 76% among salaried workers (SPa 2005b). In the majority of countries an enterprise must have a minimum number of employees (usually from five to ten) to impose compulsory coverage and deduct contributions. Employees in microenterprises, excluding professionals and technicians, ranged from 7 to 16% of the urban employed labor force in 2003–4 (ECLAC 2005a), but that figure was probably underestimated because most microenterprises are informal and it is extremely difficult to detect them and control their compliance; the cost of inspection may be similar to or higher than the revenue collected from contributions. In Mexico the law requests a minimum of only two employees for compulsory affiliation but enforcement is very weak (Castro 2003). Collecting agencies usually give priority to large and medium enterprises, because they are easier to detect and collect, hence neglecting smaller enterprises. Household surveys taken in fourteen countries in 2000–3 demonstrate that the smaller the enterprise, the lower its employees’ coverage and vice versa: 2 to 17% in small enterprises except in three countries, 29 to 80% in medium enterprises, and 56 to 96% in large enterprises except in one country. ‘Coverage in large establishments is between three and 30 times higher that in small establishments [thus supporting] the hypothesis that larger establishments tend to be formal . . . while the smaller ones are basically informal’ (Rofman 2005: 18).

3.2.2. Rural sector In the least developed countries of the region, an important part of the employed labor force is still in agriculture, for instance, 32– 37% in Guatemala, Honduras, Nicaragua, and Peru (ECLAC 2005c). Except for salaried workers employed in large plantations, this sector is difficult to cover because it is mostly comprised of peasants, seasonal workers, selfemployed, sharecroppers, and squatters. Vir-

45

tually all of them are engaged in subsistence agriculture, either without an employer or only for a few months of the year; they are dispersed and their income is very low— all substantial barriers to their incorporation. In Central American and Andean countries, indigenous people constitute a significant part of the total population and are engaged in this type of work (for analysis of the size and coverage of indigenous people see Section 8.3). In 2001–4 the regional rural labor force engaged in self-employment and unpaid family work averaged 56% and ranged from 60 to 86% (ranked higher to lower) in Bolivia, Peru, Paraguay, Brazil, Honduras, Guatemala, and Ecuador, but was only 26% in Costa Rica and 32% in Chile—no data were available for Argentina, Cuba, and Uruguay (Table 3.3). Legal coverage of all agricultural workers is obligatory in five countries: Argentina, Chile, the Dominican Republic, Uruguay, and Venezuela. Three countries have special regimes for rural workers or peasants: Brazil, Ecuador, and Mexico (see below). In Colombia there are many exceptions; in Panama compulsory coverage is only for those who work more than three months annually; in Bolivia and Costa Rica only salaried workers are covered; in Cuba state farm workers and cooperative members are mandatorily covered whereas small private farmers can join voluntarily (some groups of agricultural workers have been incorporated by law or agreements); and in El Salvador, Guatemala, and Honduras only those in large plantations are covered. Nicaragua’s reform law postponed to 2007 stipulates compulsory coverage for agricultural workers. No information was available for Haiti, Paraguay, and Peru. These workers have the lowest pension coverage among the groups difficult to incorporate or are excluded altogether: in 1996–7, 98% were not covered in Honduras, 96% in Colombia, 94% in El Salvador, 88% in Guatemala, and 87% in Nicaragua (Mesa-Lago et al. 1997a, 1997b; Mesa-Lago and Durán 1998; Ayala and Acosta 2002). Household surveys in 2000–3 showed that in nine out of eleven countries coverage of the rural labor force was from onethird to one-sixth the level of urban coverage.

46

Reassembling Social Security

The three countries with higher rural coverage were Chile and Costa Rica (41 and 44%, respectively), whereas in Brazil it was 18% but probably excluded the rural pension scheme (no data were available for Argentina and Uruguay); in the remaining eight countries rural coverage ranged from 2.5 to 12.7% (Rofman 2005).10 In the 1960s and 70s Brazil, Ecuador, and Mexico introduced special social insurance schemes that included a basic pension for rural workers or peasants who lack an employer. Brazil’s scheme for low-income rural workers has steadily expanded in recent years reaching about half of them; the scheme has a registration of pensioners but not of active workers (Interviews 2004). Ecuador’s scheme requires that nonsalaried peasants organized in cooperatives, communes, and agrarian associations request voluntary incorporation, and the decision by social insurance (IESS) is based on a minimum size of the community and viable facilities. Coverage steadily rose in the 1980s, peaked at close to one million beneficiaries in 1995 (21% of the rural population), and thereafter decreased to 18% of the rural population in 2005, leaving out 68% potential beneficiaries (Mesa-Lago 1993; IESS 2005; World Bank 2005e; González 2006). Conversely, Mexico’s peasant scheme selects the communities based on a set of indicators on marginality (e.g. poverty incidence, high proportion of uninsured, concentration of indigenous people), but also some logistic criteria as access to roads and proximity to some facilities. Coverage peaked at fourteen million peasants in 1985 (57% of the rural population), thereafter declined and stagnated around ten million or 29% of the rural population in 2005 (Mesa-Lago 1992; Munguía 2006).

3.2.3. Measures to improve coverage of difficult groups Making compulsory the affiliation of the selfemployed and other groups difficult to cover although positive would not necessarily result in effective coverage in most countries, due

to several reasons: (a) job instability, earnings usually below the minimum salary, and low capacity to contribute; (b) in addition to their own contributions the self-employed usually must pay the employer’s contribution despite lacking an employer; (c) the average self-employed percentage contribution is three times that of the average salaried worker and, although such combined percentage is sometimes imposed on the minimum salary, it is still quite a heavy burden (see Section 6.1); (d) serious obstacles for the identification, affiliation, estimation of income, collection of contributions, and inspection on compliance of these workers, because they lack an employer, are dispersed, and must take the initiative for enrolment and payments;11 and (e) lack of incentives for affiliation to the pension program, because they have other more urgent needs, as well as a potential eligibility for social assistance pensions (Mesa-Lago 2001c). Incentives (particularly fiscal subsidies), sanctions, and special programs to cover these groups have achieved some success in a few countries. The Costa Rican state makes a contribution to the sickness-maternity program, both as a third party and in lieu of the employer contribution, on behalf of lowincome self-employed and other voluntarily insured; the 2001 reform law extended such fiscal subsidy to the public pillar of the pension program, a mandate finally implemented in 2006, but is granted regardless of income hence providing a regressive subsidy to high-income professionals (Rodríguez Herrera 2006). Chile’s legal draft of reform of 2006 grants the self-employed equal rights with salaried workers and incentives to gradually incorporate them (e.g. tax deduction of contributions, granting family allowance and coverage in occupational risks, and subsidies to individual accounts) until affiliation eventually becomes obligatory in ten years (Gobierno 2006). The Fund of Pension Solidarity in Colombia is geared to extend coverage of the self-employed and other groups who lack access to social insurance due to their poor socioeconomic characteristics but coverage was less than 10% of total self-employed

Effects of Pension Reforms on Universal Coverage in 2004 (SBC 2005). Brazil sanctions since 2003–4 increased affiliation by three million: employers hiring self-employed who don’t deduct their contributions from the salary paid, face stiff sanctions and even prison, whereas high-income self-employed professionals who are not up-to-date in their contributions are not eligible for fiscal incentives (Schwarzer 2006). In Uruguay a legal draft of tax reform in 2006 obliges all self-employed to pay taxes, a measure intended to expand their coverage (Bertranou and Vázquez 2006; Murro 2006a). Some of these plausible measures must control a potential moral risk: low-income salaried workers, in conspiracy with their employers, may simulate to be selfemployed to take advantage of fiscal subsidies and for the employers to evade their contributions. In a few countries (Brazil, Costa Rica, Mexico) groups of self-employed workers and/or peasants have organized associations or trade unions and signed agreements with social insurance, taking responsibility to register their members, collect their contributions, and transfer them to the pension program (contributions and benefits are usually set lower than those in the general system). These agreements, however, only cover a small proportion of those workers (who have a preference for health care protection instead of pensions), have confronted administrative problems, and lack the needed support and incentives from social insurance institutions. Microinsurance, recommended by the ILO as an alternative to formal insurance in low-income developing countries, has been successful in some nations of Africa, as well as in Bangladesh, India, and the Philippines, but not in Latin America (ILO 2001a; Reynaud 2002).

3.3. Social assistance pensions for the uninsured and impact on poverty Poverty incidence in Latin America averaged 42% in 2001–4 and exceeded that average in

47

most countries (particularly in the latecomerlow group): 77% in Honduras, 60–69% in Nicaragua, Bolivia, Paraguay, and Guatemala, and 45–51% in Peru, Colombia, Ecuador, El Salvador, the Dominican Republic, and Venezuela (Haiti probably had the highest poverty incidence). Countries in the pioneer-high group had the lowest poverty incidence: 15–29% in Argentina, Costa Rica, Cuba, Chile, and Uruguay (data on Cuba and Uruguay are limited to the capital and urban areas, respectively) but Brazil had 38%. Panama and Mexico, both in the intermediate group, had incidences of 34% and 37%, respectively (Table 3.3). Coverage of the labor force by contributory pension programs increases with income as proved by data from fourteen countries (with very few exceptions) distributed by income quintiles in 2000–3: from 0.1 to 11% covered in the poorest quintile; 1 to 48% in the second quintile; 6 to 70% in the third quintile; 12 to 80% in the fourth quintile; and 31 to 90% in the richest quintile. Countries with the lowest coverage in the poorest quintile were those with the highest poverty incidence and with a large indigenous population in most of them: all countries in the latecomer-low group; Bolivia, Ecuador, and Peru in the intermediate group; and Argentina in the pioneer-high group due to the crisis. Conversely, countries in the pioneer-high group, which also have the lowest poverty incidence, exhibited the highest coverage in the lowest quintile: 16 to 30% in Brazil, Chile, Costa Rica, and Uruguay (based on Rofman 2005). In addition, surveys systematically show that education parallels income distribution: the higher the educational level the better coverage and vice versa. Finally, in countries with a low overall coverage, the great majority of insured are concentrated in the capital and major urban areas and very few or none in the countryside where poverty incidence is much higher. For instance, 63% of the Dominican Republic insured in 2005 were in the metropolitan area and the second city, while the rest of the provinces ranged from less than 1 to 5% of the total insured (SIPEN 2006); and 64% of Guatemalan insured were concentrated in the

48

Reassembling Social Security

capital-city department in 2004 and dropped to 3% in the second most urbanized department, and in the remaining twenty departments ranged from 0.1 to 2.4% (IGSS 2005). A Chilean survey in 2004 identified the features of nonaffiliates as follows: 66% had very low income, 46% were poor, 52% had very little education, and 75% had not worked in the last three years. When asked why they were not affiliated to the pension system they gave the following reasons by order of importance: the system is not compulsory, lack of money, ignorance of the system, never worked, distrust of the private system, being a housewife or having unstable employment, and very expensive system (Bravo et al. 2006). A recent study characterizes social assistance pensions as pure fiscal cash transfers not tied to a contributory history, intended as a safety net for the poor or persons who fall through the cracks of a contributory program, and granted usually to the elderly, but also in some Latin American countries to the disabled or widows/orphans. Social assistance pensions can be ‘supplementary’ to contributory programs with relatively high or low coverage or ‘core’ when granted to a large part of the population in countries where often there is no contributory program or it has low coverage (few countries have parallel contributory and core social-assistance schemes of similar importance). Each of these two variants can in turn assume two forms: means tested and targeted on the poor or universal and flat regardless of income (Palacios and Sluchynsky 2006). The most typical combination in Latin America is supplementary to a contributory program with relatively high coverage and targeted on the poor. In the design of structural reforms, little or no attention was paid to the poverty prevention pillar in countries that already had it in place, and even less in countries lacking such a pillar (Crabbe and Giral 2005). Before the reforms there were supplementary meanstested social assistance pensions targeted to the uninsured poor in Argentina, Brazil, Chile, Costa Rica, Cuba, and Uruguay. These are the countries with the lowest poverty incidence

and the highest coverage in the contributory system, except for Argentina during the crisis (Tables 3.2 and 3.3). Brazil also has the special pension scheme (more like the core type) for low-income rural workers, which is not means tested and only requires proof of the years of work as rural worker; very few beneficiaries are in extreme poverty, some are not even lowincome and receive a fiscal subsidy (Interviews 2004).12 Chile introduced in 2004 a scheme (Chile Solidario) that provides assistance pensions compatible with a bonus to the extreme poor and old-age indigents in a waiting list for a pension; a 2006 law grants the universal right to social assistance pensions to those aged 65 and older or incapacitated who qualify as extreme poor, aimed at the elimination of quotas and waiting lists (Ley 19,949 2004; Ley 20,102 2006). All social assistance programs are means tested (except for the rural pension in Brazil that is not truly assistance): personal or household income cannot exceed a percentage of the contributory pension or the minimum salary (see Section 4.3.4), but except for Chile there is little evidence that such a test is applied rigorously and revised periodically. In Costa Rica there are persons not truly in need who receive these pensions and the law of reform stipulates that the assistance pension must be extended to all those in need, hence it is essential to target it better and refine the means test. In Argentina, in addition to the assistance pensions, there are pensions awarded by special laws (to former Presidents, Nobel laureates, etc.) and ‘grace pensions’ (granted at congress discretion) whose beneficiaries usually have adequate resources; in 2001 both types of pensions combined were half of the total number of noncontributory pensions and their costs tantamount to 57% of total costs. Poor targeting and inadequate means tests lead to free raiders and, combined with noncontributory pensions granted to persons not in need, take precious resources needed to expand pensions to the truly poor (Mesa-Lago 2001a; Bertranou et al. 2002; ILO 2002b; Martínez and Mesa-Lago 2003; Barrientos 2006a).

Effects of Pension Reforms on Universal Coverage The Bolivian reform created a universal basic pension (Bonosol) of the core not meanstested type that awards an annual lump sum, regardless of income, to all citizens above age 65 provided they were age 21 and older by the end of 1995. Bonosol was granted to 450,114 in 2005, 84% of its beneficiaries lived in urban areas and only 16% in rural areas, which suffer the highest poverty incidence and concentrate the indigenous population.13 Most poor in rural areas and some in urban slums lack identity papers or if they hold them have irregularities that prevent collecting Bonosol; rural poor often live far away from the paying offices and incur travel expenses. Furthermore, Bonosol is granted to 80% of retired workers who also receive a contributory pension but 75% of all elderly do not receive such a pension. Therefore Bonosol is not a supplementary social assistance pension but a small universal basic pension not targeted on the poor but aimed at a certain generation only (Müller 2004; Gill, Packard, and Yermo 2005; Rofman 2005; Bertranou 2006; Palacios and Sluchynsky 2006; SPVS 2006a). The remaining thirteen countries (five with private systems and eight with public systems) don’t currently grant social assistance pensions and the immense majority endures a low coverage in the contributory program (difficult to extend) and the highest poverty incidence. Ecuador and Mexico have social insurance pension schemes for peasants, many of whom are poor; a few other countries have very minor social assistance aid schemes that cover a tiny part of the population, for instance, the Dominican Republic and Ecuador.14 The six supplementary targeted meanstested social assistance pensions are not necessarily guaranteed to all the poor, because they depend on available fiscal resources and most countries have quotas and waiting lists (Mesa-Lago 2001a). These pensions covered the following percentages of the total population in 2000–5:15 0.9% in Argentina, 1.2% in Brazil (plus an additional 3.5% by the rural program and 0.4% by other aid schemes, for a total of 5.1%, the highest coverage among

49

all countries), 1.8% in Costa Rica and Cuba, 2% in Uruguay, and 2.3% in Chile (plus an additional 1.4% in Chile Solidario for a total of 3.7%). Bolivia’s Bonosol beneficiaries represented 4.8% of the total population in 2005; it is argued that either 75 or 100% of the population aged 65 and over is covered by it (Müller 2004; Palacios and Sluchynsky 2006), but Table 3.5 shows that only 15% of the total was covered perhaps because Bonosol was excluded. Coverage by all types of social assistance pensions in the seven countries range from 1 to 5% of the total population, only a fraction of their respective poverty incidence that range from 15 to 45% (62% in Bolivia) of the total population. Despite their relative small proportion and limitations, 1997–2000 data show that social assistance pensions had a significant positive impact on reducing poverty and indigence by the following respective proportions: 31 and 67% in Argentina, 29 and 96% in Brazil, 19 and 69% in Chile, and 24 and 21% in Costa Rica (Bertranou et al. 2002; Schwarzer and Quirino 2002; Mesa-Lago 2005b).16 The crisis of 2001–2 in Argentina, however, reversed that progress: in 2003 poverty had risen to 45% of the population, and 22% of the pensioners and 50% of those aged 60 years and older who did not receive a pension were poor (MTESS 2003). A recent World Bank study concludes that well-targeted social assistance programs have an immediate benefit on reduction of poverty and inequality, albeit the latter effect is attenuated by the relatively small size of assistance transfers relative to transfers to social insurance, which are several times higher and tend to be regressive because they are accessible mainly to employees in the formal sector and financed by general taxation, while poor households work mostly in the uninsured informal sector (Perry et al. 2006). The percentage of assistance pensions relative to the total number of pensions (contributory and assistance) fluctuated from 9 to 31% in the six Latin American countries with supplementary means-tested schemes. ECLAC (2006) has estimated that granting noncontributory

50

Reassembling Social Security

Table 3.5. Percentage of the populations aged 65 and above covered by pensions in private and public systems, between 2000 and 2003 Systems/Countries

Total

Men

Women

Private Systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay

68.3 14.7 63.8a 18.6 62.0b 10.9 14.5 19.2 23.2 87.1

74.3 16.1 72.6 22.9 71.1b 15.5 18.0 17.8 27.7 76.9

64.2 12.7 57.2 13.1 54.2b 5.9 9.6 18.0 14.6 78.9

Public Systems Brazil Ecuador Guatemala Nicaragua Panama Paraguay Venezuela

85.9 15.2 11.3 4.7 45.0 19.6 23.9

80.0 17.3 17.0 n.a. 52.0 18.9 26.7

76.4 10.8 4.6 n.a. 48.2 14.5 18.0

Sources: Based on household surveys; data compiled by Rofman 2005, except Costa Rica from INEC 2005; no data available for Cuba, Haiti, and Honduras. n.a. = non available. a Coverage in Chile increases to 76% adding noncontributory pensions (Bertranou, Gana, and Vázquez 2006). b Based on the household survey of 2005, Rofman gives 36.6, 48, and 36.6% probably excluding noncontributory pensions.

pensions (both targeted on the poor and flat universal) to those aged 65 and above in seventeen countries of the region would reduce poverty by an average of 18 percentage points albeit with significant differences among countries. Several laws of reform mandated some type of social assistance for the poor but have not been implemented yet. Colombia’s first law set aid equal to 50% of the minimum wage for elder indigent people but was not enforced; and the second reform law established an account in the solidarity fund to help people in extreme poverty or indigence but we lack information on its current status

(Ayala and Acosta 2002). Costa Rica’s reform law mandated the universality of the social assistance pension for all the elderly in need, but that clause had not been implemented by March 2006 and there were 4,000 elderly in a pension waiting list (Martínez and Mesa-Lago 2003; Rodríguez Herrera 2006). The Dominican Republic’s reform law prescribes a ‘subsidized’ or noncontributory pension for poor who are aged 60 and above, severely disabled, or single mothers, its implementation planned for 2004 had not materialized by April 2006.17 The reform laws of Ecuador and Venezuela stipulate an assistance pension for those in need, but the former was partly declared

Effects of Pension Reforms on Universal Coverage unconstitutional in 2005, while the latter had not been implemented by the end of 2006. The Mexican government announced in January 2006 a new social assistance scheme that would pay 15% of the average income per capita to one million elderly in the poorest communities, tantamount to 0.9% of the total population (Palacios and Sluchynsky 2006). Chile’s legal draft of reform of 2006 grants a basic solidarity pension at age 65 for those in the 60% of the population with lower income, regardless of whether they have contributed or not, and is intended to become universal and replace all social assistance pensions (Gobierno 2006). World Bank officials state: In the last two decades, social security reforms in Latin America and elsewhere have focused mainly on restructuring the ‘income-replacement function’ of pension systems [the private pillar of compulsory savings], while their ‘povertyprevention function’ . . . has not received the attention it deserves . . . this component should be the main attraction of a social security system, not a sideshow . . . the lack of attention to this core component of government policy is a serious mistake.

The Bank officials appropriately recommend fortifying the public pillar in order to reduce poverty (Gill, Packard, and Yermo 2005: 199). Esping-Andersen and Myles (2006) make a persuasive case that a guaranteed basic pension, financed from general revenue, is an effective tool against old-age poverty, helps to diversify the financing of pension expenditures and increases progressiveness.

3.4. Coverage of the older population Measuring coverage of the elderly faces fewer problems than coverage of the labor force, because it is based on actual recipients of a pension rather than on those who may have a potential right to that benefit (ISSA 2003b; Rofman 2005). Household surveys taken in 2000–3 provide information on seventeen countries (the ten with private systems and

51

seven with public systems) on coverage of the population aged 65 and older combining all pension programs, including social assistance, and distributed by gender (Table 3.5). Five countries of the pioneer-high group have the highest coverage: 87% Uruguay, 86% Brazil, 68% Argentina, 64% Chile, and 62% Costa Rica18 (no statistics are available for Cuba19 ). The lowest poverty incidence in the population aged 65 and above was found in Brazil and Chile (Gill, Packard, and Yermo 2005). Panama had the highest coverage in the intermediate group with 45%; the remaining six countries have lower coverage, ranging from 24% in Venezuela to 15% in Bolivia (probably excluding Bonosol). With the exception of Paraguay, countries in the latecomer-low group have the lowest coverage: 5% in Nicaragua, 11% in the Dominican Republic and Guatemala, and 14.5% in El Salvador (no data are available on Haiti and Honduras but probably they have even lower coverage). Estimates on coverage of the population aged 65 and older differ significantly based on other household surveys.20 Coverage of women at age 65 is considerably lower than coverage of men at the same age in all countries except in Uruguay where the opposite is true; in Mexico coverage is virtually the same, but a more recent survey shows significantly lower coverage among women. The gap in gender coverage is threefold in the Dominican Republic and about twofold in El Salvador and Peru; such a gap is smaller in countries of the pioneer-high group, as well as in Panama in the intermediate group. In the majority of private systems it is impossible to measure the impact of structural reforms on coverage of the elderly because of the relatively short time that the reforms have been in operation as well as the lack of historical series. In Chile the only available data are for 1992–2003, hence albeit impossible to compare coverage before and after the reform it is feasible during the reform: coverage at age 65 and above increased from 75 to 79% in 1992–2000, but coverage by contributory pensions diminished from 67 to 64%,

52

Reassembling Social Security

while by social assistance pensions rose from 8 to 15%, and overall coverage declined to 66% in 2003, below the 1992 level. The contributory system protection of the elderly is eroding largely because many insured don’t qualify for a minimum pension and that trend should continue in the future (Arenas de Mesa and Gana 2003).21 In Argentina, poverty incidence among those aged 60 and older without a pension jumped from 27% in 1994 (the year the reform started) to 50% in 2002, while indigence incidence in that age cohort increased from 7 to 16% (Golbert and Lo Vuolo 2006). Coverage of the population aged 65 and older dropped from 77% in 1994 to 72% in 1999, almost at an average of one percentage point per year and two projections indicate a decrease in coverage of the elderly: at age 60 and above from 65 to 35% in 2000–30, and at age 65 and above from 70 to 45% in 2005–30, both exhibiting an annual rate of decline of about one percentage point (Bertranou et al. 2001; Grushka 2002; MTESS-ILO 2005). In Uruguay coverage of the population aged 65 and above did not change in 1995–2002 but that in the 60–4 age bracket decreased from 75 to 71% (Coimbra and Forteza 2004). World Bank officials admit that ‘in several Latin American countries the share of the elderly population receiving pension benefits is falling’, but predict that ‘the prospect of poverty during one’s retirement years will become relatively rare in Latin America over time’ (Gill, Packard,and Yermo 2005: 114, 125). Such prediction rests on three optimistic assumptions: sustained economic development, rising average life-time income, and greater opportunity to save and accumulate assets, as well as evidence from developed countries. The title of their table comparing poverty between the elderly and the overall population in eighteen Latin American countries states that poverty among the elderly is as frequent as that among other age groups, but an inspection of the table data leads to contradictory results: in twelve countries the elderly have lower poverty rates than the overall population, whereas in six of them the opposite is true (the poverty rates of elderly are

worse in five private systems and in one public system).22 And Barrientos (2006a) has found that in ten of seventeen Latin American countries old people are overrepresented among the poor. The Bank officials then argue that poverty rates (measured by current income) often fail to capture wealth and, based on a single survey on income distribution in Lima, induce a general conclusion that older people have the highest wealth accumulation among all age groups, thus compensating for their lower income (Gill, Packard, and Yermo 2005: 199–204). And yet World Bank officials acknowledge that structural reforms have focused on the private savings pillar—where evidence from Argentina, Chile, and Uruguay shows a decline in pension coverage of the elderly— and neglected the public pillar of poverty prevention: half of the countries with structural reforms don’t grant assistance pensions, and in the other half such pensions are not guaranteed to all the poor. Furthermore, according to World Bank data, between one-third and onehalf of the insured in contributory private systems will not gain the right to the minimum pension, which are insufficient to satisfy basic needs (see Section 2.3.1; Golbert and Lo Vuolo 2006).23 All of the above suggest that contrary to the Bank officials’ prediction, coverage of old people will deteriorate and poverty will rise unless rapid measures are taken to expand coverage of both the active labor force and the elderly.

3.5. Impact of the reforms on coverage 3.5.1. Coverage in private and public systems, and regional trends The most significant challenge faced by social insurance pension systems in Latin America, regardless of whether they are private or public, is to maintain and expand coverage. Contrary to the common assumption that the structural reform and the private system

Effects of Pension Reforms on Universal Coverage would increase coverage, data show a decline in protection. Private coverage based on affiliates is twice or thrice as high as those based on active contributors (averages of 60 and 25%, respectively, in 2005) because insured are unemployed or may have abandoned the labor force or shifted from the formal to the informal sector or are counted twice; coverage based on contributors in turn may underestimate real coverage because some affiliates who did not pay in the last month could do it in the immediate future and retain effective coverage. Chilean coverage based on affiliates was 116 versus 60% of the labor force based on active contributors, demonstrating the noted overestimation. Coverage based on more reliable active contributors, before the reform and in 2004, declined in all ten countries with structural reforms, and the weighted average decreased from 38 to 26%. According to a historical series based on active contributors in Chile, which has the longest functional structural reform, coverage declined between 1973– 5 (before the reform) and 2000. Comparisons in coverage between private and public systems are not precise, particularly due to different periods used to define the condition of active contributor (the last month in all private systems whereas they could be larger in some public systems), but estimates indicate that the weighted average in private systems was 26% vis-à-vis 39% in public systems (both above the ILO minimum norm of 20%); after subtracting Brazil, the average in public systems decreased to 20% (lower than the private average). Estimates of coverage based on household surveys that include the insured in separate schemes, albeit leaving out key countries and being afflicted by problems, show a weighted average of 36% in private systems smaller than the weighted average of 41% in public systems. Regardless of whether the system is private or public, the older it is and the bigger the formal labor sector, the higher its coverage and vice versa, hence countries in the pioneerhigh group have the highest coverage and those in the latecomer-low group the lowest coverage, with few exceptions. Coverage in

53

the pioneer-high group stretched from 45 to 59%, including Panama, which is in the intermediate group and significantly improved its coverage and regional ranking in the last two decades, and excluding Argentina whose coverage (24%) shrank below the regional average. In the intermediate group, coverage ranged from 11 to 28%, in Peru and Bolivia diminished and is among the four lowest in the region. In the latecomer-low group coverage oscillated from 8 to 20%, the highest in Guatemala and El Salvador (tied), followed by Honduras and Nicaragua (the last four higher than Bolivia and Peru), the Dominican Republic, Paraguay, and probably Haiti. Public systems have not only kept their ranking in the region, but some have improved it.

3.5.2. Groups difficult to cover Social insurance pension programs (particularly but not exclusively in private systems) were originally designed for workers employed in the urban formal sector with stable jobs, medium-high salary, mostly males and with high density of contribution, but in the region the majority of the labor force is informal and/or agricultural, with unstable employment, low wages, and poor density of contribution (especially among women), hence it is very difficult to extend coverage. The informal sector grew from 43 to 47% of urban employment in 1990–2002 and its most important component, self-employment, ranged from 31 to 47% in the least developed countries but is lower in the pioneer-high group. Argentina, Brazil, Chile, Costa Rica, and Uruguay (all in the pioneer-high group) have the most expanded legal mandatory coverage of the informal and rural sectors (except for the self-employed in Chile), while countries in the latecomer-low group have legal voluntary coverage or impose restrictions. In fourteen countries legal coverage of the self-employed is voluntary and one country excludes them; this sector is very difficult to incorporate because of instable jobs, low income, the addition to its own contribution

54

Reassembling Social Security

of the employer’s contribution (that they lack), and obstacles for affiliation and collection of payments. Statistical coverage is higher in countries with mandatory affiliation (30, 29, and 23 in Argentina, Uruguay, and Brazil, respectively) and lower in those with voluntary affiliation (0.2 to 5% in five countries), Costa Rica is an exception with 24% when affiliation was voluntary and should increase now that it is mandatory. In Chile 28% of the labor force is self-employed; coverage is voluntary and in the private system it fell from 12 to 5% during the reform (mostly highincome professionals). Compulsory legal affiliation helps but does not necessarily solve the coverage gap; incentives (particularly fiscal subsidies) and sanctions appear to have increased coverage in Argentina, Brazil, Costa Rica, and Uruguay, but subsidies have not been successful in Colombia. Another informal labor group with low coverage is that of domestic servants who have legal compulsory coverage in twelve countries (albeit difficult to enforce in practice) and voluntary in only five (nonenforced laws in four countries stipulate mandatory coverage); 27 and 39% of these workers are covered in Brazil and Costa Rica, respectively, both with compulsory affiliation, and no data are available from other countries. Employees of microenterprises are legally excluded in most countries; data on the employed labor force by size of enterprises in fourteen countries demonstrate that the smaller the enterprise, the lower the coverage and vice versa: statistical coverage in large enterprises is between three and thirty times higher than that in microenterprises. The rural labor force engaged in selfemployment or unpaid family work averaged 56% in the region and in the least developed countries rose to 63–86%; these workers are also very difficult to incorporate because of the nature of their work, wide dispersion, and low and unstable income. Full legal coverage of agricultural workers exists in five countries, the rest have special regimes or impose restrictions; the proportion of the agricultural labor force that is uninsured in five countries oscillate between 87 and 96%; coverage

of the rural population usually is one-third to one-sixth smaller than coverage of the urban population. Only three countries have introduced special social insurance pensions for rural workers or peasants: coverage of the rural labor force was 18% in Ecuador, 29% in Mexico, and 50% in Brazil in 2004–5; in the first two, coverage has declined in recent years while in the third it has steadily increased.

3.5.3. Social assistance and poverty Poverty incidence in Latin America averages 42% of the total population and most countries exceed that average particularly in the latecomer-low group. Coverage of the labor force by contributory pensions increases with income: countries with the lowest coverage in the poorest quintile are those with the highest poverty and vice versa. A declining coverage of the labor force in most countries is or will result in decreasing protection of the elderly, a population group rapidly growing in the region. Structural reforms emphasized the private mandatory savings pillar but paid little or no attention to the poverty prevention pillar in countries that already had it in place, and even less in countries lacking such a pillar, priorities now reversed by World Bank officials. Before the reform, social assistance pensions for the uninsured poor were established in four of the ten private systems (Argentina, Chile, Costa Rica, and Uruguay), as well as in two of the ten public systems (Brazil and Cuba); all of them are supplementary means-tested schemes in pioneer-high group countries with the highest regional coverage in their contributory programs and the lowest poverty incidence, except Argentina, after the crisis. Bolivia’s core scheme, the only one created by a structural reform, is not targeted on the poor but granted regardless of income, 84% of beneficiaries are urban residents and 80% also receive a contributory pension, whereas 75% of the elderly don’t get a pension. The number of assistance pensioners as a percentage of the total population in 2000–5

Effects of Pension Reforms on Universal Coverage was 1% in Argentina; 2% in Costa Rica, Cuba, and Uruguay; 4% in Chile; and 5% in Bolivia and Brazil (including the rural program and other aid schemes). Such coverage is only a fraction of these countries poverty incidence that ranged from 15 to 45% of the total population (62% in Bolivia). Furthermore these programs are submitted to financial constraints and most have quotas and waiting lists. Despite their limitations, however, social assistance pensions have significantly reduced poverty by 19–31% and extreme poverty by 21– 96% in four countries. Contrasting with the above restrictions, Argentina awards noncontributory pensions to nonpoor by special laws (to former presidents, Nobel laureates, etc.) and at the discretion of congress (‘grace pensions’); the combined number of these pensions was half the total number of noncontributory pensions and 57% of their total costs in 2001. In Costa Rica nonpoor persons receive social assistant pensions making it essential to target them better and refine the means test. Thirteen countries lack social assistance pensions and they endure the lowest coverage in the contributory program and the highest poverty incidence. Reform laws in four countries (Colombia, the Dominican Republic, Ecuador, and Venezuela) stipulate the creation of social assistance pensions for the poor, but they had not been implemented by the end of 2006. Chile extended coverage to all the indigents in 2006 and the legal draft of 2006 creates a universal basic pension for the population in the lowest 60% of income. The structural reforms of El Salvador, Mexico, and Peru (as well as the failed reform of Nicaragua) did not include a social assistance pension, and four poor countries with

55

public systems are in the same situation. Often high costs are alleged to be the cause of lack of social assistance but fiscal transfers to this program are a fraction of transfers to social insurance, which are regressive because they are received mainly by employees in the formal sector and financed by general taxation, while the poor work mostly in the uninsured informal and rural sectors. Estimates of costs of supplementary means-tested schemes indicate that they are financially viable and would reduce poverty by about 18 percentage points.

3.5.4. Coverage of the elderly In 2000–3 coverage of the population aged 65 and above ranged from 62 to 87% in Argentina, Brazil, Chile, Costa Rica, and Uruguay (all in the pioneer-high group); declined to 15–24% in the intermediate group (Bolivia, Colombia, Ecuador, Mexico, Peru, and Venezuela) except for Panama with 45%; and was lowest (5–14%) in four countries of the latecomer-low group. The impact of structural reforms on coverage of the elderly is difficult to measure due to lack of historical series, but partial data from three countries indicate a decline: Chile’s coverage increased slightly in 1992–2000 because of the expansion of social assistance pensions that compensated for the decline in coverage by contributory pensions, and overall coverage had deteriorated by 2003 below the 1992 level; Argentina’s coverage fell at an annual average of almost one percentage point in 1994–9 and it is projected to continue declining at one percentage point in 2000– 30; and Uruguay’s coverage decreased in 1995– 2002. Such evidence is contrary to the World Bank prediction that poverty among old people will decrease over time in the region.

Notes 1. An alleged higher affiliation in Mexico is attributed to the workers’ precise knowledge of the amount of their contributions and capital returns, the strict link between contributions and pensions, and that their individual accounts won’t be used for other purposes (Hernández 2001).

56

Reassembling Social Security 2. ‘Ironically, while a worker’s density of contributions is relatively more important in assessing whether they are covered in [a private rather than in a public] system, the private and decentralized structure of the reformed system in Chile has made that data on contribution history unavailable to government researchers’ . . . ; administrators have successfully resisted in court the state request for those data (Gill, Packard, and Yermo 2005: 193); in 2002 and 2004 the government conducted surveys of affiliates to estimate the contribution density (EPS 2004, 2006). 3. Chilean coverage statistics have been manipulated by comparing 1981 (at its trough) with that in 2005, skipping the highest coverage in 1973–5 and 1997, and giving the false impression of a significant steady increase (Arthur 2006; Larroulet 2006). 4. Cuba officially claims 100% coverage (Travieso 2002) but only 15% of workers contribute and regional comparisons are impossible; grossly estimated coverage fell from 90% to 66% in 1989–97, because the 1990s economic reform reduced state employment (covered) and increased private-informal employment (partly not covered); since 2003 there has been a reversal in that trend (ILO 1999; Mesa-Lago 2003b, 2005b). Haiti’s main program coverage was roughly 1.6% in 1999 (based on IADB 2003, ECLAC 2005b). 5. If other important schemes (e.g. civil servants in Brazil, Mexico, and Venezuela) are included, the public average increases to 46% (32% excluding Brazil), but the comparison requires estimates of separate schemes in private systems (Mesa-Lago 2004 updated with Brugada 2004 and MPS 2005b). 6. The association of superintendencies of private systems publishes since 1999 a standardized series on labor force coverage based on active contributors (AIOS 2000–6); public systems lack a similar association and series. 7. In Chile the increase is only one percentage point despite the inclusion of armed forces and policemen; in Costa Rica the increment of 3.5 points is due to the addition of teachers and judges (SPa 2005b), and in Ecuador the increment of 2.5 points appears underestimated relative to the 7 points covered by separate schemes (World Bank 2005e). Conversely, Uruguay coverage is 3.5 points lower although including four separate small schemes; the same happens in Bolivia and Guatemala. 8. World Bank estimates of labor force coverage in eight private systems based on household surveys in 1998–2000 include separate schemes but are different to those in Table 3.3 (third column) also based on household surveys: Argentina 36%, Bolivia 10%, Chile 63%, Costa Rica 23%, El Salvador 26%, Mexico 46%, Nicaragua and Peru 11% (Gill, Packard, and Yermo 2005: figure 1.1). 9. Monotributista is a small contributor (a person, cooperative member or less than three partners in a commercial society) who pay taxes sporadically; they can choose between the general tax and a simplified system to which they make a single payment for all taxes including social security (MTESS-ILO 2005). 10. Coverage in the primary sector (mainly agriculture) ranged from 2 to 18% in ten less developed countries and 28–54% in Argentina, Chile, Costa Rica, and Uruguay, while in the secondary sector it was 15–64% and in the tertiary sector 20–68% (Rofman 2005). 11. In Mexico the self-employed must present a tax-office certificate that they are up to date in their contributions (Hernández 2001). 12. Brazil also has two other assistance schemes: for the urban elderly and incapacitated and for the old and invalid in urban and rural communities that covers 0.4% of the total population (Palacios and Sluchynsky 2006). 13. In 2002, the poverty incidence was 52% in urban areas and 79% in rural areas, extreme poverty or indigence was 21 or 63%, respectively (ECLAC 2004a). 14. In Ecuador a tiny welfare scheme created in 1998 provided small payments to 1.7% of the total population: the elderly, disabled, and mothers with small children, but with inadequate targeting. Affiliation was closed to be retargeted and reopened in 2003 but its implementation has been slow and confusing and there are no data on the number of beneficiaries (World Bank 2005e).

Effects of Pension Reforms on Universal Coverage

57

15. A more appropriate indicator would be the population aged 65 and over who are poor and covered by social assistance pensions, but systematic data were not available. 16. Barrientos (2006a) shows that the elimination of such pensions would significantly increase poverty incidence in Argentina, Brazil, Chile, and Costa Rica. 17. A small monthly aid is provided to indigents aged 65 and over (0.2% of the total poor) but the program is plagued by bureaucracy and unstable in its payments (Lizardo 2006). 18. Rofman (2005) gives 36.6% for Costa Rica probably excluding social assistance pensions, while Bertranou, Gana, and Vázquez (2006) give 76% for Chile including such pensions. 19. The author grossly estimates that 66% of the population aged 65 and over was covered by contributory pensions in 2001 (based on Travieso 2002; ONE 2003). 20. Based on previous household surveys (1997–2001), World Bank officials give lower figures than those in Table 3.5 for Argentina (66%), Bolivia (12%), Chile (41%), Colombia (15%), Costa Rica (33%), El Salvador (9%), and Peru (20%) but higher figures for the Dominican Republic (15%) and Mexico (20%) (Gill, Packard, and Yermo 2005). For other divergent estimates see Bertranou 2001; Bertranou et al. 2002; Durán 2002; Bertranou and Arenas de Mesa 2003. 21. According to another series, total coverage rose from 80 to 83% in 1990–6 and decreased to 76% in 2003 (below the 1990 level), contributory pensions fell from 73 to 62% in the entire period while noncontributory increased from 7 to 14% (Bertranou, Gana, and Vázquez 2006). 22. The initial version of this Bank report had a table with 1998 data on adult income by age groups in eight private systems showing with one exception that those 65 years and over had the highest poverty rates (Gill, Packard, and Yermo 2003: 152); the report’s final version replaced that table with the one discussed in the text. 23. Esping-Andersen and Myles (2006) show a relationship, albeit not perfect, between the degree of pension privatization and old-age poverty in developed countries (Australia, Canada, Western Europe, and the United States).

4

Effects on Equal Treatment, Solidarity, and ComprehensivenessSufficiency

This chapter analyzes three principles: equal treatment, particularly unequal entitlement conditions in separate privileged schemes and gender inequality; mechanisms in favor and against solidarity and their redistribution effects; and comprehensiveness and sufficiency of pensions, including retirement ages and spans, pension formulae in public systems, sufficiency of minimum and social assistance pensions, adjustment of benefits, and a discussion on whether private pensions are higher than public pensions.

4.1. Equal treatment 4.1.1. Survival of privileged schemes and their unequal entitlement conditions Most structural reforms standardized entitlement conditions and pension formulae in public systems (closed or open) and public pillars in mixed systems, as well as within private systems, hence strengthening the principle of equal treatment. Despite the alleged superiority of private systems, however, virtually all structural reforms have excluded several groups of insured with separate schemes of defined benefit that enjoy far superior entitlement conditions than the private system: lower retirement ages, less contribution years, seniority pensions regardless of age, pensions equal to 100% of the last salary and adjusted

to the salary of the personnel in activity. Such exclusion is true of public systems also. As more of these schemes exist, the worse are the unjustified inequalities in entitlement conditions.1 All countries have preserved the scheme of the armed forces except Costa Rica (which eliminated it in the 1940s) and Panama among public systems; the reform law of Bolivia mandated the integration of the armed forces but with a special regime. This exclusion is particularly incongruous in Chile where the military dictatorship implemented the reform and the armed forces privileged scheme survives after eighteen years of democracy. Argentinian armed forces can retire at age 43 with twenty-five years of service (instead of contribution) contrasted with ages 60/65 and thirty years of contribution in the general system; in Cuba with twenty-five years of service at any age (internal security personnel can retire ten to fifteen years younger than in the general program); in Colombia with five to ten years of service less than in the public system; in El Salvador ten years younger than in the general program; in Guatemala with twenty years of contribution regardless of age; in Honduras with eighteen years and in Paraguay with fifteen years; in Mexico twenty years younger than the general program for women and ten for men; and in Uruguay twenty-two years younger or with a seniority pension regardless of age. Lower ages of retirement have been justified arguing that military careers are short and with high risks,

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 59 but a similar case could be made for mine workers who don’t enjoy such generous conditions. In virtually all countries, armed forces’ pensions are set at 100% of the past year’s or past month’s salary before retirement, with the highest ever salary in Haiti. Most countries adjust such pensions to the salary of the personnel in activity. All private systems except Bolivia have also maintained separate schemes for civil servants and/or executive, legislative and judicial branches, teachers and university personnel, policemen, oil and bank workers, and other powerful groups (see Section 5.1). This is also the case in all public systems, excluding Panama, which has one single program with standardized conditions for all insured. Brazilian civil servants hired before 1998 can retire twenty years younger than in the general program (see Section 4.3.1); Colombian congressmen and judges seven- to twelve-years younger than in the general public system; Mexican federal civil servants and oil workers ten years younger or with twenty-eight years of service at any age (retirement at 46 is common); Honduras civil servants seven to ten years younger; Nicaraguan teachers, Paraguayan congressmen and Uruguayan public notaries five years younger; and Venezuelan civil servants insured by 1991 can retire with thirty-five years of contributions regardless of age. In Peru the civil servant scheme was closed to new affiliates but still has a considerable number of insured (congressmen, judges, ministers, diplomats, etc.) who may retire with ten years of contributions regardless of age. Civil servants’ pensions are also set at 100% of the last salary at least in Brazil,2 Guatemala, Mexico, and Peru and are often adjusted to the salary of personnel in activity. Because of lower retirement ages, higher replacement rates, and better adjustment rules, pension levels of separate schemes are several times the average of the general program and collected for a much longer retirement span. Before the parametric reform in Brazil, the executive branch average pension was six times that of the general average

(RGPS), the armed forces’ was ten times, the legislative and judicial branches as well as the central bank were 22 times, and the public attorney ministry was 36 times; the cap introduced in 2004 should slowly reduce such differences. Colombian congressmen’s and judges’ average pensions are 22 times that of the corresponding public system average. Honduras civil servants’ minimum pension is ten times that of the general program, whereas teachers’ maximum pension is 28 times higher. Paraguay armed forces’ pensions average 208% of the national medium taxable salary. Peruvian civil servants’ monthly maximum pension is several times that of the maximum in the general public system. Such largesse provokes huge deficits financed by fiscal transfers and significant inequality (see Section 6.1). Virtually all countries have voluntary schemes that grant supplementary pensions to certain groups: fourteen in Bolivia and thousands in Brazil but they only cover 8% of the labor force, probably medium and high income, increasing differences in pension levels of workers in the private sector (Interviews 2004).

4.1.2. Gender inequality Women are typically discriminated against vis-à-vis men concerning pensions. Household surveys in fourteen countries in 2000–3 show that, in eight of them, pension coverage of women in the labor force was lower than that in men; it was about equal in four countries and higher in only two (Rofman 2005). According to a Chile survey of 2002, women had a slightly higher coverage than men in the private system, but those in the lowestincome quintile, no education and in domestic service and unpaid family work had significantly lower coverage than men; the 2004 survey revealed that 72% of the uninsured were women (EPS 2004; Bravo et al. 2006). In Dominican Republic private system only 42% of affiliates were women compared with 58% of men in 2003, but the proportions in the

60

Reassembling Social Security

closed public system were 54 and 46% respectively (Lizardo 2004). As already explained, coverage of women at age 65 is considerably lower than men of the same age in all countries except one. Furthermore, the average women’s pension level is inferior to that of the men because of factors external and internal to the pension system. Among the external factors those derived from the labor market are: (a) rate of labor force participation is smaller in women than in men because of permanent or temporary exit to care for children, the elderly, and sick (reproductive labor within the family is neither remunerated nor covered by social insurance in the region); (b) women’s unemployment rate tends to be higher than men’s; (c) proportionally more women are engaged in unskilled labor, domestic service, informal sector, part-time, seasonal and temporary jobs, self-employment, homework and jobs without contract than men, and those occupations receive lower salaries and normally are not covered by social insurance; (d) enterprises invest more in training men than women, due to the latter’s labor interruptions; (e) women’s average salary is normally inferior to men’s, because of the previous reasons and transgression of the legal mandate to pay equal salary for equal work; and ( f ) as a result of all the above, women have lower contribution density than men (accumulate less contributions during their working life). Another external factor is demography: women’s life expectancy is four to five years longer than men’s, hence their period of retirement is longer and this often results in lower pensions (Arenas de Mesa and Benavides 2003; Arenas de Mesa and Gana 2003; Bertranou and Arenas de Mesa 2003; ECLAC 2004d; Montecinos 2006). Inequality factors internal to the pension system exist in private and public systems but tend to be stronger in the former. In nine of the twenty systems (four in private and five in public), women’s retirement age is five years lower than men’s (Table 4.1), which combined with females’ higher life expectancy results in a pension period from nine to ten years longer

than men’s as average. Most female pensioners are usually indirect insured who earn the right as widows or orphans or mothers of direct insured (husbands, fathers, or children); furthermore, some countries subordinate that pension to women’s economic dependency on the direct insured. The significant increase in divorce has hurt women’s protection especially when they have not contributed to their own retirement, because if a divorced man remarries, the former wife can lose all or part of her survivor pension. In Chile single females are head of one-fourth of households, and one half of children born from unmarried couples are mainly the responsibility of their mothers and both often lack pension protection (ILO 2001b; Bertranou 2006; Montecinos 2006). Private systems accentuate gender inequality due to the following reasons: (a) most of them established a high number of contribution years to grant a minimum or public-pillar pension (twenty years in Chile, twenty-five in Dominican Republic contributory-subsidized regime, thirty in Argentina, and thirty-five in Uruguay) making it more difficult for women to access that benefit; (b) they are based on contributions during the entire working life instead of the past few years of work harming woman who have a lower contribution density than men; (c) they apply mortality tables differentiated by sex and age, therefore, in the calculation of the pension annuity the accumulated fund in the individual account is divided by the average life expectancy, resulting in lower pensions for women than men, particularly if women retire earlier; (d) they often charge a fixed commission that has a stronger regressive impact on women than on men, because it proportionally reduces more the smaller fund accumulated in women’s individual accounts and augments the gap to finance the minimum pension; and (e) because many women don’t meet requirements for a contributory pension, they are underrepresented among those pensioners but overrepresented among social assistance and minimum pensioners3 (Sapag and Sapag 2001; Bertranou and Arenas de

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 61

Table 4.1. Normal ages of retirement and retirement span in private and public pension systems, c. 2005 Contributory regime Retirement agea Systems/Countries

Retirement span (years)b

Women

Men

Woman

Men

60

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagesc

65

22.6

14.3

d

d

d

d

60 57e 60/65f 60g 55 65 65 60 60.5

65 62e 62/65f 60g 60 65 65 60 62.8

24.0 24.1 24.0/19.9f 20.0 25.7 18.8 17.2 23.2 21.2

16.7 17.4 19.8/17.5f 17.8 18.7 16.5 15.1 18.2 17.0

Public systems Brazil Cuba Ecuador Guatemala Haiti Honduras Nicaragua Panama Paraguay Venezuela Averagesc

60h 55 55i 60j 55 60 60k 57l 60 55 57.7

65h 60 55i 60j 55 65 60k 62l 60 60 60.2

21.9 27.2 27.8 21.5 20.3 22.2 20.2 25.1 20.0 25.6 23.2

15.8 20.5 25.5 19.8 19.2 16.2 18.5 19.0 17.6 19.4 19.2

Social assistance regime age

70 65 65 65

70

65 m

Sources: Retirement ages from legislation; life expectancy 2000–5 from CELADE 2004. a Private systems usually allow retirement when the individual account has accumulated a sum to guarantee the financing of a minimum pension; in Argentina required for minimum, public and private pensions. b Average years as pensioner at the time of retirement. c Nonweighted. d No age required but certain sum in individual account hence it is impossible to estimate life expectancy; ages 50/55 for those who contributed to closed public system. e In the public system gradual age increment; in the private system as in note a. f Normal age of retirement at 65; early retirement at 60/62 but the parametric reform of 2005 increased years of contribution and proportionally reduced the level of pensions; an average of the two life expectancies was used in the private system averages. g Contributory regime also at age 55; contributory-subsidized regime (not in force) also at 65; subsidized regime (not in force) age 60 for indigent. h In RGPS, for rural workers 55/60, for civil servants 40/53 or 55/60 depending on year of affiliation.

62

Reassembling Social Security

Mesa 2003; Arenas de Mesa and Mesa-Lago 2006). Summarizing, as women have lower contribution density, earlier statutory ages of retirement and longer life expectancy than men, the accumulated pension fund in their individual accounts not only is lower than men but it must be stretched for a five- to ten-year longer retirement span, resulting in a lower annuity pension. It is argued that the private system treatment of women is fairer because it avoids subsidies between sexes, ignoring that women usually pay the entire cost of raising their children and caring for the elder and the sick, as pension systems in Latin America don’t grant credit for that labor (Chilean law prior to the reform awarded woman one year of credit for each child raised). Montecinos (2006) contends that those unpaid activities constitute significant savings for society but de-saving for women; as the size of the family shrinks and more women participate in the labor market, fewer females would be engaged in those tasks and pensions would have to be higher to cover those services or the state forced to provide them. Five private systems set an equal age of retirement for men and women (Costa Rica,4 Dominican Republic, Mexico, Peru, and Uruguay), as well as in Ecuador partly unconstitutional law. This is also true of five public systems: Ecuador—current public, Guatemala, Haiti, Nicaragua, and Paraguay. Such homologation facilitates more contributions and a bigger fund in women’s individual accounts vis-à-vis countries where women’s retirement age is five years younger, but does not compensate for the higher female life expectancy

(for diverse views on homologation of retirement age see Martínez and Mesa-Lago 2003; ECLAC 2004d). A compensatory factor in private systems exists when the insured man has a spouse and children, because the calculation of his pension takes into account the cost of the potential pension due to the widow (including her life expectancy) and orphans, hence the man’s annuity is reduced and a transfer occurs. This provision, however, discriminates between single men and married men with children as the latter receive a smaller pension than the former. In the case of a male insured with few years of contribution, the widow is entitled to withdraw the amount deposited in her husband’s individual account, a right that usually didn’t exist in previous public systems (Bertranou 2006). Public pension systems are more neutral or positive on gender than private systems because they usually redistribute resources from men to women through several compensatory mechanisms: (a) grant the minimum pension often without requiring a minimum number of years of contribution or fewer years than in private systems; (b) calculate the pension based on the past three to seven years of work, hence disregard the lack of contributions in years when women don’t work; (c) apply unisex tables in the pension calculation (mortality tables not differentiated by sex that don’t discriminate for the female longer life expectancy);5 and (d) set women’s replacement rates, combined with the minimum pension, higher than men’s rates, because they live longer. Mixed systems tend to compensate gender inequality more than substitutive systems that are totally private, because the

← (continued) i Increased to 60 in the reform law partly declared unconstitutional and also by regulations in 2006 (other ages in Table 4.2 notes). j Law stipulated a gradual rise to 65 in 2008 but it was declared unconstitutional. k In the public system, ages were to be gradually increased by the repealed law of 2000; in the private system the age was to be 65. l The reforms of 2005–6 allow early retirement at 55–6 for women and 60–1 for men with 15 years of contribution starting in 2008 with a basic replacement rate reduced by an adjusting factor. m There is no specific age but is based on need.

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 63 first (public) pillar attenuates such inequality, while the second (private) pillar accentuates it. Such compensatory effect depends on the relative importance of the first pillar relative to the second pillar: the effect is stronger in Costa Rica because the pension paid by the first pillar is the most important and that paid by the second pillar is supplementary, while the opposite is true in Argentina. The region doesn’t have a comprehensive study measuring the actual comparative effects by gender of private and public pension systems, largely because private systems are relatively recent and lack longitudinal data on employment histories of current workers or benefits of current pensioners. World Bank simulations, however, compare the distribution of internal rates of return by gender in reformed (private) and nonreformed (public) systems with mixed results: a marginal increase of the rate earned by women relative to those earned by men in Colombia and Chile, as well as higher rates among poor women in Argentina, Chile, and Mexico; but the rate of men relative to women increased in Peru, and subsidies that women received before the reform were transferred to men in Bolivia, El Salvador, and Mexico. ‘[I]n every other country where the new retirement . . . model was adopted, women earn lower returns from the new systems relative to the returns earned by men’ (Gill, Packard, and Yermo 2005: 91–2; for other simulations see James, Cox-Edwards, and Wong 2005). In contrast with the mixed results of the above simulations, there is strong evidence in Chile (where the reform has been in effect for twenty-five years and a survey has provided longitudinal data) that the private system has adverse effects on women. In 2001–2, average contribution density in women was 44% versus 60% in men, the accumulated fund in women’s individual accounts was between 32 and 46% of that accumulated by men, the female replacement rate ranged from 52 to 57% vis-à-vis a male rate that ranged from 81 to 86%, and the average women’s pension, if retired at age 60, was 60% lower than men’s and 87% lower if retired at age

65 (SAFPb 2002b; Arenas de Mesa and Gana 2003; EPS 2004; Bravo et al. 2006). Furthermore, it is estimated that 35% of women who were in the 40–5 age bracket in 2005 will get pensions inferior to the assistance pension level; an additional 10% will receive a pension higher than the assistance one but lower than the minimum pension, and therefore, 45% will receive a pension lower than the minimum pension (Arenas de Mesa and Mesa-Lago 2006). Policies under debate to reduce the accentuation of gender inequality in private systems include: the use of unisex mortality tables,6 promotion of joint individual accounts for insured males and their spouses, mandating married male insured to retire with join annuities covering their spouses, granting credits to women who raise their children, and creating a pension solidarity fund with tripartite contributions. Costa Rica is designing a strategy to extend coverage of housewives who lack direct insurance but are included in the husband’s survivors insurance; the housewife would contribute and the husband’s contribution would be divided in equal amounts and deposited in two individual accounts, giving the wife the right to her own pension (Rodríguez Herrera 2006). Chile’s legal reform draft of 2006 proposes an annual bonus granted to mothers (affiliated or not to a pension system) for each child born alive, division of the sum accumulated in the individual account of each spouse in case of divorce, and survivor pensions to widowers (Gobierno 2006).

4.2. Solidarity and income distribution 4.2.1. Solidarity versus equivalence World Bank officials state that ‘By replacing what were often regressive single-pillar PAYG systems that frequently paid overly generous pensions to a privileged few, with [multipillar private] systems, reforms were expected to introduce a more equitable system’ (Gill,

64

Reassembling Social Security

Packard, and Yermo 2005: 90). It is true that the principle of solidarity has been eroded in many public systems, but it has been replaced in private systems by the principle of strict equivalence between the contribution and the pension level, hence the amount of pensions varies considerably according to insured salary, contribution density, capital return of the invested fund in the individual account, and macroeconomic factors, reproducing existing disparities in salary and generating a wider disparity of benefit levels (ILO 2001b). In addition, private systems by their own nature eliminate the distribution among generations and shift the distributive function to the state (external to the private system), through the guarantee of a minimum contributory pension and the concession in five countries of noncontributory pensions (mostly social assistance). Reform supporters argue that employers’ contributions create distortions in the labor market, increase the cost of labor, reduce the country’s competitiveness, and stimulate the substitution of labor by capital, hence generating higher unemployment; therefore they favor the elimination or reduction of the employer contribution, an action taken in five countries.7 Opposite to those claims, the ILO (2000b) sustains that in the long run the employer contribution is transferred to the worker (through a lower salary) and do not provoke a negative impact on employment, although that effect may occur in the short run. Several experts have questioned the assumption that private systems provide more incentives and minimize labor market distortions created by public systems, and have concluded that there is no simple dominance of one system over another in terms of both elements: the ultimate goal of a pension system is welfare not labor supply and the change from defined benefit to defined contribution may create undesirable risks that affect welfare; there is a tradeoff between incentives and redistribution, which is an important objective of public systems; it is quite complex to separate labor

market distortions created by other policies (e.g. a progressive income tax) from effects generated by a pension system; the distortion created by the payroll tax for pensions is not measured by that tax itself but by any resulting difference between the net present value of marginal benefits and the marginal tax; financing social assistance is another important goal of all pension systems that require taxation and results in labor market distortions; and a debt-financed transition to individual accounts creates distortions due to the higher taxes required to finance that debt (Orszag and Stiglitz 2001; Barr 2002; Barr and Diamond 2006).

4.2.2. Mechanisms for and against solidarity Structural reforms have introduced the following regressive distribution mechanisms: (a) the exclusion from the reform of insured in separate schemes, who generally have middle and high income, don’t contribute to the general program but enjoy generous entitlement conditions, benefits, and fiscal subsidies (see Section 4.2.1); (b) the virtual exclusion from coverage, in five countries, of the self-employed and other low-income informal workers, as well as the poor (see Section 3.2.1), and the imposition to the selfemployed of a percentage contribution two to four times higher than that of salaried workers; (c) the absence of a maximum pension; (d) the accentuation of gender inequalities (see Section 4.1.2); (e) the elimination or reduction of the employer contribution in half of the systems and the increase in the worker contribution in six of them (see Section 6.1); ( f ) the proportionally stronger reduction of the tax burden to high-income insured who benefit from deferred taxes on their contributions deposited in individual accounts; (g) the high administrative cost of the system exclusively financed by the insured except in three countries, which substantially diminish the deposit in the individual account and the

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 65 future pension, particularly harming to lowincome insured (see Section 5.5); (h) the fixed commission collected by some administrators, which takes a proportionally higher part of the total contribution of low-income workers than that of high-income workers, hence proportionally reducing more the deposit in individual accounts and the pension level of low-income insured; (i) the intergenerational inequalities caused by the burden on the oldest insured, who support the bulk of the installation costs of the new system, in favor of younger insured who support a much lower burden; and ( j) the fiscal cost of the transition that involves a transfer to middle- and highincome insured, financed by taxes, often on consumption and paid by the entire population, including the uninsured, a regressive effect that increases as coverage descends8 (Arenas de Mesa 1999; SAFPb 2002; Uthoff 2003; Kiefer 2004; Gill, Packard, and Yermo 2005). The alleged solidarity mechanisms and progressive redistributive effects of private systems are normally external to it. The minimum pension, guaranteed by the state and financed by taxes, does not generate a redistribution within the insured cohort in the private system but between taxpayers and the insured with insufficient funds in their individual account to finance such a pension; a considerable part of the current insured will receive that pension often not adjusted to the cost of living (Bolivia does not grant a minimum pension and Peru did not pay it for nine years). The social assistance pension that has a progressive poverty-reduction effect and is granted in half of the countries with structural reform is also a state responsibility and financed by the entire population. Mexico’s peasant scheme is financed by the state, within the public program, and excluded from the private pillar. In 1981–2004, the total fiscal cost of Chile’s pension system averaged 5.2% of GDP annually, distributed as follows: 4.8% to cover transition costs on behalf of insured in the closed public program and military pensions but only 0.4% for social assistance pensions (Arenas de Mesa and Mesa-Lago 2006).

There are three important solidarity mechanisms, however, one of them with caveats. Colombian insured earning a salary fourfold of the minimum wage pay a contribution of 1% to a solidarity fund and an additional contribution rising from 0.2 to 1% as income increases from 16 to 20 times the minimum salary; the first 1% is assigned to finance the extension of coverage to self-employed and other groups whose socioeconomic conditions prevent them from affiliating to social insurance, and the additional percentage will be devoted to the social assistance pension (LRP 2002). Dominican Republic employers paid 0.4% of the wage bill to a solidarity fund to finance the minimum pension in the contributory program (LDSS 2001).9 Costa Rica’s government makes a contribution to the public pillar on behalf of the self-employed (to compensate for the lack of employer) but those workers are excluded from the private pillar and the state contribution is granted regardless of income, hence it is a regressive subsidy for high-income self-employed such as professionals (Rodríguez Herrera 2006). World Bank officials ask if private systems have corrected the alleged regressive effect of public system and acknowledge: ‘Very little evidence exist one way or the other, largely because . . . the final impact of structural reforms on income inequality cannot be precisely measured until large segments of the population begin to retire with pensions financed primarily from individual retirement accounts’ (Gill, Packard, and Yermo 2005: 90). Nevertheless, such officials conducted simulation exercises showing that in eight countries with reform (excluding Costa Rica and Dominican Republic) the new private systems lowered the regressive impact of public systems. Those results cannot be checked because the Bank report presents bar graphs, without explaining the methodology used, not reproducing the base data, and sometimes engaging in obscure interpretation (see Mesa-Lago 2004). Another graph measures (with the Gini index based on 1995 data) the marginal contribution to income inequality generated by pensions relative to earned income from labor

66

Reassembling Social Security

in seven countries, three of them with private systems (Argentina, Chile, and Uruguay) and four with public systems (Bolivia—before the reform, Brazil, Paraguay, and Venezuela). Note that the reform of Uruguay began in 1996 (one year after the year of the data) and that of Argentina in 1994 (only one year before the year of the data). The graph indicates that the regressive impact of pensions was higher than that of labor income in all countries (including the private system of Chile) except in the private systems of Argentina and Uruguay, where it was slightly lower. The highest increases in inequality by pensions relative to labor income were in Brazil and Paraguay (due to separate privileged schemes for civil servants and other groups), but the difference between the regressive impact generated by pensions and labor income in Venezuela public system was proportionally much lower than in Chile private system, with the oldest structural reform (Gill, Packard, and Yermo 2005: Graphs 5.1 and 5.3). Conversely, Chile’s survey of 2002 showed that the private system has expanded existing differences in the labor market between affiliates of divergent income levels and pensioners at the same income levels: among active workers there was a ratio of 8.6 times between the maximum taxable income and the minimum salary, but among pensioners in the two income groups the said ratio will rise to 9.3 times, and to 13 times when incorporating contribution densities differentiated by sex (Arenas de Mesa and Mesa-Lago 2006). All combined pension expenditures in Uruguay are regressive (social assistance pensions are progressive and the rest are regressive) and the Gini coefficient worsened in 1999–2003 (Presidencia 2005). Among public systems there are several redistributive mechanisms of a progressive nature: (a) rural and social assistance pensions in Brazil financed by contributions paid by buyers of rural products and state transfers, social assistance pensions in Cuba financed by the state, and peasant pensions in Ecuador financed by transfers of contributions from workers and employers in the general

program, all of whom have reduced poverty; (b) pension formula with a higher replacement rate for low-income pensioners; (c) minimum and maximum pensions that reduce inequality between high-income and lowincome insured (see Section 4.3.2); (d) transfers from men to women (explained in Section 4.1.2); (e) employer contributions that average twofold the worker’s, particularly important in countries with high coverage (see Section 6.1); ( f ) state financing of the deficit in the general program in countries with high coverage (Brazil and Cuba, also true in Argentina and Uruguay public pillars); and (g) solidarity contribution of 11% imposed on civil servant pensions that exceed a certain level in Brazil. But public systems also have regressive mechanisms, some of them similar to private systems: (a) affiliates to separate schemes in most countries don’t contribute to the general program but enjoy better entitlement conditions and benefits than average insured in that program and receive significant fiscal subsidies (Mexico’s federal government finances 71% of pensions of their civil servants, who don’t contribute to IMSS, relying on taxes paid by the uninsured; ISSSTE 2006); (b) the self-employed and other low-income informal workers, as well as the poor, are excluded in most countries, and the self-employed contribution is 2–8 times higher than that of the salaried worker in four countries (see Section 6.1); (c) in countries with very low coverage (Nicaragua and Paraguay) if the employer contribution is indeed transferred to prices (instead of actually paid by workers), most of the uninsured labor force partly finances the insured minority; (d) the pension formula based on the average of the three to seven years of salary prior to retirement favors those insured whose salaries increase with age but penalizes manual workers whose wages decrease with age and physical deterioration (see Section 4.3.2); ( f ) the state does not fulfill its financial obligations to the general program in four countries but subsidizes the separate schemes; and (g) where the state fulfills its obligations and coverage is low, most of the labor force, which is uninsured,

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 67 partly finances the insured minority through consumer taxes (Mesa-Lago 2003e). World Bank officials note some of the above regressive effects and add two more; the first one is convincing: workers with low income, scarce education, and high average mortality enter the labor force earlier, contribute proportionally more and collect a pension for less time than workers with higher income, better education, and lower mortality who postpone their entrance to the labor force, contribute for a shorter period and collect a pension for a longer time. The second alleged regressive effect is that ‘pension benefits [in public systems] are based on earnings rather than on need’, an argument inconsistent with the Bank officials’ criticism of public systems because of their tenuous or nil relationship between contribution and pension level, and support for the strict application of the principle of equivalence in private systems (Gill, Packard, and Yermo 2005: 90, 96, 98). There are no studies measuring the net impact of progressive and regressive mechanisms in public systems. Segmentation, however, is a strong negative factor as two public systems with high coverage and fiscal transfers to vulnerable population groups illustrate. Brazil’s general program (RGPS) has 88% of the pensioners (two-thirds of whom collect a pension equal to the minimum salary) and endures a deficit financed by the state equivalent to 30% of the total fiscal deficit generated by the entire pension system; conversely, civil servants’ schemes that only have 12% of pensioners (receiving pensions between 9 and 35 times higher than the RGPS’s average pension) suffer a deficit financed by the state equivalent to 70% of the total fiscal deficit in the entire pension system (Schwarzer 2004). In Cuba the cost of the armed forces’ scheme that covers about 4% of the labor force equaled the total deficit financed by the state of the entire general program that probably covers most of the labor force (Mesa-Lago 2003b). But these inequalities are also found in separate schemes of private systems: Chile’s military pensions averaged 1.4% of GDP annually in 1981–2004, half the cost of all ongoing public pensions, 13

times that of minimum pensions, and fourfold that of social assistance pensions (Arenas de Mesa and Mesa-Lago 2006), whereas Bolivia’s price tag was 0.5% of GDP in 2003, 44% more than the cost of minimum pensions (Garrón 2004). These regressive effects are probably worse in countries with low coverage, more segmentation, and lacking fiscal subsidies to vulnerable groups. Although more research is needed to measure the comparative impact of private and public systems on income distribution, the above analysis indicates that countries with higher coverage, most unified systems and fiscal transfers to low-income groups and the poor have more solidarity than countries with low coverage, very fragmented systems and none or low fiscal transfers to vulnerable groups.

4.3. Comprehensiveness and sufficiency of benefits Private systems are characterized by undefined (uncertain) benefits because the insured pension will be based on the amount accumulated in their individual accounts, which in turn will depend on several factors, some of them unrelated to the insured wage and contribution density. Chile, El Salvador, and Peru offer three retirement options: an annuity paid by an insured company; a programmed retirement paid by the old-age program administrator (planned withdrawals until the fund is exhausted); or a combination of both, first the programmed and then the annuity. In Bolivia and Uruguay only an annuity is allowed; in Argentina, Colombia, Costa Rica, and Mexico the combination is not permitted. A crucial promise of structural reforms is that they will pay better pensions than in the public system, a key incentive in the transfer of insured from the public to the private system in countries that grant such choice. The benefit is defined in public systems because the law establishes the formula for the calculation of the pension, which is usually

68

Reassembling Social Security

generous by international standards although varies significantly among countries. The formula starts with a base salary (an average of wages during a period of the insured working life); a replacement rate (a percentage) is then applied to the base salary, and the result is the pension. Both in public and private systems there are separate public privileged schemes with pension formulae more generous than that of the general program. This section examines ages of retirement and their impact on access and length of pensions, describes the pension formula in public general programs and separate schemes, reviews the adjustment of pensions, compares social assistance pensions in both systems and their relationship with average contributory and minimum pensions, and inquires if the promise of a better pension in private systems has materialized.

4.3.1. Retirement ages and spans, and contribution years in all systems The majority of the structural reforms laws increased the statutory (‘normal’) ages of retirement: Argentina, Bolivia, Chile, Peru, Uruguay (for woman), and Ecuador partly declared unconstitutional law. In Colombia’s public system the retirement age is gradually being lifted until 2015. Ages of retirement in Peru were raised from 55 for women and 60 for men to 65 for both sexes, significantly reducing the average length of life expectancy at retirement (retirement span) now the shortest in the region. Costa Rica, Dominican Republic, El Salvador, and Mexico did not boost the age of retirement. Costa Rican entitlement conditions are equal in the private and public pillars; the parametric reform of the public pillar neither eliminated nor raised early retirement ages (60/62 versus 65 for the normal age), but increased years of contribution from 20 to 25, and proportionally reduced the amount of the pension to discourage early retirement (Table 4.1). Years of contributions were also increased for the minimum pension in the private

systems or pillars in mixed systems of Colombia, Dominican Republic, El Salvador, Mexico, Peru, and Uruguay, but most of them don’t require specific ages and contributions for retirement, only that the insured individual account accumulates certain amount determined by law to finance a pension equal or superior to the minimum pension. Insured without the required years of contributions don’t receive the minimum pension but are entitled to withdraw the fund accumulated in their individual accounts; in Argentina and Uruguay mixed systems the insured are not entitled to the public-pillar pension but withdrawal of the private pillar fund.10 Only 16% of the insured in Uruguay’s private sector would reach the 35 years of contributions at age 60 (32% at age 65), most of them will have to postpone retirement to 70 or won’t be eligible at all (Murro 2006). Mexican unions forced the government to grant affiliates at the time of the reform the right to choose at retirement the best pension between the stipulated in the closed public system and that resulting from the individual account in the private system. Structural reforms also raised years of contribution counted in the calculation of the base salary in most public systems/pillars, and reduced replacement rates. Most public systems have not increased ages of retirement in the general program and years of contribution are not required for minimum pensions or are considerably lower than in private systems. Actually increasing those years is one of the most opposed measures in parametric reforms. There is wide diversity in ages of retirement and other entitlement conditions in public systems; only Panama has fully standardized them. Brazil probably has the most diverse and complex system. The program for privatesector workers (RGPS) has old-age and seniority retirement. The former is at ages 60/65 (women/men) with 15 years of contribution, which results in retirement spans of 21/14 years (rural scheme ages of 55/60 are 5 years less than in the general program plus 15 years of work instead of contributions). Seniority retirement requires 30/35

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 69 years of contribution at any age, hence boosting the retirement span. The schemes of federal, state, and municipal civil servants concede even more generous entitlement conditions but have been restricted by the parametric reform of 2003,11 which set an equal plateau for all pensions in the country, laying the bases for the future standardization of entitlement conditions of civil servants with the general program; such a plateau had not been implemented in 2006, waiting for approval of a law regulating civil servants’ supplementary pensions (Schwarzer 2004, 2007; Pinheiro 2005, 2006; see also Section 4.3.2). Ages of retirement in Cuba’s general program are 55/60, among the lowest in the region, and retirement spans are 27/20 years, the second largest among all twenty countries (four other countries have similar ages but lower retirement spans). Domestic and foreign experts have recommended an increase in retirement ages and years of contributions (Peñate 2000a; Barros 2003; Mesa-Lago 2003b; Sabourin 2003), but a reform draft in 2003 kept ages unchanged and proposed increments in the replacement rate for each year of postponed retirement (Sandó 2003). Among countries in the intermediate group, Ecuador sets age 55 for both sexes, the lowest age for men in the twenty countries, which leads to the largest male retirement span of 25 years, but the required 30 years of contribution are the second highest.12 Panama’s retirement ages are 57/62 with retirement spans of 25/19 years; the required 15 years of contribution, average in public systems, are gradually increased to 20 years by the ongoing reforms (Ley 51 2005; Asamblea 2006). Venezuelan retirement ages in the general program are identical to Cuba (55/60) and retirement spans are 25/19, the female is the third largest in the region, while the 14.5 years of contribution is the second shortest among all public systems.13 Liberal entitlement conditions in these four countries are very difficult to finance and require an urgent parametric reform to avoid severe financial imbalance.

Entitlement conditions in the five public systems of the latecomer-low group tend to be the most lax except in Honduras. Haiti’s general program has, together with Ecuador, the lowest ages of retirement for men in the region (55), but requires 20 years of contributions, quite difficult to meet in those countries. Ages of retirement in the general programs of Guatemala, Nicaragua, and Paraguay are 60 for both sexes, years of contribution are 15 in the first two but 25 in Paraguay, also difficult to fulfill taking into account the characteristics of its labor market. In Honduras’ general program ages of retirement are 60/65 (the same or higher than in the pioneerhigh group) and the female retirement span is among the shortest. Comparisons of entitlement conditions between private and public systems are not precise because of their diversity and the special characteristics of some private systems (entitlement not based on age and years of contribution but on the amount accumulated in the individual account). Nevertheless, Table 4.1 indicates that average ages of retirement for men and women in private systems (60.5 and 62.8 years) are higher than the corresponding averages in public systems (57.7 and 60.2 years). It is therefore logical that average retirement spans for men and women in private systems (21.2 and 17 years) be lower than averages in public systems (23.2 and 19.2 years). This apparent advantage, however, would be meaningless if public systems are not in financial and actuarial equilibrium, a potential problem aggravated by the generosity in the formula for calculating public pensions (see Section 4.3.2). The required age for noncontributory pensions (mostly social assistance) is usually higher than that in the general contributory program, although beneficiaries of the former have a lower life expectancy (Table 4.1): 65 in Bolivia (10 and 15 years more than in the closed contributory public system), as well as in Chile (5 years more for women in the contributory), and 70 in Argentina (5 and 10 years more) and Uruguay (5 years more). Brazil’s assistance pensions are granted at age 65 for

70

Reassembling Social Security

both sexes, 5 years more for women than in the RGPS. Age for social assistance pensions in Costa Rica is also 65, the same as the normal age of retirement for the contributory regime, but the latter allows early retirement at 60/62. Ecuador’s law (partly declared unconstitutional) sets the age of retirement for peasants at 65/70, 5 to 10 years more than in the general contributory program. Conversely, retirement ages for rural pensions in Brazil are reduced by 5 years, which is fair because of the lower life expectancy of these workers and Cuba does not require a specific age for assistance pensions. Colombia’s law of reform not fully enforced reduces the age for assistance pensions by 3 years relative to the general program, and the not-enforced reform law of the Dominican Republic sets an equal age. For humanitarian and equity reasons the age at which to grant old-age assistance pensions should be the same as that of contributory pensions, but it is argued that it would create incentives to evade in the contributory program, hence a higher age is needed for social assistance (Palacios and Sluchynsky 2006).

4.3.2. The pension formula in public pensions The pension formula in public systems is very generous (Table 4.2).14 The base salary in the general program is calculated as an average of the past 3 years of salaries in Paraguay, 3 to 5 years in Honduras, five years in Cuba, Ecuador, Guatemala, and Nicaragua, 5 or 10 years (the most beneficial) in Venezuela (reduced to 2 years by the 2002 not yet enforced law) and 7 years in Panama. The average of 5 years in public systems not only is too short but tends to be the highest in the working life of the insured. That practice provokes various negative effects: (a) encourages under declaration of wages during the earlier and longest part of the working life, because those wages are not counted in the calculation of the base salary, while it stimulates over declaration of wages in the latest

working years with the objective of increasing the base salary and the pension (such rational albeit illegal behavior minimizes the contribution and maximizes the pension); (b) penalizes the insured who honestly report the full salary and punctually pay their contributions during their entire working life, as well as manual workers whose salary, opposite to that of nonmanual workers, normally declines at the end of the working life due to physical deterioration; (c) diminishes the link between contributions and the level of the pension; and (d) introduces an unpredictable factor because the years of salary used in the calculation are seldom adjusted to inflation (Table 4.2). Minimum replacement rates are 30% in Haiti, 40% in Venezuela,15 50% in Cuba, Guatemala, Honduras, and Nicaragua, and 60% in Panama. Maximum replacement rates are 80% in Guatemala, Honduras, and Nicaragua (100% for low-income insured), 90% in Cuba, and 100% in Panama and Paraguay. The minimum range (50–100% except Haiti and Venezuela) and maximum range (80–100%) exceed the ILO minimum norm of 45%. Such rates are very difficult to finance and, combined with the too liberal base salary, could lead to financial bankruptcy in the long run. In Brazil old-age retirement program (RGPS), the base salary is 80% of the average of the highest years of salary annually adjusted to the CPI since 1994 until the time of retirement. The replacement rate is calculated applying a ‘pension factor’ to the base salary, which incorporates the age of retirement, the rate and period of contribution, and life expectancy, as the lower the age and the rate and period of contribution the lower the replacement rate, and hence there is a disincentive for early retirement. All public systems have a minimum contributory pension and most have a maximum pension too. The level of the minimum pension varies among private and public systems and countries but is usually fixed in relation to the minimum wage. In Brazil it is equal to one minimum salary for RGPS and rural workers

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 71

Table 4.2 Other entitlement conditions for old-age pensions in private and public systems, 2005 Years of contribution Systems/Countries

Privatea

Public

Base salary (Years)b

Replacement (Rate %)c

Adjustment of general pension

30

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

Discretionary US Dollar UFp CPI CPI Noq Discretionaryr CPI CPI Wage index

k

k

5 5 5 10 3 or 5 5 7m 3 5 or 10o

50/90 75/h 50/80 30/ n.a. 50/80 50/80 60/100m 100n 40/n.a.o

CPI No CPI CPI No Minimum wages Discretionaryt Discretionaryu CPI No

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay

20 20 20e 25 & 30f 25 25g 20 35

30i 15 10–20 20–25j 20e 25 & 30f 25 9.5g 20 35i

Public systems Brazil Cuba Ecuador Guatemala Haiti Honduras Nicaragua Panama Paraguay Venezuela

N.A. N.A. N.A.h N.A. N.A. N.A. N.A. N.A. N.A. N.A.

12–15k 25l 30h 15 20 15 15 15m 25n 14.5o

d

Sources: Legislation. n.a. = non available. N.A.= non applicable. a In most private systems, years required for minimum pension; retirement allowed regardless of years of contribution when the individual-account fund is sufficient to finance the minimum pension. b In public systems, years of salary (usually the latest) to estimate the base salary for the pension. c In public systems, percentage applied to base salary to set minimum and maximum pensions; public pillars in parallel and mixed models have base salary and replacement rates not shown in the Table. d Fund equal at least to 70% of the base salary, no right to minimum pension, no years of contribution required. e Needed in both public and private pillars; the 2005 reform raises contribution years to twenty-five and years for base salary to twenty. f Contributory regime thirty years; contributory-subsidized regime twenty-five years, and subsidized regime no contributions required (last two not in force). g Insured in the closed public system at the time of the reform can choose the best pension between publicsystem rules and individual-account fund. h Public also with any age and forty years of contributions or ages 65/70 with fifteen/ten years of contribution, replacement rate is base salary plus an annual factor for every contribution year; the private system, not in force, set thirty years for minimum pension. i If not met there is no public pension but withdrawal from individual-account fund. j Years of contribution rising from twenty to twenty-five in 2005–15.

72

Reassembling Social Security

and is received by two-thirds of all pensioners. In Cuba the minimum pension received by half of the pensioners was US$7.50 monthly in 2005 (based on the rate of 20 pesos per US $1 in state exchange agencies); before the crisis it was supplemented with price subsidies for rationed goods, free health care and low transportation and public-utility rates, but in 2006 rationed goods only covered needs for seven to ten days per month and the remaining food had to be bought at very high prices in free agricultural markets and hard-currency shops, whereas health care and transportation had deteriorated, and utility rates raised. In Dominican Republic the minimum pension equals the lowest minimum wage, in Mexico the minimum wage of the Federal District, in Honduras 75% of the minimum wage, in Chile 70% of the minimum wage, and in Cuba about 67% of the minimum wage. Most public systems also set a maximum pension or ceiling on the benefit. In Brazil the maximum pension in RGPS is equal to eight minimum salaries with a ceiling of US$800 monthly that will be imposed on civil servant pensions also when the reform is fully implemented. In Ecuador the minimum pension was US$25 and the maximum US$125 in 2003, both very low and the minimum barely sufficient to cover food needs (the pension paid by the

peasant scheme is even lower). In Argentina the minimum monthly pension was US$134 (in both the public and the private systems) and the maximum pension in the public system was US$1,065 in 2006. In Panama the maximum pension was US$1,500 monthly in 2004, similar to the top pension in the United States, still some pressure groups claim that is low for high-income insured. Extending the number of years used to calculate the base salary (therefore increasing the years of contribution required) would lessen or eliminate the noted adverse effects and, combined with a reduction of replacement rates (particularly the maximum), would make the systems more financially viable in the long run. On the other hand, the extension in the number of working years used to calculate the base salary should use annual salaries adjusted by the CPI. These changes should be accompanied by a rise in the age of retirement in countries with very high life expectancy (Cuba, Ecuador, Panama, and Venezuela).

4.3.3. Adjustment of private and public pensions The adjustment of general program contributory pensions varies among countries

← (continued) k RGPS’s old age: years rising in 2005–11; upon the base salary (80% of the average of the best years since 1994) a pension factor is applied for replacement rate; seniority: thirty/thirty-five years of contribution at any age; others conditions for civil servants. l Years of work. m The ongoing reform gradually raises contribution years to twenty in 2013 and years for calculating the base salary to ten, and modifies replacement rates. n Also with thirty years of contribution and 75% replacement rate. o The nonenforced 2002 law sets twenty-five years of contribution and a ceiling of 85% on the average of the past two years of salary. p Monetary units constantly adjusted to inflation. q Salaried workers based on minimum wage and self-employed based on CPI (not enforced). r At government discretion based on wage index and available fiscal resources. s Adjusted at 75% of minimum salary. t Law of 2005 (postponed to 2007) annually adjusted if a sizeable change in cost of living occurs. u Minimum and maximum pensions adjusted every three years depending of fiscal situation; ongoing reforms stipulate an increase of US$10 every five years after 2010.

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 73 (Table 4.2, last column) and there are no standardized historical series on the evolution of real pensions.16 Among private systems, Chile has the best method because contributory pensions are set in inflation-indexed accounting units (UF). In Colombia, Costa Rica, Ecuador, Mexico, and Peru the government or the social insurance institute annually adjusts pensions to the cost of living by the consumer price index (CPI); in Uruguay they are adjusted by the wage index, and in Bolivia the government adjusts them based on an index tied to the US dollar.17 In the Dominican Republic, pensions in the contributory regime (salaried workers) should be adjusted relative to the minimum wage of the public sector but this had not been done at least until 2006, pensions in the contributorysubsidized regime (self-employed, not yet in force) should be based on the CPI. Adjustment is at the government discretion in Argentina (after the devaluation of the peso, based on a set peso–dollar exchange rate) and in El Salvador (based on the taxable wage index and available fiscal resources). The minimum pension in Chile is adjusted at government discretion based on the CPI of the previous year; in the Dominican Republic it is equal to the lowest minimum wage; in Mexico it is equal to the Federal District minimum wage annually adjusted by the CPI, and in El Salvador it is not adjusted (legislation; Gana 2002; Pérez Montás 2006). Among public systems four countries adjust pensions annually to the CPI: Brazil (RGPS), Ecuador, Guatemala, and Paraguay. In Honduras pensions were very low until 2000 because they were capped by a contribution ceiling that was unchanged for 28 years and were not adjusted to inflation, hence they lost half of their real value in 1994–8; in 2001 the contribution ceiling was increased and pensions must be adjusted to guarantee 75% of the minimum wage. Panama does not adjust pensions to CPI; there is a minimum and maximum pension adjustable every three years provided that the financial situation allows it, hence the state has ample flexibility and adjustments have been usually the result of

political and trade union pressure; the 2005–6 reforms don’t guarantee adjustment to inflation (Ley 51 2005; Mesa-Lago 2005a; Asamblea 2006). Cuba, Haiti, and Venezuela do not periodically adjust general program pensions. The lack of annual adjustment of pensions to the CPI, wage index, or another institutionalized mechanism in eight countries (three private and five public) results in a deterioration of the real value of pensions.18 Countries that have generous entitlement conditions (like Cuba, Panama, and Venezuela) but lack that mechanism are in the long run negating such apparent liberality. For instance, the ILO projected that a 1998 pension in Panama, without indexation, would lose 77% in its real value by 2050, increasing poverty among the elderly. In Cuba, because of the crisis, the real value of the average pension fell 41% in 1989– 98; there was an increase in 2005 but this did not compensate for the previous decline; the average monthly pension in 2005 was US$9 (Mesa-Lago 2005a, 2005b, 2006d). In those countries it would better to tighten the generous entitlement conditions and annually adjust pensions either to CPI or the wage index.

4.3.4. Levels of noncontributory pensions in private and public systems The importance of noncontributory pensions (social assistance except for Bolivia) has been already stressed and one of the major recommendations of the World Bank is to fortify the poverty prevention first public pillar. But only seven countries provide these pensions: five in private systems (Argentina, Bolivia, Chile, Costa Rica, and Uruguay) and two in public systems (Brazil and Cuba). To be entitled to an assistance pension, the individual or family income normally cannot exceed a percentage of the contributory minimum pension or the minimum salary, submitted to a means test. In some countries the noncontributory pension is set very close to the minimum contributory pension (even to the average

74

Reassembling Social Security

contributory pension), hence, generating perverse incentives.19 The level of noncontributory pensions, relative to the average contributory pension, in various years between 2002 and 2005 was: 60% in Uruguay, 55% in Cuba, 41% in Argentina, 30% in Chile, 32% in Costa Rica (half of the minimum contributory pension in both), and 31% in Brazil RGPS (equal to the minimum contributory pension). Dominican Republic subsidized regime not yet in force in 2006 stipulates a noncontributory pension equal to 60% of the minimum salary in the public sector. Bolivia Bonosol is about 10% of the minimum salary. In 2002–5 the monthly noncontributory pension was: US$134 in Uruguay, US$84 in Argentina, US$77 in Brazil, US$75 in Chile,20 US$40 in Costa Rica, and US$5 in Cuba. Bolivia’s annual Bonosol was US$225 in 2005 equivalent to US$19 monthly. Dominican Republic noncontributory pension not yet in force in 2006 is projected at US$46 monthly.21 Barrientos (2006a) argues that the level of noncontributory pensions is set below the poverty line to minimize disincentives to participate in contributory pension programs. Although a thorough study is needed on the sufficiency of noncontributory pensions, those paid in Uruguay and possibly Argentina, Brazil, and Chile seem to cover basic needs; Costa Rica’s covered the basic cost of the food basket; Cuba’s was well below basic food needs even with the eroded supplements, as was possibly Bolivia’s. Chilean noncontributory pensions brought out of poverty 40% of recipients 60 years and older in 2003 and another 60% were lifted from indigence to poverty. It is crucial to maintain an adequate gap between the minimum contributory pension and the noncontributory pension to avoid incentives to evade affiliation and compliance with social insurance as well as simulation of need to obtain assistance. In Chile the minimum pension was 1.8 times the noncontributory pension and half the average contributory pension in 2005; but public contributory pensions rose at an annual average of

2.3% in 1990–2005, minimum pensions at 2.9%, private contributory pensions at 3.6%, and noncontributory pensions at 6.6%. If such trends continue, the gap between noncontributory pensions on one hand and minimum and contributory pensions on the other will shrink. Venezuela’s constitution of 1999 set the level of the noncontributory pension not less than the urban minimum wage, which created a serious obstacle to its implementation because such a level would be equal to the contributory minimum pension and require an increase in the latter as well as considerably more funds; in 2005 a law of social services included an aid to the elderly in need tantamount to 60–80% of the urban minimum wage.

4.3.5. Are private pensions higher than public pensions? The reforms have pledged that the private system will pay better pensions than those of the public system, because of punctual payment of contributions, lower administrative costs, and higher capital returns. Barr (2002) has theoretically refuted the claim that private pensions will be higher than public ones based on their higher capital returns, arguing that an adequate comparison of the two rates of return requires the inclusion of administrative costs and risks in both systems, as well as transitional costs. Empirically it is challenging to verify the said promise because of the absence of up-to-date comparable statistics between the two systems. Two Chilean experts stated in a study of 2001 that ‘the latest information published by the [superintendence] is for June 1992’ (Acuña and Iglesias 2001: 27). According to that information, almost a decade old, average private pensions were higher than average public pensions in most branches: 43% in old age, 68% in disability, and 42% in widows, but 9% lower in orphans. These figures are contradicted by the following data on the average level of private pensions (March 2002) as compared with the average for public pensions (December 2001): private old-age

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 75 pensions (63% of total pensions) were 24% lower than public ones; disability pensions (7% of the total) were 15% higher; survivors’ pensions (28% of the total) were 110% higher, and the weighted average for all private pensions was only 3% higher than the average for public pensions (based on INP 2001; SAFPb 2002b).22 The Chilean survey of 2002–3 found that 9% of the insured who moved from the public to the private system had the expectation that they would get a better pension but 41% of pensioners in the private system had to work because the pension was very low, compared with only 23% of pensioners in the public system who had to work to supplement their pensions (EPS 2004). Argentinean projections in 2002 indicated that the changes made during the 2001–2 crisis (including the halving of contributions and the conversion of financial instruments expressed in dollars into devalued pesos) could reduce by 65% the benefits of an average pensioner with 30 years of contributions (ILO 2002b). The average private old-age pension in 2003 was 22% higher than the average public pension, but in 2005 the former was 6% lower than the latter, and a projection in 2005 estimated that the public pension will be 2% higher in 2050 (MTESS-ILO 2005). In Colombia, the superintendence publishes data on the level of public pensions but not on private pensions (SBC 2005). Public pensions have a higher rate of return than the capital return in the private system, which was one key reason for the majority of insured persons to stay in the public system until 2002, but that year a second reform tightened the pension formula in the public system to reduce the amount of its pensions (LRP 2002; Kleinjans 2004). The Peruvian association of pension administrators reports that average public pensions were 19–59% lower than private pensions in 2006 (AAFP 2007) but, according to Morón and Carranza (2003), all low-income insured with 20 years of contributions (and half of those with 30 years of contributions) will have a higher pension in the public system because the private system did not grant the recognition bond in the first ten

years of the reform due to lack of regulations. In the Dominican Republic 18% of private system insured (mainly those who were older than age 45 at the time of the reform and shifted from the public to the private system) will get a pension lower than the one they would have received if they had stayed in the public system (Pérez Montás 2006). It is soon to predict whether private pensions will be higher than public ones in the future, because the private system is not yet mature, for instance, in 2004 it paid only 26% of total pensions in Chile,23 2% in Colombia, and 0.5% in Uruguay (BPS 2005a; SBC 2005). Determining the replacement rate (average pension as proportion of average working life salary) in closed public systems is relatively easy, since it is based on defined benefit, thus in Chile such rate stretched from 61 to 80% in 2000. But it is much more difficult to estimate the replacement rate in private systems, since it depends on multiple variables: age of entry in employment, growth rate of wages, contribution density, and rate of return on the pension fund. Until recently, estimates of private system replacement rates in Chile relied on assumed contribution densities that oscillated from 70 to 90% (depending on pessimistic or optimistic assumptions) and averaged 80%. Recent simulations of replacement rates are more accurate than previous assumed rates because they use actual contribution densities from 2002 survey data. Based on densities differentiated by age, the observed replacement rates for all groups decreased to a range from 53 to 64% contrasted with the previous average rate of 80%, those rates are lower than public replacement rates that range from 61 to 80% (EPS 2004; Arenas de Mesa and MesaLago 2006). Argentinian estimates for 1994–2003 show that 52% of total affiliates had less than 32% contribution density, 30% of them had densities between 32 and 80%, and only 20% of affiliates had 80% or higher contribution density. The average replacement rate in the private system was 72% in 2005 and was projected to decline to 53 or 42% in 2050 (based on two different assumptions); projections for

76

Reassembling Social Security

the public system exhibited a decline from 75 to 55% (MTESS-ILO 2005). A Dominican Republic actuary has calculated that the average replacement rate of insured in the private system with 30 years of contributions and adequate density will be between 35 and 45% and a considerable proportion of them will only qualify for the minimum pension (Pérez Montás 2006). The access and level of the minimum pension can be used as indicators of the current and future sufficiency of pensions. Chilean recipients of the minimum pension (combining the public and private system) in 2000 accounted for 43% of the total number of pensioners in the two systems; the minimum pension averaged 62% of the minimum wage and 23% of the average wage in the private system in 2005 and both percentages showed a downward trend in 1990–2005. The contribution density projected when the reform was designed was higher than it actually is: more accurate and recent estimates show that about half of affiliates in the private system will receive a minimum pension, 35% of men and 60% of women (Arenas de Mesa and Hernández 2001; Arenas de Mesa and Benavides 2003; Bertranou, Gana and Vázquez 2006). According to Argentinian surveys in 2001, within the population of productive age, 33% of men and 45% of women had little or no hope of meeting the requirements to obtain a minimum pension (Bertranou and Arenas de Mesa 2003; see also Estrada 2004). The majority of Dominican Republic insured will only receive a minimum pension, equal to the lowest minimum wage and granted only for old-age pensions (Pérez Montás 2006). Based on surveys in the metropolitan areas of Santiago and Lima in 2000, the World Bank estimates that 30% of Chilean males and 50% of females affiliated in the private system do not meet the requirements to get a minimum pension, and 30% and 60% respectively in Peru (Gill, Packard, and Yermo 2005). These percentages would have been even higher if the surveys had had a national scope and included rural and smaller urban areas.

4.4. Impact of the reforms on equal treatment, solidarity, and comprehensiveness/ sufficiency 4.4.1. Equal treatment 4.4.1.1. Survival of schemes with privileged entitlement conditions Despite the alleged superiority of private systems, virtually all structural reforms excluded several powerful groups of insured in separate schemes with more generous entitlement conditions: retirement at 10-to-22 years younger, less years of contribution and seniority pensions regardless of age; such exclusion is similarly found in public systems. The armed forces in all countries (except in Costa Rica and Panama, and partly in Bolivia) have successfully resisted integration albeit they implemented Chile’s structural reform. There are also separate schemes with superior conditions for civil servants in thirteen countries, as well as for other powerful groups in seventeen countries, exceptions are Bolivia and Panama (the latter unique in having one fully standardized program for all insured). Pensions of the armed forces and civil servants are usually equal to the past year’s or month’s salary and adjusted to the personnel in activity salary, hence their pension levels (as well as those of congressmen, judges and teachers) are from 6 to 36 times higher than the average pension in the general program in at least five countries.

4.4.1.2. Gender inequalities In most countries women’s coverage is lower than men’s both among active workers and elder persons whereas female pensions are usually lower than male because of factors external and internal to the pension system. External factors are (relative to men): a lower rate of labor force participation and concentrated on uninsured occupations), higher unemployment, lower training at work, smaller salary for equal work, lower

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 77 contribution density (due to all previous factors and exit from work to raise children), and 4-to-5 years longer life expectancy. Ten systems (half private and half public) set an age of retirement for women five years younger than men, resulting in women’s average retirement span being 9-to-10 years longer than men’s. The other ten systems (equally divided) set equal retirement ages for both sexes, which facilitate women to accumulate more contributions and a bigger fund in their individual accounts, but don’t compensate for higher female life expectancy. In all systems, most women are indirectly insured through their husbands and, if the couple divorces and the man remarries, the former wife may lose all or part of her pension. Women’s work at home raising children and caring for the elderly is neither remunerated nor protected by pensions despite the considerable savings these tasks generate for society. Although gender inequalities exist in private and public systems, the latter are relatively more neutral or positive because they grant the minimum pension with none or fewer years of contribution, base the pension formula on the latest years of the working life, and use unisex mortality tables, hence transferring subsidies from men to women; nevertheless, if women retire earlier their pension could be smaller. Private systems are not neutral on gender and accentuate inequalities because they require more years of contribution to grant the minimum pension, are based on contributions during all the working life, and apply mortality tables differentiated by sex, all of which result in lower women’s pensions (inequality is attenuated for married women because the men’s annuity takes into account the spouse’s life expectancy). Mixed systems tend to compensate gender inequality more than substitutive systems that are totally private, because the public pillar attenuates such inequality, while the private pillar accentuates the inequality, but this depends on the relative importance of the public pillar relative to the private pillar. Scarce and debatable simulations on the impact on gender of

private and public systems offer mixed results but acknowledge that, despite the structural reforms, women earn lower returns than men. Longitudinal data from Chile, after 25 years of reform, show women having lower funds in individual accounts, replacement rates and average pensions relative to men and 45% of female insured getting a benefit lower than the minimum pension.

4.4.2. Solidarity 4.4.2.1. Solidarity replaced by equivalence in private systems The adherence to strict equivalence in private systems eliminated solidarity among income groups, generations and genders, reproducing existing inequalities in the labor market and salaries. Supporters of structural reform claim that there is solidarity in private systems but actually it is exogenous to them and the state responsibility. Exceptions are Colombia and Costa Rica (and Dominican Republic nonenforced law) that have solidarity contributions paid by employers, high-income workers, or the state to help finance low-income selfemployed or assistance pensions. The ILO asserts that social security is an important tool to promote solidarity and redistribution, and that private systems should not weaken those systems where the risk is distributed or pooled among all the insured.

4.4.2.2. Mechanisms in favor and against solidarity Although solidarity in most public systems has been eroded in practice, still they have more mechanisms favorable to solidarity than in private systems, where mechanisms against solidarity prevail: the worker contribution is much higher in private systems; the employer contribution has been retained in all public systems but has been eliminated or reduced in five private systems that have increased workers’ contributions; private systems charge high

78

Reassembling Social Security

administrative costs to the insured (including regressive fixed commissions that affect proportionally more low-income insured diminishing their deposit in the individual account and the future pension) but not public systems; private systems defer tax payment on workers’ contributions that proportionally favor those with higher income, not granted by public systems; years of contributions for minimum pensions have been augmented in private systems (Bolivia does not grant such pensions) but they are not required or are fewer in public ones; private systems do not impose a ceiling on pensions as public systems do; there is intergenerational solidarity in public systems but not in private systems that also generate inequality against older insured who bear higher installation costs of infrastructure than younger insured; and gender inequalities are mollified in public systems but accentuated in private systems. Common mechanisms against solidarity in private and public systems are: the exclusion of privileged schemes that do not contribute to the general program but enjoy generous benefits and fiscal subsidies (except in Panama) aggravated in five public systems by the state failure to honor its obligations with the general program, the exclusion of the majority of self-employed workers and other low-income groups, the lack of social assistance pensions for the poor (in five private but eight public systems), and in countries with low coverage the uninsured, that is majority, partly finances coverage of the minority insured often through consumption or ad hoc taxes (in private systems to finance transition costs) and by the potential transfer of employers’ contributions to prices. There are no rigorous studies to measure the net effect of progressive and regressive mechanisms on income distribution in private and public systems, and to convincingly determine which of the two systems have a more progressive/regressive impact. The few available simulations are limited to a few countries, mostly before or right after the structural reforms, and their results are questionable.

4.4.3. Comprehensiveness and sufficiency A fundamental promise of structural reforms is that a private system will pay better pensions, a pledge that considerably helped in the shift of insured from public systems, and yet such assumption has not been proven so far. Several private systems have increased the years of contributions to gain the right to a minimum pension hence excluding about one-third of women and half of men in three countries; all public systems grant the right to a minimum pension and with fewer restrictions. In one third of all systems, pensions are either not adjusted or the government has discretionary power to do so (more in public than in private systems). On the other hand, regardless of the type of system and with a couple of exceptions, separate schemes enjoy entitlement conditions and benefits more generous than in the general program. Social assistance pensions are not granted in 70% of the countries, 60% in private systems and 80% in public systems, the difference seems to be explained by the level of development of the countries and social security systems rather than the private or public nature of the pension system (except for Bolivia).

4.4.3.1. Retirement ages and spans Structural reforms increased statutory ages for retirement in five countries and the years of contribution required for the minimum pension in seven countries, which makes it more difficult to gain the right to the latter. In Peru, one of the least developed countries in the region, the retirement age was raised to 65 for both sexes, 10 years more for women and 5 more for men than before, resulting in the shortest retirement spans in the region. Most public systems have not increased the ages of retirement but, in four of them, ages are very low relative to life expectancy at retirement (Ecuador, Cuba, Venezuela, and Panama); conversely, Honduras ages are too high relative to male life expectancy. Although

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 79 comparisons are not precise, average retirement ages in private systems are higher than in public systems, while retirement spans are lower. The median of 15 years of contribution in public systems is too low, compared with a median of 20 to 25 in private systems, but 30 years are required in Argentina and 35 years in Uruguay, and if not met there is no right to the public-pillar pension but only to withdraw the fund accumulated in the private-pillar. Brazil’s parametric reform increases the very low retirement ages of civil servants, while that of Venezuela (not in force yet) stipulates a gradual raise of retirement ages relative to life expectancy. In countries with structural reform, only a minority of the insured with high income and contribution density will be able to accumulate a minimum amount in the individual account, retire before the statutory age, and receive a pension with an adequate replacement rate. But the majority will have to continue working until reaching the statutory age of retirement that has been excessively augmented in some countries. The stricter conditions in private systems are justified by the necessity to fortify their financial sustainability, but they create disincentives for affiliation and compliance and have a negative effect on coverage and poverty. Social assistance pensions require from 5 to 10 years of age more than contributory pensions, despite the lower life expectancy of the poor. The laws of two countries (Colombia and Dominican Republic), not yet implemented, cut the retirement age for assistance pensions, whereas Cuba doesn’t require a specific age and retirement ages in Brazil for rural pensions are reduced by 5 years and only 15 years of work are needed in the absence of contributions.

4.4.3.2. Liberal formula to calculate public pensions The pension formula in public systems sets the base salary on the average last five years of salary, which is too short a period and

provokes negative effects: stimulates under reporting of salaries during most of the working life and overreporting in the last years before retirement (in order to minimize the contribution and maximize the pension), penalizes those who honestly report their full salary as well as manual laborers who suffer a decline in salary due to physical deterioration toward the end of their working life, undermines the link between contributions and pension levels, and submits the pension level to the hazard of inflation. Replacement rates tend to be high: the minimum ranges from 50 to 70% (with two exceptions) and the maximum ranges from 80 to 100%; both exceed the ILO minimum norm of 45%. These entitlement conditions are financially unsustainable and, if not tightened, will eventually lead to financial bankruptcy in many public systems.

4.4.3.3. Minimum and maximum pensions In all private systems, for those insured who meet entitlement conditions but not the required accumulation in their individual accounts, the state (or social insurance in mixed models) finances the difference so that they receive a minimum pension, except in Bolivia. Most private systems, however, tie eligibility to the minimum pension to the number of contributions paid to the second pillar of mandatory savings, and some have increased the number of such contributions making it more difficult to gain the right, thus in Argentina, Chile, and Peru 30–33% of affiliate men and 45–60% of affiliate women won’t be eligible for a minimum pension. All public systems legally grant a minimum pension with fewer restrictions than private systems. The level of the minimum pension is usually fixed in relation to the minimum wage: 100% in Brazil, Dominican Republic (lowest minimum wage), and Mexico (minimum wage of the Federal District), 75% in Honduras, 67% in Cuba, and 64% in Chile. In private systems there is no ceiling imposed on the pension, which is based on the fund accumulated in

80

Reassembling Social Security

the individual account without limit. All public systems have a maximum monthly pension which together with the minimum pension generates progressive distribution effects, ranging from US$125 in Ecuador to US$1,500 in Panama.

4.4.3.4. Adjustment of pensions Private systems tend to have better adjustment mechanisms than public systems. Eight systems (four private and four public) adjust their pensions to the CPI; three to either the wage index, an automatic mechanism to the currency unit, or to the US dollar (all private); and one based on the minimum salary (public). Eight systems, three private (Argentina, Dominican Republic, and El Salvador) and five public (Cuba, Haiti, Nicaragua, Panama, and Venezuela), lack an institutionalized mechanism for annual adjustment, or the government has ample discretionary power to do the adjustment, often depending on available fiscal resources. The minimum pension in the large majority of countries is either adjusted based on the minimum wage or left to the government’s discretion. The lack of pension adjustment has resulted or will result in significant loss in the real value of pensions and increase poverty in countries that otherwise have generous entitlement conditions (for instance, Cuba and Panama). In separate schemes (armed forces and civil servants), both in private and public systems, the pension is often adjusted to the salary of the personnel in activity or annually adjusted to the CPI or above it.

4.4.3.5. Noncontributory pensions Despite the recognized importance of noncontributory pensions, they are granted in only seven systems (five private and two public). Such pensions range from 31 to 32% of the average contributory pension in Brazil, Chile, and Costa Rica; 41–47% in Argentina and Cuba; and 60% in Uruguay, the relatively small gap therefore should generate incentives for evasion and payment delays in

the contributory program. Despite that closeness between the two pensions, however, the monthly level of the noncontributory pension is generally small; it ranges from US$5 in Cuba to US$19 in Bolivia; to US$40–49 in Argentina, Costa Rica, and Dominican Republic (not yet in force); to US$75–77 in Brazil and Chile; and to US$134 in Uruguay. In Cuba and Bolivia the level is insufficient to cover minimum food needs, while it probably meets those needs in most of the other countries, and seems adequate to meet basic needs in Uruguay. In Argentina half of the total noncontributory pensions are approved by congress to persons with resources and in Costa Rica some social assistance pensions are granted to nonpoor.

4.4.3.6. Nonconfirmed assumption of higher private than public pensions The promise that private pensions are and will be higher than public pensions cannot be empirically proven at least for the time being, due to the lack of comparative statistics on the average amount of pensions in both systems, the small percentage of current private pensions relative to the total number of pensions, and difficulties in projecting replacement rates in private systems. In Chile, which has the longest period of reform operation, the average private old-age pension (that accounted for 63% of total pensions) in 2001–2 was 24% smaller than the corresponding average public pension, and the weighted average of all private pensions was only 3% higher than the average of all public pensions, but in 2004 the average pension was 16% higher than the average public pension. In Argentina, the average private pension in 2003 was 22% higher than the public average but in 2005 the former was 6% lower than the latter. Colombia’s public pensions have been higher than private pensions because of the generous formula of calculation, an incentive to stay in the public system that is being dismantled by the reform of 2002 that tightened such formulae. In the Dominican Republic, 18% of

Equal Treatment, Solidarity, and Comprehensiveness-Sufficiency 81 the insured in the private system will get a pension lower than if they had stayed in the public system. It is tough to compare future replacement rates in both systems, because private systems are not yet mature and their rates will depend on multiple factors difficult to predict. In Chile, new projections based on observed contribution densities from survey data denote replacement rates between 53 and 64% compared with a previously assumed average of 80%, and contrasted with estimated public replacement rates between 61 and 80%; furthermore, half of current affiliates will only receive the minimum pension.

In Argentina, the average replacement rate in the private system is projected to decline from 72 to 53% or 42% in 2005–50, worse than in the public system. In the Dominican Republic, replacement rates are projected to range from 35 to 45% and the majority of insured will only receive a minimum pension. Surveys taken in metropolitan Santiago and Lima confirm that 30% of male affiliates and 50–60% of females will not meet the requirements for a minimum pension, and such percentages must be higher at the national level given that those not living in metropolitan areas are predominantly rural.

Notes 1. Sources for this section are legislation; Ayala and Acosta (2001); ILO (2002b); IPSFA (2002, 2006); ISSA (2003b, 2005); Morón and Carranza (2003); Saldaín (2003); SHCP (2004); Herrera (2005); MPS (2005b); and Arenas Monsalve (2006). 2. Brazil’s constitutional amendment of 2005 stipulates that to get a pension equal to 100% of the last salary, civil servants must be 55/60 years of age, have 30/35 years of contributions, 25 years of service in the public sector, 10 in the career, and 5 in the same post. 3. In Chile, 37% of women aged 65 and older receive contributory pensions versus 71% of men in the same age bracket, while 17 and 12% receive noncontributory pensions, and 23% and 0.3% survivor pensions respectively (Montecinos 2006). 4. For normal ages of retirement, but reduced to 60/62 for early retirement. 5. Chilean women retiring at the same age as men will receive a pension 10% lower based on different mortality tables but the same based on unisex tables (Bertranou and Bravo 2006). 6. World Bank officials reject unisex mortality tables because of their supposedly adverse effects on private insurance markets (Gill, Packard, and Yermo 2005), although such adverse effects occur in voluntary rather than mandatory schemes (because men can decide to opt out). The United States, European Union, and part of Central and Eastern Europe mandatory social security systems use unisex tables, as well as many enterprise pensions. 7. From an economic perspective, the distinction between worker’s and employer’s contribution is said to be irrelevant as both are part of the ‘gross wage’ (nonwage remuneration). The employer’s contribution could either be transferred to the worker or the consumer; reformers argue that the employer actually pays it. 8. The repeal of Nicaragua’s structural reform law gave as a reason: ‘the social inequality created by forcing the entire population to pay for the transition cost of a system that would have benefited only the formal labor sector’ (Ley 568 2006). 9. Peru’s 1% solidarity contribution initially paid in the private system was later abolished. Chile’s legal draft of reform of 2006 introduces a progressive solidarity pension that is gradually reduced according to the percentage of the pension received from the contributory system until it is zero with a pension of US$375, and the state provides a solidarity contribution (Gobierno 2006). 10. In Argentina, a decree of 2005 allows retirement in the public system with the required age but not the thirty years of contributions, deducting the needed contributions from the pension. The reform law of 2007 guarantees a minimum pension in the public and private systems (Ley 26,222 2007).

82

Reassembling Social Security 11. Civil servants hired before 1998 can retire at ages 40/53 with 25/30 years of contributions and 5 years in the post; the seniority option (with 30/35 years of contribution regardless of age) is eliminated; future civil servants get their ages for old-age retirement increased to 55/60, the years of contribution to 30/35, and the years in the job to 10; and those retiring before those ages suffer a discount for each year thus creating incentives to postpone retirement. 12. Ecuador partly declared unconstitutional law revises the public system minimum retirement age every five years to maintain the retirement span at fifteen years; if the law is implemented such ages should be raised from 55 to 66 for men and 70 for women (World Bank 2005e). 13. Venezuela’s reform law of 2002 (not implemented by the end of 2006) stipulates that by 2007 another law must gradually increase the ages of retirement relative to life expectancy and years for contribution (LOSSS 2002). 14. Sources for this section are legislation; Mesa-Lago (2000c, 2000e, 2001a, 2005a, 2005b, 2006d); Durán and Cercone (2001); ILO (2001b, 2003a; MPAS (2002b); MPS (2003); Saldaín (2003); Sandó (2003); Schwarzer (2004); Ley 51 (2005); and Asamblea (2006). 15. The law of 2002 not enforced by the end of 2006 sets a maximum rate of 85%; it has been estimated that the minimum pension would be equivalent to the minimum salary and would not be financially sustainable (Villasmil 2006). 16. In Brazil the real pension (RGPS) rose 22% in 1993–2002; in Chile (private) it increased 66% in 1990–2005; in Cuba it fell 41% in 1989–98; in Ecuador jumped 340% in 1995– 2004; in Guatemala rose 26% in 1990–2000; in Honduras decreased 50% in 1994–1998; in Venezuela declined 60% in 1982–94, and in Uruguay fell 25% in 2000–4 (Mesa-Lago 1995, 2005b; Durán and Cercone 2001; MPS 2003; World Bank 2005e; Bertranou, Gana, and Vázquez 2006; Lagomarsino 2006). 17. A law of 2002 changed the adjustment of public pensions to a unit of account of local currency transactions similar to the UF in Chile but with a lag of one year; private pensions continued to be indexed to the US dollar (Garrón 2004; Bonadona 2006). 18. Adjusting benefits to the CPI keeps the purchasing power of pensioners but wages usually increase faster than prices hence creating a widening gap in future generations between wage earners and pensioners. Adjusting pensions to the minimum salary generates obstacles to salary negotiations because of the extra expenditure on pensions (Bertranou, Gana and Vázquez 2006). 19. Sources for this section are Müller (2001, 2004); Bertranou, Solorio, and Ginneken (2002); BPS (2002); ILO (2002b); Martínez and Mesa-Lago (2003); Mesa-Lago (2000e, 2006d); MTESS (2003); SSS (2003); Garrón (2004); Schwarzer (2004); Bertranou, Gana and Vázquez (2006); Lizardo (2006); Rodríguez Herrera (2006); Conte Grand (2006); and SPVS (2006). 20. Chile Solidario pays a monthly bonus of US$20 to the extreme poor aged 65 and above, gradually halved in two years. The noncontributory pension was increased by 10% to US$83 in 2006. The legal draft of reform of 2006 introduces a monthly basic pension starting at US$110 in 2008 rising to US$140 in 2009 (equivalent to 68% of the minimum salary), this pension will gradually replace social assistance and minimum pensions (Gobierno 2006). 21. The Dominican tiny aid scheme for the poor that covers 0.2% of the population pays US$9 monthly, grossly insufficient to meet basic food needs. 22. Average public pensions are given in pesos while average private pensions are given in UF (currency units automatically adjusted to inflation) (SAFPb 2005; SSS 2005). A comparison of aggregate average private and public pensions in 2004 indicates that the former was 16% higher than the latter, but no desegregation was given by type of pension (Bertranou, Gana and Vázquez 2006). 23. The total number of pensions was 1.8 million in 2004, out of which 54% were public contributory, 20% public noncontributory and 26% private; the latter would be smaller if minimum pensions paid by the state were included (SAFPb 2005; SSS 2005; Bertranou and Bravo 2006).

5

Effects on Unity, State Responsibility, Efficiency, Costs, Social Participation, and Reform Goals

Structural reforms pursue the elimination of the state or social insurance monopoly in public systems and claim that with private administration and freedom of choice of the insured to select administrators, competition will be promoted and will improve efficiency and reduce administrative costs. But a proper function of private systems requires that the state exerts fundamental functions: making affiliation mandatory, regulating the system, creating a public superintendence to oversee it, and finance substantial fiscal costs during the transition. This chapter measures the degree of unification or segmentation in public and private systems, discusses the overall role of the state in coordination and supervision; examines freedom of choice in private systems and estimates the degree of privatization reached; tests whether competition truly functions, affiliates have information and skills for rational selections, and efficiency has improved; calculates and contrasts administrative costs in private and public systems, and compares social participation in both systems.

5.1. Unification versus segmentation Some structural reforms incorporated separate pension schemes in private systems, but the armed forces (with two exceptions), civil servants in six countries, and other powerful

groups in nine countries resisted integration and kept their schemes. In all public systems (except one) the armed forces have also preserved their schemes; as well as civil servants in seven countries and other groups in eight countries. Substitutive reform models closed public systems and often placed them under a single public management entity, whereas in parallel models there are two separate systems: public and private. In mixed models there are two approaches: Argentina and Uruguay have a public system and a mixed one with two pillars, but Costa Rica has only one mixed system with its two pillars. All private systems have voluntary supplementary pensions (third pillar), and virtually all public systems also have voluntary supplementary funds for certain groups (the largest number in Brazil). These factors generate significantly different degrees of unification/segmentation in both private and public systems without a clear distinctive pattern between the two. There have been no unification processes in five public systems and one private system. Private systems have more decentralized management than public systems. The administration of the six noncontributory pensions is separated from that of contributory pensions in three countries but combined in the other three. The process of unification of multiple contributory pension programs and schemes in the region has been difficult, prolonged, and uneven; there are successful examples of total or partial unification into a general program in both structural and parametric

84

Reassembling Social Security

reforms. Among private systems, Bolivia integrated the major program, 28 supplementary funds and 5 separate schemes (Garrón 2004); Chile unified 33 of 35 schemes into a public institution; Costa Rica incorporated 17 of 19 independent pension funds for civil servants; and Dominican Republic integrated the schemes for civil servants, but teachers halted their contributions to the private system and transferred them to their own scheme in 2005 (Pérez Montás 2006). In Argentina 10 of the 24 schemes for civil servants in the provinces were incorporated into the Union general program through agreements but the other 14 still have independent schemes; in addition, 18 of 20 schemes of municipalities and numerous provincial schemes for professionals remain separated (MTESS-ILO 2005). Colombia’s initial reform law (1993) allowed the survival of multiple schemes for civil servants and oil workers postponing their incorporation until 2014, but the second reform law (2002) ordered the affiliation to the general program of new civil servants and oil workers, as well as the reform of schemes for teachers, armed forces, and former presidents; the third reform law (2005) stipulates shutting down all separate schemes by 2010 but exempts the armed forces, teachers, and presidents (Ayala and Acosta 2002; Arenas Monsalve 2006). El Salvador’s reform law closed the scheme of civil servants and incorporated new entrants into the general program, but allowed the survival of other pension schemes with special laws provided they submit to rules established in the said law. Peru’s civil servant scheme continues separated but closed to new members although some groups have tried to reenter it; the armed forces–policemen and fishermen schemes stay independent. Prior to Uruguay’s structural reform there was a process of unification of multiple schemes into the social insurance institute but keeping diverse regimes (civil servants, industry and commerce, rural workers, and domestic servants) and the reform did not integrate six separate schemes. Mexico has not had a unification process except for the

integration of a few minor schemes into the general program. Panama is the most unified public system in the region as one single program covers all workers, without any special regime. There are unification processes in three countries. Brazil has a general program (RGPS) for salaried urban workers in the private sector, state enterprises and two-thirds of the municipalities, and a scheme for rural workers; additionally there are thousands of schemes for federal civil servants (including the three government branches and the armed forces) in the 27 states and one-third of the municipalities. A process of partial unification or convergence (particularly for new civil servants) promoted by the parametric reforms of 1998– 2004 has integrated all schemes of civil servants within each state and set the bases for a future standardization of entitlement conditions of federal, state, and municipal employees; but the full unification of all civil servant schemes was blocked by states and municipalities that requested that the federal government should finance fiscal transition costs (MPS 2003; Interviews 2004; Pinheiro 2005). Cuba unified 51 separate pension schemes in 1959–63 (Peñate 2000b). Venezuela’s reform law of 2002 (not enforced in 2006) gives five years to integrate all existing schemes into one single system, but coexisting as special regimes until all insured rights are extinguished, but the constitution excludes the armed forces from the integrated system.1 Ecuador, Guatemala, Haiti, Honduras, and Paraguay have not had unification processes; in Ecuador attempts are being made to create additional municipal schemes without a national framework (IADB 2001; Balsells 2002; Durán 2003; Saldaín 2003). Table 5.1 classifies private and public pension systems by their current degree of unification or segmentation in twenty countries and identifies separate schemes for the armed forces, civil servants, and other groups: the three government branches, policemen, teachers and university professors, professionals, journalists, banking employees, oil, railroad, electricity and printing workers, and

Table 5.1. Unification versus segmentation in private and public pension systems, 2005 Separate programs/schemes Systems/ countries Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Public systems Brazil Cuba Ecuador Guatemala Haiti Honduras Nicaragua Panama

Degree of unification or segmentation

Number of programs/ schemesa

Armed forces

Civil servants

Others

Highly segmented Unifiedb Relatively unified Highly segmentedc Relatively unified Segmented Highly segmented Highly segmented, without coordination Segmented Highly segmented

9+e 1 4 9 3 6 7 5+e

Yes Nob Yes Yes No Yes Yes Yes

Yes No No Yes No Yesf Yesf Yes

Policemen (2), provinces (14), municipalities (18), professionals, etc. No Policemen Presidents, congress, teachers, oil, national university Teachers, judiciary, supplementary funds Policemen, teachers (since 2005) Policemen, congress, judiciary, central bankf State and municipal employees, oil, supplementary funds

5 7

Yes Yes

Yesf No

Fishermen Banking, public notaries, university professionals, policemen

Highly segmented, without coordination Relatively unified Segmented without coordination

4+e

Yes

Yes

3 5+e

Yes Yes

No Yes

3 5+e 7

Yes Yes Yes

Yes Yes Yes

3 1

Yes No

No No

States, municipalities (one-third), rural workers, numerous supplementary funds Internal security Policemen, peasants, special regimes for teachers, printers, construction, etc. No; but 14 supplementary funds Various in autonomous agencies, banks, etc. Executive, judiciary, congress, teachers, university, central bank, journalists; supplementary funds Policemen No; but supplementary funds (cont.)

Relatively unified Segmented without coordination Highly segmented, without coordination Relatively unified Unified (the most)

Table 5.1. (Continued) Separate programs/schemes Systems/ countries Paraguay Venezuela

Degree of unification or segmentation Highly segmented without coordination Highly segmented (the most) without coordination, expected to unifyd

Number of programs/ schemesa

Armed forces

Civil servants

7

Yes

Yes

13

Yes

Yes

Others Executive, judiciary, congress, policemen, teachers, banking, railroads, printing, domestic service, self-employed State, municipal, judiciary, congress, comptroller, teachers, oil, central bank, policemen, electricity, universities and 400 more

Sources: Legislation; other sources in the text. a The major program is always counted in addition to the other programs/schemes identified in the table; in Argentina, Chile, Colombia, Dominican Republic, El Salvador, and Peru two programs are counted: the public system (either open or closed) and the private system. b The reform law integrated the armed forces into the private system but with special norms. c The initial reform law allowed the continuation of schemes for civil servants and others until 2014, the second law mandates the incorporation of new civil servants and oil workers, and the reform of schemes for presidents, teachers, and armed forces, the third law orders to shut down all separate schemes but exempts civil servants, armed forces, teachers, and presidents. d The law of 2002 not yet in force at the end of 2006 stipulates the unification of all separate programs except for the armed forces. e Principal programs, excludes many programs in states, provinces, municipalities, etc. (at least fifty in Argentina, thousands in Brazil); within Ecuador’s major program in the public system there are thirteen special regimes (including civil servants and listed in ‘others’); the reform law was expected to partly unify them; in Haiti there are many schemes for banking employees, autonomous agencies, etc. f Scheme closed but still have insured who decided to stay, all new civil servants must join the private system.

Effects on Unity, State Responsibility, Efficiency fishermen. The two ‘unified’ systems are those of Bolivia (private) and Panama (public) that have one sole program although Bolivian armed forces have a special regime. The five ‘relatively unified’ systems have three or four programs, the general program (private or public—including the closed program) and separate schemes: Costa Rica, Cuba, Chile, Guatemala (without counting 14 supplementary pension schemes), and Nicaragua. The four ‘segmented’ systems have 5 to 6 programs: Dominican Republic, Ecuador (without counting 13 special regimes within the major program), Haiti, and Peru. The nine ‘highly segmented’ systems have from 7 to 13 programs, all of them with separate schemes for the armed forces and civil servants (in three countries at the federal, state/provincial and municipal levels) plus schemes for many other groups: Brazil (the most segmented counting the thousands of separate schemes in the federal government, states, and municipalities); Argentina (the second most segmented counting some 50 schemes in provinces, municipalities, and professionals); Colombia; El Salvador; Honduras; Mexico (the third most segmented counting the separate schemes in states and municipalities); Paraguay; Uruguay; and Venezuela (the fourth most segmented with at least 13 schemes and 400 special regimes of civil servants alone). The above analysis does not denote a clearcut separation between private and public systems concerning their degree of unification or segmentation (public systems are slightly more unified than private systems), but shows that 13 of the 20 systems are either segmented or highly segmented, 7 private and 6 public. Segmentation without coordination often results in lack of portability, loss of contributions of workers who change jobs and impossibility of gaining a pension. The management of private systems has been decentralized, albeit with diverse degree, through two to sixteen administrators; however, the remaining public systems or schemes have traditional centralized administration, as in Argentina, Colombia, Costa Rica, and

87

Mexico. Public systems tend to have a more centralized administration than private systems, Cuba being the extreme case; Brazil is much more decentralized because of the segmentation by Union, states, and municipalities (partly true in Argentina also), and Honduras, Paraguay, and Venezuela because of the fragmentation into numerous autonomous schemes, but within each of them there is a centralized structure. The administration of noncontributory pensions is completely separated from that of contributory programs and in charge of the state in Argentina (provinces and municipalities have their own schemes) and in Chile. Bolivia’s Bonosol is financed by a separate fund managed by the two private pension administrators. A separate administration of noncontributory pensions is appropriate to avoid potential transfers from insurance to assistance. For instance in Costa Rica and Uruguay, social assistance pensions continue to be managed by social insurance, legally separated from the contributory program but, in practice, transfers from insurance to assistance have occurred in the past when the state did not fulfill its financial obligations with the assistance scheme. Among public systems, in Brazil the federal government manages both the main contributory program and social assistance pensions, and in Cuba both contributory and assistance pensions are managed and paid by the state (Mesa-Lago 2001a).2

5.2. The role of the state The role of the state in the administration and financing of pensions varies according to systems and countries; reformers adduce that such a role is subsidiary in private systems. According to theory there should not be an important role for the government where consumers (insured) in markets are well informed, competition works properly and there are no technical impediments to efficiency. Actually the insured lack information, competition does not work properly and efficiency

88

Reassembling Social Security

gains have been uneven, hence there is an important role for the state in private pension systems. Current state functions are fundamental rather than subsidiary: making mandatory the affiliation, regulating and supervising of the system, financing all transition costs including pensions from closed public systems, guaranteeing minimum pensions and minimum average capital returns, taking care of bankrupt pension funds, and providing social assistance pensions. Public, minimum, and social assistance pensions, all paid by the state, accounted for 80% of all pensions in Chile in 2002, whereas private pensions only accounted for 20% (based on Arenas de Mesa and Benavides 2003; SAFPb 2003; SSS 2003). In the long run private pensions will increase while public pensions will decrease and disappear, but the state will continue financing minimum and assistance pensions (see Section 6.3). Additional state functions are needed to correct current inefficiencies and flaws in private pension systems (for a comparison of state functions see Müller 2002). In Argentina and Uruguay the state manages the public systems/pillars and finances about half of their deficit; in Bolivia and Mexico the state treasury became the financier of all ongoing pensions after their public systems were totally shut down; Chile and El Salvador placed closed public schemes under public management; in Colombia’s parallel model the public system has almost half of the insured and is managed by social insurance, and disability and survivors’ branches are managed by social insurance in Colombia, Costa Rica, and Mexico. In all private systems there is an overseeing public agency; in six of them there is an ad hoc superintendence of pensions (Argentina, Chile, Costa Rica, Dominican Republic, El Salvador, and Mexico), while in four countries the superintendence has a wider scope, overseeing also insurances, banks, and/or other financial institutions (Bolivia, Colombia, Ecuador—partly unconstitutional law, and Peru), and in Uruguay it is the central bank. A single superintendence in Colombia and Costa Rica oversees both the private and

the public systems/pillars; Chile has two separate superintendencies but the legal reform draft of 2006 proposes their unification (Gobierno 2006). The degree of autonomy of the superintendence, a crucial attribute for an adequate performance of the pension system, varies among the countries (see Sections 5.4 and 5.5). The superintendence is financed by the state in Chile, Colombia, Costa Rica, Mexico, and Uruguay; by the fund administrators in Argentina, Bolivia, El Salvador, and Peru; and by the insured and employers in the Dominican Republic. There are no superintendencies of pensions in public systems; the overseeing functions are divided into several state agencies often generating conflicts of jurisdiction, overlapping, and lacunae. Among public systems the role of the state is obviously higher than that in private systems, reaching a maximum in Cuba, where it administers and finances the entire system: a subordinate agency to the ministry of labor and social security manages the general contributory program and the overall social assistance program; armed forces and internal security personnel schemes are managed by their respective ministries; and there are no supplementary or private funds. In Brazil the general program (RGPS), as well as those of rural workers and social assistance pensions are administered by the federal government (Union), while the schemes of civil servants at the three geographical levels (including the armed forces) are managed by the Union, the 27 states, and the municipalities; the states subsidizes all these programs; 68% of the total pension fund in voluntary supplementary pensions is managed by public institutions and 32% by private ones; various public agencies supervise these funds and a law creates a unified agency for their regulation and supervision (MPAS 2002b; MPS 2003). In the remaining public systems the role of the state is considerably lower. Ecuador’s major program (IESS) is autonomous, as are the separate schemes for the armed forces and policemen, there are several minor supplementary funds; the system lacks a single

Effects on Unity, State Responsibility, Efficiency public agency in charge of policy, coordination, and supervision—three entities share such functions with unclear demarcation (World Bank 2005e). Guatemala’s general program (IGSS) is autonomous, while the civil servants and armed forces schemes are directly administered by the state; supplementary pension funds are managed by financial entities and insurance companies, regulated and supervised by the banking superintendence. Honduras’ seven pension programs are autonomous, a national commission of banking and insurance supervises investments of the major one (IHSS). Panama’s CSS is autonomous and the supplementary pension funds are managed by private entities; in Paraguay all seven programs are autonomous and there is no general supervisory agency; and in Venezuela the thirteen programs and schemes are also autonomous and until now there has been neither state regulation nor one public supervising agency, although they are planned in the reform law not implemented yet at the end of 2006. Protecting and enhancing the autonomy of social-insurance institutes is important in order to reduce state and political interference, but some of them enjoy excessive independence and power that obstructs their control and supervision. For instance, Ecuador’s IESS has autonomy far beyond the usual elsewhere: drafting of regulations, approval of benefit increases without an explicit indexation rule, and the implementation of the reform law (World Bank 2005e). Such problems are aggravated when there are multiple additional autonomous schemes that lack central coordination and supervision as in Brazil, Ecuador, Haiti, Honduras, Mexico, Paraguay, and Venezuela.

5.3. Reform goals: freedom of choice and privatization A goal of structural reforms is freedom of choice: because the private system is better than the public, the argument goes, most

89

insured in the public system shift to the private system. Statistics at the end of 2004 seemed to confirm such assumption: within the ten countries, an average of 84% of total contributors were in the private system or the private pillar of the mixed system and only 16% in the public system/pillar. Moreover, 100% of contributors were in the private system in Bolivia, Mexico, and Costa Rica; 95–96% in Chile, Dominican Republic, and El Salvador, 86% in Argentina, and 76% in Peru. Nevertheless, the distribution of the total number of contributors, adding private and public systems in nineteen countries, shows that 66% were in the public system and 34% in the private system; largely because half of total contributors are in Brazil’s public system (Table 5.2).3 The shift from public to private system among the ten countries does not result exclusively from the goodness of private systems but from other factors also: (a) restrictions to the freedom of the insured to stay in the public system or move to the private, including their age in some countries; (b) legal measures, incentives, and guarantees granted by the state, as well as advertisement and illusory promises geared at promoting the change; (c) default rules for new entrants to the labor force who are automatically enrolled in the private system if they do not make a choice; (d) decisions and influence by employers; (e) the time the reform has been in operation; and ( f ) the rate of return of the public system compared with the rate of capital return in the private system (see Table 5.3). Examples of the role of these six factors in promoting a change to the private system in the ten countries follow. In Bolivia and Mexico where 100% of the insured are in the private system, as well as Costa Rica’s mixed system, the law obliged all insured to change, hence there was no freedom of choice. In the five countries with the substitutive model and in three with the mixed model (Costa Rica,4 Ecuador—partly unconstitutional law, and Uruguay), those who enter the labor force must affiliate to the private system/pillar. Argentinian new

90

Reassembling Social Security

Table 5.2. Degree of privatization: distribution of contributors between private and public pension systems c.2004 Public system Systems/countriesa

Thousands

Private system

Total

%

Thousands

%

Thousands

%

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay

600 0 157 1,955 868b 25 27c 0 422 594d

14 0 4 47 100b 6 5 0 24 63

3,541 394 3,572 2,241 868b 568 489 12,573 1,357 346d

86 100 96 53 100b 95 95 100 76 37

4,141 394 3,729 4,196 868b 593e 516 12,573 1,779 940

100 100 100 100 100 100 100 100 100 100

Subtotal

4,648

16

25,081

84

29,729

100

Public systems Brazil Ecuador Guatemala Haiti Honduras Nicaragua Panama Paraguay Venezuela Subtotal

38,500 1,103 868 118 393 348 677 170 2,174 44,351

100 100 100 100 100 100 100 100 100 100

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

38,500 1,103 868 118 393 348 677 170 2,174 44,351

100 100 100 100 100 100 100 100 100 100

Total

48,999

66

25,081

34

74,080

100

Sources: Contributors to private system, all in December 2004, from AIOS (2005); except Dominican Republic SIPEN (2005). Contributors to public system or pillar connected with private systems, all in December 2004, from: Argentina SAFJP (2005); Chile SSS (2005); Colombia SBC (2005); Costa Rica CCSS (2005a); Dominican Republic SIPEN (2005); El Salvador UPISSS (2006); Peru Morón (2006); Uruguay BCU (2005) and BPS (2005a). Contributors to pure public systems in 2004 (except Brazil 2003, Haiti 1999, Honduras 2001, and Paraguay 2000) from: Brazil MPS (2005b); Ecuador IESS (2006a); Guatemala IGSS (2005); Haiti IADB (2001); Honduras IHSS (2003); Nicaragua INSS (2005a); Panama Mesa-Lago (2005a); Paraguay ILO (2003b); Venezuela IVSS (2006). a Ecuador and Nicaragua still have public systems because their structural reforms have not been implemented; Dominican Republic refers to the contributory regime; there are no data on Cuba (only 15% of the workers contribute to the system). b All the insured are in the mixed system: second pillar (private) and first pillar (public); the total and subtotal only counts Costa Rica once within the public system. c December 31, 2005. d Those in the mixed system are also in the public system; refers to jobs instead of active contributors and the total is overestimated by about 10%. e Includes 3% without information that were assigned proportionally to the private and public system.

Effects on Unity, State Responsibility, Efficiency

91

Table 5.3. Freedom of choice of the insured in private pension systems, 2005–6 Between systems Private systems Argentina Bolivia Chile Colombia

Costa Rica Dominican R.a

Ecuadorb El Salvador Mexico Peru Uruguay

Between administrators

Insured at the time of the reform

Insured after the reform

Years to change

Can change from public to mixed, but not return to public All had to change to private Had a period to choose between public and private Can change from public to private every 5 years but not 10 years before retirement All had to join the mixed Divided by age; younger than 45 must change to private, older have option Divided by age and income Divided by age, only the middle age group had choice All had to change to private Can change from public to private, but not return to public Divided by age; young changed to mixed, old could choose between public and mixed; also required certain income to join mixed

Can choose between public and mixed Must join private Must join private

2

AFJP

c

2d

AFP AFP

Can choose between public and private

2

SAFP

Must join mixed Must join private

1 1

OPC AFP

Must join mixed Must join private

n.a. 1

EDAP AFP

Must join private Can choose between public and private Must join mixed after a given salary level, optional for smaller earners

Abbreviated name

1e 2

AFORES AFP

2

AFAP

Source: Legislation; ISSA (2005). a Contributory regime. b Not in force yet as several articles of the law have been declared unconstitutional. c No changes were allowed until 2000, thereafter new workers are permitted to choose between two existing AFPs but change is restricted among those already insured. d There is no legal limit but in practice two are possible due to the time that the change takes. e More than once per year since 2005.

entrants who don’t make a choice within a given period are automatically assigned to the private system. Five countries divided the insured at the time of the reform according to their age, the younger had to move to the private or mixed system, the older either had to stay in the public system or had a choice to move (Dominican Republic, El Salvador, Uruguay, and Ecuador—partly unconstitutional law). Most of the insured

in Dominican Republic and El Salvador were young persons forced to move, which largely explains the 95% affiliated to the private system in those two countries (Table 5.2). When the reform has been for a long time in operation (25 years in Chile), the obligation of new workers to join the private system, together with the gradual retirement of insured left in the public system, leads to a high proportion affiliated in the private

92

Reassembling Social Security

system. Chilean insured at the time of the reform had a short period to decide either to stay in the public system or move to the private, but the latter contribution was reduced (the contribution to the public system was not) to stimulate the move; in addition 56% of affiliates in the private system reported they had been forced to move by their employers (EPS 2004); finally, the Chilean state granted the most generous guarantees in the private system during the transition. It is because of all of these reasons (not only for the goodness of the private system) that 96% of insured are in the private system. In Peru, initially the percentage contribution and age of retirement in the public system were lower than in the private system and there was freedom of choice between the two systems, therefore not enough insured were moving to the private system despite a 13.5% increment in salary to those who changed. To promote the shift the government raised the retirement age in the public system to the level of the private age (from 55/60 to 65), raised from 5–10 to 20, the years required for a minimum pension in the public system, and increased the contribution in the public system above that in the private one: 13 and 8%, respectively (Morón and Carranza 2003). Workers who enter the labor force in Peru and Argentina can choose between the public and the private or mixed system respectively, but once in the private or mixed system they cannot return to the public;5 this option has contributed to a lower proportion in the private/mixed system than in six substitutive systems: 76–86% vis-à-vis 95–100%. Advertisement has also played a crucial role in promoting the shift, as well as promises of better pensions, lower administrative costs, and insulation from politics and government interference in private systems relative to public ones, promises that generally have not materialized. In contrasts with the other eight countries, Uruguay’s public system keeps 63% of the total insured while Colombia’s retains 47% (Table 5.2). Colombia’s public system was financially fortified by a parametric reform and its reserves were similar to those of the

private system until 2002 (SBS 2003). Entitlement conditions, benefits, and contributions were the same in both systems, the rate of return of the public system until 2003 was higher than the rate of capital return of the private system,6 and the insured had freedom to shift between the two systems every three years. But the second reform (2002) gradually tightens entitlement conditions in 2005– 14: increases the retirement age by two years, expands the contribution years by five years, reduces the replacement rate by five percentage points, and augments the period to change systems from three to five years banning shifts ten years before retirement (LRP 2002). Finally, the public system has been left with the burden of paying all ongoing pensions and the state charged it with the payment of two extra monthly pensions. As result of all these changes, the level of public pensions will decline and eventually become lower than that in the private system, while public system finances have deteriorated (its reserves in 2004 were half of those in the private system), thus reducing the previous incentives to stay in it (SBC 2005).7 Uruguay gave a period to insured older than 40 to choose between the public and the mixed system, and the majority preferred the former; in addition only the insured with a certain level of income can join the mixed system; in 2005 those over 40 who shifted voluntarily to the mixed system and faced a low pension were allowed to return to the public system (ISSA 2005). Within private systems the insured have freedom to select and change an administrator more than once a year in Mexico (providing that the shift results in a lower commission), once a year in Costa Rica, Dominican Republic, and El Salvador, but only every two years in Argentina, Chile, Colombia, Peru, and Uruguay (Table 5.3). In Bolivia there was no freedom to change between the two administrators until 2000 and it was mostly limited to new workers; in 2005 only 0.4% of total affiliates changed administrators (SPVS 2006a, 2006b). In Argentina 80% of workers who don’t make the selection of an administrator are assigned to the private system and a

Effects on Unity, State Responsibility, Efficiency specific administrator (Conte-Grand 2004).8 Peruvian employers enroll their employees in the private system and to the administrator where the majority of their workers are affiliated. Since 2005 Mexico has made an important and positive change: workers who don’t choose an administrator are assigned to the one that charges the lowest commission, 77% of those with new accounts in 2004–6 selected the cheapest as well as 83% who shifted administrators in the first quarter of 2006 (Budebo 2006). The freedom to move is costly for the administrators and has been curtailed by rules, for instance in Chile restrictions on shifts reduced them from 40% of total contributors in 1994 to 4% in 2005 (SAFPb 2002, 2005). In Dominican Republic, Peru, and Uruguay only from 0.1 to 0.3% of the affiliates changed administrators in 2005 (AIOS 2005).

5.4. Reform goal: competition Structural reformers maintain that competition between pension fund administrators is

93

the essential foundation of the private system because it promotes efficiency and hence reduces administrative costs and increases capital returns. The underlying assumption is that administrators9 compete for the insured who in turn have the needed information and skills to rationally select the best, meaning those charging the lowest commission and paying the highest capital return, hence increasing the individual account and the pension level (security considerations are less important due to legal guarantees). This section and Sections 5.5–5.7 attempt to answer crucial questions: does competition really function, do the insured have the information and skills to make rational selections, has efficiency improved and competition reduced administrative costs? Competition is a positive element but largely depends on the size of the insured market. Table 5.4 suggests that the higher the number of insured the higher the number of administrators and vice versa:10 Mexico had 35 million insured and 16 administrators in 2005; Argentina 10.6 million and 11; Chile

Table 5.4. Competition in private pension systems: size of insured market, number of administrative firms, and concentration, December 2005 % of affiliates concentrated in Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagesa

Affiliates (thousands) 10,621 934 7,395 6,362 1,431 1,275 1,280 35,276 3,639 687 6,890

Number of administrative firms 11 2 6 6 8 7 2 16 5 4 6.7

Two biggest firms

Three biggest firms

39 100 55 51 61 60 100 39 57 74 64

53 100 79 68 66 85 100 38 76 86 75

Sources: AIOS (2006), except last column author’s estimates based on BCU (2005a); CONSAR (2005); SAFJP (2005); SAFPb (2005); SBC (2005); SBS (2005); SIPEN (2006); SPa (2005a); SPb (2005); SPVS (2005). a Nonweighted.

94

Reassembling Social Security

7 million and 6; Colombia 6 million and 6; Peru 3.6 million and 5; and Bolivia and El Salvador about 1 million and 2. There are three exceptions: Costa Rica and Dominican Republic had about 1 million insured and 8 and 7 administrators respectively, and Uruguay 687,000 insured and 4 (twice as many as El Salvador with half of the insured). The main explanation for the first two countries is that their systems have been established very recently and statistics show that in all countries the number of administrators increase in the early years, peaks and later declines due to closings and mergers: in Argentina the number fell from a peak of 25 to 11; in Chile from 21 to 6; in Colombia from 10 to 6; in Peru from 8 to 5; in Uruguay from 6 to 4; and in El Salvador from 5 to 2. Another potential reason is that these three countries (together with Colombia) have multiple types of administrators that might facilitate entry and help to maintain the number; in Uruguay the largest administrator is the state bank. In Mexico, administrators declined from 17 to 12 and then increased to 16 in 2005 (17 in 2006), due to policies to facilitate entry like a cut in the capital required, and the superintendence involved in organizing new administrators (Budebo 2006). Because of the small number of Bolivian insured, the government decided that there was a market only for two administrators, divided all the insured by their place of residence between the two, and banned changes until 2000; the superintendence approved the entry of a third administrator in April 2002 and announced a bidding but by mid-2006 there were still only two administrators and a de facto oligopoly (Müller 2003; Aponte 2004; SPVS 2006a). Argentina’s superintendence has asserted ‘Although it cannot be affirmed categorically that the market is an oligopoly, it presents numerous characteristics that constitute the principal features of an oligopoly model’ (SAFJP 2002b). World Bank officials find serious flaws in competition: the ‘oligopolistic industry’ has created perfect conditions for high and increasing concentration, a cause of growing concern; tight

restrictions to switch administrators have fomented collusion, ‘created a captive clientele for each pension fund administrator and institutionalized what was already an oligopoly’; and ‘pension fund industries in Latin America are anything but good examples of competition’ (Gill, Packard, and Yermo 2005: 68, 146, 233, 238). Countries with a small insured market should not carbon copy the structural reforms of big countries, because among other things they will incur in a high risk that competition, the essential foundation of the private system, will not work at all or do it improperly. After Nicaragua reform law was enacted in 2000 there were five foreign corporations interested in entering the market, later the number decreased to three and in 2004 no data was reported; the implementation of the system was suspended that year and the law repealed in 2005. Ecuador will face similar problems if its reform law is ever implemented, as well as the public systems of Guatemala, Haiti, Honduras, Panama, and Paraguay if they decide to adopt a private system. These countries would have to decide whether to have only private administrators, as in half of the existing private systems or multiple administrators (private, public, and mixed) as in the other half, to facilitate entry in the market and more competition. Another important decision is whether administrators should be exclusively devoted to manage pensions and forced to develop a costly national infrastructure. To circumvent this obstacle, countries with a small insured market could authorize administrators to use the existing infrastructure of financial institutions (banks, insurance companies, savings institutions, etc.) or outsource some functions to them, with due caution and separation, in order to reduce installation costs and promote competition.11 The law of the Dominican Republic included that measure and by mid-2005 had seven administrators, although its labor force was similar to that of Bolivia (Suárez 2004). World Bank officials discuss the opening to mutual funds and insurance companies, particularly for the third voluntary pillar (Gill, Packard, and Yermo 2005).

Effects on Unity, State Responsibility, Efficiency Even in those countries having a considerable number of administrators, competition is affected by excessive concentration. Table 5.4 (last two columns) shows concentration of the affiliates in the biggest two and three administrators in 2005. Among the three largest: 100% in Bolivia and El Salvador, 85–86% in Dominican Republic and Uruguay, 76 and 79% in Peru and Chile, 66 and 68% in Costa Rica and Colombia, 53% in Argentina, and 38% in Mexico. The latter has the smallest concentration because the law established a ceiling of 17% of total affiliation to each administrator during the first four years of the private system and 20% since 2001, in addition to the entry of new administrators in 2005–6 and policies to increase competition (see below). An Argentinian decree of 2000 also set a ceiling of 27% but was later abolished (ILO 2002b). Concentration in the ten private systems averaged 75% among the largest three administrators and 64% in the largest two. Chilean concentration steadily grew from 64 to 79% in 1982–2005, and the biggest administrators are largely controlled by foreign corporations (Arenas de Mesa and Mesa-Lago 2006).12 According to World Bank officials the increasing process of concentration will put investment decisions into even fewer hands (Gill, Packard, and Yermo 2005). If the biggest administrators are the ones offering the lowest administrative costs and highest capital returns, concentration should not be a cause of concern except for oligopolies, but this has not always been the case systematically and in the long run as proven in Chile. Three reasons explain why the insured select those administrators despite not being necessarily the best: (a) most insured lack the information and/or skills needed to make a rational decision (see Section 5.5);13 (b) advertisement influences the decision by selling an image of security and solidity but does not provide the insured with data on commissions and capital returns to identify the best administrators; and (c) salesmen collect a commission paid by the administrator every time they change an affiliate,

95

not necessarily serving the insured’s best interests (Mesa-Lago 2001b). Chile’s peak ratio of 1 salesman per 160 active contributors in 1998 drastically fell thereafter because of measures that discouraged such changes;14 Argentina’s ratio of 1 salesman per 131 insured in 2002 increased to 1 per 388 insured in 2004, and Peru’s ratio was 1 salesman per 447 insured in 2005 (SAFPb 2002; SAFJP 2002a, 2005; SBS 2005). Salesmen often entice the insured to change administrator with gifts and some countries have specifically banned that practice. Administrators often compete to steal insured from their rivals instead of making an effort to affiliate noncovered workers. Argentina’s superintendence has accepted that ‘competition through increases in advertisement and marketing appears to be more effective than the reduction in commissions and the raise in capital returns’ (SAFJP 2002b: 20, 2003). Mexico’s recent policies to increase information and reduce commissions led to a 23% decline in 2003–6 in salesmen effectiveness to get affiliates, whereas the effectiveness in the cut of the commission increased 41 times (Budebo 2006). Chile’s legal reform draft of 2006 introduces an annual bidding among administrators for new affiliates, assigned to that guaranteeing the lowest commission for eighteen months (Gobierno 2006).

5.5. Information Barr and Diamond (2006) argue that consumer sovereignty (competition and choice) only enhance welfare where the insured are well informed, but the latter is not the case in practice and the worst-informed insured are disproportionately the least well-off. Furthermore, even if the necessary information is provided, pensions are too complex for most insured to make rational choices. The lack of insured’s information on the private pension systems is appalling and confirms that they don’t select administrators based on their performance and hence there

96

Reassembling Social Security

is little welfare-enhancing. According to an Argentinian survey in 2000, 57% of the insured labeled as insufficient or very insufficient the information they had of the system; 50% had superficially read the periodic report on their individual accounts and 14% did not read it at all; 48% selected the administrator by a salesmen suggestion, decision of the employer or the trade union or a lottery, only 26% made such selection based on capital returns and 15% on the administrative commission charged; 55% of those who reported paying a commission ignored its amount; 47% considered the contribution a tax; and 50% had little or no confidence in the management of such contributions (SEL 2000). Administrators don’t inform the insured of the amount needed in their individual accounts to get a given pension at the time of retirement (MTESS-ILO 2005). In Chile’s survey among private system affiliates in 2002–3, 40% said they didn’t receive the report on their individual accounts and, of the 60% who acknowledged receiving it, only 16% used it to make decisions; 56% didn’t know the sum in their individual account; 93% ignored what the administrative commission was or thought that there was none; 90% lacked knowledge or had incorrect information on how the pension fund was invested; 57% was unaware of the existence of voluntary contributions; 79% didn’t know the requisites to obtain a minimum pension (54% believed poverty was the requisite) and 78% ignored the value of that pension; and 12% of those who shifted to the private system did it because of publicity or salesmen but only 2.6% to get higher capital returns and 0.5% based on knowledge of the recognition bond (EPS 2004). A follow-up survey taken in 2004 contrasted percentages of respondents who said they knew key issues and how many gave the right answer: 51 and 34% on the contribution; 2 and 0% on the commissions; 54 and 36% on the balance in the individual account; 34 and 0% on the requisites of the minimum pension; 49 and 4% on the amount of the minimum pension; and 18–21% and 4–11% on the

requisites and amount of the social assistance pension. Comparisons on the same questions between the two surveys suggest a declining insured knowledge (EPS 2006). In Mexico, in 2001, ‘important groups of the population are still sceptical of pension administrators because they believe, sometimes correctly, that the government or the administrators or the financial system are not trustworthy in general’ (Hernández 2001: 34). A survey of affiliates in 2003 confirmed that point: only 35% of them believed that their funds were secured, 62% said capital returns were low, and only 5% that they were high.15 Since 2003 new policies have significantly improved information among affiliates: programs on the television and radio, a web page, workshops with workers, and pamphlets written in simple language; obligation of administrators to provide standardized comparative data on commissions, capital returns, projected balances, etc.; calculators provided free to insured that allow them to check the amount in their individual accounts and project the pension levels at retirement in different administrators; and information to workers who transfer about which is the administrator charging the lowest commission. These policies seemed to have paid according to a survey of 2005: 74% of affiliates believed that their fund was secure and only 23% considered capital returns low. Conversely, the percentage of respondents who reported knowing the amount of the commission decreased from 93% in 2003 to 71% in 2005, still high but there was no information on whether a question was asked to check if indeed respondents knew how much the commission was (Budebo 2006). Administrators spent huge sums in advertisement but little or nothing on educating the insured. Nevertheless the debate and education of the insured in Chile concerning the selection between multiple investment funds (organized by the association of administrators and the workers union confederation) was important, received substantial resources and is an example to follow. Mexican policies to

Effects on Unity, State Responsibility, Efficiency improve information among the insured seem to have been successful also.

5.6. Efficiency A Chilean study on the overall technical efficiency of the industry in 1982–99 led to the following results: (a) at the start of the first year (1982) pension fund administrators could have provided the same services with 43% of the resources utilized; (b) efficiency improved thereafter and peaked at 79% in 1989; (c) there was a noted decline in efficiency in 1990–4 and in the last year it was at the same level as 1982; (d) efficiency improved in 1995–8 but in the last year was still below the peak of 1989 (albeit higher than in the initial year) and the same services could have been provided with 65% of the resources utilized; and (e) there was significant variation in the period but no continuous trend in the improvement of efficiency (Barrientos 2005). Private management has improved efficiency in aspects such as handling the individual account and providing periodic information to the insured on the state of such accounts (albeit little used and understood). Chile has notably simplified the process of requesting and granting a pension and significantly reduced the time to start payments: a provisional pension is paid in a week after presenting the request and is checked and adjusted in one month. An electronic information network setup in 2004 by the superintendence provides comprehensive data to affiliates including comparison of pension plans and projection of future pensions (ISSA 2004). Colombia has reduced the average processing time to grant a pension from more than a year to a couple of months. On the other hand in Argentina it still takes between eight and twelve months to receive a pension due to the difficult proof of the thirty years of contributions (ILO 2002b); in a survey of 2000, 68% of respondents said that the administrators were doing a poor or mediocre

97

job or didn’t give an opinion (SEL 2000). In the Dominican Republic there were 5,000 pensions under evaluation in 2005, tantamount to 15% of the total number of ongoing pensions in the public system (Pérez Montás 2006). Many private systems have hired fairly well-qualified personnel, in some countries not submitted to legal restrictions on salaries, etc. Systematic data on public systems related to the length of the procedure to obtain a pension and other efficiency indicators are very scarce but the available data suggest that they are poor with some important exceptions. Costa Rica’s public pillar has electronic systems for registration, archives, individual accounts, collection of contributions, and detection of payment delays, but some countries lack these facilities particularly in the latecomer-low group. In Brazil the law gives a period of forty-five days to process a pension and the average time is reported to be thirty-two days; insured records since 1994 are computerized but from 1976 to 1994 must be documented (the computing of such records has begun); affiliates have access to the Internet but only a minority use it to request the pension (Interviews 2004). Uruguay’s public system grants pensions in 30 days providing the needed data are available, otherwise more time is needed; opinion and users’ satisfaction surveys in 2004 reported that 75–76% of insured processing or collecting pensions labeled the service good or very good and 78.5% considered the waiting time as satisfactory (BPS 2005b). Most public systems are besieged by excessive high-turnover personnel, occasionally with low skills, and costly fringe benefits that sometimes include a special scheme for their personnel with more liberal entitlement conditions and benefits than those of the insured in the general program they work for. Many public systems could learn from efficiency improvements in some private systems, particularly in the rapid process of granting the pension (still a major hindrance in most public systems) and keeping up-to-date individual accounts

98

Reassembling Social Security

particularly important in partly funded public systems or pillars.

5.7. Administrative costs The assumption that the private system and competition reduce administrative costs relative to public systems has been theoretically refuted with three arguments: decentralization results in a loss of economies of scale; profit seeking and advertisement are significant costs absent in public systems; and competition only impedes excessive rent but doesn’t ensure low costs because its level is determined by the structure of accounts (Orszag and Stiglitz 2001; Barr 2002). ‘The evidence that administrative costs of individual accounts are higher—often considerably higher—than PAYG schemes is wellestablished’ (Barr and Diamond 2006: 36). Having demonstrated that competition does not exist or function properly in many private systems, the corollary of reduction in administrative costs is left without a premise. Private system’s total administrative cost (total commission) encompasses a net commission and a premium both paid to administrators; the premium is transferred to a commercial insurance company to cover the risks of disability and death (managed by the public system in Colombia, Costa Rica, and Mexico) whereas the net commission is kept by the administrator for managing the old-age scheme, individual accounts, investment of the fund, etc. Comparisons of the net commission between private systems are complex because it is mostly fixed as a percentage of salary but also in a few countries as a fixed sum on the salary or a percentage of the balance in the individual account (assets) or on capital returns; such differences obstruct the standardization of the total net commission into a unified average.16 The total commission is financed only by the insured, except in Colombia (the employer pays 75% and the insured 25%), Dominican Republic (50% each on the net commission), and El

Salvador (entirely paid by the employer since 2006). Table 5.5 exhibits the total administrative cost (net commission plus premium) as percentage of the salary at the end of 2005, standardized by the association of superintendencies of the ten private systems (AIOS). The total administrative cost in Bolivia was 2.21%, the second lowest among all private systems, because there is neither competition nor publicity, hence the net commission was only 0.5%, but the premium was 1.71%, the second highest. Total costs are also low in Dominican Republic (1.6%) because the net commission (0.6%) is similar to Bolivia and the premium is relatively low (1%) but, in addition, 30% is charged on capital returns that exceed the interest of deposit certificates in commercial banks. Total costs are higher in other countries: 2.3% in Chile, 2.6% in Argentina, about 3% in Colombia, El Salvador, Peru, and Uruguay, and almost 4% in Mexico (because the premium is the highest among all countries). In Costa Rica there is no charge on salary but a commission of 8% on gross capital returns, as an incentive to improve them (SPa 2005b). The two components of the total administrative cost have diverse magnitude and trends. The premium fluctuates between 0.76% and 1.71% (excluding Mexico), is the lower component in five countries and for several years exhibited a declining trend but in recent years has been rising in several countries.17 The net commission ranges from 1.23% to 1.99% (except in Bolivia, Costa Rica, and Dominican Republic for the reasons explained) and is the higher component in most private systems; it has oscillated through time but not exhibited a declining trend in most countries. The net commission size largely explains total administrative costs and its reduction is one of the greatest challenges faced by private systems. But the rise in the premium in some countries is a matter of concern; in 2005 the average net commission was 1.38% and the average premium was 1.32% hence virtually closing the previous gap (Table 5.5). Chilean administrators should periodically hold an open bid to select the

Effects on Unity, State Responsibility, Efficiency

99

Table 5.5. Administrative costs in private systems as percentage of wages, December 2005

Systems/ countries Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averageb

Deposit in individual account

Net commission

Premium

Total

4.41 10.00 10.00 10.50 3.97 6.40 10.01 7.16 8.00 12.17 8.74

1.23 0.50 1.54 1.59 0.28c 0.60d 1.71 1.34 1.99 1.85 1.38

1.36 1.71 0.76 1.41

2.59 2.21 2.30 3.00

c

c

1.00 1.28 2.50 0.91 0.98 1.32

1.60 2.99 3.84 2.90 2.83 2.70

Administrative costsa Total deduction

Costs as % of total deduction

7.00 12.21 12.30 13.50 n.a. 8.00 13.00 11.00 10.90 15.00 11.43

37.0 18.1 18.7 22.2 n.a. 20.0 23.0 34.9 26.6 18.9 24.4

Sources: Based on AIOS (2006); averages by author. n.a. = nonavailable. a Net commission charged by the administrator of the old-age program, premium for the insurance company to cover disability and survivor risks, and the sum of both. b Nonweighted average of nine private systems (excludes Costa Rica). c There is no commission on wages but 8% over the gross capital return; the risks of disability and survivors are covered by the public pillar; total is not comparable. d In addition, 30% is charged on the amount that exceeds the average capital return.

insurance company but they usually get the same one (Martínez 2003). Peruvian administrators have contracts with insurance companies and, although legally the insured can choose a different company, in practice they don’t and the lack of competition pushes up the premium (Morón and Carranza 2003). At the end of 2005, the total administrative cost as percentage of the total discount from the salary was: 18–19% in Bolivia, Chile, and Uruguay; 20–23% in Colombia, Dominican Republic, and El Salvador; 27% in Peru; and 35–37% in Argentina and Mexico (Table 5.5, last column). Costa Rican costs cannot be estimated because the commission is not charged on salaries but on capital returns. Nonweighted averages of the nine private systems show that 8.7% was deposited in the individual account, 2.7% was the total administrative cost, 11.4% was the total discount

and 24% was the administrative cost as percentage of total deduction. Because the first generations of pensioners support the bulk of installation costs of the new system, in Chile the cost of the net commission alone took about half of the total amount contributed by a worker who earned the average salary, contributed to the system every year and retired in December 2000 (Gill, Packard, and Yermo 2005: 147–8). Projections in several countries indicate that many insured will not save enough to finance their pensions; if administrative costs were reduced, a higher percentage of salaries could be deposited and help to self-finance pensions and cut fiscal costs on minimum pensions (Uthoff 2002). Some supporters of structural reforms now acknowledge that administrative costs are too high and that competition by itself (where it effectively functions) does

100 Reassembling Social Security not ensure their reduction (Holzmann 2001; Valdés-Prieto 2001). Restrictions imposed on changes of administrators have reduced operation expenses, but such savings usually are not passed to the insured through lower commissions.18 Fixing administrative costs as a percentage of the salary does not generate incentives to cut such costs and only two countries (Bolivia and the Dominican Republic) have set low ceilings on that percentage. The potential advantage of fixing costs as a percentage of capital returns as in Costa Rica and partly in the Dominican Republic have not been assessed. World Bank officials argue that Latin American private pension systems have ‘been generally successful at reducing costs’, although with several important caveats: commissions are still ‘unacceptably high for a large percentage of the population [and] only a small portion of the decline in operating expenses are being passed on to affiliates as lower commissions.’ In Peru, ‘management fees have remained steady even though the ratio of operating costs to fees fell by almost half between 1998 and 2002.’ In Chile, ‘for a worker earning the average wage that retired in 2000 and had contributed each year to the system . . . management fees would have consumed approximately one-half of his total contribution.’ But such officials also assert that ‘in Chile and a few other countries, commission rates are slowly coming down to reasonable levels (less than 20% of contributions or 1% of assets)’ (Gill, Packard, and Yermo 2005: 8, 147, 233). Chilean official data, however, show that the total administrative commission was 2.44% of wages in 1981, peaked at 3.6% in 1984 and declined to 2.37% in 2006, only 0.07 points below the starting rate after twenty-five years of reform.19 On the other hand, Peru’s net commission increased from 2.36% to 2.39% in 1999–2001 but decreased to 1.83% in 2006, a reduction of 0.53 points in seven years, and Mexico’s rose from 1.79% to 2.19% in 1998–2001, but decreased to 1.25% in 2006 a decline of 0.54 points in eight years, reportedly because of improvement of information (AIOS 1999–2006; Budebo 2006). We

have seen that only in three private systems are administrative costs less than 20% of the total deduction (18–19%), while in the remaining private systems such cost ranges from 20 to 37% and average 24%. Commenting on the World Bank officials’ claim that commission rates in several countries are 1% of assets, Barr (2007) considers that too high compared with 0.3% in Sweden.20 Barr and Diamond (2006: 36) add: ‘Under plausible assumptions over a working life, an annual administrative charge of 1% of a person’s pension accumulation will reduce the total accumulation by about 20%.’ There is strong support now in Argentina, Chile, and Uruguay to control administrative costs, and World Bank officials properly set as a priority the need for private systems to cut their costs. Administrative costs of private and public systems in Latin America are compared for the first time in Table 5.6, which contrasts both standardized as percentages of total taxable wages.21 In 2005 private system costs ranged from 0.28% and 0.58% in Costa Rica and Dominican Republic (both underestimated because they exclude the commission on capital returns) to 3.36% in Colombia. The weighted average administrative cost of the ten countries was 1.63%,22 which means that in present value terms over a full career, about one-third of the contributions go to pension administrators rather than to the workers’ pension. The average of 1.63% is slightly higher than the nonweighted average of 1.38% resulting from the net commission as a percentage of wages in the ten countries estimated in Table 5.5. Average administrative costs of the ten private systems in 2005 based on total taxable wages (1.63%) were divided into 1.01% operational expenses (59% for administrative expenses such as salaries and computer costs, and 41% for marketing-like commissions to salesmen and advertisement) and 0.62% profits, the latter accounting for 38% of total administrative costs (author’s estimates based on raw data from AIOS 2005). According to AIOS (2006), the average system profit as percentage of the net commission was 23% in

Effects on Unity, State Responsibility, Efficiency

Table 5.6. Administrative costs as percentage of total taxable wages in private and public pension systems, between 2001 and 2005

Systems/countries

Total taxable wages (million dollars)a

Administrative costs (million dollars)b

Private systemsc Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Totals and averaged

1,306.7 123.5 2,193.1 672.2 388.9 172.0 145.8 6,240.8 834.4 131.2 12,208.6

19.0 1.0 36.3 22.6 1.1 1.0 4.1 97.1 15.2 1.3 198.7

Public Systemse Argentina Brazil Costa Rica Ecuador Guatemala Nicaragua Totals and averaged

27,662.7 114,130,728.9 4,379.5 3,900.0 2,190.5 859.2 114,170,320.8

108.9f 3,577.1 10.0 6.7 0.5 13.7 3,974.1

Percentage (%)

1.45 0.80 1.65 3.36 0.28 0.58 2.81 1.56 1.82 0.99 1.63 0.39 0.003 0.22 0.17 0.02 1.59 0.003

Sources: Private systems total taxable wages (author’s estimates) and net commissions based on AIOS (2005); public systems based on Durán and Cercone (2001); ANSES (2006); CCSS (2005a); INSS (2005a); MPS (2005a); IESS (2006b). a Taxable wages in private systems are author’s estimates based on the average wage and number of active contributors in December 2004, both given in US dollars; taxable wages in public systems are annual (2001 in Guatemala, 2004 in Brazil, Costa Rica, and Ecuador, and 2005 in Nicaragua), converted into US dollars by the author at the official average exchange rate for the year. b Private systems administrative expenses are net commissions (exclude premium for disability and survivors) in December 2004 given in US dollars; public systems administrative expenses converted in US dollars by the author, only for pension program in Argentina (estimate), Brazil, Costa Rica, Ecuador, and Guatemala but for all social-insurance programs in Nicaragua. c Costs of Costa Rica and Dominican Republic are underestimated; in the former they are charged over capital returns, and in the later there are charges both as net commission on wages and on capital returns. d Totals in first two columns and weighted average in the third column. e Brazil, Guatemala, and Nicaragua public systems; Argentina and Costa Rica public pillars; no data available on public systems in Cuba, Haiti, Panama, Paraguay, and Venezuela, as well as on the public parallel systems in Colombia and Peru, and the public pillar in Uruguay. f Estimate of the pension part of administrative expenses, if total expenses are used (including other schemes) the total is US$155 million and 0.56% of total taxable wages.

101

102 Reassembling Social Security 2005–6 but ranged from 32 to 61% in Bolivia, Chile, Colombia, El Salvador, and Uruguay. According to the World Bank, Chilean profits averaged 30% on assets (ROE) in recent years, twice the rate of commercial banks (cited by García 2006). A heated debate on the size of such profits occurred in 2006: two wellknown scholars estimated them as 53% in 1999–2003 on average, whereas a study commissioned by the AFP industry found them equivalent to those in similar financing companies; both sides exchanged accusations of methodological errors.23 In Mexico, profits reportedly decreased 32% in 2003–6 due to policies to improve competition and cut commissions (Budebo 2006). Administrative costs as percentage of total taxable wages estimated in Table 5.6 for four public systems and two public pillars in 2004– 5 were: 0.003% in Brazil, 0.02% in Guatemala (2001), 0.17% in Ecuador, 0.22% in Costa Rica, 0.39% in Argentina, and 1.59% in Nicaragua. The latter was overestimated because it is based on total administrative expenses in all social security programs including sicknessmaternity, the one with highest costs, while in the other five countries percentages are for the pension program alone. The weighted average of the six public systems was 0.003% because of the huge weight of Brazil; if that country is excluded the average increased to 0.48% still less than one-third of the 1.63% average in private systems. The shift from public to private system in El Salvador resulted in a sixfold increase in administrative costs: the average cost in the two major public programs was 0.5% of the salary in 1996 (before the reform) and jumped to 2.99% under the private system in 2005 (Mesa-Lago and Durán 1998; Table 5.5). The above analysis supports the argument that public systems have considerable lower administrative costs than private systems because they don’t pay commissions for salespeople or have advertisement and profits, and they can take more advantage of economies of scale. Most administrative costs in public systems result from excessive personnel and expenditure in salaries and

fringe benefits. Nevertheless for a full comparison of administrative costs between the two types of systems, data are needed from the public systems of Cuba, Haiti, Paraguay, and Venezuela, as well as the public parallel systems of Colombia and Peru, updating figures of Guatemala and on separate administrative expenditures of the pension program in Nicaragua.

5.8. Social participation in the administration Before structural reforms virtually all public pension systems had tripartite administration or representation in their management (workers, employers, and government), although not always effective. The administration of private systems is done by profit-seeking corporations exclusively dedicated to manage pensions. Nevertheless in six countries (Argentina, Colombia, Costa Rica, Mexico, Dominican Republic, and Uruguay) the administration could also be public (social insurance, state banks) or mixed (associations, cooperatives, and family allowances). Uruguay’s major administrator is the state bank that has 37% of total insured and in Costa Rica half of the administrators are public with 80% of the insured (SPa 2003; BCU 2005). Chile’s reform commission recommended the creation of a public administrator: the state bank (Consejo Asesor 2006). Workers and employers finance contributory pensions in private systems (exclusively workers in three countries) whose fundamental objective is to grant benefits to the insured; therefore, it is logical that workers (and employers where they contribute) participate in the administration. And yet, in nine of the ten private systems neither the insured nor employers have any representation in management, supervision, or advising (Table 5.7). An exception is the Dominican Republic’s national council of social security with seventeen members: five from the government, four

Effects on Unity, State Responsibility, Efficiency

103

Table 5.7. Social participation in the administration of private and public pension systems, 2005–6 Systems/countries Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Public systems Brazil Cuba Ecuador Guatemala Honduras Haiti Nicaragua Panama Paraguay Venezuela

Social participation

No; a tripartite National Pension Council stipulated by the law has not been created No No Private: No. Publica : Yes, tripartite Private: No. Publica : Directive Board, 10 members: 4 state, 3 workers, and 3 employers National Social Security Council, 17 members: 5 state, 4 each workers and professionals, 3 employers, and 1 indigents No Private: No, except in the administration of disability and survivor risksa No Private: No. Publica : Directive Board, 7 members: 4 state, 1 each workers, employers, and pensioners National Council, 15 members: 6 state, 3 workers, 3 employers, and 3 pensioners No Publica : Directive Board, 3 members: 1 state (presides), 1 workers, and 1 employers. Private: No, various articles of the law declared unconstitutional Directive Board, 6 members: 2 state, 1 each workers, employers, physicians, and universities Directive Board, 9 members: 3 workers, 3 employers, 2 state, 1 physicians Council, 9 members: 3 workers, 3 employers, and 3 state (not operational in 2000) Publica : Directive Board, 9 members: 3 state, 2 workers, 2 employers, and 2 pensioners. Private: No, law repealed Directive Board, 11 members: 4 workers, 3 employers, 2 state, 1 pensioners, and 1 health professionals Council, 5 members: 3 state, 1 workers, and 1 employers Councils, 7 and 9 members: 4–6 state, 1 workers, 1 employers, and 1 pensioners (not in force yet)

Sources: Legislation. a Public refers to the public parallel system in Colombia; the first public pillar in Costa Rica and Uruguay; the administration by social insurance of the disability/survivors pensions in Mexico; and the still public socialinsurance systems in Ecuador and Nicaragua.

each from workers and professionals, three from employers, and one from indigents, disabled and unemployed, hence the government is in minority (LDSS 2001). Chile’s legal draft of reform of 2006 stipulates the creation of a commission of users with represen-

tation from workers, pensioners, and administrators but limited to expressing opinions and monitoring the reform goals (Gobierno 2006). In Mexico there is no representation in the administration of the old-age program but there is tripartite representation in the

104 Reassembling Social Security disability and survivors branches managed by social insurance. There is also tripartite representation in the public systems or pillars in Colombia, Costa Rica, and Uruguay, but not in Argentina and Peru. The reforms therefore have eliminated social participation at least in the private systems with one exception. The administrators make all decisions concerning the investment of the pension fund without insured participation. An exception is Chile which in 2002 introduced five ‘multifunds’ with various types of instruments that have diverse potential capital returns and risks; the insured can select one of such funds, for instance, a young affiliate may invest more in instruments with higher returns and risks, while an older affiliate closer to retirement may prefer instruments with minimum risks and returns (SAFPb 2002). The Chilean example has influenced changes in three countries: Peru authorized multi-funds in 2003; Costa Rica’s superintendence authorized two investment funds in 2004 based on the profile of the insured; and Mexico introduced two investment funds in 2005, one with variable returns for younger workers and another with fixed returns for older workers (ISSA 2005; SPa 2005b). World Bank officials argue that the option of the insured to select among various investment funds, combined with the state guarantee of a minimum pension, may lead to moral hazard: the low-income insured can invest in instruments with the highest potential capital returns and risks, because in the last instance they will be protected by the minimum pension, thus transferring the risk to the state and increasing fiscal costs. Such problems would be aggravated in Mexico, because of the option that the insured have at the time of retirement, but not in mixed systems (Argentina, Costa Rica, and Uruguay) because they have separated the two pillars and that of mandatory savings does not guarantee a minimum pension (Gill, Packard, and Yermo 2005). Contrary to private systems, all public systems (except Cuba), including most public systems/pillars in parallel and mixed models, have tripartite participation in the major pen-

sion program by representatives of workers, employers and the government, and also from pensioners in three countries. The joined participation of workers and employers is in a majority in Costa Rica (public pillar), Ecuador, Haiti, Nicaragua, Panama and Honduras, as well as in Brazil when combined with representatives from pensioners. Workers and employers representatives are tied with government representatives in Guatemala, and they are in a minority in Paraguay, Venezuela, and Uruguay (public pillar) where the government has the majority24 (Table 5.7). The ILO Conference of 2001 recommended a ‘social dialogue’ among workers, employers, and the government, when debating a pension reform, advice ignored by most structural reforms (Mesa-Lago and Müller 2002). Colombia was one of two countries where the initial reform was widely debated leading to several legal drafts, but subsequent reforms did not follow that healthy practice. Costa Rica’s two reforms were preceded by a social dialogue: prior to the structural reform, a national consensus forum in 1998 with ample participation from all social sectors signed 96 agreements that became the basis for the reform law, whereas the parametric reform of the public pillar in 2005 was debated by a social commission also with representatives from all relevant sectors, supervised by the people’s defender (Martínez and Mesa-Lago 2003; CCSS 2005c). There was no social dialogue in Argentina’s structural reform but the severe economic crisis prompted the creation of a commission for pension reform in 2002, with fourteen representatives from workers, employers, government, pensioners, pension administrators, and experts, which delivered recommendations to the government that influenced the law of reform of 2007 (MTESS 2003; Ley 26,222 2007). In Brazil a council of economic and social development, consisting of 100 representatives from trade unions, employers, scholars, and social groups publicly debated the parametric reform of 2003 (Interviews 2004). Chile’s structural reform was approved by the military government without discussion but President

Effects on Unity, State Responsibility, Efficiency Bachelet appointed in 2006 an advisory council with 15 pension experts from diverse positions (including advisors of trade unions and pension funds) that held 49 hearings with 250 people from 70 relevant organizations, an international seminar and a meeting with foreign experts; the council submitted a report to the government in 2006 that was the basis for the legal draft of reform (Consejo Asesor 2006). These are examples to be followed to ensure that all interested sectors are consulted, give opinions that are taken into account, and become part of the reform. A reform made without social dialogue would lack legitimacy and could be eventually rejected by the population and changed.

5.9. Impact of the reforms on unity, state responsibility, efficiency, costs, social participation, and reform goals 5.9.1. Most systems are still segmented or highly segmented The classification of the degree of unification in the twenty countries shows total unification in two (Bolivia and Panama), relative unification in five (Costa Rica, Cuba, Chile, Nicaragua, and Guatemala), segmentation in four (Dominican Republic, Haiti, Peru, and Uruguay), and high segmentation in nine (Argentina, Brazil, Colombia, Ecuador, El Salvador, Honduras, Mexico, Paraguay, and Venezuela). No clear-cut separation was found between private and public systems concerning their degree of unification or segmentation, but 13 of 20 systems remain either segmented or highly segmented, 7 private and 6 public. The process of unification of multiple pension schemes has been difficult and uneven, and there are successful examples of complete or partial unification in countries with either structural or paramet-

105

ric reforms. Among private systems notable examples are Bolivia that integrated 6 schemes and 28 supplementary funds; Chile that incorporated 30 of 32 schemes, and Costa Rica that integrated 17 of 19 schemes for civil servants; among public systems Cuba unified 51 schemes, and Panama has one unified system. Partial uncompleted processes of unification are those of Argentina, Brazil, Colombia, Dominican Republic, El Salvador, Peru, and Uruguay (all private with one exception); Venezuela’s unification is stipulated by law but not implemented yet. There has not been unification process in five public systems (Ecuador, Guatemala, Haiti, Honduras, and Paraguay) and in one private (Mexico). The armed forces maintains a separate scheme in all private and public systems with three exceptions: Bolivia (albeit with a special regime), Costa Rica, and Panama; civil servants’ schemes continue in 12 countries, 5 in private systems, and 7 in public; and there are separate schemes for numerous groups subsisting in all systems but three of them. All private systems has voluntary supplementary pensions (third pillar), and virtually all public systems also have voluntary supplementary funds for certain groups (Brazil has the largest number). The management of private systems has been decentralized with diverse degree through two to sixteen administrators, whereas, the remaining public systems, pillars, or schemes have centralized administration, as in Argentina, Chile (closed system), Colombia, Costa Rica, Mexico, and Peru. Public systems tend to have a more centralized administration than private systems, Cuba being the extreme case; Brazil is more decentralized because of the segmentation by Union, states and municipalities, and Honduras, Paraguay and Venezuela because of the fragmentation into numerous autonomous schemes. The state administers and finances noncontributory pensions, completely separated from contributory pensions in two private systems (Argentina and Chile), but in Costa Rica and Uruguay they are managed by social insurance and, although legally separated from the contributory program, in practice the latter

106 Reassembling Social Security has made transfers to the former when the state did not fulfill its financial obligations to social assistance. Bolivia’s Bonosol is financed by a separate fund managed by the private pension administrators. In the public systems of Brazil and Cuba the government manages both the main contributory program and social assistance pensions but from different budgets.

5.9.2. The state role continues to be fundamental rather than subsidiary Contrary to reformers’ allegation that the state plays a subsidiary role in private systems, its administrative, regulatory, and financial functions are vital: mandatory affiliation, regulation and supervision, financing of all transition costs, guarantees of minimum pensions, minimum average capital returns and protection in case of bankruptcy of private administrators, and provision of social assistance pensions. Despite the reforms, 66% of total contributors in the region (combining private and public systems) are in public systems and only 34% in private systems. The state fully finances the shut down public systems of Bolivia and Mexico, manages and subsidizes the public system/pillar in Argentina and Uruguay, the open or closed public systems in the other private systems, and handles disability and survivors risks in three countries. In public systems the role of the state in administration and financing is virtually total in Cuba; in Brazil is important in the administration of all public programs that also receive fiscal transfers; but such role is diluted in countries with high segmentation that have numerous autonomous schemes lacking coordination and public supervision (Ecuador, Haiti, Honduras, Paraguay, and Venezuela); actually in these countries the role of the state is similar or even smaller than in most private systems. Protecting the autonomy of socialinsurance institutes is important to reduce state and political interference, but some of

them enjoy an unwarranted degree of autonomy and autarchy that makes their control or supervision very difficult (e.g. Ecuador). In all private systems there is a public superintendence (in most only dealing with pensions, in a few with wider scope), with divergent degrees of autonomy, that regulates and oversees the private system; it is financed by the state in five systems, by the fund administrators in four and by the insured and employers in one. There are virtually no public systems with a sole superintendence of pensions, a serious problem in those highly segmented; Costa Rica is unique in having a single superintendence for the two pillars, in other countries the overseeing functions are divided into various state agencies often generating confusion, conflicts of jurisdiction, overlapping, and lacunae.

5.9.3. Reform goals: freedom of choice and privatization Reformers allege that the goodness of the private system combined with freedom of choice promotes the shift of insured from the public system, and indeed an average 84% of total contributors in the ten countries with structural reforms (86–100% in seven) are in the private system/pillar and only 16% in public systems/pillars. But the shift of insured has been promoted by causes other than the goodness of the private system: the law forced the shift of all insured in three countries, cut the contribution in the private system (or increased the contribution in the public system) or introduced incentives for the transfer in another three, divided the insured by age and obliged the younger to move in six countries, and obliged all new entrants in the labor force to join the private system in all countries except in Argentina and Colombia (in the former, however, those who don’t make a selection were assigned to the private system until 2007). Employers have forced or influenced the shift at least in Chile and Peru. Reformers’ promises, enhanced by advertisement, prompted the move also: the private

Effects on Unity, State Responsibility, Efficiency system would pay better pensions, cut administrative costs and insulate the pension fund from state interference, but such promises generally have not materialized. Conversely, Uruguay’s public system keeps 63% of the insured because those above age 40 could choose to stay and most of them did, and lowincome insured normally cannot join the private system. Colombia’s proportion of insured in the public system was higher than in the private one until 2003 because the former paid better pensions and there was freedom to change every three years, but the 2002 reforms tightened the public system pension formula and increased the period for change to five years, still 47% of total insured were in the public system in 2004. Insured in private systems have freedom to change administrators more than once a year in Mexico, once a year in four countries, but only every two years in five other countries. In Bolivia there was no freedom to change between the two administrators until 2000; in other countries, if the insured don’t make a decision, the state, the employer, or a lottery assign them to a specific administrator. The freedom to move is costly for the administrators and has led to restrictions in most countries, significantly reducing the percentage of shifts and fomenting collusion in the industry.

5.9.4. Reform goal: competition does not work properly or does not exist Competition is the foundation of private systems but in order for it to function properly, an adequate number of administrators must exist, which largely depends on the size of the insured market. Countries with the largest number of insured have the highest number of administrators (Mexico and Argentina), while those with the smallest number of insured have only two administrators (Bolivia and El Salvador), an exception seems to be countries that have multiple types of administra-

107

tors (private, public, and mixed) that might facilitate entry. Bolivia has a virtual monopoly because the state divided the insured by their place of residence between the two administrators; changes were banned until 2000 but only 0.4% of the insured shifted in 2005, and still there were only two administrators in 2006. Initially the number of administrators increases but later they merge and their number declines. Even countries with an appropriate number of administrators have a high and increasing concentration in the biggest three (averaging 75% in the ten countries). Such concentration is usually not the result of an informed decision by the insured based on best performance by administrators (low commissions and high capital returns), but of other causes such as advertisement and marketing without providing useful data, and the work of salesmen who are paid a commission for each worker they move. The World Bank acknowledges that there is not real competition as the industry is an oligopoly with a captive clientele and rising concentration.

5.9.5. The insured lack information Choice plus competition only enhance welfare where the insured are well informed. There is ample evidence from surveys in Argentina, Chile, and Mexico that the insured lack knowledge or have incorrect information on key elements of the private system (Chilean surveys suggest declining knowledge), that they don’t select administrators based on performance (commissions and capital returns) but due to influence from salesmen or employers, and that they don’t have the skills to select the best administrators. Administrators spent huge sums in advertisement but little or nothing on educating the insured. Recent policies in Mexico, however, appear to have improved insured information on some issues, whereas Chile’s educational campaign for selection of investment funds seemed to be successful.

108 Reassembling Social Security

5.9.6. Uneven gains in efficiency A study on overall technical efficiency in Chile’s pension fund industry in 1982–98 showed significant variation but without a continuous improvement in efficiency and that the same services could have been provided with 65% of the resources utilized. Most private systems have improved efficiency in providing periodic information to the insured on their individual accounts (albeit little understood and used), some have also developed a simpler and more expedite process of requesting and granting pensions (Chile and to a less extent Colombia) but not others (Argentina and Dominican Republic). Many private systems have also hired fairly wellqualified personnel. Most public systems lack such advances but Brazil’s and Costa Rica’s public systems/pillar have notably improved electronic techniques and Brazil’s average time to grant a pension is shorter than in Argentina and Colombia. Some public systems (particularly in the latecomer-low group) lack electronic techniques for enrolments, archives, individual accounts and collection of contributions, and many of them are besieged by excessive personnel, sometimes with low skills.

5.9.7. Higher administrative costs in private than in public systems Reformers criticize the virtual monopoly in public systems for its lack of incentives to reduce administrative costs and argue that competition in private systems would reduce such costs. It was demonstrated that competition does not exist or function properly in most private systems and that they have costs that do not exist in public systems: substantial expenses on advertisement, marketing and salesmen, earning of profits and loss of economies of scale. Total administrative costs as percentage of salary range from 2% in Bolivia (the net commission is very low due to lack of competition) and 4% in Mexico, for an average of 2.7% in the ten private

systems. Administrative costs as percentage of salary jumped sixfold after the structural reform in El Salvador. The insured pay all those costs, except in Colombia, Dominican Republic, and El Salvador where they are shared with the employer. As a percentage of the total deduction (deposit in the individual account, net commission for the oldage program and premium for disability survivors) administrative costs took from 18% in Bolivia to 37% in Argentina and averaged 24% in the ten countries in 2005, which reduces the sum deposited in the individual account and the eventual pension. A Chilean worker earning the average wage and contributing each year loses about half of his total contribution because of administrative costs. The net commission is the major component of the administrative cost and has not significantly declined in most private systems (in Chile it only diminished 0.07 points in 1981–2006); the premium is the smaller component and fell for several years but it is increasing in some countries and by 2006 it had closed the gap with the average net commission. Reductions in operational costs resulting from restrictions imposed on changes between administrators have not been transferred to the insured as lower commissions. Fixing administrative costs as percentage of salaries does not generate incentives to cut such costs and only two countries have set low ceilings on those percentages. Comparisons of administrative costs as percentage of assets show that they are higher in Latin America than in other countries and could reduce accumulation in individual accounts by 20%. The comparison of administrative costs in private and public systems, standardized as a percentage of total taxable wages, resulted in a weighted average of 1.63% in the ten private systems versus a weighted average of 0.003% in six public systems or pillars (0.48% excluding Brazil). Profits in private systems averaged either 23% of commission revenue or 38% of administrative cost. Public systems’ considerably lower administrative costs are explained by their lack of profits, commissions to salespeople, and advertisement, as well as taking

Effects on Unity, State Responsibility, Efficiency better advantage of economies of scales. Most administrative costs in public systems result from excessive personnel and expenditure in salaries and fringe benefits. More data from public systems are needed to strengthen the comparison.

5.9.8. Absence of social participation in private systems Before structural reforms virtually all public pension systems in the region had tripartite administration or representation in their management (workers, employers, and government), although it was not always effective. The reforms totally transferred the management to profit-seeking private enterprises albeit the administration is multiple in several countries (public administrators have most affiliates in Costa Rica and Uruguay). In Colombia, Costa Rica, and Uruguay, the tripartite representation has been kept in the public system/pillar, but there is none in the private system/pillar. The administrators manage the old-age program and commercial insurance companies are in charge of disability and survivors’ risks (except in Colombia, Costa Rica, and Mexico where they are managed by social

109

insurance). There is no participation of any kind by workers, employers and pensioners in the regulation, management, supervision and provision of pensions by fund administrators, insurance companies and superintendencies, except in Dominican Republic. The insured can neither make any decision nor have an input on the investment of its own fund, except recently in Chile and to a lesser extent in Costa Rica, Mexico, and Peru where the insured can choose between two and five investment funds. Conversely, in all public systems (except in Cuba) there is tripartite representation from workers, employers and the government in the administration (also from pensioners in three countries). Joined representatives from workers and employers have a majority in the councils of Costa Rica (public pillar), Ecuador, Haiti, Nicaragua, Panama, and Honduras (as well as in Brazil when pensioners are included); they have a tie with government representatives in Guatemala, and are in a minority in Paraguay and Venezuela. Structural reforms were approved without a social dialogue except in Colombia and Costa Rica. Recently, Argentina and Chile appointed commissions and opened a public debate on pension reforms that have influenced legal drafts; Brazil and Costa Rica parametric reforms also had a social dialogue.

Notes 1. The integrated system would have four administrative units, none of them created in 2006: the direction in charge of an unidentified ministry; the social security treasury responsible for collection of contributions and investment of pension funds; the national institute to manage all pensions and monetary benefits; and the superintendence of social security to regulate and oversee the entire system (Villasmil 2006; González 2007). 2. The World Bank (2005e) properly notes that integration of the administration of contributory and noncontributory pensions has the advantages of economies of scale and crosscontrol to avoid fraud, but the risks of political manipulation and lack of transparency, omitting financial risks. 3. The degree of privatization herein is based only on contribution to a public or private system; other variables used are the model of reform, the type of administrator, the contribution of the employer, and state guarantees (Mesa-Lago and Müller 2002). 4. Costa Rica is unique because all insured (at the time of the reform and in the future) are obliged to join the mixed system, hence all insured are in both the first public pillar (that pays the bulk of the pension) and in the second private pillar (that pays a supplementary pension).

110 Reassembling Social Security 5. Argentina reform law of 2007 allows changes between the two systems every five years (Ley 26,222 2007). 6. Kleinjans (2003, 2004) who proves that the private system is more attractive for youngestand highest-income insured, while the older prefer the public system because it pays a higher pension. 7. The Colombian government intends to terminate the social-insurance pension program (ISS) with the state guaranteeing payment of its pensions (El País, Bogotá, August 28, 2006). 8. Argentina’ reform law of 2007 changes that practice by an automatic affiliation to the public system of those who do not make a decision, also applied to affiliates with insufficient contributions to get the minimum pension in the private system (Ley 26,222 2007). 9. The original title of pension fund administrators was Administradoras de Fondos de Pensiones (AFP) in Chile, later copied by Bolivia, Dominican Republic, and El Salvador; other names are: Administradoras de Fondos de Jubilaciones y Pensiones (AFJP) in Argentina, Administradoras de Fondos de Ahorro Previsional (AFAP) in Uruguay, Administradoras de Fondos de Retiro (AFORES) in Mexico, Sociedades Administradoras de Fondos de Pensiones (SAFP) in Colombia, and Operadoras de Pensiones Complementarias (OPC) in Costa Rica (Table 5.3). 10. Müller (2006) gives another explanation: countries where insured income is higher tend to have more administrators than countries where income is lower; the former is true in Argentina, Costa Rica, and Mexico (but not in Chile and Uruguay) while the latter is true in Bolivia and El Salvador (but not in Dominican Republic). 11. Chile’s reform commission recommended opening the system to banks and insurance companies to facilitate entry, and to subcontract collection of contributions, administration of individual accounts and branches, ending the AFP exclusive dedication (Consejo Asesor 2006). 12. An alternative to measure concentration is the Herfindahl-Hirschman Index but it has important flaws (see Morón and Carranza 2003). 13. Even consumers in the United States and the United Kingdom are ignorant about financial markets and concept of risk, aggravated by scandals in capital markets (Barr 2002). 14. Barrientos (2005) questions whether a high level of transfers is an indicator of healthy competition: in Chile commissions and rates of return had only a small impact on transfers while the dominant factor was the number of salesmen. 15. A Peruvian survey in 2001 reported that 56% of the insured did not trust the private system and the percentage rose to 74% according to socioeconomic characteristics (Morón and Carranza 2003). Lack of the system transparency in Bolivia has led to loss of trust among people (Crabbe and Giral 2005). 16. Morón and Carranza (2003) discuss comparability problems and a more standardized method to combine various net commissions. 17. In Argentina, the premium fell from 2% to 0.5% in 2001–2 but rose to 1.4% in 2005. The reform law of 2007 standardizes the premium in the system charging it to administrators based on the industry average capital returns (AIOS 2001—5; Ley 26,222 2007). 18. Until 1990 Chile’s net commission and premium were clustered and administrators pocketed savings from the reduction in the premium instead of passing them to the insured. 19. Chile’s average total commission rose 4.8% in 1982–2003 (Arenas de Mesa and Mesa-Lago 2006) and operational costs per active contributor steadily rose in 1982–98 (Barrientos 2005). 20. The US pension fund for university and college professors (TIAA-CREF) charged 0.65% of net assets for all administrative expenses including annuities in 2005. 21. Taxable wages in private systems were estimated by multiplying the average number of active contributors times the average taxable salary in December 2004; administrative costs are equal to the annual net commission (excluding the disability-survivors premium) divided by the twelve months (raw data from AIOS 2005). 22. Similar estimates of administrative costs in eight Latin American private systems in 1999 resulted in a nonweighted average of 1.48% (James, Smalhout, and Vittas 2001).

Effects on Unity, State Responsibility, Efficiency

111

23. The first study was conducted by S. Valdés Prieto and I. Marinovic, and the second by C. Bonilla and C. Maquieira (‘Ganancias de las AFP Abren Debate Técnico,’ El Mercurio, May 16, 2006). 24. Venezuela’s reform law of 2002 (not enforced by the end of 2006) stipulates that the social security treasury and the national pension institute have councils integrated by nine and seven members respectively, with one representative each from workers, employers and pensioners, and a government majority, but leaves for future legislation the determination of the ways of participation.

6

Effects on Financial Sustainability and Reform Goals

Structural reform supporters allege that the shift from a public to a private system make pensions financially sustainable, improve punctual payments of contributions, reduce and eventually eliminate pension deficit and fiscal costs, cope with population ageing, promote national savings, capital markets and returns, and protect from state and political interference. Such beneficial effects are based on the change from PAYG to full funding and from defined benefit to defined contribution, as well as capital accumulation, efficient investment, and private management. This chapter tests those claims: compares contributions in both systems; evaluates the alleged improvement in compliance; examines the components of fiscal costs during the transition and beyond; contrasts initial projections of fiscal costs in private systems with reality and recent projections; evaluates financial-actuarial equilibrium in public systems; and tests if private systems indeed insulate pensioners from population ageing and against state-political interference, while promoting national saving, capital markets, portfolio diversification, and capital returns.

6.1. Contributions Before structural reforms, public systems were financed with tripartite contributions and the employer’s averaged two thirds of the total, thus meeting the ILO minimum norm (Convention 102) that the worker should not finance more than 50% of the total contribu-

tion. With few exceptions, structural reforms violated this norm. Three countries eliminated the employer contribution (Bolivia, Chile, and Peru) and two reduced it: Uruguay cut it slightly while increasing the proportion of the worker’s, and Argentina cut it by half through exemptions (the worker’s contribution was also halved but since 2003 is gradually being increased). Five private systems raised the worker’s contribution: Bolivia, Colombia, Dominican Republic, El Salvador, and Peru. Only Colombia and Dominican Republic augmented the employer’s contribution. Two countries did not legally or in practice change the contribution of the worker or the employer: Costa Rica reassigned previous contributions from other programs and Mexico raised the state contribution on the wage bill. The elimination or reduction of the employer’s contribution has provoked an increase in the worker’s contribution or fiscal costs or both. Workers also pay commissions for the administration of their pension funds, except in Colombia, Dominican Republic (law not yet enforced), and El Salvador, where they are shared by the employer. An average of 56% of the current total contribution in private systems is paid by workers and 44% by employers. In Bolivia, Chile, and Peru the worker pays 100% of the total contribution infringing the ILO minimum norm; also true in Uruguay where the worker contributes 55%; the remaining countries meet the ILO norm. In Costa Rica and Mexico, the state contributes a percentage of the wage bill to the private system (Table 6.1).1

Financial Sustainability and Reform Goals

113

Table 6.1. Contributionsa by employers, workers, and state in private and public pensions systems, c.2005 (% of wages or income) Systems/Countries

Employers

Workers

Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagec

10.8–16.00h 0 0 9.75 6.50 5.00 7.00 6.20 0 12.50–16.50h 6.24

7.00–11.00 12.21 12.27 3.25 5.00j 2.00j 6.00 2.59 11.19 15.00 7.85

Public systems Brazil Cuba Ecuadord Guatemala Haiti Honduras Nicaragua Panamae Paraguayf Venezuelag Averagec

20.00i 12.00 1.10–3.10 3.67 2.00–6.00 2.00 6.00 2.75 7.60 5.91 6.60

8.00–11.00i 0k 6.00–8.00 1.83 2.00–6.00 1.00 4.00 6.75 4.90 0.84 3.98

Stateb

0.25

2.21

1.83 0.50

1.50 1.50

Total

Self-employed

17.80–27.00 12.21 12.27 13.00l 11.75 7.00m 13.00 11.00 11.19 27.50–31.50 14.09

11.00o 12.21 12.27 13.00 4.75p 2.00q 13.00 8.79 11.19 27.50 11.57

28.00–31.00 12.00 9.10–11.10 5.50 4.00–12.00 3.50 10.00 9.50 12.50 6.75n 10.58

20.00 12.00 9.10 5.50 4.00–12.00 3.50 10.00 9.50 r

6.75n 9.37

Sources: Legislation; US-SSA (2004); AIOS (2005); MTESS-ILO (2005). a Private systems include deposit in individual account, commission and premium; mixed systems include contributions to both pillars. b Only third-party contributions as percentage of wage bill, paid in Costa Rica and Mexico, not usually paid in Guatemala, Honduras, Paraguay, and Venezuela (and excluded from the total); the state should pay a lump sum in Panama but usually does not; in Ecuador the state must finance 40% of pension costs but only pays part of it; financing of deficit in various countries excluded. c Nonweighted average; total excludes state contributions. d Based on 2005 regulations, contributions of the 2001 reform law were not in force in mid-2006. e The reforms of 2005–6 gradually increase contributions until 2013: employer to 4.25%, worker to 9.25%, total to 13.5%, state annual lump sums, and self-employed to 13.5%. f In general program (IPS); different contributions in other schemes. g The 2002 law did not set contributions and maintained the previous ones, the pending reform law increases contributions: employer to 10.5%, worker to 3.5%, and total to 14%. h Argentina’s employer in private sector average 10.8% and state employer 16%; Uruguay’s employer private sector 12.50% and state employer 16%. i Average employer’s contribution 20%; worker’s average 8%.

114 Reassembling Social Security None of the ten public systems have eliminated or reduced the employer contribution; an average of 38% of the total contribution is paid by workers and 62% by employers, hence complying with the ILO minimum norm. In seven countries the employer’s contribution is between 1.5 to 7 times the worker’s (a minority of them contributes in Cuba); exceptions are Ecuador and Panama where the worker contribution is twice that of the employer (contravening the ILO norm) and Haiti where the two contributions are the same. Separate schemes enjoy better financial terms than in the private system: the armed forces usually contribute a nominal amount if at all (Chile, Mexico, and Peru) or their contributions are smaller than in the general program (Uruguay); civil servants pay a lower contribution (Peru closed scheme); oil workers often don’t pay as their schemes are entirely state financed (Colombia and Mexico) and the same is true of congressmen (Argentina and Colombia); and banking employees contribute less than in the general program (Uruguay). Mexico’s federal civil servants scheme (ISSSTE) has a higher contribution but still insufficient to finance the generous benefits and the resulting deficit covered by the state equaled 71% of the scheme expenditures in 2004; the same happens in the scheme of employees of the principal program (IMSS) whose deficit equaled 93% of the scheme expenditures, financed by contributions from insured and employers to IMSS.

In most public systems the state is legally obliged as third party to pay a contribution on the wage bill or finance part of pensions in the general program but in seven countries doesn’t honor such obligation or does it irregularly, incurring substantial debts (see Section 6.2); conversely, the government fulfils its financial obligations with privileged schemes and covers their deficit. Brazil federal civil servants contribute 2–5 percentage points less than insured in RGPS and the state subsidy to cover the deficit in the two programs is five to one. Cuban schemes for armed forces and internal security are entirely state-financed. Guatemalan civil servants and armed forces contribute more than in the general program but their generous entitlement conditions and benefits provoke a deficit financed by the state. Honduras state contributions as employer to separate schemes are six to nine times higher than to the general program, and the legal draft of reform makes the state responsible for financing pensions of such schemes once their reserves are depleted. Venezuelan civil servants contributions are lower than those in the general program and the state finances the deficit. Panama contributions are the same for all the insured, a unique case among public systems. With some exceptions, the total contribution on the wage bill is higher in countries of the pioneer-high group (due to matured programs, aged populations, and

← j In Costa Rica excludes the commission, not charged on wages but on capital returns (there is no premium); the Dominican Republic charges an additional commission on capital returns. k Workers contribute 5% in only 15% of state enterprises. l Total contribution to rise to 15.5% in 2006 divided in the same proportions. m Total contribution should gradually rise until 2008. n The 2002 law raises the total contribution to 14%; the state may finance part of the contribution of lowincome self-employed. o ‘Autónomos’ based on presumed income; ‘monotributistas’ pay fix sums. p Public pillar, the state contributes 0.25% and 7.25% is divided between the worker and the state; selfemployed are excluded from the private pillar. q In the contributory-subsidized regime not enforced in 2006; the state subsidizes the contribution of lowincome self-employed. r Fixed sum according to income.

Financial Sustainability and Reform Goals low active workers/pensioner ratios): 18–31% in Argentina, Brazil, and Uruguay, but 12% in Chile, Costa Rica, and Cuba explainable respectively by no employer’s contribution, no worker’s contribution and relatively young program-population. Contributions in the intermediate-group oscillate from 10–13% except 7% in Venezuela despite its generous entitlement conditions. Contributions are the lowest in the latecomer-low group (because they have younger programs and populations and higher active workers/pensioner ratios): 3.5–10% in Dominican Republic, Guatemala, Haiti, Honduras, and Nicaragua, but 13% in El Salvador and Paraguay, a heavy burden in these two countries. The nonweighted average total contribution in private systems is 14% compared with 10.6% in public systems (Table 6.1). Within private systems, the smallest contributions to the savings pillar are: 4.25% by the worker in Costa Rica (but excluding part of the net commission and the entire premium), 7% by the worker in Argentina (reduced from 11% due to the crisis and gradually increasing again), and 7.5% in Dominican Republic (2.5% worker and 5% employer) but excluding the additional charge on capital returns. In the other seven private systems the worker contribution stretches from 11% in Mexico to 15% in Uruguay. Mixed systems have an additional contribution to the public pillar: 10–16% by the employer in Argentina; 7.5% in Costa Rica (4.75% employer and 2.5% worker), and 12.5% by the employer in Uruguay (Table 6.1). In public systems, contributions of the two pioneer-high countries are quite diverse. Brazil’s general program (RGPS) total contribution ranges from 28–31%, the highest in the region (8–11% workers and 20% employers); the state subsidizes the deficit and contributes to the rural program (largely financed with a tax on marketing of rural products paid by buyers). Cuba’s total contribution is low: 12% and paid by employers; salaried workers didn’t pay for forty years and began to contribute 5% in 1999 but only when employed in state enterprises under the ‘perfecting system’, which accounted for only 15% of total

115

enterprises and 2% of total workers in 2003; cooperatives members pay 5% of revenue. In the three public systems of the intermediate group, the total contribution to the general program is lower than in the pioneer-high group but higher than in the latecomer-low group: Ecuador 9–11% depending on whether it is the private or public sector (6–8% worker and 1–3% employer);2 Panama 9.5% (6.75% worker and 2.75% employer); and Venezuela 6.57% (0.84% worker and 5.9% employer), the third lowest in the region despite its generous entitlement conditions, when the 2002 reform law is implemented contributions will rise more than twice. With one exception, latecomer-low countries have the lowest contributions in the general program in the region: Honduras 3.5% (1% worker and 2% employer); Guatemala 5.5% (1.83% worker and 3.67% employer) and Haiti 4–12% based on four income brackets (equally divided between worker and employer). But Paraguay has 12.5% (4.9% worker and 7.6% employer) the highest total contribution in public systems after Brazil, despite its strict entitlement conditions. Self-employed contributions usually combine the percentage of salaried workers plus the employer that they lack, hence the self-employed percentage contribution is between twofold and eightfold the salaried worker contribution in twelve countries: Argentina,3 Brazil, Colombia, Ecuador, El Salvador, Guatemala, Haiti, Honduras, Mexico, Nicaragua, Uruguay, and Venezuela (the law of 2002 stipulates that the state may finance half the contribution of low-income workers including the self-employed). Cuban selfemployed contribute 12% of their declared income whereas 85% of wage earners don’t contribute; Panama’s ongoing reform gradually increases the self-employed contribution from 9.5–13.5%, making even more difficult their affiliation. In the three private systems that eliminated the employer contribution (Bolivia, Chile, and Peru) the contribution is the same for salaried and selfemployed workers. Costa Rican self-employed

116 Reassembling Social Security

Table 6.2. Affiliates in private pension systems that contributed in the past month, December 1998 to December 2005, and June 2006 (in percentages) Private systems

1998

1999

2000

2001

2002

2003

2004

2005

2006

Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagesa

48.9

44.3

39.1

b

b

b

52.8 n.a. N.O. N.O. 67.2 63.4f 45.6 67.4 57.9

53.4 51.6c N.O. N.O. 63.7 60.2f 45.7 58.7 55.5

50.9 48.5c N.O. N.O. 55.2 57.9f 41.7 53.9 51.0

29.0 47.0 53.7 48.7c n.a.d N.O. 53.2 44.7f 41.2 53.2 43.5

33.2 46.9 51.0 47.6c n.a.d N.O. 47.6 41.7f 39.4 45.1 42.1

35.2 39.0 51.9 48.7c 73.1 65.5e 46.3 39.3f 41.9 52.7 42.3

35.4 44.9 50.4 39.0c 68.1 53.5e 41.9 38.8f 39.9 52.5 41.6

40.6 47.5 51.2 50.6c 65.3 47.2 39.5 37.6f 36.2 55.7 41.7

39.6 47.1 50.7 51.6c 65.1 52.6 38.3 37.4f 37.5 61.6 41.8

Sources: Author’s series based on AIOS (2000–6); and Mesa-Lago (2006b). n.a. = nonavailable. N.O.= not in operation yet. a Based on total affiliates and total contributors in all countries with available data. b Before 2001 a contributor was an affiliate who had at least one contribution since the start of the system. c In 1999–2003 a contributor was an affiliate who had at least one contribution in the past six months; data for 2004 was originally published as a contribution in the last month (as shown in the Table) but revised to 49.5% in 2005 suggesting that there was a return to the 1999–2003 estimation. d In 2001–2 a contributor was an affiliate who had at least one contribution in the past year. e Original data were given as 58.5% in 2003 and 49.8% in 2004 and increased as shown in the Table. f A contributor is an affiliate who had a contribution in the past two months.

percentage contribution is smaller than the salaried worker’s and the state adds a subsidy, similarly to that in Dominican Republic under the contributory-subsidized regime pending implementation. As an average, selfemployed contribute 11.6% in private systems versus 7.8% salaried workers and 9.4% in public systems versus 4.0% salaried workers (Table 6.1). Although the percentage paid by the self-employed is usually applied on the minimum wage, still such burden is substantially heavier than among salaried workers, because the self-employed earn less than the minimum wage in many countries, although professional self-employed have fair to high income. The heavy load on low-income selfemployed is an insurmountable barrier to coverage unless the employer contribution is matched by the state or subsidized with fiscal transfers.

6.2. Compliance Structural reformers assert that ownership of the individual account and the principle of equivalence in private systems are strong incentives for the worker to affiliate and punctually pay contributions in order to get higher accumulation in the individual account, capital return, and pension.4 Table 6.2 contests the alleged incentives of the private system to improve compliance: the percentage of affiliates that actively contributed (in the past month) in the ten private systems declined in all of them in each year in 1998–2006, with few exceptions (for comparability see Table notes). In 2006, only between 37 and 53% of the affiliates were active contributors in all countries except in Costa Rica and Uruguay. The weighted average of active contributors in the ten private systems steadily decreased

Financial Sustainability and Reform Goals from 58% to 42% in 1998–2004 and virtually stagnated in 2005–6. Despite twenty-five years of reform and the most successful private system, Chile compliance fell from 76 to 51% in 1983–2006 (SAFPb 1983; AIOS 2006). World Bank officials now qualify previous reform assumptions: ‘despite the time that has passed since reforms in several countries, the expected improvement in incentives attributable to the introduction of individual retirement accounts has not been rigorously tested,’ Chile has ‘falling numbers of active contributors in the labor force’ and Latin America endures ‘low density of contributions to the mandatory pillar.’ Surveys in Santiago de Chile and Lima (instead of the national level) led Bank officials to infer that contribution density is higher after the reforms, but these results ‘cannot be used as evidence that a transition to private individual . . . accounts is the only way to improve incentives and achieve greater [compliance]’, because a similar improvement may arise from linking contributions and pension levels within PAYG as in Brazil. The officials also found that Chilean affiliates tend to contribute just enough to qualify for the minimum pension (and the state subsidy) and thereafter usually stop contributing to the private system and rationally give priority to less risky and costly alternatives like housing, household enterprise, and education; similar results were found in Peru (Gill, Packard, and Yermo 2005: 4–6, 101, 166, 184, 188, 194). The survey conclusion of higher contribution density after the Chilean reform is contradicted by two panel surveys taken among the same group of affiliates in that country: the contribution density declined from 52% in 1980–2002 to 48% in 2000–4 and 45% of respondents did not contribute in 2000–4, 38% men and 53% women (Bravo et al. 2006). ‘Rational’ persons do not join the private system despite its alleged incentives for several reasons: (a) low-income entrepreneurs have a high opportunity cost to invest scarce capital in insuring themselves and their employees; (b) small employers, farmers, and selfemployed cannot withdraw funds from their

117

individual accounts to cope with hardships before retirement; (c) low-income workers judge immediate consumption more important than differed consumption, especially because they suffer a higher mortality; (d) workers (regardless of their income) perceive a high political risk of government unfulfilling its promises and guarantees; (e) workers in general endure a higher risk in a private system than in a public one, because of the uncertainty on the amount of the pension and capital returns; and ( f ) the strategy of mixing mechanisms of formal protection (pensions, investment in housing, etc.) and informal protection (traditional family help) may generate higher returns and less risks than in a simple pension system (Gill, Packard, and Yermo 2005: 168–170; see also Section 3.1). Some experts maintained for years that the high level and increment in the worker contribution created incentives for noncompliance stronger than the alleged incentive of the private system to affiliate and contribute. Right after the approval of El Salvador reform, the ILO (1997) warned that with the increase in contributions by almost fourfold, it will be more difficult to avoid the generalized phenomenon of evasion and under-declaration of salary despite the incentives of the new system and sanctions. The author added that the jump in contributions could also expand the informal sector and provoke a fall in collected revenue and coverage in private systems, as have indeed occurred (Mesa-Lago and Durán 1998; Mesa-Lago 2001b). World Bank officials now adhere to this argument acknowledging that individuals easily avoid the mandate to affiliate and contribute: ‘It is likely that the extent of evasion is partly linked to the level of mandated contributions . . . if the mandatory contribution rates were decreased and the maximum taxable wage reduced, the contribution frequency is likely to increase . . . especially among poorer and younger workers who . . . have many consumption competing demands’ (Gill, Packard, and Yermo 2005: 231). Another important cause of noncompliance overlooked by the World Bank is that many

118 Reassembling Social Security employers deduct workers’ contributions but don’t transfer them with their own or delay such transfers to the pension administrator or central collecting agency. This legal transgression is sanctioned in most countries with the payment of such contributions adjusted to inflation, plus interests and fines, and even the shut down of the enterprise. The failure to make the transfer and deposit in the individual account should appear in the periodic report that administrators provide to insured, but surveys show that workers don’t examine or don’t understand this report, and probably ignore the right to denounce the transgression and sue the employer. Administrators are entitled to request payment from employers, collect an additional charge as incentive and appeal to the courts if they are not successful, but the judicial process is frequently cumbersome, costly, and prolonged. The Chilean debt for payment delays alone jumped sevenfold in 1990–2002, reaching 1% of the total value of the pension fund and 43% was uncollectible due to enterprise bankruptcy. New policies introduced in 2004–5 annul dismissals of workers whose employers have retained their contributions, expedite the procedure for debt collection, and create ad hoc courts to handle claims; tougher sanctions against noncompliance have been proposed (SAFPb 2002b; ISSA 2005; Programa 2005). In Argentina 40% of employers evaded in 2003 (Conte-Grand 2004), and Colombia’s 36% employer evasion was a key reason to introduce a centralized collection system in 2005 (ISSA 2006). Uruguay’s social insurance institute (BPS) didn’t comply until 2006 with the legal mandate to annually collect the employers’ debt from noncompliance (95% of workers who denounced the transgression was fired) largely because the BPS must go to regular courts, hence special courts for collection of contributions are planned (Murro 2006).5 Costa Rican employer payment delays to the public pillar were high and trade unions forced the inclusion in the reform law of the most comprehensive and strongest regulations and sanctions of noncompliance in the region (including the ‘social

security crime’) but an evaluation of results is needed (Martínez and Mesa-Lago 2003). In all countries the employer is responsible for deducting and transferring the contributions but collection and record keeping is centralized in six private systems (Argentina, Colombia, Costa Rica, Dominican Republic, Mexico, and Uruguay), through a special public entity or social insurance or a private notfor-profit agency. Mexico’s collecting agency is public (the IMSS) but a private agency distributes the collected funds among administrators, which accredit the corresponding sum to individual accounts, the private agency also determines, collects, and distributes the federal government contribution, processes public and private pensions, and controls transfers among administrators (Budebo 2006). In Argentina a state agency collects and accredits the funds, it is more powerful than a private collecting agency and should be more efficient in reducing costs, but the crisis makes difficult to evaluate if it has improved compliance. In the other four private systems collection is decentralized: the employer must fill several payment forms one for each type of insurance and administrator, cumbersome and costly for small enterprises, although not for large enterprises that use electronic means. The centralized collection and a single form simplify the payment process for small enterprises, should improve control of punctual payment of contributions and reduce administrative costs due to economies of scale. But private administrators have opposed it arguing that would delay contribution transfers, induce errors, and violate confidentiality (McGillivray 2001; ISSS 2006). The World Bank proposes to give employers a prominent role in the administration of the private system, acting as intermediaries between the insured and pension administrators through three functions: collect contributions and keep records and archives, offer advice to the insured in the selection of administrators and investment funds, and provide enterprise plans of voluntary savings with defined contribution, through group contracts negotiated by employers with

Financial Sustainability and Reform Goals administrators. Such functions, supported by employers’ records and deduction/transfer of contributions, would increase competition and efficiency, and reduce administrative costs, but could be afflicted by some important problems: risk of conflicts of interest, need of stronger regulation and inspection than currently and with higher costs due to the large number of employers relative to the small number of fund administrators (Gill, Packard, and Yermo 2005). Enterprise voluntary savings plans are widely spread in Europe but require proper state regulation and control, which leads to the problems noted above. The proposed collection and advice functions are controversial and risky because of the documented employers’ past history of illegal retention of workers contributions. Such difficulties and risks demand extreme caution in the proposed policies, particularly in countries where entrepreneurial functions are not well developed, employer noncompliance is common and state supervision is poor. Unlike private systems, there are no standardized statistics on compliance in public systems, and data are very scarce and often out of date. The ILO (2003a) has estimated the percentage of salaried workers that contributed to all social insurance programs (not exclusively to pensions) in four public systems in 2002: Panama 74%, Brazil 67%, Venezuela 58%, and Paraguay 29%. Figures were available also for seven private systems: Uruguay 98%, Costa Rica 76%, Chile 74%, Colombia and Mexico 65–66%, Peru and Argentina 52% (see Section 11.1.3). The highest compliance was in the private mixed systems of Uruguay and Costa Rica; Panama tied in third place with Chile, and Venezuela had a higher compliance than Argentina and Peru. Information on compliance in other public systems is scarce. Cuba doesn’t publish data but compliance is probably high because the immense majority of enterprises are state owned-managed. Nicaragua’s payment delays in the main program were reduced from 50 to 10% in 1991–2005 (INSS 2006) but in Honduras accumulated payment delays were higher than the cost of all pensions in 2002

119

(IHSS 2003). Although specific data on pension compliance as well as from other countries are needed, the above figures suggest that compliance is highest in the pioneer-high group (and Panama), declines in the intermediate group and is lowest in the latecomer-low group. In most public systems the state has not fulfilled its financial obligations to contribute as employer and third party to the principal program or has delayed such contributions, used the pension fund reserves to cover fiscal deficit, and is the major debtor. Exceptions are Brazil and Cuba where the state has financed the significant deficit of the general program, also true of the public system/pillar of Argentina and Uruguay. Ecuador’s government didn’t pay the constitutionally mandated contribution of 40% of social insurance (IESS) pensions in 1985–2000 and thereafter made payments with delays; IESS estimated the cumulative state debt as US$2,744 million (9% of GDP) in 2005 but the ministry of finance acknowledged only 20% of that figure, due to divergent conversion of national currency to dollars, and no agreement had been reached in 2006 (World Bank 2005e; IESS 2006a, 2006b). Panama’s state is responsible for 26% of total payment delays and the CSS had to deposit its funds (at least until 2005) in the national bank that pays lower interest rates than commercial banks. In Paraguay and Venezuela the state has not fulfilled the legal contribution of 1.5% on the wage bill to social insurance and is the principal debtor. Guatemala’s state has not paid 25% of IGSS pension costs for fifty-six years, accumulating a debt of US$1,315 million that has depreciated because it has not been adjusted to inflation, and IGSS reserves are deposited in the central bank or invested in public debt paper with very low interest (Herrera 2006). Honduras’ state did not honor for many years its contribution of 0.5% of the wage bill to the general program (IHSS), thus accumulating a huge debt, the government is also responsible for 64% of total payment delays, and most of the IHSS reserves are deposited in the central bank that pays an interest rate below the market

120 Reassembling Social Security rate. Paraguay’s state contribution of 1.4% of the wage bill to the IPS has rarely been paid, and the government is responsible for a significant part of noncompliance.

6.3. Components of fiscal costs in private systems in the transition and beyond A major claim of structural reforms is that they will reduce and eliminate fiscal cost in the long run. World Bank officials assert that ‘ensuring fiscal stability was the primary impetus behind the region’s pension reform’ and their simulations predict that the reforms are ‘likely to deliver substantial reductions in public liabilities’. However such simulations are afflicted by important problems (see Section 6.3.1) and the Bank officials warn that ‘fiscal sustainability is far from certain’ and that ‘a positive impact of pension reform on solvency perceptions is not as obvious as theoretical models claim’ (Gill, Packard, and Yermo 2005: 7, 44, 54). Furthermore, fiscal costs are difficult to measure, project, and compare between countries because of diverse components included and methodologies utilized (Mesa-Lago 2000a). The three main components of fiscal cost during the transition and beyond are financed by the state with very few exceptions: the operational deficit of the public system, the recognition bond and the minimum pension. Table 6.3 provides that information on the ten private systems implemented, as well as Ecuador partly declared unconstitutional law. The state also grants guarantees and social assistance pensions in some countries that increase fiscal costs.

6.3.1. Operational deficit of the public system In all but one private system the state must pay the deficit generated in the public system (operational deficit) except in Costa Rica

because its partly funded (collective partial capitalization: CPC) public pillar finances all current pensions (see below). The operational deficit varies according to the treatment given in the three models of structural reform to the implicit pension debt (IPD), that is, the present value of all long-term obligations, including the payment of ongoing and future pensions.6 In the substitutive model, the public system is totally closed and all the IPD is made explicit immediately, hence ongoing pensions and those generated by the few insured left in the public system are financed by the state. A deficit results because either 100% of the insured (Bolivia and Mexico) or 95–96% of them (Chile, Dominican Republic, and El Salvador) moved to the private system and stopped contributing to the public system, which is left with virtually all pensions to pay but without contributors or very few of them. The Chilean state has financed the operational deficit for twenty-five years. Conversely, the Dominican government delayed granting 5,000 public pensions (15% of all current public pensions), together with its contributions as employer, and there is no data on state transfers of contributions to the solidarity fund (Pérez Montás 2006). El Salvador’s government ran out of money to pay ongoing pensions in 2006 and forced pension administrators to buy debt pension paper (LFOP 2006). PAYG or partly funded (CPC) always have an IPD, but the model of reform may turn it explicit and generate an immediate fiscal cost or postpone it totally or partially. In the parallel model the IPD is only made explicit in the private system, but postponed in the public system that keeps contributors that reduce fiscal costs but not the IPD. In Colombia, 53% of the insured moved to the private system and their contributions and those of their employers were lost for the public system, which was left with all ongoing pensions and payments of two extra monthly pensions, hence reserves that were plentiful right after the reform declined sharply and the public system operational deficit eventually will soar and require fiscal transfers. In

Financial Sustainability and Reform Goals

121

Table 6.3. Fiscal costs of the structural reform in private pension systems, 2006 Financial responsibilities of the state Private systems

Covers public system deficit

Pays recognition bond

Guarantees a minimum pension

Argentina

Yes

Yes (paid by the first pillar, the public system)

Bolivia

Yes

Chile

Yes

Colombia

Yesb

Costa Rica

Noc

No ceiling, adjustable to inflation, 30 years of previous contributions needed (paid by the first pillar, public) Ceiling, adjustable to US dollar, does not earn interest, 5 yrs of previous contributions needed No ceiling, adjustable to inflation, earns 4% real annual interest, previous contributions needed Ceiling, adjustable to inflation, earns 3% real annual interest, 3 years of previous contributions needed No

Dominican R.a

Yes

Ecuadora

Yes

El Salvador

Yes

Mexico Peru

Yes Yes

Uruguay

Yes

No ceiling, adjustable to inflation, earns 2% real annual interest, previous contributions needed, not accredited at the end of 2005 No No ceiling, adjustable to inflation, does not earn interest, 4 years of previous contributions needed No Ceiling, adjustable to inflation, does not earn interest, 3 years of previous contributions needed; granted only after 2002b No

No

Yes

Yes (with limitations)

Yes (paid by the first pillar, the public system) Yes (but the lack of the bond will reduce those with the right) Yes (paid by the first pillar, the public system) Yes

Yes Only after 2002 and for affiliates born before 1945 Yes (paid by the first pillar, the public system)

Sources: Based on legislation; see also text. a The reform law of Ecuador was partly declared unconstitutional in 2005; the Dominican Republic contributory-subsidized and subsidized regimes had not been implemented in 2006. b Paid by the public system instead of by the state. c There is no financial deficit.

Peru 76% of the insured moved to the private system hence most IPD became explicit (more than in Colombia); the public system must finance all ongoing pensions as well as the

recognition bond (paid by the state in other countries) therefore public system reserves are projected to be depleted by 2009 (Morón and Carranza 2003).

122 Reassembling Social Security In mixed models the IPD is made partially explicit in the second (private) pillar of mandatory savings but postponed in the first (public) pillar. In Argentina and Uruguay 86 and 37% of the insured respectively moved to the mixed system and although employers’ contributions still go to the public pillar they were reduced (considerably more in Argentina than in Uruguay) and the insured contributions are deposited in the private pillar hence part of the IPD was made explicit, the public pillar endures deficit and requires significant state transfers. There is no financial deficit in Costa Rica’s public pillar because all contributors were obliged to stay in it; that pillar has substantial reserves but faced an actuarial disequilibrium that was restored in 2005 (see Section 6.5). The operational deficit cost increases in the transition first stage, reaches a peak and eventually should decrease and disappear. In Chile it peaked at 6.9% of GDP in 1984, is projected to diminish to 1.8% in 2010 and disappear in 2050, for a transition period of seventy years (Arenas de Mesa and Mesa-Lago 2006).

6.3.2. Recognition bond In eight countries the state is legally bound to pay a recognition bond, equivalent to the value of the contributions accumulated in the public system, by all insured who transferred to the private system. Normally the bond is estimated and a certificate emitted by the state and placed in the insured individual account but cannot be cashed until retirement. Four countries don’t grant the bond: Mexico (because of the choice given to those insured when they retire, as already explained), Costa Rica, Ecuador, and Uruguay because in a mixed model the insured do not move, but remain in the first pillar, which pays them a basic pension (Argentina pays a compensatory benefit in lieu of the bond). A minimum period of contribution that ranges from one to thirty years is required to gain the right to the bond; its value is annually adjusted to inflation in six countries (Bolivia

adjusts it to the US dollar) and earns an annual interest rate from 2 to 4% in Chile, Colombia, and Dominican Republic. Four countries do not set a ceiling to the value of the bond (Argentina, Chile, Dominican Republic, and El Salvador), but such ceiling is imposed in Bolivia, Colombia, and Peru. Bolivia’s formula to calculate the bond reduces it (Bonadona 2006), and the bond calculation and issue in El Salvador were five years behind in 2002 (Mesa-Lago 2004). In Peru the bond requires three contribution years prior to affiliation in the private system and within ten years prior to the reform, it is not paid by the state but by the public system, its emission was delayed ten years and in 2005 had been granted to only 25% of the insured who changed to the private system (Morón and Carranza 2003; SBS 2005). The Dominican Republic bond had not been calculated and accredited to the individual accounts in 2005 (more than two years after the legal deadline) creating a serious obstacle to retirement; due to high inflation in 2001–4 the bond ought to be adjusted by 115% significantly increasing fiscal cost (Pérez Montás 2006). The cost of the bond is null at the start of the system but later grows as the system matures and more insured retire and cash the bond, eventually the cost of the bond peaks and then decreases and disappears as all those entitled to this benefit die. Chile’s bond costs steadily increased to 1.3% of GDP in 2000 and are projected to end in 2038 (Arenas de Mesa and Mesa-Lago 2006).

6.3.3. Minimum pension Affiliates in private systems (at the time of the reform and thereafter) whose individual account is insufficient to finance a minimum pension at retirement receive the difference from the state but they need to have from twenty to thirty-five years of previous contributions. Bolivia does not grant a minimum pension to insured in the private system; due to pensioners’ public demonstrations, a minimum pension was awarded in 2001 but only

Financial Sustainability and Reform Goals for old-age and to those insured in the public system, submitted to a cap and an annual recalculation (Garrón 2004; Bonadona 2006). Peru has granted the minimum pension to insured in the private system since 2002 and only to those born before 1945 that have at least twenty years of contributions and meet other requirements, hence most insured are not eligible (Morón and Carranza 2003). El Salvador places considerable restrictions on eligibility (Mesa-Lago, 2003a). In the mixed models of Costa Rica and Uruguay (also in Ecuador nonenforced law), the minimum pension is actually the pension paid by the first public pillar, which is the majority of the total pension in Costa Rica, but less so in Uruguay and a minority in Argentina. Because contribution density is low in virtually all private systems, the insured won’t accumulate enough in their individual accounts to finance their minimum pension forcing the state to pay the difference (Uthoff 2002). Half of Chilean insured are expected to receive the minimum pension, the majority in Dominican Republic and 30–60% in Peru (see Section 4.3.1). A Solidarity Fund helps to finance the minimum pension in Colombia and the Dominican Republic; in the former the public system is largely responsible for paying that pension; in the latter the fund disburses the subsidy only when a programmed pension is chosen not when an annuity is selected (Pérez Montás 2006). The cost of the minimum pension is null or low at the start of the transition and later increases and does not end because the right is guaranteed not only to those insured at the time of the reform but to future affiliates in the private system also; a reduction of that cost would require a significant increase in formal employment, real wages, and contribution density.7

6.3.4. Other state guarantees In Argentina, Chile, Colombia, and Uruguay the state offers two additional guarantees: if a pension administrator is unable to pay the

123

minimum annually guaranteed capital return on the individual account with the reserves created for that purpose, the state makes up the difference, and if a fund administrator or insurance company goes bankrupt, the state assumes responsibility for the payment of its pension obligations (in Uruguay, these guarantees are restricted to the insured in the state bank, which partly explains why it has 38% of total affiliates). The Dominican Republic state is responsible for any fault or nonfulfilment that occurs in the private system. Although not a direct transition cost, in half of the private systems the state finances social assistance pensions, also stipulated in laws of other four. Such pensions are financed out of the general budget with three exceptions: in Bolivia partly from a fund created with proceeds of seven large privatized enterprises; in Costa Rica half comes from employers’ contributions to a special fund that often has suffered deficit due to state failure to transfer the legally mandated resources, and in Colombia with insured and employers contributions plus state transfers. In 2000–2, the cost of social assistance pensions as percentage of GDP was: 0.2% in Argentina, 0.3% in Costa Rica, 0.4% in Chile, and 0.6% in Uruguay; relative to the total cost of pensions the corresponding percentages were 3.6%, 5.5%, 7%, and 5.5% (Bertranou et al. 2002). Compared with the high cost of the transition, social assistance pensions take a very small proportion; furthermore they reduce poverty and have a progressive impact on income distribution (see Section 3.3), strong reasons to allocate more resources to these pensions.

6.4. Estimates and projections of fiscal costs in private systems According to IADB none of the structural reforms ‘used a high-quality model capable of accurately predicting when small changes in costs of benefits are introduced . . . a significant flaw that led to a considerable disparity

124 Reassembling Social Security

Table 6.4. Domestic estimates and projections of fiscal costs in private systems, at the start of the reform or later, compared with those of the World Bank, 2005 (deficit as % of GDP) Domestic initial and later projections Private systemsa Argentina Bolivia Chile Colombia El Salvador Mexico Peru Uruguay Averageb

World Bank projections

Initial Year

2000

2020

2040

2001

2020

2040

−2.5 −0.2 −3.8 −0.9 n.a. −0.9 n.a. −5.1 n.a.

−1.8 −2.2 −6.1 −1.5c n.a. n.a. n.a. −4.5 n.a.

−0.3 −0.9 −3.6 −2.2 n.a. −1.0 n.a. −3.8 n.a.

+0.2d −0.2 −3.3 −2.0e n.a. n.a. n.a. −3.6 n.a.

−2.5f −3.5g −7.2h −1.6 −1.4 −0.5i −0.7i −4.0j −2.7

−2.3 −2.1 −3.4 −1.0 −3.2k −0.7 −0.9 −2.1 −2.0

−3.6 −1.7 −0.5 −3.4 −2.6 −0.7 −0.8 −2.5 −2.0

Sources: Domestic from: Argentina Rofman, Stirparo, and Lattes (1997); Bolivia Gersdorff (1997); Chile Arenas de Mesa (1999); Colombia Schmidt-Hebbel (1995); Mexico Grandolini and Cerda (1998); Uruguay Noya and Laens (2000); El Salvador and Peru Mesa-Lago (2000a). World Bank from Gill, Packard, and Yermo (2005: 42). n.a. = nonavailable. a There are no data on Costa Rica and Dominican Republic. b Nonweighted average estimated by the author for World Bank figures; not feasible for national figures due to lack of data in half of the countries and different methodologies. c A later study estimated it as −3% in 2000 (Ayala and Acosta 2001). d A posterior study estimated it as +1% (Bertranou, Grushka, and Schulthess 2000). e Year 2025, the projection does not extend to 2040. f Actually averaged −4.6% yearly in 1995–2001 (Grushka 2002). g The World Bank also estimated it as −5% in 2002 (Gill, Packard, and Yermo 2005). h Annual average of −5.7% in 1981–2000 and was 6% in 2001 (Arenas de Mesa and Mesa-Lago 2006). i Higher estimates are −1.1% for Mexico and −2.7 for Peru both in 2003 (Crabbe and Giral 2003). j The deficit was −4.5% in 2000 and showed an increasing trend since 1996 (Ferreira-Coimbra and Forteza 2004). k In 2006 the fiscal cost took 2.9% of GDP, 14 years before the peak projected by the World Bank for 2020.

between actual and anticipated results in some countries.’ Most of them failed to conduct a thorough diagnosis of the existing pension system, its obligations, assets, administrators’ skills, and quality of record keeping on worker’s contributions and years of employment (Crabbe and Giral 2005: 14). World Bank officials have estimated and projected transition fiscal costs (operational deficit, recognition bond, and minimum pension) relative to GDP in eight countries for 2001–40 (Gill, Packard, and Yermo 2005), which are contrasted in Table 6.4 with domestic estimates and projections in six countries

at the time of the reform or a few years later. Bank estimates for 2001 on Argentina, Bolivia, and Colombia are 0.7, 1.3, and 1.1 percentage points higher than domestic estimates, whereas Bank projections for 2040 are respectively 3.3, 1.5, and 1.4 percentage points higher than domestic projections;8 and Bank projections for 2040 in Argentina, Colombia, and El Salvador9 are significantly higher than in 2001, whereas in Mexico and Peru are slightly higher, and only in Bolivia, Chile, and Uruguay are lower, forty-two to fifty years after the start of the reform. Bolivia initially projected that fiscal costs would virtually

Financial Sustainability and Reform Goals disappear in 2037, but the Bank projects 1.7% for 2040, blaming fraud due to poor regulation for the superior and more prolonged cost. Conversely a Bolivian economist estimates that fiscal costs steadily increased from 3.1 to 5.3% of GDP in 1997–2003, exceeding all previous projections, at an annual average rate of 1.6% higher than pre-reform levels, caused in 19% by unexpected liabilities such as minimum and armed forces pensions, better adjustment and a jump in retirements under the old system (Garrón 2004). The Bank estimate of Chilean fiscal cost for 2001 was one percentage point higher than the domestic estimate, despite that the later included 1.3% for the armed forces pension deficit probably excluded by the Bank; conversely, the Bank projected cost for 2040 is three points less than the domestic projection. Bank cost estimates for Mexico, Uruguay, and Peru in 2001 were from 0.5 to 1 percentage points lower than other estimates (Crabbe and Giral 2005). The Bank excluded two private systems from its estimates/projections. Costa Rica doesn’t have fiscal costs because the public pillar finances all ongoing pensions and its equilibrium was enhanced by the 2005 parametric reform. Dominican Republic experts estimate the cost of the recognition bond from 3 to 5% of GDP and the solidarity pension cost from 1.5 to 2%, a total of 3.5 to 7% excluding the operational deficit for which there are no estimates (Lizardo 2004; Pérez Montás 2006). World Bank officials’ simulated projections for the eight countries in 2000–50 indicate that in six of them the fiscal cost without the reform would have been smaller through the entire period in Argentina, for thirty years in El Salvador and Peru, and for twenty years in Bolivia, Chile, and Mexico;10 conversely, the IPD would have been higher without the reform. Said officials do not show the assumptions, methodology, and calculations for their simulated projections but frankly note several flaws that reduce their certainty and usefulness: (a) simulations can diverge substantially from actual transition costs, hence must be interpreted very carefully and can-

125

not replace thorough specific analysis of the reforms; (b) they assume correct implementation of the reforms and adequate management of the new systems thus cannot capture probable administrative difficulties that could unexpectedly balloon fiscal costs as happened in Argentina and Bolivia; (c) they don’t precisely compare the IPD before and after the reform; (d) the macroeconomic impact of converting the IPD into an explicit debt and the effect of the reform on country risks are uncertain, as Argentina also illustrates, and global crises and external factors may create other unpredicted risks; (e) the reform has introduced new implicit and explicit obligations with corresponding fiscal costs in some countries, which can happen elsewhere and change simulation results; ( f ) fiscal sustainability is far from been secured and the positive impact of the reform on solvency is not as obvious as theoretical models claim; and (g) ‘empirical evidence shows that pension reforms can provoke severe cash-flow problems in excess of initially projected transition costs and hence seriously constrain public sector liquidity’ (Gill, Packard, and Yermo 2005: 39, 44–54). Chile’s reform is the longest in operation and the only having historical statistical series (1981–2004) on fiscal costs disaggregated by components as percentage of GDP. The total deficit averaged 5.7% annually and in 2004 still was 5.5% and took 33% of the nation fiscal burden. The operational deficit decreased from 6.9 to 2.5% but the recognition bond rose from 0.2 to 1.3%; social assistance pensions averaged 0.4% and are stagnant; the minimum pension increased from zero to 0.1%, and the trend in military pensions is difficult to assess but took 1.3% of GDP in 2004 (25% of the total pension deficit). Projections for 2005–10 (thirty years after the reform) indicate that the annual average total deficit will remain high at 5% (only 0.5 points less than the 1981–2004 average), proving that projections made in the 1980s were rather optimistic significantly underestimating fiscal costs because they assumed a higher contribution density. The operational deficit average

126 Reassembling Social Security will decrease from 3.3 to 2% of GDP in the period, whereas the average cost of the recognition bond will double from 0.6 to 1.2%, and other costs will remain basically unchanged (Arenas de Mesa and Mesa-Lago 2006). The previous analysis and the transition length as much as seventy years demand that the projection of fiscal costs be done in a professional and extremely careful manner, preferably by more than one entity and based on various scenarios, and their results published and opened to public debate by national and international experts. These reasonable steps were omitted by several structural reforms whose initial projections seriously underestimated fiscal costs, were not published, and in El Salvador manipulated by the government to show false lower fiscal costs (Mesa-Lago and Durán 1998). The long time of the transition and the effect of its fiscal deficit on public finance also demand a previous determination of the means to cope with such heavy burden. In Chile the deficit in the pension system was accompanied by an adjustment in the fiscal balance of the central government (excluding pensions) that generated an annual average surplus of 8.5% of GDP in 1981–2004, which allowed the government to finance the pension deficit. If Chile had not had fiscal discipline and a surplus, the reform would have very difficult to finance and provoked significant fiscal imbalances and instability or forced substantial increases in taxes to finance the deficit (Arenas de Mesa and Mesa-Lago 2006). The Chilean example has not been replicated by other structural reforms. The heavy fiscal costs during the transition have induced a trade off in all countries. Because of Chile’s relative generosity of benefits during the transition, its fiscal costs are the highest of the eight countries that have available data and 2.4 times their average (Table 6.4). In other countries the state has reduced fiscal costs by denying the recognition bond (Ecuador—nonenforced law, Peru—during the first nine years of the reform) or denying the granting of the minimum pension (Bolivia) or granting those benefits with restrictions,

for instance, not adjusting the value of the recognition bond to inflation, imposing a ceiling on its value (Bolivia, Colombia, and Peru), requesting excessive years of contributions to grant the right to the bond (Argentina and Peru) or postponing its calculation and credit to the individual accounts (Dominican Republic) or delaying and limiting the right to the minimum pension to those born after a certain date (Peru) (Mesa-Lago 2000a; Arenas de Mesa and Benavides 2003; Table 6.3). All of these restrictions are imposed at the cost of insured welfare and some have been proposed by the World Bank: ‘the government will have to renege on some PAYG commitments’ (cited by Müller 2003: 16).

6.5. Financial and actuarial equilibrium in public systems Among public systems those in the pioneerhigh group face the most severe financial and actuarial deficit, problems significantly reduced in two countries of the intermediate group, and more so in the latecomerlow group, because their programs have not matured and the population is relatively young. Cuban and Venezuelan general programs are on PAYG, lack reserves, confront serious financial and actuarial deficit and receive fiscal transfers. Brazil major program (RGPS) is on notional defined contribution (NDC);11 Ecuador and Panama are partly funded (CPC) and their reserves are substantial, whereas the postponed reform in Nicaragua stipulates a CPC. The general programs of the four countries in the latecomerlow group (Honduras, Guatemala, Nicaragua, and Paraguay) are on CPC and have reserves. Haiti may be on PAYG or CPC. Costa Rica’s public pillar is on CPC and has considerable reserves (see Section 6.7). All separate schemes are based on PAYG and confront financial deficit. Five indicators to measure financialactuarial sustainability in nine public systems

Financial Sustainability and Reform Goals

127

Table 6.5. Indicators of financial sustainability in public pension systems, between 2000 and 2004 Surplus (+) or Deficit (−) as % of: Public systems Brazil Costa Ricab Cuba Ecuador Guatemala Honduras Nicaragua Panama Paraguay Venezuela Averagesc

Pension expenditure (% GDP) 10.5 1.8 6.5 3.9 0.4 0.2 1.8 4.0 0.7 0.5 3.3

Contribution (% on wages)a

Revenue

GDP

Current

Equilibrium

−61.1 +24.3 −50.0 +21.7 +40.0 +66.1 +25.6 −8.3 +37.7 −26.7 +6.9

−4.8 +1.2 −2.3 +1.4 +0.2 +0.2 +1.2 −0.3 +0.4 −2.4 −0.5

28–31 7.5 12.0 9–11 5.5 3.5 10.0 9.5d 12.5 6.8 n.a.

n.a. 10.5 18–39 e

3.9 3.5 15–16f 16.2 n.a. 11.0g n.a.

Ratio Actives per pensioner 1.7 6.3 2.8 4.2 7.5 22.0 4.6 5.9 7.5 7.5 5.3h

Sources: Brazil IPEA (2001, 2002), MPAS (2002a, 2002b), MPS (2003), Schwarzer (2004), Pinheiro (2005); Costa Rica CCSS (2005a); Cuba ILO (1999), Sabourin (2003), Mesa-Lago (2005b); Ecuador World Bank (2005e), IESS (2006a); Guatemala Durán and Cercone (2001), BG (2005), IGSS (2005); Honduras Durán (2003), IHSS (2003); Nicaragua INSS (2005, 2006); Panama Mesa-Lago (2005a); Paraguay ILO (2003b), Saldaín (2003); Venezuela González et al. (2002), Armas (2005). GDP from ECLAC (2005c). No data available for Haiti. n.a. = nonavailable. a The methodology and projection periods are different in the countries. b Public (first) pillar. c Nonweighted. d The reforms of 2005–6 gradually increase the contribution to 13.5% in 2013. e According to projections made in 2003 the current contribution will maintain financial balance until 2022, if the state debt is not paid and investment returns are not taken into account, and until 2050 if such debt is paid and investment returns included. f Reserves in 2004 were 6.% of the amount required based on actuarial valuation, the equilibrium contribution is a rough estimate. g The law of 2002 not enforced by 2006 increases the contribution to 14%. h Excluding Honduras, if included the average ratio is 7; the ratio in Haiti is 50 that would increase the average ratio to 10.9.

and in the public pillar of Costa Rica are compared in Table 6.5. As more mature a pension system, older its population, more generous its entitlement conditions and higher its population coverage, greater the share of pensions in total security expenditures and vice versa, the highest shares being in the pioneerhigh group and the lowest in the latecomerlow group. As a percentage of GDP pension

expenditure in 2000–4 were: 10.5% in Brazil, 6.5% in Cuba, 4% in Ecuador and Panama, 1.8% in Costa Rica (public pillar), and 0.2 to 0.7% in Nicaragua, Paraguay, Venezuela, Guatemala, and Honduras.12 Many social insurances did not separate finances of their various risk branches; the pension fund transferred resources or made loans to its sicknessmaternity branch or invested in its health care facilities, the latter laudable from a social

128 Reassembling Social Security view point but negative financially because they did not have capital returns and hospitals could not be sold. With the gradual maturity of the pension systems their share of total social security expenditures grew, previous surpluses shank and the transfers, loans and investment in sickness-maternity stopped. Many social insurances now have separate branches but the damage done to the pension fund is difficult to reverse in some countries. The financial balance (current revenue minus current expenditures) was a significant financial deficit as percentage of the general program revenue in 2000–4 in Brazil −61%, Cuba −50%, and Venezuela −27%. Conversely there were surpluses in Costa Rica (public pillar) 24%, Ecuador 22%, and Panama 20% (the latter turned into a deficit of −8.3% in 2004). The four latecomerlow countries had higher surpluses: Honduras 66%, Guatemala 40%, Paraguay 38%, and Nicaragua 26%.13 Similar tendencies were observed in the financial balance as a percentage of GDP: deficit in Brazil −4.8%, Cuba and Venezuela −2.4%, and Panama −0.3%, but surpluses in Guatemala, Honduras, and Paraguay from 0.2 to 0.4%, Costa Rica (public pillar) and Nicaragua 1.2%, and Ecuador 1.4%. The nonweighted average balance as a percentage of program revenue was a surplus of 7% but became a deficit of 0.5% as percentage of GDP because of the heavy deficit of Brazil, Cuba, and Venezuela, which averaged −3%, whereas the remaining countries (excluding Panama) had an average surplus of 0.8%. Eight public systems and Costa Rica’s public pillar have data on their actuarial situation, but statistics scope and date vary and the contribution required for actuarial equilibrium is not technically comparable due to diverse methodologies. Table 6.5 indicates that Cuba and Venezuela suffer a severe disequilibrium (Brazil too although we lack figures); Panama endures a growing but still relatively small disequilibrium; Nicaragua has a significant disequilibrium, and the lowest imbalances are in Guatemala, Honduras, and probably Paraguay, whereas Ecuador appears

to have the longest period of equilibrium. A summary of the most relevant information on these countries follows. In Brazil, despite the parametric reform of RGPS, the entire pension system had in 2003 the highest deficit among public systems: 4.8% of GDP, 1.7% the RGPS, and 3.1% the civil servant schemes, even higher without the state contribution. The RGPS projected financial deficit ranges from 3.5 to 7.5% of GDP in 2050, whereas that of all civil servants schemes is 4% of GDP in 2032, and their IPD 302% of GDP in the next sixty years. The parametric reform of civil servant schemes in 2003–4, however, reduced their financial deficit to 2.4% in 2005 and the IPD 270% of GDP in the next sixty years, still a huge burden considering they cover only 5.7% of the labor force. The equilibrium contributions of the RGPS and civil servants scheme were not available (Schwarzer 2004, 2006; Pinheiro 2005). Cuba’s contribution of 12% is grossly insufficient and the ensuing financial deficit covered by the state rose from 35 to 50% of program revenue and from 1.3 to 2.3% of GDP in 1989–2003; the pension deficit is the major component of the total fiscal deficit and will keep rising with the ageing process and low ages of retirement. To eliminate such deficit the contribution had to be increased from 12 to 18% of the wage bill in 2003 and gradually climb to 20% in 2010, and reach 39% or 86% later, according to different scenarios.14 The deficit of the armed forces scheme, entirely financed by the state, was similar to that of the general program in 1995 (ILO 1999; Mesa-Lago 2003b, 2005b; Sabourin 2003). Venezuela’s general program equilibrium contribution had to be increased from 6.75 to 11% in 2000 and projected to 24% in 2035, but the law of 2002 not enforced by 2006 raised the contribution to 14%. The schemes of civil servants, armed forces, and most other groups suffer a worse disequilibrium (Mesa-Lago 1995; González et al. 2002; Villasmil 2006). In Panama ILO actuarial projections in 1998 and 2001 showed a growing actuarial disequilibrium in the pension program. The

Financial Sustainability and Reform Goals cumulative actuarial debt for 2050 jumped from 10 to 68% of GDP between the two projections, and the fiscal subsidy to finance the 2050 deficit rose from 75 to 87% of revenue. A financial deficit occurred for the first time in 2001 and the reserve was projected to be depleted in 2012 and the actuarial deficit estimated at US$10,600 million. The current contribution of 9.5% had to be increased to 16% in 2001 to balance the program until 2050, without adjusting pensions to the CPI (to 19% doing such adjustment). The ongoing reforms stipulate a gradual increase in the contribution in 2007–13, without pension adjustment and not rising the ages of retirement; it is unknown if such increase will restore actuarial equilibrium (ILO 2001a; Mesa-Lago 2000b, 2003d, 2005a; Ley 51 2005; Asamblea 2006). Ecuador’s general program (IESS) had a declining financial surplus in 1993–9 and a significant rising surplus in 1999–2002 (resulting from a jump in real wages and contribution revenue) reaching 2% of GDP, but declined to 0.5% in 2004, due to the sharp increase in real pension value. These figures are based on contribution revenue alone and exclude state legal transfers and IESS investment returns, if these two additional sources were accounted the actual surplus would be 1.4% of GDP in 2004. World Bank simulations in 2003 show that the current contribution of 8.5% (actually 9–11%) will maintain the equilibrium until 2022 if the state does not pay its debt and investment returns are not counted, and until 2050 if both sources are included. If parametric reforms are not implemented, however, reserves would be depleted by 2055 and a fiscal subsidy required on top of the state contribution of 40% of pension costs (World Bank 2005e; IESS 2006a). Nicaragua general program actuarial reserves were 6% of the amount required in 2004 and its actuarial deficit has grown aggravated by the system maturity. The current contribution of 10% is grossly insufficient (it may have to be increased to 15% or 16% to restore equilibrium) but the actuarial valuation done in 2004 failed to recommend an increase (INSS 2006;

129

Ruiz 2006). If the private system had been implemented, fiscal costs were projected to increase threefold in 2005–2015 and benefit only 10% of workers, the reason why the reform law was suspended in 2004 and repealed in 2005 (ISSA 2006; Ley 568 2006). In Paraguay the ILO projected in 2003 the actuarial deficit of the general program as 0.1% of GDP in 2025 rising to 0.8% in 2050; albeit not providing the equilibrium contribution the ILO judged the disequilibrium relatively small and feasibly corrected by a parametric reform. The scheme for civil servants had an actuarial deficit equal to 100% of GDP in 1997, whereas those for congressmen and railroad workers suffered financial disequilibrium and the state financed their increasing deficit (ILO 2003b; Saldaín 2003). Guatemala’s general program equilibrium contribution in 2001 was 3.9%, 1.6 percentage points less than the current contribution of 5.5%, but will have to be increased to 6.3% in 2020 to avoid a deficit, even so that contribution would be low compared with that in other countries.15 Civil servants and armed forces schemes suffer a substantial and growing financial deficit that require increasing fiscal transfers (Durán and Cercone 2001; Apén 2002; Balsells 2002). In Honduras, an actuarial study of the general program in 2003 projected that its current contribution of 3.5% (the lowest equilibrium contribution in the five public systems with data) could continue until 2030, based on current low population coverage and meager pensions and not adjusting the contribution ceiling, but that a financial deficit would occur in 2038. If coverage is increased and the ceiling adjusted, the 3.5% contribution could be extended until 2042 and the deficit postponed until 2050. Three separate schemes confront a severe actuarial disequilibrium (Mesa-Lago 2000c; Irías et al. 2002; Durán 2003). The actuarial equilibrium in the public pillar of mixed systems and in public systems of parallel systems varies. Argentina and Uruguay lack data on the actuarial equilibrium, but taxes and state transfers to the public system of Argentina increased from 16 to 52% of total

130 Reassembling Social Security expenditures in 1991–2005,16 and in Uruguay rose from 16 to 47% in 1992–2004 (INARSS 2002 and ANSES 2006; BPS 2005a). Costa Rica’s parametric reform of 2005 restored the equilibrium of the public pillar at least until 2040: the current total contribution of 7.5% to that pillar will be gradually raised starting in 2010 to 10.5% in 2035; those who take early retirement before the statutory age of 65 will get an actuarially adjusted pension (CCSS 2005b, 2005c). In Colombia, in 1998 (four years after the reform), the ILO projected that the public system would maintain its financial equilibrium until 2025 without touching its reserves and until 2032 using such reserves. But the public system finances were badly harmed, its reserves declined 67% in 2000–5 (still were half of the capital accumulated in the private system) and the equilibrium period shrank, largely because the state charged it with the payment of two extra monthly pensions, and the young contributors transferred to the private system caused significant revenue loss (Jaramillo 1999; SBC 2000–5). The Peruvian state financed 80% of the public system in 2002; simulations based on state financing 40% of pension expenditures indicate that the reserves will be depleted by 2009 (Morón and Carranza 2003). Mexican transition fiscal costs in the private system are small compared with the huge deficit in public schemes: that of federal civil servants (ISSSTE) was 59% of expenditures in 2003 and to achieve financial equilibrium the current contribution of 7% had to be increased to 24% in 2006 and to 72% in 2050; if the current retirement age of 55 is raised to 65 still the contribution would have to be increased to 27%; states civil servants schemes have a deficit equal to half of ISSSTE; the scheme for employees of IMSS had a deficit equal to 93% of expenditures in 2004 and is using reserves from the health care program that will be depleted by 2020 unless a radical reform is implemented (see Section 11.1.5). The projected combined obligations of those and other public schemes (oil, electricity, universities) and the private system was 117% of GDP in 2003 and is increasing fast; the share

of the private system was only 21% of GDP (Mesa-Lago 2000g; ISSSTE 2004, 2006; SHCP 2004). The nonweighted average of the financial balance in the nine public systems (plus Costa Rica’s public pillar) was a deficit of −0.5% of GDP, while the average fiscal cost in eight private systems was a deficit of −2.7% of GDP (Tables 6.4 and 6.5). These two indicators are not technically comparable due to several reasons: the financial balance in a private system is calculated using the actual value of the future contribution and pension flows, whereas in a public system uses the current value; public systems have not made explicit the IPD and it’s not completely fair to compare their balances with those of private systems that have done so, and the deficit of public pillars in Argentina and Uruguay are unknown and have not been included in the calculation of the average public deficit. Just for the sake of comparison, however, the average private deficit is more than fourfold the public average deficit although that gap would be significantly reduced if proper adjustments were made. There are no standardized projections of the balance of public systems similar to those of World Bank officials on the fiscal cost in private systems, but we have analyzed actuarial projections in seven public systems and one public pillar that contrast the current contribution with that projected to achieve equilibrium in the long run. Guatemala’s contribution of 5.5% is actually excessive and only 3.9% in required; Honduras’ contribution of 3.5% does not demand an increase; Ecuador’s contribution can maintain the financial balance of the system until 2022 or 2050 depending on whether the state pays its debt and investment returns are excluded or included; Costa Rica’s contribution to the public pillar will gradually rise from 7.5 to 10.5% to equilibrate the system until 2040; but current contributions in the other three countries need to be increased: Cuba’s from 12% to at least 18%, Panama’s from 9.5 to 16%, and Venezuela’s from 6.8 to 11%. The projected equilibrium contributions of Costa Rica and Venezuela

Financial Sustainability and Reform Goals are smaller than the 13.5% average of current total contributions in private systems, while those of Cuba and Panama are lower than the current total contributions of Argentina and Uruguay (see Table 6.1).

6.6. Population aging Population aging results from the decline in fertility and mortality rates, as well as life expectancy increase, phenomena more significant in the pioneer-high group (with a couple of exceptions) but less important in the latecomer-low group. In 2000 there were significant differences between Latin American countries regarding the demographic transition, observable in the percentage distribution of the population among three age cohorts: 0–14 years (young), 15–64 (productive) and 65 and over (old). In the pioneer-high group, Uruguay had the highest old cohort (13%), followed by Argentina and Cuba (10%), Chile (7%), Brazil (6%), and Costa Rica (5%). The first three experienced a decrease in the young cohort while the productive cohort was still growing except in Uruguay where it declined since the 1990s. Cuba’s aging process is accelerating: its rate of population growth was 0.26% in 2003 (the lowest in the region); its population will start to decline in 2015 and become the oldest by 2025; its young cohort decreased in 1970–2001, whereas the productive cohort rose but decreased since 1991, and the old cohort steadily climbed from 9 to 14.5%. The intermediate group had higher proportions in the young and productive cohorts than in the pioneer-high group (both still expanding), as well as lower old cohorts; the latter in six countries ranged from 4 to 4.7%, but Panama had 5.5% (same as Brazil). The latecomer-low group had the smallest old cohorts: from 3 to 3.6% in Guatemala Honduras, Nicaragua, and Paraguay, but 5% in El Salvador and 4.3% in Dominican Republic (respectively above and comparable with countries in the intermediate group); this group had the highest young cohorts and the lowest productive cohorts although the latter

131

were expanding and will continue to do so for some time (ECLAC 2003c; Cuba from Barros 2003, Benítez 2003, ONE 2004). As more aged the population and more mature the pension system, lower the ratio of active insured workers per one pensioner, hence, more difficult to maintain financial sustainability in public systems; on the other hand, increasing life expectancy will make more difficult to maintain the pension level in private systems. ECLAC (2003c) projects that the population age 60 and over in the region will increase in 2000–25 from 40 to 96 million, and accelerate in 2025–50 to reach 180 million. The old cohort will grow at a five-year rate of 2.9% in 1995–2000 but 3.7% in 2020–25 (compared with population growth rates of 1.6% and 0.9% respectively), jumping from 8% of the total population in 2000 to 23% in 2050, hence, one of each four persons will be elderly. These regional projections vary significantly among countries, which ECLAC classifies in four demographic stages (advanced aging, moderate advanced aging, moderate aging, and incipient aging), the first two include most countries in the pioneer-high group, and the other two roughly correspond with the intermediate and latecomer-low groups. In developed countries the aging process has taken between six and ten decades, but only two or three decades in Latin America, hence the need to adapt social protection systems to demographic changes much faster than in developed countries (Bertranou and Bravo 2006). Historically pioneer-high countries have been the most affected by aging and had to significantly increase their contributions (Argentina and Uruguay), because of strong worker’s resistance to rise retirement ages or tighten the pension formula; on the other hand latecomer-low countries are the youngest and have the lowest contributions. As the structural reform eliminated the principle of solidarity among generations and each insured person finances it own pension, arguably the private system won’t suffer a deficit, will maintain the same contribution and insulate the pension system from the

132 Reassembling Social Security aging process (James and Brooks 2001). We have seen, however, that the reform-generated fiscal deficit is substantial and prolonged. Furthermore, due to rising life expectancy, pensioners will live longer and collect a pension for a larger retirement span, forcing either an increase in the retirement age or the contribution or a combination of both. As the population ages and the pension system matures, more people will retire and withdraw pension funds, fewer workers will contribute and capital accumulation in the total pension fund will decline. In Chile there were 1.7 million people within the ages 55–69 in 2005 but their number will increase to 3.9 million in 2050, they will retire and withdraw pensions from their fund, but the population within contributory ages 20–54 will decrease from 9.2 to 8.6 million (Bertranou and Bravo 2006). Argentina pension fund is projected to peak between 2020 and 2040 and decline in 2040–50 (MTESS-ILO 2005). ‘The argument that funding insulates pensioners from demographic change should not be exaggerated [it] is not a strong argument for a shift [from PAYG] to funding’ (Barr 2002: 10). In public systems the ratio of active insured workers per one pensioner is highest in the latecomer-low group (incipient aging) and lowest in the pioneer-high group (advanced and moderate-advanced aging). Brazil entire system ratio was 1.7, the lowest in the region after Uruguay whose ratio was 1.3 in 2004. Due to the aging process and low retirement ages, Cuba’s ratio was 2.8 in 2002 and is projected to fall to 1.5 in 2025. The following ratios pertain to the general program in 2000– 4 (lower ratios in separate schemes are given in some countries): Ecuador 4.2 but 1 in the armed forces and policemen; Nicaragua 4.6 (low for this country, due to stagnation in labor force coverage); Panama 5.9; Costa Rica 6.3 (high because it is relatively young population); Guatemala 7.5 but 3.5 in civil servants and armed forces; Venezuela 7.5 but projected to decline to 3.3 in 2035; Honduras 22, the second highest, but 11 in civil servants, 4.7 in teachers and 3 in armed forces; Haiti 50,

the highest in the region, but 2.8 in civil servants (sources in Table 6.5; Guatemala Apén 2002; Haiti IADB 2001; Uruguay BPS 2005a). The nonweighted average ratio of the public systems excluding Haiti and Honduras was 5.3 (7 including Honduras and 10.9 also including Haiti) which is high but, if weighted, it would be much smaller due to the low ratio of Brazil and its huge population (Table 6.5) Non weighted average ratios for the three groups are: 3.6 in pioneer-high, 5.9 in intermediate, and 19 in latecomer-low. Within public systems, pioneer-high countries have the lowest active/pensioner ratios, highest contributions, and little possibility to keep rising them, therefore they must increase retirement ages, extend the period of contribution, and tighten the pension formula (as Brazil has partly done). Conversely, countries in the latecomer-low and some in the intermediate group have relatively young populations, a productive cohort that should grow for some time and the smallest old cohort that will take time to expand significantly, hence they have the lowest active/pensioner ratios; because their contributions are the smallest and retirement ages lowest they have more time and space to increase both. This does not mean, however, that these countries don’t need a reform; on the contrary, as sooner it is introduced, easier the solution will be.

6.7. Reform goals/assumptions: private systems promotion of national saving, capital markets and returns, and immunity against political interference Four important assumed beneficial financial effects of structural reforms and private pension systems are: (a) significant capital accumulation in the pension fund; (b) increase in national saving; (c) development of capital and financial markets and new financial instruments that lead to portfolio

Financial Sustainability and Reform Goals

133

Table 6.6. Stock of financial assets accumulated in private pension systems in 2005 and reserves in public pension systems, between 2000 and 2005 Accumulated fund (2005) Private systems Argentina Bolivia Chile Colombia Costa Ricaa Dominican R. El Salvador Mexico Peru Uruguay Averagesb

Million US dollars

% of GDP

22,565 2,060 74,756 16,015 711 381 2,896 55,205 9,397 2,153 18,614

12.9 21.6 59.4 17.2 3.7 1.3 18.3 7.0 12.1 15.3 13.8

Reserves (2000–5) Public systems (with CPC) Brazilc Costa Ricaa Ecuador Guatemala Honduras Nicaragua Panama

Million US dollars

% of GDP

149,600c 1,071 2,122d 498 322 267 1,681

18.1c 6.2 7.8d 2.4 16.0 5.6 13.0

22,223

16.8

Sources: Private systems from AIOS (2006). Public systems from: Brazil ABEFPC (2006); Costa Rica CCSS (2005a); Ecuador from IESS (2006a), ECLAC (2004c); Guatemala Durán and Cercone (2002); Honduras IHSS (2003); Panama Mesa-Lago (2005a). Percentage of GDP in public weighted by the author based on GDP from ECLAC (2005c). No data available on reserves of Paraguay, Cuba, Venezuela, and Haiti. CPC = Collective partial capitalization (partly funded). a Second pillar (private) in first segment and first pillar (public) in second segment. b Weighted; public average subtracting Brazil declines to about US$1,000 million and 7.5% of GDP. c Brazil programs are PAYG, data refer to accumulated capital in voluntary supplementary pension funds. d Invested funds in 2003; reserves should be much higher and state debt is not included.

diversification; and (d) high capital returns greater than in public systems. These alleged effects, in turn, promote economic growth, employment, and better pensions (World Bank 1994; Vittas 2000; Walker and Leford 2002; Corbo and Schmidt-Hebbel 2003). In addition, private systems supposedly are protected against state and political intrusion. This section tests those five assumptions with data from all systems in the region.

6.7.1. Capital accumulation Table 6.6 confirms the first reform financial assumption, albeit with notable differences among countries. The accumulated capital in pension funds in million US dollars at the end of 2005 was: Chile 74,756, Mexico

55,205, Argentina 22,565, Colombia 16,015, Peru 9,397, El Salvador 2,896, Uruguay 2,153, Bolivia 2,060, Costa Rica 711, and Dominican Republic 381. All countries increased the fund in 2000–5, except for 2001–2 in Argentina (−44%) and Uruguay (−15%), due to the crises especially in the former (AIOS 2001–6). The pension fund is managed by a small number of administrators that concentrate an enormous and increasing economic power. The accumulated pension fund varies according to the time the system has been in operation, the number of insured, the size of the economy, the wage level, the rate of capital return, and macroeconomic performance. Chile’s reform has twenty-five years and has accumulated the biggest fund. Mexico reform is only ten years old but has the second biggest fund (74% of the Chilean fund) because the

134 Reassembling Social Security country has the highest number of insured and second largest economy in the region.17 Argentina’s economy is the major third and its pension fund was 60% of Chile’s in 2001, but the crisis reduced it to 30% in 2005.18 The smallest accumulations of Costa Rica and Dominican Republic result from the small size of their economies and number of insured and less time of reform operation. Capital accumulation has grown so fast partly because very few insured are retiring but it will eventually decline as explained above. The percentage of the pension fund relative to GDP depends not only of the sum accumulated but also on the size of the country GDP: 59% in Chile, 15–22% in Bolivia, Colombia, El Salvador, and Uruguay, 12–13% in Argentina and Peru, 7% in Mexico (the second biggest GDP) and 1–4% in Costa Rica and Dominican Republic (Table 6.6). Accumulation in public systems is considerably smaller than in private systems because, as already explained, some of them being on PAYG lack reserves (Brazil, Cuba, and Venezuela), whereas others are partly funded (CPC): Costa Rica (public pillar), Ecuador, Guatemala, Honduras, Nicaragua, Panama, and probably Paraguay. The highest capital accumulation in the region, however, is not in Chile but in Brazil voluntary supplementary pension funds: US$149,600 million and 18% of GDP in 2005 and such accumulation will be higher when supplementary pension funds for civil servants are implemented. Ecuador invested funds of the general program were US$2,122 million in 2003, similar to those of three private systems, and 7.8% of GDP exceeding three private systems. Panama’s fund had accumulated US$1,681 million in 2003, a sum higher than in two private systems/pillars, and its percentage of GDP (13%) was above five private systems. Honduras’ fund equaled 16% of GDP in 2000 surpassing six private systems, and Costa Rica’s first public pillar had accumulated US$1,071 million in 2004, 50% more than the accumulation in its private pillar. Albeit comparisons are distorted by Brazil’s huge accumulation in supplementary pension funds, the average

of six public systems and Costa Rica’s public pillar was US$22,223 million, 19% higher than the private average of US$18,614 million; in percentage of GDP the public weighted average of 16.8% was higher than the 13.8% private average; subtracting Brazil, public averages declined sharply to US$1,000 million and 7.5% of GDP (Table 6.6). In summary including Brazil, partly funded public systems fairly compare with private systems on capital accumulation, but excluding Brazil the claim of higher capital accumulation in private systems holds. World Bank officials admit that the only way to capital accumulation is not through the ‘heavy reliance’ on a mandated second pillar as in private systems; ‘countries such as Brazil that have reasonably well-developed capital markets may well choose to change the parameters of their public PAYG pension system rather than switch to a mandatory funded program’ (Gill, Packard, and Yermo 2005: 277).

6.7.2. Increase in national saving The second reform financial assumption has been theoretically refuted in the sense that the accumulation of the fund does not necessarily result in an increase in national saving: (a) when increases in savings occur they take place in the early years after the creation of the fund but, when the system matures, savings are equal to the cost of pensions; (b) increments in mandatory savings in the accumulation stage of the private system may be compensated, at least partially, by reduction in voluntary private savings,19 and could be accompanied by an equal rise in government borrowing; (c) savings do not lead per se to new investment, as could be in gold or debt paper; and (d) the increase in investment does not necessarily augments production because it depends on its productive use (Barr 2002; Barr and Diamond 2006).20 Empirical studies also question the assumption that private systems increase national saving because the accumulation in the

Financial Sustainability and Reform Goals pension fund must be related with the fiscal cost of the transition (see Section 6.3). Chile is the only country whose reform has had a sufficiently long period of time (twenty-five years) to test such assumption and most studies reach negative conclusions. Holzmann (1997) subtracted the reform’s fiscal costs (negative) from savings in private pensions (positive) and concluded that the impact of the reform on national saving was negative in 1981– 8. He was unable to prove a positive direct impact in 1989–96, and therefore advised Latin American countries against entertaining high expectations that the reforms would increase national saving. Arenas de Mesa (1999) measuring results in annual percentages of GDP in 1981–97 found that the capitalized savings averaged 2.7% but fiscal cost averaged 5.7% and the net result was −3% (dissaving). He projected that in 2000–5 savings would be slightly higher than fiscal costs and such net positive outcome will continue to grow, but it would take twenty years to offset the negative balance occurred in the first twenty years.21 Acuña and Iglesias (2001) subtracted the ‘the social security pension deficit during the transition’ from private pension savings but excluding the cost of minimum, armed forces, and social assistance pensions, still they calculated an average negative net result of −2.7% in 1982–97, somewhat lower than Arenas de Mesa’s, who included all fiscal costs of the reform. Observing private saving components in 1980–94, Barrientos (1998) found that corporate savings explained almost all the increment, while household savings were negative in all years except three, hence the pension reform didn’t have a significant impact on household savings. In turn, Uthoff (2001: 32) disaggregated the components of national saving in 1980–99 and concluded that ‘The direct effect of pension reform on national saving is at best very weak. In that period, national income increased 10 percentage points of GDP but only 2.5 points could be attributed to the reduction in the deficit of pensions.’ A supporter of the Chilean reform states that the assumption that the private system increases

135

national saving ‘is not correct in general’, hence he mollifies such assumption: ‘Moving the pension plan from PAYG to prefunding offers a unique opportunity to raise national saving’ (Valdés-Prieto 2001: 80, 82; emphasis by the author). Conversely, Haindl Rondanelli (1997) estimated that the reform had a positive impact on national saving in 1990–4, but based his conclusion on the general tax burden instead of the direct fiscal cost of the reform. If he had subtracted the average cost of the public system deficit of 4.6%, from the average private pension savings of 3.1%, he would have obtained a negative result of −1.5% even excluding other fiscal costs of the reform. In a paper commissioned by Chile’s association of pension-fund administrators, Corbo and Schmidt-Hebbel (2003) took into account only the operating deficit and the recognition bond thus excluding minimum, armed forces, and assistance pensions (similarly to Acuña and Iglesias). Upon that partial basis they estimated that the reform increased national saving by an annual average of 2.3% of GDP in 1981–2001, ranging from 0.7% to 4.6% contingent upon the type of financing of the fiscal deficit, the increase in obligatory savings in the pension fund and the behavior of voluntary savings. The authors noted that pension reform was part of wider structural reforms such as fiscal adjustment, financial liberalization, and labor and capital market changes, which made difficult to properly isolate the impact of pension reform.22 Notwithstanding, they generalized their results for Chile to the entire region albeit conditional on the depth of the reform and structural features of the economies: ‘As more radical the replacement of the PAYG system by one of capitalization and greater the financing of the transition deficit through fiscal adjustment, stronger will be the effects on accumulation . . . ’ (p. 37). World Bank officials judged this study econometrically sound but relying on assumptions about the nature of pension reform and economic growth that are far from certain, and noting that the impact of pension reform must be isolated from other structural reforms

136 Reassembling Social Security (Gill, Packard, and Yermo 2005: 49, 53; for other views see Kiefer 2004).

6.7.3. Development of the capital markets and portfolio diversification The third financial assumption of structural reforms is that private systems contribute to develop capital markets and create new investment instruments, which lead to diversification of the portfolio, helping to compensate risks among various instruments. Holzmann (1997) study on Chile concluded that the pension reform had helped to make financial markets more liquid and mature, and that empirical evidence coincided with the assumption that pension reform had contributed to the development of financial markets and a more diversified portfolio. Nevertheless, he cautioned that such evidence was not sound proof that the pension reform had been the decisive factor in the development of such markets since mid-1980s, because it could have been caused by other factors unrelated to the reform. On the contrary, Corbo and Schmidt-Hebbel (2003) sustained that the contribution of pension funds to Chilean financial market development was ‘quite robust’ (between 31% and 46% of financial development in 1981–2001) and recommend the most radical structural reform possible, to maximize this effect. World Bank officials assert that ‘the evidence that pension reform contributes to capital development is tenuous’ because the impact of such reform would have to be separated from other structural reforms: ‘capital market development in some countries has been driven largely by statesponsored modernization of the financial sector infrastructure, tax and bankruptcy reform, and the regulatory structure developed for pension funds and other financial institutions’ (Gill, Packard, and Yermo 2005: 49, 85). The percentage distribution of the pension fund portfolio by investment instrument in December 2005 shows that most countries are far from achieving adequate diversification (Table 6.7). In Argentina, Bolivia, Costa Rica,

El Salvador, Mexico, and Uruguay, from 60 to 82% of the portfolio was invested in public debt securities and 47% in Colombia. Only in Chile and Peru that share was a minority (16% and 20%) and zero in the Dominican Republic, but it took Chile twenty-four years to reduce the share from 46 to 16%, thanks largely to the superintendence role. An average of 46% of the portfolio in the ten private systems was in public debt instruments, but increased to 66% if Chile were subtracted because it accounts for 40% of total invested funds. Capital returns has been influenced by high interest rates for state debt paper (for instance, in Argentina until the end of 2001), but that is costly to the economy and cannot be sustained in the long run.23 The second most important investment, in financial institutions, averages 16%. Dominican Republic legal restrictions led to 97% of the private pension fund invested in shortterm certificate deposits from commercial banks and savings associations that pay considerably lower interest rates than the central bank hence commercial banks reinvested those funds in central bank certificates and earned juicy profits; the public system fund has a better portfolio distribution (Pérez Montás 2006). In Chile and Uruguay, 29 and 37% respectively of total investment is in financial institutions. Stocks are a favorite instrument to diversify the portfolio and fairly developed capital markets should offer multiple stocks for investment of the pension fund. But only in Peru investment in stocks was significant at 36%, it was 11–15% in Chile, Argentina, and Colombia, 6% in Bolivia and nil in the rest; the average share of the portfolio in stocks in the ten countries was 10.6% but 7.8% excluding Chile. Latin American capital markets don’t have large and highly liquid stocks and bond markets and trading volume are very small relative to market capitalization; stock value relative to GDP in 2003 was 12% in Brazil, 9% in Chile, 4% in Argentina and Mexico, and 0.5% in Colombia and showed a declining trend in all countries since 1994 (Matijascic and Kay 2006). The volume traded in Chile’s stock exchange is not high because of the vast

Financial Sustainability and Reform Goals

137

Table 6.7. Percentage distribution of pension fund in private systems and three public systems by investment in financial instrument, December 2005 Systems/ countries Private systems Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averagesa Public systemsb Brazilc Nicaragua Panama

Government debt

Financial institutions

Nonfinancial

Stocks

Mutual funds, etc.

Foreign emissions

60.9 70.0 16.4 47.3 72.1 0.0 81.0 82.1 20.3 59.5 46.4

5.1 6.8 28.9 10.4 13.2 96.8d 12.7 4.2 11.1 36.8 15.9

1.8 13.5 6.8 14.4 5.3 3.2 6.3 11.8 10.7 2.7 8.5

13.4 6.3 14.7 11.3 0.2 0.0 0.0 0.4 36.4 0.1 10.6

8.1 0.0 2.8 2.0 3.2 0.0 0.0 0.0 2.8 0.0 2.4

8.9 2.5 30.2 10.4 2.7 0.0 0.0 1.5 10.1 0.0 15.1

1.8 0.9 0.2 4.3 3.4 0.0 0.0 0.0 8.7 0.9 1.1

12.1 33.5 51.6

1.1 66.1 42.8

4.2 0.0 5.2

20.2 0.4 0.4

56.9 0.0 0.0

0.0 0.0 0.0

5.5 0.0 0.0

Others

Sources: Private systems from AIOS (2006). Public systems from Brazil ABEFPC (2006); Nicaragua INSS (2006); Panama Mesa-Lago, (2005a). a Weighted average distribution of total fund; Chile’s fund accounts for 40% of the total and heavily influences the averages. b Cuba, Haiti, and Venezuela are on PYGO, as well as Brazil whose data are on the supplementary pension funds; no data available on Guatemala, Honduras and Paraguay. c Other completely different distribution is: 50.6% government securities, 5.6% financial institutions, 6.8% nonfinancial, 30% stocks-equity, and 7% others (Pena 2007). d The private fund is virtually all invested in short-term CDs from commercial banks and savings associations.

concentration of the entrepreneurial sector, thus 45% of the most traded stocks are owned by only three holders hence the liquidity of the stock market is well below what should be expected in a country with such developed pension fund and the second highest stock-market capitalization relative to GDP. In smaller economies, the number of firms that trade in liquid markets is even lower. Peru’s five largest firms in the primary market concentrate 33% of total emissions and the ten largest 51%; only one stock is liquid enough to attract foreign investors. Chilean pension funds have played an important role in the expansion of the mortgage bond market, the most developed in the region, but this has not occurred elsewhere (Morón and Carranza

2003; Kiefer 2004; Gill, Packard, and Yermo 2005). Some experts assert that the existence of a mature domestic capital market is not an indispensable prerequisite for the privatization of the pension system (Iglesias and Acuña 1991). Others contend that in countries with very limited institutional capacity, the existing financial structure is too weak to absorb the flow of funds from fully funded systems (Barr and Diamond 2006). There is evidence in the region against the first position and supporting the second. Chile had a capital market many decades prior to the reform, albeit less developed than now, but after twenty-five years of reform, serious flaws subsist: administrators can invest, jointly or separately, in

138 Reassembling Social Security the same enterprise and influence the value of its stock; insurance companies that are shareholders in pension firms can also invest in their stock, and the huge capital inflow vis-àvis the relatively limited instruments available in the market generates overvaluation of some instruments (Arenas de Mesa and Mesa-Lago 2006). Even in Brazil and Mexico capital markets are too small to absorb the investment from a mature fully funded regime. Peruvian pension funds have 36% of market securities and the fund is growing at a much faster pace than the capital market. Despite considerable time elapsed after the reform in other seven private systems, their portfolio remains significantly concentrated on public debt with zero or little investment in stocks. Costa Rica pension superintendence acknowledges that its efforts to diversify the portfolio have failed because of the lack of financial instruments in the market (SPa 2005b). Legislation sets percentages (often minimum and maximum) of the pension fund portfolio to be invested in specific instruments, and in several countries the maximum percentage established for public debt securities is very high. In Mexico 100% of the portfolio can be invested in public debt, but only 35% in private debt and 10% in financial entities, leaving little space for the private sector (Rubalcava and Gutiérrez 2000). El Salvador’s law permits investment from 80 to 100% in public debt and up to 85% in other public securities, but only 20% in stocks of private enterprises and exempts treasury bonds from risk rating (Mesa-Lago 2003a). Lacking adequate domestic instruments for investment, foreign instruments should be an alternative but some countries forbid it because is judged detrimental to national economic interests and provoke strong political opposition. After a long and heated debate on this issue, in 1992 Chile approved foreign investment with a ceiling of 3% of the portfolio, a limit gradually lifted to 30% (to 80% by the legal draft of reform in 2006; Consejo Asesor 2006). The share invested in foreign instruments in 2005 was: 30% in Chile, 9– 10% in Argentina, Colombia, and Peru, 2.5% in Bolivia and Costa Rica, and virtually zero

in the remaining five countries. The average in the ten countries was 15%, higher than the average share in domestic stocks (10.6%), but subtracting Chile said average decreased to 5%. Costa Rica’s reform law of 2000 authorized a limit of 25% of investment in foreign instruments, while the regulations allow as much as 50% but at the end of 2005 only 2.7% was in that instrument. Mexican laws of 2002 and 2004 authorized similar investment with ceilings of 12 and 20% respectively (ISSA 2003a, 2005), but only 1.5% was invested in 2005. Argentina’s economic crisis provoked a fall in the pension fund value and an increment in the portfolio concentration on public debt. In 2001 the government pressured fund administrators to convert instruments valued in dollars and traded in international markets into ‘guaranteed loans’ with lower interest; later various decrees forced administrators to invest earnings from banking CDs and cash into debt paper. The superintendence collaborated with these state measures by lifting the ceiling of investment in state debt paper and not counting the indirect investment in such instruments by bank funds so not to exceed the legal limit. In 2002 the government converted the ‘guaranteed loans’ and other dollar instruments into Argentine pesos and further reduced interest rates. The peso was par with the dollar but in mid-2002 fell to 3.5 pesos per dollar, due to the severe crisis, the suspension of payments on the foreign debt and the elimination of the exchange parity, however, debt instruments held by pension funds were exchanged at the rate of 1.4 pesos per dollar adjustable to inflation. The percentage of the portfolio invested in public debt jumped to 78% at the end of 2002. As a result, the value of the portfolio lost 65% based on dollars and devaluation projections albeit a reported gain based on pesos adjusted to inflation or tied to the wage-index variation. In 2003 the government decided to pay only 25% of the nominal value of the debt in default, possibly having a further negative effect on the value of the pension fund (ILO 2002b; Alós, Muiños, and van Cauwlaert 2003; Bertranou, Rofman, and Grushka 2003; González

Financial Sustainability and Reform Goals 2003; Hujo 2004). Salvadorian experts fear a potential default a la Argentina due the excessive and growing dependence of the pension portfolio on state debt (Rosales 2006). World Bank officials have evaluated the positive effects of private pension funds on capital and financial markets, as well as on investment, but separating such effects from other factors and policies, and have also identified pension fund flaws: (a) private systems have achieved some of the highest standards in the region in asset evaluation, risk rating, and disclosure, whereas transparency and integrity of financial markets have significantly improved but those advances ‘could have taken place independently of the pension reform’; (b) ‘in terms of investment and performance goals, pension funds are hardly different from mutual funds, hence, it is possible that similar benefits could have been obtained had the pension fund regulatory framework also been applied to the mutual fund industry’; (c) the transition fiscal costs burden has forced many governments to impose restrictions on stocks and foreign securities and channel pension funds toward public debt securities, hence there is low investment in financing the private sector, ‘pension funds are effectively used by governments as captive sources of finance’; (d) the illiquidity of pension investments coupled with conservative strategies have brought stability in capital markets but ‘much of this stability is artificial [because] it is driven at least in part by portfolio rules that force pension funds to hold mainly domestic assets’ particularly state bonds; and (e) ‘the instability created by a large transition debt is an obstacle to the deepening of financial markets; in the absence of a sustained fiscal effort, therefore, transition costs can severely curtail the positive impact of pension funds in capital markets.’ The officials ask if the reform economicfinancial goals should be ‘ends in themselves or simple means to improve the welfare of citizens’, that is, ‘are people in the reforming countries better off because of the adoption of [private systems]?’ Their answer is that ‘workers may not judge the new system quite as kindly as do fiscal and financial specialists’

139

(Gill, Packard, and Yermo 2005: 53–4, 68–70, 85–6). There are scarce, incomplete and contradictory data on the portfolio distribution in partly funded public pension systems/pillars that have reserves and investment (see Table 6.7). In 2005, Brazil supplementary pension funds were invested 57% in mutual funds, 20% in stocks and only 12% in government debt, as diversified as Chile and Peru, but investment in foreign emissions was banned (but other distribution gave 51% in government debt and 30% in stocks). On the other hand Nicaragua’s distribution was 66% in CDs, 33.5% in public debt and only 0.4% in stocks. Panama’s portfolio in 2004 was 52% in state debt, 43% in financial institutions (bank deposits mostly in the central bank), 5% in nonfinancial institutions (mainly mortgage and personal loans to insured and pensioners) and only 0.4% in stocks; investment in foreign instruments was banned. Costa Rica’s public pillar had 82% invested in government debt and Colombia public system had 89%, both percentages higher than in their private counterparts (CCSS 2005a; SBC 2005). Part of Ecuador reserves were deposited in the central bank without earning interest in 2003, 72% was invested in financial instruments and 28% in nonfinancial investment mainly personal and mortgage loans to insured and pensioners (that had disastrous results in the past), as well as administrative offices and facilities, all with very low real returns (World Bank 2005e; IESS 2006a). Guatemala’s portfolio was 85% in public debt or deposited in the central bank in 2004, earning very low interest rate (IGSS 2004). Honduras’ was mostly invested in bank deposits in 1999, partly in the central bank that paid interest similar to inflation; the reform legal draft allows 40–60% of investment in public debt but zero to 15% in each stocks and foreign instruments (Congreso 2005). Complete and comparable statistics on public systems/pillars are needed to reach solid conclusions, but available data indicate that portfolios in private and public systems, with few exceptions, are little diversified and excessively concentrated on state debt instruments,

140 Reassembling Social Security

Table 6.8. Average annual real gross rates of capital returns in private pension systems and in public systems with invested financial reserves (partly funded: CPC) Private systems

Average capital return (%)a

Argentina Bolivia Chile Colombia Costa Rica Dominican R. El Salvador Mexico Peru Uruguay Averaged

9.2 9.1 10.1 5.9 6.5 −1.0 8.6 7.5 8.4 11.6 7.6

Public systemsb Costa Ricae Guatemala Honduras Panama

Average capital return (%)c 7.7 10.4 6.2 5.0

7.3

Sources: Private systems from AIOS (2006); except Chile SAFPb (2006). Public systems from Mesa-Lago (2004, 2006c); Costa Rica author’s estimates based on World Bank (2003) and CCSS (2005a). a Real annual average return from the inception of each system until June 2006 (averages of Argentina, Colombia and Peru exclude the first 2–3 years). b These four countries are not on PAYG but CPC (also Ecuador, Nicaragua, and Paraguay but no data were available); Brazil, Cuba, Haiti, and Venezuela have PAYG and no financial reserves; Brazil supplementary pension funds have substantial reserves and investment but no data were available on capital returns. c Annual average: Costa Rica 1992–2005, Guatemala 1999–2000, Honduras 1994–2002, and Panama 1996–2004. d Nonweighted. e Public pillar.

problems intensified in small countries without or with incipient capital markets.

6.7.4. Capital returns The fourth financial assumption of structural reforms is that private systems have higher capital returns than public systems. Orszag and Stiglitz (2001) theoretically refute that assumption while Barr (2002) notes that a proper comparison requires the inclusion of transition costs, as well as risks and administrative costs of both systems. Holzmann (2001: 58) partly accepts these counterarguments: ‘Rates of return in systems of individual accounts are not necessarily higher [than in public systems], but they could be if administrative costs can be controlled [or] transition

costs are low or nonexistent in systems that are in an initial stage.’ Nevertheless, it has been demonstrated that administrative costs are higher in private than public systems and that transition fiscal costs are high and prolonged in most private systems; this section shows that the average capital return in four public partly funded systems is similar to that in the ten fully funded private systems. The average real capital return (adjusted by the CPI), since the inception of the private system until mid-2006 was: 12% in Uruguay, 9–10% in Argentina, Bolivia, Chile, and El Salvador, 6–8% in Colombia, Costa Rica, Mexico, and Peru, and −1% in Dominican Republic; the nonweighted average for the ten countries was 7.6% (Table 6.8). These are gross rates of return without subtracting the cost

Financial Sustainability and Reform Goals of the net commission for old-age administration, hence net returns are lower; the regional association (AIOS) does not publish them because of the difficulties involved: commissions and capital returns may vary from month to month, and commissions are mostly charged as a percentage of salaries, whereas capital return rates are based on the fund (see Morón and Carranza 2003). However, when the cost of the commission is subtracted in Chile, the total average return in 1982–98 was 5.1%, not the 11% calculated by the superintendence based on gross rates. The average worker would have done better, earning a real average return of 7.2%, depositing all his contributions in a savings account (cited by Matijascic and Kay 2006). In 1981–2000, gross capital returns in Chile averaged 11.9 percentage points less than the returns of the Selective Stock Price Index in Santiago stock exchange (Acuña and Iglesias 2001). Argentina’s net capital returns (deducting administrative costs) were lower on average than returns from fixed-term deposit certificates in 1995–2003 (MTESS-ILO 2005). Peru’s average returns in 1993–2000 were lower and less stable than the interest rate paid by bank deposits and Brady bonds; and the pension fund cumulative return in 1994–2002 was 20% lower than NASDAQ and 48% lower than SPP (Morón and Carranza 2003). The Dominican Republic had a negative capital return (−22%) in the first year of its system operation due to high inflation; the abrupt decrease in inflation prompted a better return but still negative (−3%) by mid-2005 (Pérez Montás 2006). The monthly rate of return in five private systems since their inception until the end of 2000 had one of the highest standard deviations in Latin America and the return by unit of risk was the lowest in the region (Gill, Packard, and Yermo 2005). The above average capital returns are for the entire period of operation of the reforms, but measured from the system inception to the mid-1990s, the average was much higher and thereafter significantly lower, due to economic and capital market crises in 1995, 1998, and 2001. For example, Chile’s averaged

141

14% in 1981–94 and 5% in 1995–2004 with −2.5% in 1995 and −1% in 1998, whereas Argentina’s averaged 20% in 1994–7, 7% from mid-1997 to mid-2001 and −14% in 2001 (AIOS 1999–2006; SAFPb 2002–5). Such fluctuations involve an important risk: if the insured retire during a capital market boom the pension will be good, but if there is a prolonged crisis the pension will shrink drastically as occurred in Argentina. Therefore two insured persons with similar wage and contribution histories may receive significantly different pensions depending on capital returns and the time of retirement (Barr 2000). Such risk might be attenuated in mixed systems because they combine two pillars, one of defined benefit and another based on the individual account. Gill, Packard, and Yermo (2005) note that high gross returns on pension funds portfolios resulted from elevated interest rates on state debt-issues raising two concerns: how long such high returns can be sustained as fiscal adjustment lowers the spreads on government debt, and part of the high returns reflects the risk of default where fiscal adjustment is slow, as Argentina illustrates. Performance of mandatory and voluntary savings in the second and third pillars has not been good from the worker perspective because of capital return volatility, high commissions, poor development of insurance markets and adverse risk selection and rigidity in annuities, all of which increase the cost and relative risk of private systems compared with public systems. Virtually all countries have rating-risk agencies (normally public although private too) to evaluate instruments, crucial for investment by administrators and for security to the insured. On the other hand, the existing guarantee of the annual minimum capital return in most countries, puts pressure to invest in short-term instruments and is a disincentive for diverse behavior because, if the return is negative, the administrator must use its reserves and replenish them, which generates a ‘herding effect:’ all portfolios tend to be similar because administrators avoid

142 Reassembling Social Security venturing into instruments with high risk of negative return in the short run although may have potential for high return in the long run. That problem could be tackled if the period for the calculation of the minimum return is extended from one to two or three years, although such policy would have opposite effects: potential for higher competition, portfolio diversification and returns, but also higher risk for the insured. The previous discussion demonstrates that the structural reform and a private system do not ensure by themselves the development of the capital market, portfolio diversification and high capital returns; other conditions and policies are necessary. Evidence also questions the assumption that a capital market is not a necessary precondition to structural reform because it provides the needed impetus for such market. The Chilean success has been partly due to its capital market being founded at the end of the nineteenth century and later tightly regulated to ensure minimum competition for long-term investment, as well as transparency, confidence, and opening to foreign instruments. Conversely, small countries like Bolivia, El Salvador, Costa Rica, and Uruguay had no securities markets before the reform or they were very small or incipient with few investment instruments and highly concentrated. El Salvador hastily promulgated a law creating and regulating the securities market just before the start of the reform. These countries have the least portfolio diversification and several of them obstructed potential diversification by banning investment in foreign instruments. Even countries with more developed capital markets, such as Argentina and Mexico, have been unable to diversify their portfolios because of legal rigidity, political pressure and in Mexico prohibiting investment in foreign emissions until recently and without tangible results. The heavy concentration of the portfolio in public debt in most countries, submits capital return to the interest rate set by the state, a high rate increases the capital return but cannot be maintained in the long run, while a low rate leads to poor capital returns.

The economic environment for the reform is crucial also. Chile’s macroeconomic policy and its results have been excellent for a long time: high economic growth, controlled inflation, strict fiscal discipline, a successful indexation mechanism, a positive and moderate interest rate, a realistic exchange rate, and virtual elimination of the state and central bank external debt. Three years before the reform, the government generated a fiscal surplus of 5.5% of GDP, thus forestalling the adverse impact of the reform fiscal cost on the budget balance and such surplus has been maintained subsequently except for a couple of years, a unique case in the region. Finally, the superintendence acted independently from the government, developed and enforced strict regulations to impede excessive risks and fraud, promoted portfolio diversification, and pushed for the authorization to invest in foreign instruments and gradually expand it. Argentina is the adverse opposite case: poor fiscal discipline, unrealistic exchange rates, lack of resources to finance the transition fiscal costs, high interest rate paid by public debt that was unsustainable in the long run, and a superintendence that coalesced with the government to increase pension fund investment in public debt that was later devaluated (Uthoff 2003; Mesa-Lago 2004; Gill, Packard, and Yermo 2005). Capital return rates are available only for four public systems (not for Ecuador, Nicaragua, and Paraguay) that are partly funded with reserves and financial investment, for different periods between 1992 and 2005: 10.4% in Guatemala, 7.7% in Costa Rica (public pillar), 6.2% in Honduras, and 5% in Panama, for an average of 7.3%, slightly lower than the 7.6% average of private systems. Guatemala’s rate was only surpassed by Uruguay’s and was similar to the Chilean rate; Costa Rica’s rate was identical to those in Mexico and Peru and higher than its own rate in the private pillar, as well as the rates of Colombia and Dominican Republic; the Honduras rate was similar to those in Colombia and Costa Rica and higher than in Dominican Republic; but Panama’s rate was the lowest

Financial Sustainability and Reform Goals except for Dominican Republic (Table 6.8). It should be recalled that four public systems are on PAYG and hence lack reserves and financial investment; Brazil although on PAYG has a huge fund in voluntary supplementary pension schemes but aggregate data of its rate of return were not available. Honduras’ pension fund was mostly invested in bank deposits, part in the central bank paying interest similar to inflation, conversely Guatemala’s is largely invested in public debt or deposited in the central bank and should earn adequate interest in view of its high return rate (Durán 2003; IGSS 2005). In Panama 33% of the fund was deposited in the central bank in 2004 and earned an interest rate of 1.5% versus 6.4% in the savings banks, 6–9% in public debt and 8% in mortgage bonds or loans, thus pulling down the overall capital return rate; capital returns would increase significantly if central bank deposits were invested in commercial bank CDs or other instruments (Mesa-Lago 2005a). Although public pensions are less affected to market volatility than private pensions, because of the guaranteed defined benefit, they are not immune to lack of portfolio diversification and sustained low rates of returns that may eat into their reserves and provoke actuarial disequilibria.

6.7.5. Immunity against state and political interference Before the structural reforms, the government played a harmful role in many public pension systems: (a) used their funds to finance the fiscal deficit, forcing the investment of reserves in public debt or their deposit in the central bank, without adjusting the principal to inflation and often paying interests below the market rate; (b) was the main debtor to social insurance (both as an employer and third party) and, when it signed debt payment agreements, usually didn’t adjust the principal to inflation or honored the accords; (c) secured a majority of state representatives

143

in social insurance boards and controlled or influenced key decisions; and (d) removed and appointed employees in the institutes, often without proper skills or assigned to superfluous activities (Sections 5.8 and 6.2). Structural reform supporters claimed that the ownership of the individual account, the principle of equivalence and the private administration of pension funds would impede the traditional state and political interference in public systems. Such attractive promise, embellished by advertisement, was instrumental in countries that granted freedom of choice to promote the insured shift from the public to the private system, and to justify such shift in those countries where the state forced all insured to move to the private system. And yet the World Bank (1994: 203) was initially conscious of two potential problems: If governments have mismanaged their centrally administered pension plans how they can be counted on to regulate private funds effectively?, and if government regulates and guarantees the plans, won’t it eventually end up controlling these funds? To confront these dilemmas the Bank set as a condition for structural reform success that the state behaves well. Reacting to the World Bank report of 1994, the ILO deepened the second concern: the important role of the state in a private system as regulator, supervisor, and financier of the transition can open the door to the government negative intervention, for instance, exerting pressure to invest the fund in public debt and impeding portfolio diversification (Beattie and McGillivray 1995). Barr (2002) warned that fiscal imprudence could generate inflation and decapitalize private pension funds while inadequate regulation of financial markets could harm the stability of such funds; furthermore the government could cut the capital return rate in private systems requiring their administrators to invest in public debt and paying interest lower than the market rate or abolishing or reducing tax benefits. Orszag and Stiglitz (2001) added that least developed countries lack capital markets or they are incipient, whereas public

144 Reassembling Social Security agencies regulatory capacity is weak therefore the potential for state intrusion is greater. Before the Argentine crisis, the author noted that high interest rates paid by public debt securities were responsible for the abnormally elevated capital returns in the private system, and alerted that such rates were unsustainable in the long run (Mesa-Lago 2001b). Barr and Diamond (2006: 21) conclude: ‘Political risks affect all pension schemes because all depend critically—albeit in different ways—on effective government.’ The Argentine crisis demolished the assumption of political immunity of private pension systems: not only governments can behave badly pressuring and forcing administrators to invest most pension funds in public debt paper, as experts had predicted, but also can convert dollar-valuated instruments into peso-valuated instruments, then devalue the currency and cut the state interest rate, plummeting drastically the pension fund value. A survey by the pension superintendence before the crisis found that 66% of the insured believed that the private system was dependent on political changes (SEL 2000). But Argentina is not unique, thus Bolivia’s government bonds denominated in US dollars had an 8% interest that was unsustainable and the rate was cut to 3.6% in 2003; the state also forced pension administrators to buy shares from the fund that finances Bonosol to get resources to pay the latter; and recently has tried to convert public debt instruments denominated in dollars held by fund administrators into national currency-denominated bonds with lower value and higher risk (Garrón 2004; Müller 2004). The Colombian government charged two extra pensions annually to the public pension system. Dollarization of El Salvador economy provoked a strong fall in the state interest rate and capital returns in 2002; furthermore, transition fiscal costs have been higher than officially and erroneously projected, the government has not created the legally stipulated amortization fund to finance such deficit, and budget resources cover only 12% of current costs, hence

delaying the emission of the recognition bonds (Mesa-Lago 2003a). Before Ecuador private system was scheduled to start, the state debt to social insurance was about US$3,000 million and critics asserted that if the debt was not paid the first pillar would start bankrupt; the Supreme Court ruling of unconstitutionality of the reform law halted the system implementation (Universidad 2003). More than a decade after the World Bank report of 1994, Bank officials acknowledge: It is often claimed that [privately managed pension systems offer] a greater degree of protection [than] they would under a public PAYG regime . . . The ability of the multi-pillar model to isolate the pension system from abuse by governments may have been oversold by reformers . . . the degree of protection against policy risk offered by privatizing . . . pensions can be exaggerated . . . the crisis in Argentina illustrates how any government organized retirement security system . . . can fell prey to politicians [but] similar threats to the viability of funded pension schemes can emerge in other countries of the region (Gill, Packard, and Yermo 2005: 5, 53, 133).

The new policies proposed require a ‘benevolent policymaker whose aim is to maximize the welfare of the population’ but ‘unfortunately, reality is often far from this ideal scenario, as governments in Latin America have been particularly apt at proving’ (Gill, Packard, and Yermo 2003: 202). The foundations of such policies are exposed to the danger of state interference: neither the poverty prevention (public) first pillar nor the mandatory and voluntary savings (private) second and third pillars can be completely insulated from policy risk. The first pillar requires a government-mediated transfer of resources between workers and ‘risk will always be present . . . , especially when governments are unable to maintain fiscal discipline and macroeconomic stability’; on the other hand, ‘privately managed programs, as the Argentine example clearly shows, can also easily fall prey to cash-strapped governments’ (Gill, Packard, and Yermo 2005: 261–3).

Financial Sustainability and Reform Goals

6.8. Impact of the reforms on financial sustainability and reform goals/assumptions The replacement of a public by a private system doesn’t necessarily solve the financial problems in the former: more than half of total affiliates in the ten private systems do not pay their contributions punctually; the move from PAYG to fully funded increases fiscal expending in the short-medium term if the system continues to meet its accrued PAYG obligations during the transition; projections of fiscal costs during the long transition have underestimated such costs in most countries, and rising life expectancy will force an increase in contributions, the retirement age or both. On the other hand, the generalized vision that public systems are financially unsustainable should be illuminated with sophisticated differentiation among countries: the pioneer-high group faces such predicament unless urgent parametric reforms are implemented, but the situation of some countries in the intermediate group and most in the latecomer-low group is better.

6.8.1. Private systems impose higher contributions on workers An average of 56% of the total contribution on the wage bill in private systems is paid by workers and 44% by employers, whereas in public systems the proportions are 38 and 62%. The employer contribution was eliminated in Bolivia, Chile, and Peru, and the worker pays 100% of the total contribution (55% in Uruguay), hence violating the 50% maximum set by the ILO (the remaining countries) meet such norm; half of the structural reforms augmented the worker contribution. None of the public systems have eliminated or reduced the employer contribution and they comply with the ILO norm, except in Ecuador and Panama; in seven public systems the employers’ contribution is from 1.5 to 7 times the workers’. The nonweighted average

145

total contribution in private systems (excluding the state) is 14% compared with 10.6% in public systems. Regardless of the type of system and with few exceptions, total contributions are highest in pioneer-high countries, due to the maturity of their programs and aged populations, and lowest in latecomerlow countries because they have less mature programs and younger populations. In the private systems of Costa Rica and Mexico the state contributes to the general program as a third party though a percentage of the wage bill. In two public systems the state finances the significant deficit of the pension programs (Brazil and Cuba) but in six of them the state is legally obliged to contribute a percentage on the wage bill or pay part of pension costs to the general program but does not meet that obligation (Ecuador, Guatemala, Honduras, Panama, Paraguay, and Venezuela). Separate schemes in private and public systems have contributions often smaller to those of the general program or do not pay at all and receive substantial state subsidies. The self-employed percentage contribution is from twofold to eightfold that assigned to the salaried worker in twelve countries; but is the same in the three private systems that eliminated the employer contribution. Selfemployed contribute as average 12% versus 8% salaried workers in private systems and 9% versus 4% in public systems, an insurmountable barrier to their coverage in both unless the employer contribution is matched by the state or subsidized with fiscal transfers targeted on the low-income.

6.8.2. Compliance deterioration The assumption that ownership of the individual account and the principle of equivalence in the private system results in higher pensions thus creating strong incentives for punctual payment of contribution was rejected: in eight of the ten private systems only between 37 and 53% of affiliates are active contributors, and the weighted average of the ten systems declined from 58 to 42%

146 Reassembling Social Security in 1998–2006. The World Bank sustains, however, that after the reforms there was a higher contribution density in the cities of Santiago and Lima (not at the national level), although acknowledges that the improvement in incentives attributable to the private system has not been rigorously proved. Actually panel surveys in Chile show that the national contribution density decreased from 52% in 1980–2002 to 48% in 2000–2. Explanations given for noncompliance are: exits from the labor market, unemployment, changes from formal to informal employment, temporary jobs, perceived higher risks in private than public systems, high and increasing contributions imposed on workers, employer retention of worker’s contribution (Chilean employer’s debt increased sixfold in the 1990s and equaled 1% of pension fund value), as well as declining or halting probability to contribute after the worker gains the right to the minimum pension (in Chile and Peru), because of higher priority to other alternatives that are less risky, costly, and illiquid (housing, business, life insurance, children education, and health care). Collection of contributions is centralized in all public systems and six private systems, and decentralized in four private systems; there is no evaluation of how the different type of collection may affect compliance. No systematic data on compliance are available on public systems, an important vacuum that impedes proper comparisons. The state legal contributions to six public systems are usually not honored or done irregularly, and the state is a major debtor. Estimates of the percentage of salaried workers that contribute to all social insurance programs in eleven countries (seven private and four public) show that two public systems performed better than four private systems (Panama and Chile were identical), and another public system performed better than two private systems. Disaggregated data on pensions and from other countries are needed but available figures suggest that, regardless of the type of pension system, compliance is higher in the pioneer-high group, declines in the intermediate group and is lowest in the latecomer-low group.

6.8.3. Projection of fiscal costs in private systems The World Bank and supporters correctly argue that structural reforms reveal the implicit pension debt (IPD) hidden in public systems. But they also affirm that fiscal costs in the transition (operational deficit, recognition bond, and minimum pension) will gradually decline and end. The disappearance of most fiscal costs (operational deficit and recognition bond) is estimated to take between forty and seventy years, but the cost of minimum and assistance pensions is expected to increase. Furthermore, ample evidence indicates that such costs are considerably higher than initially projected in some countries, will take longer than anticipated, and will grow in the next forty years in five of the eight countries with World Bank projections. In the three countries where fiscal costs are expected to decrease (Bolivia, Chile, and Uruguay) domestic data contradict the Bank estimates and projections. In 2004 Chile has the highest fiscal cost (5.5% of GDP) after a quarter century of the reform, largely because it offers the most generous rights/benefits during the transition, but such cost has been financed by steady annual substantial fiscal surpluses, not replicated elsewhere. Argentina projected fiscal costs as 2% of GDP but they actually ballooned to an average of 4.6% before the crisis and the country macroeconomic policy failed, whereas Bolivia’s initially projected cost was 2% in 2000, but actually was 5.3% in 2003. Five reforms after Chile tried to reduce fiscal costs by sacrificing benefits of the insured and pensioners: Bolivia, Dominican Republic, Ecuador (partly unconstitutional law), El Salvador, and Peru.

6.8.4. Financial and actuarial equilibrium in public systems Six public systems are partly funded (Ecuador, Honduras, Guatemala, Nicaragua, Panama, and Paraguay), as well as Costa Rica’s public pillar, whereas the systems of Brazil, Cuba,

Financial Sustainability and Reform Goals and Venezuela are on PAYG (Haiti’s status is unknown). We lack standardized projections of the balance of public systems similar to those of the World Bank on the fiscal cost in private systems, but there are actuarial projections in seven public systems and one public pillar. The two countries in the pioneer-high group face the most severe problems of financial sustainability, while those difficulties are attenuated in two countries in the intermediate group, and the best situation is enjoyed by countries in the latecomerlow group. Brazil parametric reforms introduced notional defined contribution in the RGPS and tightened entitlement conditions and benefits in civil servants schemes, which have improved but not achieved financial sustainability; Cuba faces the worst financial and actuarial disequilibria and so far measures to correct them have not been passed; Venezuela’s parametric reform has been enacted but not implemented. The financial balance is a deficit in Brazil, Cuba, and Venezuela (as well as in Panama since 2003) and a surplus in Ecuador, Guatemala, Honduras, Nicaragua, and Paraguay (as well as Costa Rica’s public pillar). The equilibrium contribution in Cuba, Nicaragua, Panama, and Venezuela would have to be increased significantly (still below the level of current total contributions in Argentina and Uruguay); the current contribution in Ecuador and Honduras can sustain the system for a long period, and that of Guatemala is estimated to be excessive. The average deficit resulting from the financial balance in the ten public systems (plus Costa Rica’s public pillar) is −0.5% of GDP, one fourth of the average deficit of −2.7% of GDP resulting from fiscal costs in eight private systems, but these two indicators are not technically comparable.

6.8.5. Population aging All pioneer-high countries are in advanced or moderate-advanced aging demographic stages, while intermediate countries are in the moderate aging stage and most latecomer-

147

low countries in the incipient aging stage. In public systems of the pioneer-high group the advanced aging process, pension scheme maturity, and decline in the ratio of active workers per pensioner, worsen the financial disequilibrium, and forced an increase in their contributions in one; such problems are not currently serious in latecomer-low public systems in the pioneer-high group because they have younger populations and pension schemes and a higher active/pensioner ratio (ratios average 3.6 in pioneer-high group, 5.9 in intermediate and 19 in latecomer-low). Structural reformers claim that private systems won’t be affected by the aging process because they have eliminated solidarity among generations, are on defined contribution, each insured person finances its own pension and the system won’t suffer a deficit, a claim that helped to shift insured from public to private systems. But rising life expectancy (particularly in the pioneer-high group) will force an increase in the contribution in private systems or a reduction in the level of the annuities or both. Furthermore, the high and prolonged fiscal cost of the transition could generate pressure for the state to default its obligation of financing public pensions or restrict some guarantees in the private system such as the recognition bond or the minimum pension.

6.8.6. Reform goals/assumptions: promotion of national saving, capital markets and capital returns, and immunity against state-political interference Reformers claim that contrary to public systems, private systems promote capital accumulation, national saving, capital market development, portfolio diversification and higher capital returns, and are immune to government-political interference. The only assumption confirmed by evidence is capital accumulation and only if Brazil supplementary pension funds are excluded, whereas portfolio diversification, capital returns, and

148 Reassembling Social Security political immunity were rejected, and there is no solid evidence regarding national saving and capital markets.

6.8.6.1. Capital accumulation There has been a notable capital accumulation in private systems, although with significant variation between countries resulting from period of reform operation, number of insured, size of the economy, wage level, rate of capital return, and macroeconomic performance. At the end of 2005 capital accumulation equaled 59% of GDP in Chile, 12– 22% in other six private systems, 7% in Mexico (the second largest economy), and 1–4% in the remaining two systems. The biggest accumulation measured in US dollars, however, was that in the voluntary supplementary funds to the public system of Brazil (twice that of Chile), that ranked third for its 18% of GDP (2.5 times that of Mexico) despite being the largest economy in the region. Two partly funded public systems (Honduras and Panama) had accumulations of 13–16% of GDP higher or comparable to six private systems; Costa Rica’s public-pillar reserves were 50% higher than its private-pillar capital accumulation, and the other three partly funded public systems had accumulations comparable to those in three private systems. The weighted average accumulation and percentage of GDP in the ten private systems were lower than corresponding averages in seven partly funded public systems/pillars including Brazil; subtracting Brazil, however, accumulation and percentage wise, private averages were considerably superior to public ones.

6.8.6.2. Net impact on national saving and development of capital markets The net impact of pension fund accumulation on national saving, done annually subtracting fiscal costs, has been measured by six studies in Chile, the only country with a period of reform operation long enough to sustain such analysis, as well as the most successful reform. The majority of such studies concluded that

there was negative savings averaging about −3% of GDP in the first sixteen years of the reform and one of them warned Latin American countries against expectations that the reform would increase national saving. Conversely, two studies estimated that the impact was positive, albeit with divergent magnitude, excluding components of fiscal costs, and based on assumptions about the nature of pension reform and economic growth that are far from certain. One of those studies concluded that the pension reform had helped to make financial markets more liquid and mature, and that empirical evidence coincided with the assumption that pension reform has contributed to the development of financial markets, but cautioned that such evidence was not sound proof that the pension reform has been the decisive factor in the development of such markets. On the contrary, another study sustained that the contribution of pension funds to financial market development was quite robust and recommend the most radical structural reform possible, to maximize such effect. The World Bank, however, considers that evidence that pension reform contributes to capital development is tenuous because its impact have to be separated from other parallel structural reforms that could have contributed to such development and it is extremely difficult to isolate the effects of various reforms.

6.8.6.3. Portfolio diversification Most private systems have failed to diversify their pension fund portfolios: between 60 and 82% of investment in six of the ten systems was concentrated on public debt paper in 2005. It took twenty-five years in Chile to cut that type of investment from 46 to 16%, largely due to the positive role of the superintendence. Only in Peru the investment share in stocks was significant (36%), in four other countries it was 6–15%, and zero in the remaining five; some countries ban investment in foreign instruments and only Chile had a share of 30%, in five other countries the share was 2–10%, and virtually

Financial Sustainability and Reform Goals zero in the rest. Brazil’s supplementary pension funds portfolio might be as diversified as Chile’s and Peru’s, but partly funded public systems/pillars suffer from lack of diversification as most private systems: 52–89% of the pension fund is invested on public debt paper in four countries and a sizeable share deposited in the central bank that pays interest below the market rate or in CDs in three countries, only 0.4% is in stocks in two countries and investment in foreign instruments is banned, miniscule or nil. Most countries in the region lack capital markets or they are incipient and are not properly regulated or they don’t trade sufficient instruments, problems that particularly afflict small countries, regardless of their pension system nature.

6.8.6.4. Capital returns In the ten private systems annual average real capital returns since the start of operations until 2006 ranged from −1% in Dominican Republic to 12% in Uruguay and the average for all systems was 7.6%. These are gross rates of return, without subtracting the administrative cost of the old-age program hence the net average rate is lower but data are not available. Historical series on gross capital returns show significant volatility and declining trend in the long run; economic crises have negatively affected such returns, thus the average in Argentina and Chile until mid-1990s was much higher than the average thereafter; high interest rates paid by public debt paper contributed to high capital returns in the past but either have decreased sharply (Argentina and El Salvador) or will in the future. Fluctuations in capital returns involve a risk: if the insured retire during a boom in the capital market the pension will be good, but if the fund in the individual account falls badly during a prolonged crisis the pension will shrink drastically. Such risk could be attenuated in mixed systems because they combine two pillars, one of defined benefit and another based on the individual account. The average return of the pension fund was lower than the average in the stock market in Chile and

149

the interest paid by banking deposits in Peru. Rates of return in four partly funded public systems/pillars for different periods in 1992– 2005 (no data were available for Brazil supplementary funds) ranged from 5% in Panama to 10.4% in Guatemala; the rate of Costa Rica’s public pillar was higher than in its own private pillar; the average of the four countries was 7.3% similar to the 7.6% average of ten private systems. Because of the guaranteed defined benefit, public pensions are less affected to market volatility than private pensions, but they are not impervious to lack of portfolio diversification and sustained low rates of returns.

6.8.6.5. Immunity against state-political interference Prior to structural reforms, the state played a harmful role in most pension systems in the region (even today, with few exceptions, the government performs an adverse financial function in the general program of public systems). In view to that negative state behavior, the widely diffused reform assumption that the private system would be immune to state and political interference was quite attractive to the insured and a strong incentive to their move from the public to the private system. But as many experts had warned, there is no guarantee that the ownership of the individual account combined with private administration of the funds would impede the state from behaving badly, because the government fundamental functions in private systems would open the door to such intrusion. For instance the state has to finance transition costs and at the same time has regulatory powers on investment, hence it was easy to guess that governments would set high ceilings to investment in public debt securities in order to finance such transition costs or even worse cover fiscal deficit from nonpension programs. Argentina, Bolivia, Colombia, Ecuador, and El Salvador prove that, contrary to reformers expectations, the state conducted itself in a financially incorrect manner, something that could happen in other countries.

150 Reassembling Social Security

Notes 1. Sources for this section are: legislation; Mesa-Lago 2004; Argentina MTESS 2003, Conte Grand 2006; Cuba Mesa-Lago 2003b; Ecuador World Bank 2005e, IESS 2006a, 2006b; Guatemala Durán and Cercone 2001, Balsells 2002; Honduras IHSS 2003, Congreso 2005; Mexico ISSSTE 2004, 2006, SHCP 2004, Herrera 2005; Panama Mesa-Lago 2005a, Asamblea 2006; Uruguay Saldaín 2003, Lagomarsino 2006; Venezuela Armas 2005, González et al. 2002, RBV 2003. 2. The reform law of 2001 unified contributions (3% worker, 8.5% employer, and 11.5% total), but they were not in force by mid-2006. 3. Argentinean low-income micro-entrepreneurs (monotributistas) pay about US$35 monthly on pensions to the public system and an optional $33 to the public or private system. 4. According to projections before the reform in Argentina, contributing affiliates would increase from 56% to 210% in 1995–2025 (Grushka 2002). 5. Brazil social insurance institute can go for the assets of enterprises to collect their debt and fines but many companies have declared bankruptcy and others claim that if the assets are seized they will go bankrupt; the institute can’t freely pact a payment agreement with indebted enterprises. 6. Barr (2002) questions why paying the IPD, not always a good policy, is required in pensions but not in health care and education. The IPD doesn’t have to be funded fully, in the same way that the explicit national debt doesn’t have to be totally paid; the IPD ‘is a useful concept but should not be given excessive weight’ (Barr and Diamond 2006: 29–30). 7. Morón and Carranza (2003) sustain that new generations of insured won’t need the minimum pension because their pension fund will accumulate enough to finance that pension and even more, but Chile’s experience of twenty-five years and projections for the next forty years contradicts such optimistic prediction. 8. Argentina’s fiscal cost averaged 4.6% annually in 1995–2001, partly due to the halving of the employers’ contribution (Schulthess et al. 2000); costs are projected to continue until 2035 but end twenty years earlier if the employer’s contribution were restored to its original level (Grushka 2002). 9. El Salvador actuarial deficit is projected to peak in 2020 and disappear in 2072, a transition of seventy years (Mesa-Lago 2003a). In 2006 the financial deficit took 2.9% of GDP, very close to the peak of 3.2% projected by the Bank for 2020 (Rosales 2006). 10. Mexican fiscal costs exclude the huge deficit of the federal civil servants scheme (ISSSTE) whose projected cumulative deficit in 2001–12 will equal all the reserves of Banco de México in 2003 (Moreno 2004; ISSSTE 2006). 11. According to the World Bank the main NDC advantage over fully funded systems is that is cheaper to manage (not higher than PAYG) because doesn’t have investment (Gill, Packard, and Yermo 2005). 12. Venezuela’s low share is explainable by the drastic fall in the value of real pensions. Uruguay’s entire pension system has the highest share in the region: 13.8% (Presidencia 2004). 13. In Haiti general program the balance of assets and obligations was a deficit equal to 18% of assets in 1999, but the balance of contribution revenue and benefit expenditures was a surplus of 92% of contributions (IADB 2001). 14. Cuban experts differ on the pension deficit: one considers it a fiction (Peñate 2000b), others stress its importance (CIEM 1999) and estimate that the equilibrium contribution should have been 22% in 1998 rising to 33% in 2010 (Sabourin 2003). 15. The projected equilibrium was based on a gradual increase of the retirement age from 60 to 65 but it was declared unconstitutional in 2005, shortening the period of equilibrium (Herrera 2006). 16. The worsening of the deficit was caused by transferring to the private system the worker’s contribution, cutting the employer’s contribution, and financing the deficit of provincial schemes (Golbert and Lo Vuolo 2006). The public system could achieve equilibrium in 2033 and generate a surplus of 0.3% of GDP in 2040–50; the equilibrium would be in 2025

Financial Sustainability and Reform Goals

17. 18.

19.

20.

21.

22.

23.

151

and the surplus increased to 1% in 2035–50 if undecided insured are assigned to the public instead of the private system, which was authorized by the reform law of 2007 (MTESS-ILO 2005; Ley 26,222 2007). Mexico’s pension fund was projected to reach 60% of GDP in 2027 (Guillén 2000), but its average growth in 1997–2006 does not support that optimistic prediction. Argentina’s pension fund is projected to peak at 26% of GDP in 2040 and fall to 25% in 2050; if the workers’ contribution is not restored to its original level, the fund would peak at 15% in 2030 and fall to 14% in 2050; if undecided insured are assigned to the public instead of the private system (as approved in 2007), the fund would peak at 17% in 2020 and fell to 4% in 2050 (MTESS-ILO 2005). Dominican Republic suffered a decline in national saving in the first year of the reform because fifty corporate pension plans were dissolved and US$3,000 million in reserves returned to participants, equivalent to the total reserves accumulated in the private system in 2004 (Pérez Montás 2006). A high increase in national saving can coincide with a low level of capital accumulation in the pension fund and vice versa: at the start of the 1990s Japan national saving/GDP rate was 34%, the highest in eleven developed countries, but its pension fund/GDP rate was 8%, the third lowest; conversely, the United Kingdom national saving/GDP rate was 14%, the lowest in the eleven countries, but its pension fund/GDP rate was 73%, the second highest (ILO 2002a: 102). Actual data for 2000–4 and new projections for 2005–10 show indeed declining fiscal costs from 6% to 4.7% in 2000–10, at an annual average of 5.3% vis-à-vis an average of 5.7% in 1981–99; such small average reduction suggests that Arenas’ projections were too optimistic (based on Arenas de Mesa and Mesa-Lago 2006). In a 1997 study Schmidt-Hebbel found evidence that part of the increase in national saving could be traced to the pension reform, but most of it was induced by an increment in public savings (that were not fully offset by private dissaving) as well as by the tax reform that generated a strong increase in corporate savings (cited by Gill, Packard, and Yermo 2005: 119). El Salvador’s real average capital returns fell from 14% to 2.4% in 1999–2002, due to dollarization and the cut in state interest rates (Mesa-Lago 2003a).

This page intentionally left blank

PART III

HEALTH CARE REFORMS AND THEIR EFFECTS

This page intentionally left blank

7

Health Care Reforms: Taxonomy, Objectives, and Actors

This chapter briefly compares health care reforms with those of pensions in Latin America, defines health reform and identifies its objectives, summarizes their history (years of start, later reforms, and current status), elaborates a typology of health care systems and reforms in the region, and discusses the influences of internal and international actors, as well as the political environment of the reforms. This book is limited to health care; in social insurance it focuses on sickness-maternity excluding monetary benefits (e.g. maternity and sickness pay-leave) and occupational risks and diseases.

7.1. Health care and pension reforms compared In many Latin American countries, health care reforms were promoted as part of a general reform or modernization of the state and influenced by IFO. Health care reforms took place after structural pensions reforms in Argentina, Bolivia, Chile, Colombia, Dominican Republic, El Salvador, and Peru (see their dates in Tables 2.1 and 7.1), due to the connection between both programs within social security. Health care reforms preceded those of pensions in Brazil, Costa Rica, Mexico, and Nicaragua and there have been health care but not pension reforms, at least structural, in Guatemala, Honduras, Panama, and Paraguay. Uruguay implemented a structural pension reform but has not really had a health care reform; several attempts to decentralize the

public health sector had minimal progress and social insurance has not been truly reformed. Cuban health care and pension reforms were almost simultaneous at the start of the 1960s, later there were more health than pension reforms. Haiti health reform of 1996 was not implemented and there have been no pension reforms. Venezuela approved pensions and health care reforms in 1998–9 but they never took effect and the government changed the orientation of both reforms at the start of 2000 (see below). Health care is more difficult to reform than pensions because it affects a greater number of people, their benefits are immediate instead of deferred, provision of health services is more complex than those of pensions, the health market is highly imperfect and have greater asymmetries of information, some health services have significant externalities and involve public goods (control and treatment of contagious diseases), the clientele of health care organisms is diverse, there is a larger number of health care than pension employees and they require more skills, are very wellorganized and strongly resists reform, and a health reform may contain costs but cannot generate national savings (Nelson 2004; Homedes and Ugalde 2006). Contrasted with ten structural pension reforms implemented, virtually all countries in the region have reformed their health care systems, albeit with considerable differences on their comprehensiveness, depth, and degree of progress. Health reforms are more diverse than their pension counterparts regarding principles, sector segmentation, population coverage, benefits,

Table 7.1. Years of start of the reform and subsequent changes, health system types, and main features of the reforms, 2005–6

Countries

Years of reforma

Argentina Bolivia Brazil Chile

1993–6, 2000–2 1994–8 1990 1981, 1991, 2004–5

Colombia

1994d

Costa Rica

1994–8

Cuba Dominican Republic

1960–9, 1984, 1990s 2001

Ecuador

1993, 1998–2001e

El Salvador Guatemala Haiti Honduras Mexico

1999–2000f 1995 1996g 2000–1h 1984–8, 1994–2005i

Integration/ coordinationb

Separation of functions

Tripartite: public, social insurance, and private (in expansion) Tripartite: public, social insurance, and privateo Dual: public (three levels) and private (mainly supplementary) Dual: public-social insurance and private (important)

Very low Very low Low Medium

Quadripartite: contributory, subsidized, private, and ‘tied’ to public (transitory) Dual: social insurance and private (small), public only direction-regulation Single: Public (there is no private) Tripartite: public, social insurance, and private; in reform law: contributory, contributory-subsidized, and subsidizedp

Medium

Low Very low Very low Almost total Total

Total

Partial

Total Very low now; high in law Low

No Not now; Total in law

Very high Medium

No

Low

Low Low Low Low Very low

Low Low No Very low No

Low Medium Low Low Very high

Health system types

Tripartite: public, social insurance (with peasant insurance), and privateo Tripartite: public, social insurance, and private (small) Tripartite: public, social insurance, and private (mainly NGOs)o Dual: public and private (three types) Tripartite: public, social insurance, and private (very small) Tripartite: public, social insurance, and privateo

Population coveragec Very high Low Very high Very high Low or very high Very high

Nicaragua Panama

1991–7j 1996k

Paraguay Peru Uruguay Venezuela

1996–8l 1995–7, 2001 1987m 1999, 2002–5n

Tripartite: public, social insurance (through private), and private Virtually dual: social insurance (principal), public, and private (small) Tripartite: public, social insurance, and privateo Tripartite: public, social insurance, and privateo Tripartite: public, social insurance (various providers), and private Tripartite: public, social insurance, and privateo

Low Very high

Partial No

Medium Very high

Very low Low Very low Very lown

No Partial Partial No

Low Medium Very high n.a.

Source: Tables 8.2, 10.1, and 10.2 and text. n.a. = non-available a Years of start of the reform and subsequent reforms in some countries. b Very low and low integration also lack coordination, except Argentina and Brazil (low coordination); in Chile medium integration and high coordination, and in Colombia medium integration without coordination; totally integrated regimes are coordinated. c Low 51% to 66%, medium 68% to 84%, and very high 97% to 100%; rankings of Dominican Republic, Guatemala, and Haiti are based on questionable estimates; Colombia is based on two different estimates. d Legal drafts of reform in 2005–6. e Drafts and laws not implemented in 1993–8, social security reform law of 2001 (partly declared unconstitutional) and health reform law of 2002 not implemented in 2006. f A consensus on the reform was aborted in 2003, new attempt in 2006. g Not implemented; new strategy in 2004; framework for external cooperation in 2004–6. h Ongoing reforms of the health code and the social security law. i Reforms of the health law in 2003–5 created Seguro Popular de Salud. j The social security law of 2000 (including health care) was annulled and a new law approved by congress in 2005 has been postponed to 2007. k Reform limited to one hospital; the social security reform law (including health care) approved by congress in 2005 was suspended and was on debate in 2006. l Proposal to extend coverage and separate functions in 2003. m Attempt at decentralization with little success (not a real reform), new attempt of reform in 2006. n Three failed attempts at reforms since 1987, social security law of 2002 and legal draft of health reform in 2005 (none of the two implemented in 2006) set the bases of the reform, the second stipulates the creation of a national public health system that closes and integrates social insurances, which would create a dual model. o There is an important sector of traditional indigenous medicine. p Only subsidized regime partly operational, other two regimes not in force by end of 2006.

158 Reassembling Social Security regulation, insurance, provision, and financing, thus it is more difficult to identify general health reform models and their characteristics (Sojo 2001a). In pension reforms we distinguished between structural that totally or partially replace a public or social insurance system by a private one, and parametric that reinforce social insurance without replacing it. In half of the countries, structural pension reforms have advanced private coverage to an average of 84% of total contributors (see Section 5.3); in health care reforms those distinctions are difficult to apply because the private sector covers at most only 18–25% of total affiliates. It is extraordinarily complex to classify the twenty health reforms into a few general models; a tripartite model exists in twelve countries (public, social insurance, and private sectors); ten diverse models resulting from health reforms are identified herein, compared with three general models in pension reforms (see Sections 2.1 and 2.2). Health reforms have not materialized in some countries despite several attempts or have not been global but limited to changes in certain aspects or have been divided into stages and most of them are pending or their laws have not been enforced. On the other hand, the second or third generation of health reforms is more advanced than those of pensions: in 2003–6 eight countries approved or were debating laws and changes in their health care systems (see Section 7.3).

7.2. Definition of health care reform For the Panamerican Health Organization (PAHO 1997) a health care reform introduces ‘substantive changes’ in different levels and functions of the system, aimed to increase equity, managerial efficiency, and effectiveness of its interventions to meet the population health needs. Londoño and Frenk (2000) consider ‘structural’ reforms those that reduce, coordinate, or restructure the three health care

sectors, while label ‘partial’ reforms those that mainly change the organization or efficiency of one sector but don’t pursuit the coordination or restructuring of all sectors. The World Bank (2005d) differentiates between ‘structural’ reforms that transform the current health care structure and ‘incremental’ reforms that accept such structure but improve its functions by introducing relatively minor changes; the latter therefore are less complex, more easily managed, create less dislocation and don’t require changes in the roles of existing institutions and creation of new ones. La Forgia (2006) distinguishes ‘systemic’ reforms, which are not limited to a single sector, address the previous tripartite segmentation of the health system and attempt to integrate, unify, or formalize links among at least two sectors, and reforms that implement innovations but neither introduce system-wide changes (or if they do still remain on paper) nor try to reduce segmentation. In seven countries the reforms have had a global character, introduced deep transformations that affected most functions of the health system, substantially modified relations between the public, social insurance, and private sectors, opened spaces for the creation of new actors, and achieved some degree of privatization. These reforms could be labeled ‘structural’ based on the above definitions, as in Argentina, Brazil, Chile, Colombia, Dominican Republic (not fully operational yet), and Peru. Cuba’s reform was also global and deep but with a totally public system, without any space for the private sector. Costa Rica also had a global comprehensive reform in the 1970s but not privatization, although further changes in the 1990s were of an ‘incremental’ or ‘partial’ nature and involved minor delegation of services to private providers. Other reforms have been more limited in their scope of application to the sectors, strategy of implementation, or affected health functions, for example, only provision modified through a new management model in entities of the ministry of health or social insurance (Nicaragua in the 1990s); some isolated projects but not a

Health Care Reforms global reform (Guatemala provision of services in remote areas); a nutrition program (Honduras); and only a hospital in the capital city decentralized (Panama in the 1990s) (La Forgia 2005).

7.3. Historical summary of the reforms Table 7.1 shows the approximate start year of the reform in the twenty countries and years of important changes later introduced until 2005–6. At the start of the 1980s only Chile, Costa Rica, and Cuba had implemented reforms (of divergent orientation); Chile’s had just begun, Costa Rica’s had ended a first stage but not yet begun a second stage and Cuba’s was completed; in 1984 Mexico initiated part of its reform but was interrupted in 1988; and in 1990 Brazil began its reform.1 The World Bank Investing in Health published in 1993 proposed reforms and has had strong influence in the region.2 In 1993–9, ten countries began reforms and Mexico retook hers; only three countries initiated reforms in 2000– 1. At the end of the 1990s most reforms were in the design or first stage of implementation, a minority had advanced in their execution and a few were discussing a ‘second generation’ of reforms or had implemented them (Infante, de la Mata, and LópezAcuña 2000; PAHO 2002a). In 2003–6, Chile entered the third generation of reforms, Haiti retook the reform interrupted by the civil war and the political-economic crisis, Panama was debating a reform to integrate its system, and Paraguay was considering a proposal for changes. At the end of 2006 four reforms had not been fully implemented: Dominican Republic reform law of 2001 was operational in only one of three regimes and the process was paralyzed due to opposition and debate of a counter-reform.3 Ecuador organic law of 2002 was not enforced; the social security law of 2001 was partly declared unconstitutional and the needed regulations had not been enacted, and congress was dis-

159

cussing a health code and a new reform proposal (IESS 2006b). Nicaragua social security law of 2000 (including health care reform) was annulled in 2005; a new law approved by congress that year was partly vetoed by the president and postponed by congress to 2007. Venezuela reform has been in process for eighteen years (three reforms in 1997–8 were not enforced), and the essential institutions and mechanisms mandated by the 1999 constitution had not been implemented pending approval of a legal draft of reform4 (Table 7.1).

7.4. Reform goals Albeit with important differences, most reforms have pursued the following goals: (a) de-monopolize and decentralize service provision by the public and social insurance sectors, through multiple insurance firms and/or providers, with a crucial role of the private sector and greater hospital autonomy or self-management; (b) separate the functions of direction/regulation, insurance or management, financing and provision; (c) grant freedom of choice to insured/users to select insurance firms and/or providers; (d) expand population coverage, extend primary care through a basic package of benefits, and improve the quality of services; (e) promote equity, reassigning resources to reduce gaps within regions, and within municipalities and target fiscal subsidies on the poor; (f ) increase efficiency and accountability of the system; (g) change the traditional financing from the fixed budget toward reimbursement for service provision, from supply subsidies to demand subsidies; (h) reduce public expenditures and recover costs via co-payments and user fees; and (i) encourage greater social participation and control.5 These goals, however, have not been pursuit in toto and with equal emphasis by all reforms; external and internal groups and organizations with divergent interests have pushed certain objectives instead of others (see Section 7.6).

160 Reassembling Social Security

7.5. Taxonomy of health care systems and their reforms The reforms have shaped different types of health care systems (although tripartite in a large majority) with diverse roles of the public, social insurance, and private sectors; degrees of population coverage; integration and coordination; and separation of functions (insurance, financing, and provision— the direction and regulation functions are normally exercised by the ministry of health and the supervision mostly by the health ministry or a superintendence). Table 7.1 classifies the twenty countries of the region into ten health types and identifies their major features, ordering them roughly based on their higher degree of population coverage and integration-coordination; separation of functions varies greatly (for other taxonomies see Londoño and Frenk 2000; Madies, Chiarvetti, and Chorny 2000; Medici 2000; Nelson 2004; Baeza and Packard 2005). 1. Cuba: Public system totally unified (SNS), no social insurance, and private practice banned; no separation of functions as the state is entirely in charge of financing and direct provision; virtual universal free coverage. 2. Costa Rica: Virtually unified social insurance (CCSS), totally integrated (all public facilities and services were transferred to CCSS; the ministry is in charge of direction, regulation, and supervision) and a small private sector; partial separation of functions, CCSS in charge of insurance, financing (through contributions and fiscal transfers), and provision mostly through contracts with its own service branches and a minority via private providers, cooperatives, and mixed medicine; virtual universal coverage. 3. Chile: Dual with medium degree of integration and very high coordination, public-social insurance (FONASA) and private sectors (mainly insurance

4.

5.

6.

7.

firms: ISAPREs); almost total separation of functions: insurance-financing split between FONASA and ISAPREs (financing by contributions, fiscal transfers, copayments), and provision separated in public-social insurance and mostly subcontracted by ISAPREs to providers; virtual universal coverage. Brazil: Dual with very low integration and some coordination, public sector (Sistema Único de Saúde: SUS), divided into federal, state, and municipal levels, no social insurance (all their facilities were transferred to SUS), and a significant voluntary supplementary private sector; very low separation of functions, each of the three levels of the public sector exerts functions of insurance, financing (all though taxes), and provision, and the private sector provision is done with subcontracting with the public sector; virtual universal coverage. Panama: Virtually dual, integrated functionally, with a central role of social insurance (CSS), a secondary role of the public sector and a marginal private sector; no separation of functions, CSS concentrates insurance, financing (through contributions), and provision, and the same is true of the public and private sectors; virtual universal coverage. Colombia: Quadripartite not integrated but well-coordinated, with a publicsocial insurance sector (divided into two regimes: contributory and subsidized), a private sector, and a transitory public sector (tied to public hospitals); total separation of functions: insurance through firms (EPS), financing (mix of contributions and state subsidies), and provision (through multiple providers); coverage from about one-half to virtually universal (if the population tied to public hospitals is taken into account). Uruguay: Tripartite with very low integration and without coordination, various social insurance schemes, public, and private sectors (including

Health Care Reforms not-for-profit providers—Institutos de Asistencia Médica Colectiva: IAMC—that play a major role and are financed by premium); partial and very complex separation of functions; virtual universal coverage. 8. Argentina: Tripartite with very low integration and coordination, social insurances (Obras Sociales—OS, segmented along occupational lines and largely managed by trade unions cover most of the population and are financed by contributions), public sector (in charge of the provinces and some large municipalities), and private sector (including prepaid health care plans: EMP); low separation of functions, each of the three sectors has its own insurance, financing, and provision; virtual universal coverage. 9. Haiti: Dual with low integration and without coordination, the public sector plays the fundamental role, social insurance is virtually nonexistent and there is a tiny private sector; no separation of functions; low coverage. 10. Eleven countries: Tripartite with very low or low integration and without coordination between the three sectors; either no separation of functions or low/partial, each sector performs all functions; coverage mostly low and medium: Bolivia, Ecuador, El Salvador, Guatemala, Honduras, Mexico (very high coverage), Nicaragua, Paraguay, Peru, and Venezuela. The Dominican Republic currently still have a tripartite system, but the reform law stipulates its substitution by three new regimes: contributory for all salaried workers; contributory-subsidized for all self-employed above certain income; and subsidized for low-income selfemployed, unemployed, disabled, and indigents; only the latter was partly operational by the end of 2006; if the reform is fully implemented, the system would be highly integrated with total separation of functions.

161

The ten types can be loosely clustered into four mayor models: (a) tripartite (public, social insurance, and private) in thirteen countries (Argentina, Bolivia, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Paraguay, Peru, Uruguay, and Venezuela), all with low or very low degree of integration, either no separation of functions or low/partial separation, and the large majority with low-medium population coverage (except Argentina, Mexico, and Uruguay that have very high coverage); (b) dual or virtually dual in four countries (Brazil—public and private, Chile— public-social insurance and private, Haiti— public and private, and Panama—social insurance and public with a tiny private sector), quite diverse concerning integration and separation of functions, but all with very high coverage (except Haiti); (c) quadripartite in Colombia (contributory, subsidized, private, and ‘tied’ to public), good coordination, total separation of functions, and medium to high coverage, if the ‘tied’ sector disappears this model would become tripartite and join the majority; and (d) unified with very high coverage in two countries: Cuba in the public sector, the state entirely in charge of financing and direct provision (private practice banned) and no separation of functions; and Costa Rica in social insurance, responsible for financing and most direct provision (contracts a minority with private and mixed providers), with partial separation of functions. It is not possible in this section to describe in detail the health systems of the twenty countries (see Sections 8.1, 10.1, and 10.2).

7.6. External influences, internal actors, and political environment A critical stand argues that IFO (World Bank, IMF, and IADB), usually in domestic alliance with ministries of finance have supported health reforms via abundant loans tied to conditions and with an ‘economic vision’,

162 Reassembling Social Security emphasizing market mechanisms like the expansion of the private sector, competition, freedom of choice, cost reduction, efficiency, and financial sustainability, as well as separation of functions. Conversely, said stand maintains, international health and labor organizations (WHO, PAHO, and ILO), usually in domestic collaboration with health ministries have a ‘health vision’ and placed more emphasis on the universalization of coverage, equity, health promotion/primary care, and social and community participation (see Lloyd-Sherlock 2000a; Ugalde, Homedes, and Zwi 2002; Levcovitz 2006). Notwithstanding, reform objectives espoused by IFO have included equity and coverage-extension goals, as well as better targeting of public resources on poor and low-income groups, whereas health and labor international organizations have endorsed objectives such as financial sustainability and efficiency, and both types of organizations support cost control and decentralization. This book tests if the reform objectives have been accomplished. A PAHO-published study argues that health reforms initially had a ‘predominance of pure financial considerations and little influence of the ministries of health in their design’, albeit since the mid-1990s they seem to have attained ‘a better balance and certain recovery of national health authority leadership’ (Infante, de la Mata, and López-Acuña 2000: 17–18). A more recent report, however, affirms that the reforms have been centered mainly in promoting financial, structural, institutional, and managerial changes: Much less attention has been devoted to enhancing the performance of health systems and services, with the focus on reducing inequities in health conditions and in access to health care; to reducing social vulnerability in the area of health; to boosting the effectiveness of health care interventions; to promoting quality care; and to strengthening the steering function of sector and public health authorities (PAHO 2002a: 138).6

Health professionals, hospital administrators, social insurance employees, and trade unions have often opposed the reforms (sometimes

allied with political parties), particularly the expansion of the private sector, and pressed for greater public health expenditures.7 Due to the contraposition of forces, the reform emphasis has been different in the countries: market-economic principles initially predominated in Chile (freedom of choice, competition, and efficiency) but equity concerns rose significantly under democratic governments; in Argentina and Mexico emphasis was on fiscal balance and efficiency; in Costa Rica equity has been dominant, but also hospital deconcentration. Colombia (and in smaller degree Brazil and Peru) has reached a better equilibrium among goals, as a result of political compromises, on one hand conventional principles of social security (universality, equity, solidarity, sufficiency, and social participation) and on the other hand marketeconomic principles, such as competition, efficiency, and freedom of choice (Kaufman and Nelson 2004). The political environment of the reform and its principal actors has diverged also. Chile’s initial reform was carried out by the military government without any public discussion; when democracy returned an ample debate surged and a second generation of reforms passed; five legal reform drafts were submitted to parliament in 2003 and three of them approved in 2004–5.8 Colombia reform took place under a democratic government with strong national debate, a new constitution was enacted and several laws discussed in congress; attempts to reverse the process and two modifications to the original reform have occurred.9 The health ministry normally directed the initial process of design and negotiation of the reform, but in several countries IFO and ministries of finance played a central role (PAHO 2002b, 2002c). Three experts affirm: In almost all the countries the role of the international financial organizations is more and more important. The health authorities not always seem to have played the role that would correspond to them in the formulation of the national policies of financing and expenditures or in the definition of the priority areas for the granting of international

Health Care Reforms loans . . . [In some countries] the health reform did not start up until the loans with the international banks were signed (Infante, de la Mata, and LópezAcuña 2000: 14, 18).

Two scholars add: ‘by the end of the 1980s, the World Bank had became the major international health lender and began to prepare health reforms based on neoliberal principles’, and eventually ‘has become the leading international health policy maker’ (Homedes and Ugalde 2005a: 83, 94). Conversely Kaufman and Nelson (2004) question the effectiveness of IFO influence.

7.7. Measuring the effects of health care reforms Evaluating the impact of health reforms on social security principles is more difficult relative to pension reforms for several reasons: (a) considerable more variety in designs and characteristics of health reforms, which complicate the comparative evaluation of their models, policies, and results; (b) lack of a regional association of superintendencies of health, as is the case in pensions that publishes twice a year standardized statistics on the performance of the reforms, PAHO publishes every four years data on health systems and has elaborated health profiles of the twenty countries evaluating reforms results (updated about every two or three years) but without standardized comparative and evaluation statistics (see below); (c) most academic studies focus on a single country, the few comparative studies fail to use equal indicators for all countries and usually do not evaluate all effects of the reforms, some try to measure equity, others competition, efficiency, quality of services, etc.; (d) the reforms have been introduced in different years (with a significant span between them) and in countries with diverse health standards, which makes difficult to compare their results; (e) the time elapsed since the beginning of the reform is not sufficiently long (except in Chile and Cuba) to detect relevant changes in health

163

standards that may take years to show, half of the countries introduced the reform in stages instead of at once, and decisions made years before the reform could affect the outcomes; (f ) several factors difficult to separate from a health reform can influence its results, for instance, environment, genetics, individual and collective behavior, war or peace, refugees, and good or bad economic performance; and (g) usually there is not a monitoring system within each country that establishes a base line and follows up key indicators and, in some countries, the information is dispersed and not standardized, which prevents comparisons through time (Medici 2000; Castaño et al. 2001; PAHO/WHO 2001). There have been several remarkable efforts to develop methodologies and indicators to evaluate the reform outcomes. In 1997, PAHO and USAID launched a regional initiative on health sector reforms in Latin America and the Caribbean, an eight-year project geared to implement the methodology for monitoring and evaluation of the reform impact on five objectives: equity in coverage and access; effectiveness and quality; efficiency in allocation and management of resources; sustainability; and social participation and control. In 1998–2002, health profiles were developed for the twenty Latin American countries (and for fifteen non-Latin Caribbean countries), that included twelve ‘contents’ of the reforms; such profiles became available through Internet in 2005, including monitoring the twelve ‘contents’ and evaluating results on the five reform objectives, with editions ranging from 2000 to 2002 (PAHO 2005a). The profiles of the systems as well as the monitoring and evaluation of the reforms are very useful, but do not contain standardized statistical indicators for all the countries to allow systematic comparisons within the region. PAHO has also undertaken comparative analyses of health reforms and their results in the Andean and Central American subregions (thirteen countries including the Dominican Republic), but they do not offer statistics on results and several countries failed to provide information on key aspects. All Andean

164 Reassembling Social Security countries except one argued that ‘it is impossible to establish direct and unequivocal causal relationships between actions of health reform and modifications in the indicators relative to equity, efficacy, quality, efficiency, sustainability and social participation and control.’ The countries ‘try to evaluate the effect of the reforms on the general performance of the sector instead of the health situation’ but ‘the displayed results show that the reforms do not seem to have obtained a globally positive effect on the health sector’ (PAHO 2002c: 22, 29–30). With one exception, the Central American countries ‘did not define criteria of evaluation at the beginning of the reform . . . throughout the process some evaluations on the development and impact of the reforms were done, although quite limited . . . still it is not possible to have an integral evaluation of the advances of the reform made’. Two countries did not provide information on the evaluation of results giving as reasons the lack of needed data to assess the impact of the measures and the short period of time of the reform (PAHO 2002b: 8, 10, 36). PAHO explains that ‘the follow-up and assessment of health sector reforms pose conceptual and methodological problems that are far from being resolved’ (PAHO 2002a: 134). Five reasons explain the lack of information on the reform results: (a) the initial proposals usually neither included mechanisms to evaluate the reform impact or to oversee its development, nor set a term to achieve the proposed objectives; (b) the selection of indicators to evaluate the results is technically complex and causality attributions are difficult; (c) available data are often incomplete, dispersed, dated, and little reliable, (d) the reform is three years old or even less in thirteen countries and it is too soon to evaluate its results; and (e) the principal actors (including the health authorities) not always show enough interest in evaluating results (PAHO 2002a: 134, 136). According to another study: ‘Most countries don’t do the evaluation of reform results in a systematic manner and, in the few where it is done, results are seldom used as an important input to redesign the

content and strategies of reform implementation’ (Infante, de la Mata, and López-Acuña 2000: 18). The WHO 2000 report developed a new methodology to rank and evaluate health systems in 191 countries (including the twenty of Latin America) based on several indicators with 1997–9 data: (a) general level of health of the country and equality in its distribution within the population; (b) level and distribution of the responsiveness of the system to the expectations of the population (with respect to dignity, autonomy, timely care, confidentiality, quality of the basic services, access to networks of social support, and selection of the provider); (c) level and distribution in fairness (impartiality) of the financial contribution in the system; and (d) overall health system attainment and health performance. These indicators were merged into a composite index that was used to rank the countries on their performance on health level, and the general health system (WHO 2000a). But the report neither analyzed the variety of reforms in the world nor evaluated their effects. PAHO, WHO, and other entities conducted in 2001 a regional consultation on the evaluation of health systems performance, with seventy experts from nineteen countries, but the group did not agree on the feasible of designing a common frame of reference to evaluate and compare results, due to diverse definitions of the health systems; the group also estimated that the elaboration of a measuring methodology and suitable indicators could take ten years or more, and made the following recommendations: (a) set the criteria and indicators to assess performance through national and international consensus, otherwise controversies would emerge and diminish the value of such exercise; (b) reexamine WHO methodology of performance evaluation in collaboration with the countries and experts, and design a system to reinforce the collection of standardized data (the methodology and collection system should be transparent and replicable); (c) include indicators on access, equity, sustainability, competition, efficiency, and acceptability by users, cluster

Health Care Reforms them on the basis of four dimensions, and set intermediate and final goals; and (d) entitle the users to judge if the evaluation of performance shows that progress has been achieved in the fulfillment of goals and undertaking the appropriate actions to that end (PAHO/WHO 2001). The explained methodological problems and, specially, the absence of statistical indicators standardized and universally accepted to evaluate the performance of health reforms impede herein to make a comparative quantitative assessment of their impact on health standards (efficacy of the reform) therefore this book concentrates on the assessment of health reform impact on social security principles. Nevertheless Section 11.2 describes changes occurred in health indicators between

165

1990 and 2002 in the twenty countries, but not pretending to determine causality by the reform. The analysis on the effects of the reforms on social security principles will generally follow the same structure that in pension reforms, albeit with some new aspects such as access and quality of services. Most of the indicators designed by PAHO for the evaluation of regional reforms coincide with the social security principles described in this book hence whenever feasible such indicators are used in the analysis; some WHO concepts and rankings will be similarly utilized and discussed. Finally, this book tests whether the reform goals have materialized: separation of functions, decentralization, competition, privatization, freedom of choice, and so forth.

Notes 1. Uruguay made some reforms in the 1960s; an attempt at decentralization failed at the end of the 1980s, a legal draft of reform in 1995 was not approved, later there were modest advances in decentralization but without depth; all left untouched the fundamental system features (Médici 2000; PAHO 2005a). 2. Brazil, Chile, Costa Rica, and Colombia introduced the reforms before the World Bank book of 1993 and other countries had reform designs. A new World Bank book on health care in Latin America in preliminary state at the end of 2005 is discussed herein (Baeza and Packard 2005). 3. The subsidized regime began to function in 2004 but only in some regions by the end of 2006; the contributory regime was to start in 2002 but was postponed eight times, next for June 2007; no date has been set for the contributory-subsidized regime (Lizardo 2005, 2006, 2007). 4. Social security and health care reform laws enacted in 1998, just before the presidential elections, were not implemented and a new constitution approved in December 1999. The organic law of social security enacted in 2002 had its regulations still pending in 2006. A legal draft of health care reform approved in first discussion at the National Assembly in 2005 was waiting for final approval in 2006 (Rodríguez 2001; González 2004, 2006, 2007; RBV 2004; Andrade 2006). 5. According to La Forgia (2006) decentralization and the basic package were not reform goals; the first is a transfer of decision-making authority to subnational levels that could leave untouched key features of the system; the second is a tool to extend coverage targeted on groups with poor access. He also argues that hospital autonomy was not a goal of most reforms. 6. PAHO, World Bank, and IADB signed an agreement in 2000 to develop a joint health plan for the Americas aimed at: support health care reforms and reinforce public health programs and the directive role of the ministry. Implementation of these targets is monitored by a coordinating group from the three institutions (PAHO 2002a). 7. Opposition of these groups in Argentina, Bolivia, Colombia, Costa Rica, Ecuador, El Salvador, Mexico, Nicaragua, Peru, and Venezuela have argued that: decentralization divides and weakens the unions, privatization changes the remuneration systems and threatens the stability of

166 Reassembling Social Security jobs, and private enterprises restrict the autonomy of physicians and force them to use more economic than sanitary criteria (Kaufman and Nelson 2004; Ugalde and Homedes 2005). In 2003 there were thirty-seven health care strikes in twelve countries, 68% of them due to resource allocation, payment of benefits, budgets to public network facilities, and salary increases; only 11% were specifically related to health-reform issues (Scavino 2004). 8. The five original legal drafts were: system of universal access with explicit guarantees (AUGE); new systems of care, rights, and duties for users; financing and regulation of insurance firms (ISAPREs), and health authority and management system (MINSAL 2004). AUGE and the health authority laws were approved in 2004, and the regulation of ISAPREs in 2005. 9. On the political economy of health reform, see Ortiz et al. (1999); Giordano and Colina (2000); Gonzales-Rosetti and Ramírez (2000); PAHO (2002b, 2002c, 2005a); Nelson (2001); Kaufman and Nelson (2004); Weyland (2004); Homedes and Ugalde (2005a, 2005c); and Borzutzky (2006).

8

Effects of Health Care Reforms on Universal Coverage

This chapter distinguishes between coverage and access to health care services. Coverage means affiliation to social or private insurance, as well as protection by the public sector, and may be according to the law, based on statistics from health care institutions or measured by surveys, none of which necessarily conveys effective access or utilization of such services. Legally proclaimed universal coverage and access often is not enforced in practice. Finally, coverage and access do not measure sufficiency and quality hence two countries with the same social-insurance coverage could provide completely different packages of benefits with divergent quality. Generally, the degree of coverage is directly related to the level of socioeconomic development and the proportion of the labor force that is formal whereas is inversely related with poverty incidence. The health system organization is not neutral regarding coverage but a determinant factor (ILO 2000b; PAHO 2002a). Usually the formal urban sector in the region is covered by social or private insurance whereas the informal sector is excluded from both but may has partial access through the public sector and/or out-of-pocket expenses; rural areas in the least developed countries are excluded, as well as poor and indigenous people (Mesa-Lago 1992; Londoño and Frenk 2000; Roberts, Stafford, and Ashworth 2002). In its 1978 meeting in Alma-Ata, the WHO set goals for the year 2000 on universal access to health care, endorsed by all Latin American countries but without a serious, coordinated effort to evaluate the progress achieved. The WHO report of 2000 evaluated health

systems performance in the world but didn’t include statistics on population coverage and advances in the universality goal. A PAHO study informs that the 2000 goals were not fulfilled; the lack of equity in coverage worsened and more than 20% of the region population was unprotected (Madies, Chiarvetti, and Chorny 2000). According to one expert, the neoliberal approach falls short of the universality goal because it limits coverage to a basic package of services to the poor (Lloyd-Sherlock 2003). Targeting fiscal resources on the poor and vulnerable groups can result in greater equity but requires identification of these groups based on their socioeconomic status, payment capacity and epidemiological profile according to risk, as well as the adaptation of the basic package to their necessities (Medici 2000). WHO (2000a) supports a ‘new universality’ (the best possible care for all and the most basic care for the poor), meaning the provision of essential care of high quality for all, defined by criteria of effectiveness, cost, and social acceptability; although it may be necessary and efficient to ration health services, it is inadmissible to exclude any population group from coverage. PAHO evaluations of the impact of health reforms in the region conclude that most countries specifically guarantee the right to health care, even in their constitutions, but the population in half of them is neither aware of such right nor of the possible actions available to ensure it; half of the programs to expand coverage are geared to the provision of basic services at the primary level to those

168 Reassembling Social Security with worst access. PAHO most recent report indirectly estimates aggregated exclusion in the region but doesn’t offer statistics on coverage in individual countries (PAHO 2002a, 2002b, 2002c; see Section 8.2). Similarly to pension reforms, health reforms assumed they would increase coverage and even reach universality, but very few countries established a timetable to accomplish such goal and when they did the goal was not met. Reformers argue that economic growth, the market and private enterprise will increase employment and income, thus creating conditions that enable individuals to satisfy health needs with their own resources. Reforms also enforce equivalence between contributions and health benefits, providing incentives for insurance affiliation, and target fiscal resources on the poor and low-income populations thus helping to extend coverage (Cifuentes 2000). Promises on health coverage are more difficult to assess than in pensions due to multiple obstacles to compare coverage before and after the reforms (see Section 8.2.1). Coverage and access have been measured by surveys on the labor force, households or living conditions, taken in most countries in 1986–2005 (see IADB 2005). Their results demonstrate that access is strongly influenced by income, available information, population knowledge of the system (more complex in those fragmented), and simplicity or complexity of the process to gain the right to health care or be reimbursed for the payment of services. Household surveys have two advantages over labor force surveys: they embrace all the population instead of only the labor force or part of it, and report the characteristics of families instead of only the workers’. Living condition surveys have the most detailed questionnaires on family access to health services (Gasparini and Bertranou 2005). This chapter compiles all the available information and statistics on population coverage and access in the twenty countries: summarizes legal coverage in the three health sectors and by population groups and examines whether coverage targets have been met; calculates coverage of the total population, as

well as by sector, location, income, and groups difficult to incorporate; analyzes the problems to compare coverage before and after the reforms; estimates access and utilization of services, and explores the causes of the failure to meet the universality goal in some countries (for sufficiency and quality see Chapter 9).

8.1. Legal coverage and targets Constitutions and reform laws of health and/or social security in virtually all countries establish the right to health coverage, usually universal and often free in the public sector for the uninsured, but in eleven countries (the seven in the latecomer-low group and the four less developed in the intermediate group) such mandate has not been enforced.1 A more recent and effective approach is to grant the right to explicit entitlements or specific health benefits, guaranteed and financed by the state and claimable by users (see Section 9.3.1), which forces a decision between ensuring a basic or a comprehensive package, an assessment of which of the two is financially feasible and a careful projection of their costs (Baeza and Packard 2005). This section summarizes legal coverage in the three sectors (social insurance, public and private; for a full analysis see Section 10.1) and the groups covered by each; identifies the few changes in legal coverage introduced by the reform and analyzes whether its targets for coverage has been met (sources for this section are in Table 8.1).

8.1.1. Coverage by the three sectors As before the reforms, social insurance provides legal coverage to all or the majority of the population in the pioneer-high group and the most developed intermediategroup countries whereas it covers a minority of the population in the latecomer group and the least developed intermediate-group countries. Social-insurance organization varies

Health Care Reforms on Universal Coverage 169

Table 8.1. Legal mandatory coverage of health social insurance by type of occupation, dependents, and pensioners, 2005–6 Salaried workers Countries

All

Argentina Bolivia Brazile Chilee Colombia Costa Rica Cubae Dominican R. Ecuador El Salvador Guatemalae Haitie Hondurase Mexico Nicaragua Panama Paraguay Peru Uruguay Venezuela

X

Part

X X X X X X

Agriculturea

Spouseb

Childrenc

Pensionersd

g

X

X

X X X X X

X X n.a. X X

Xf

X

X X X X X X X X X X X

X X X X X X X X X X X

X X X X X X X X X X X

g

g

X

g

Xf X Xf X X X

X X X X X X X X

X X X X X X X X

X X X X X X X X

g

g

X X X X

g g

X

X X X X X X

g

g

X X Xf X X

Domestic servants

X

f

Dependent family

Selfemployed

g g g g

X X X

Sources: Legislation; ISSA (2003a, 2004, 2005, 2006a); OISS (2004); US-SSA (2004); PAHO (2005a); and MesaLago (2006a). a Usually excludes peasants, sharecroppers, and other nonsalaried workers; there are separate schemes for peasants in social insurances of Ecuador and Mexico, and agrarian insurance for salaried and self-employed in Peru; Nicaragua’s law postponed to 2007 mandatory covers permanent salaried agricultural workers. b Only in maternity in Dominican Republic, El Salvador, Guatemala, Honduras, Nicaragua, and Uruguay; only in sickness in Ecuador. c Only covers ages below: 1 year in Dominican Republic and Ecuador (6 by unenforced law), 5 in Guatemala, 9 in Uruguay, 11 in Honduras, and 12 in El Salvador and Nicaragua. d Only old-age pensioners and with a mini-package in Nicaragua; only pensioners whose income is less than the maximum annually set in Uruguay. e Brazil and Cuba coverage is by the public sector, and Chile by public-social insurance; in Haiti sickness social insurance has not been fully implemented; in Guatemala and Honduras not all departments are covered. f The mandatory coverage of all salaried workers or domestic servants is not implemented in practice; Paraguay only covers domestic servants in the capital city. g Voluntary affiliation, little effective (not enforced in Paraguay); also for small entrepreneurs who pay taxes in Argentina; Costa Rica’s law stipulated mandatory coverage in 2006; Colombia’s contributory regime covers those who earn more than two minimum wages and the rest covered in the subsidized regime; legal drafts in Nicaragua, Panama, and Venezuela make it mandatory; in Dominican Republic is mandatory by law but has not been implemented; in Uruguay part of the self-employed has voluntary affiliation and others are protected by the public sector.

170 Reassembling Social Security significantly between countries and affects coverage. In the pioneer-high group, Brazil and Cuba don’t have social insurance but public systems. Argentina’s multiple social insurances (OS) mandatory cover salaried workers in a given occupation; there are national OS for salaried workers in the private sector of the economy and one specific for pensioners, other OS cover managerial personnel, civil servants in each province (provincial OS), etc. Chile’s public-social-insurance sector (FONASA) automatically covers all who don’t choose to be affiliated to private insurance firms (ISAPREs), provides free services to the poor2 and fiscal subsidies to low-income affiliates. Costa Rica’s social insurance (CCSS) received all public facilities and services transferred by the ministry in exchange for free coverage to all uninsured poor. Uruguay’s principal social insurance (BPS) has two separate branches: sickness covers private-sector workers mainly through mutual-aid private providers (IAMC), as well as private for-profit partial and emergency insurances, whereas maternity is directly covered by BPS own facilities or contracting with public or IAMC services. In the intermediate group, Colombia has two regimes: the contributory covers those with paying capacity and the subsidized those with fewer resources and the poor freely covered; part of the population is still cared by the public and private sectors. Mexico has five social-insurances programs; the principal (IMSS) covers salaried workers in the private sector and those in the public sector not affiliated elsewhere, it also provides family health insurance and has a special scheme for peasants; the federal civil servants scheme (ISSSTE) also covers functionaries in some states while the rest have their own ISSSTEs. Venezuela’s principal social insurance (IVSS) covers salaried workers but since 2000 was been open to the uninsured. In the latecomer-low group social insurance covers formal salaried workers only, basically urban. Haiti has separate socialinsurance programs for sickness and maternity both with little implementation and coverage.

Table 8.1 exhibits the mandatory legal coverage of social-insurance sickness-maternity in 2005–6 (identifying where it is voluntary and laws in process that make it compulsory) concerning the following six groups: formal salaried workers (totally or partially covered); self-employed and domestic servants in the urban informal sector; the ruralagricultural sector; spouse and children economically dependent of the insured, and pensioners (for pre-reform legal coverage see Mesa-Lago and Bertranou 1998). Social insurance mostly covers the formal sector, usually in the middle-income stratum; the following information refers to the principal program, but all countries except two have separate schemes with superior legal coverage for the armed forces, and most countries also for civil servants, policemen, teachers and university personnel, oil workers, and other groups who often have their own facilities (see Sections 9.1.1 and 10.1). In the six pioneer-high countries all salaried workers have mandatory coverage (including domestic servants and agricultural workers). This is also true with few exceptions in five countries of the intermediate group (Colombia, Mexico, Panama, Peru, and Venezuela); Bolivia excludes both domestic and agricultural workers, and Ecuador does not enforce coverage of domestic servants. Conversely in all latecomer-low countries compulsorily coverage is restricted to part of salaried workers, because they exclude domestic and agricultural workers legally or in practice or social insurance is not operative in all geographic areas (in Honduras to seven out of eighteen departments and in Guatemala to nineteen out of twenty-two). The self-employed have full mandatory coverage only in Colombia and Costa Rica (they are covered by public systems in Brazil and Cuba), are excluded in El Salvador, Guatemala and Haiti, and have voluntary coverage in the twelve remaining countries. Argentina also voluntarily covers small taxpayers (monotributistas) who earn less than a maximum annual income; most selfemployed are covered by the public sector.

Health Care Reforms on Universal Coverage 171 Chilean self-employed who voluntarily contribute to the pension system (private or public) are automatically affiliated to public-social insurance, the rest have voluntary coverage. Colombian self-employed are divided in two groups: those earning less than two minimum wages are covered by the subsidized regime and those earning more are in the contributory regime. Costa Rica affiliation was voluntary until when it became mandatory but lowincome self-employed receive fiscal subsidies. Uruguayan self-employed may voluntary join social insurance through IAMC and those who cannot due to lack of resources are legally covered by the public system gratis (submitted to a means test) or subsidized according to income. Recent laws not yet implemented or legal drafts make coverage compulsory for the self-employed in Dominican Republic, Ecuador, Panama, and Venezuela. Domestic servants have social insurance mandatory coverage in eleven countries, although it is little effective in Ecuador, Nicaragua, and Paraguay—limited to the capital city; coverage is voluntary and little effective in Mexico, and they are excluded in the six least developed countries: Bolivia, Dominican Republic, El Salvador, Guatemala, Haiti, and Honduras. These workers have an employer but many of them lack a labor contract and, when it does exists, it is difficult for the domestic servant to demand the employer contribution or when wages are very low often both parties jointly evade affiliation. Agricultural workers are excluded in half of the region; they are mandatory covered in the six pioneer-high countries and in three of the intermediate group (Mexico, Panama, and Peru); legal coverage is mainly granted to wage earners in large plantations and members of cooperatives in some countries; nonsalaried rural workers usually have voluntary coverage (e.g. in Argentina and Costa Rica). Self-employed peasants, sharecroppers, squatters, and the like are excluded in most countries. Exceptions are social-insurance ad hoc schemes for peasants and their families in Ecuador (Seguro Social Campesino: SSC) when organized in agrarian cooperatives or

associations that voluntarily request affiliation, and in Mexico (IMSS-Oportunidades); less important is Peru’s agrarian insurance for selfemployed rural workers. The least developed countries have the largest rural population in the region, and it is usually the poorest and not covered. Pensioners are covered in all countries albeit with limitations, for instance, only oldage pensioners are eligible in Ecuador and Nicaragua, and only pensioners whose income is below the maximum annually in Uruguay. The insured dependent family (and usually pensioner’s family too) is compulsory covered in all countries but with restrictions in several of them. The spouse (usually woman) and children who are economically dependent on the insured are compulsorily covered in all countries (except Haiti) but with important limits: the female spouse in maternity but not sickness in Dominican Republic, El Salvador, Guatemala, Honduras, Nicaragua and Uruguay, while the opposite occurs in Ecuador; and they only have optional insurance in Uruguay IAMC or other providers not obliged to deliver the basic package. Seven countries exclude children above the following ages: 1 in Dominican Republic and Ecuador, 5 in Guatemala, 6 or 9 in Uruguay (depending on the type of care), 11 in Honduras, and 12 in El Salvador and Nicaragua. In contrast with such restrictions some countries provide coverage to dependent parents in Bolivia, Mexico, Paraguay, and Venezuela, and also siblings in Bolivia. In Costa Rica dependent family of social assistance pensioners are entitled to care providing they lack health insurance, can’t pay the contribution and their family income is insufficient to satisfy basic needs. Separate schemes of the armed forces and policemen cover all dependents of both active insured and pensioners without restrictions. Other workers are excluded from social insurance or have voluntary coverage at best. In most countries the law determines the minimum number of employees of small enterprises (usually between five and ten) required for mandatory affiliation; in Mexico

172 Reassembling Social Security employers with two to three employees are obliged to insure them but few do it. Unpaid family members, home, temporary, seasonal and part-time workers, as well as those without a labor contract, unemployed and housewives are excluded from coverage or can join optionally under certain conditions in a few countries. Unemployment averaged 10% of the regional labor force in 2004 and ranged from 11 to 17% in half of the countries (ECLAC 2004b), but they are covered in only a few countries: Colombia and Dominican Republic (both in the subsidized regime), Chile, Uruguay and Venezuela (when collecting compensation), Nicaragua (voluntary), and Peru (insured with a minimum number of contributions). Employers are covered in Dominican Republic and Uruguay, and owners of small enterprises voluntarily in Honduras, Mexico, and Nicaragua. Voluntary affiliation is open to housewives in Costa Rica, Honduras and Peru, home workers in Venezuela, unpaid family members in Costa Rica and Mexico, and immigrants for emergencies in Costa Rica, but very few of these workers have affiliated (see Section 8.6). The public sector, mainly the federal or central health ministry but also states-provinces and large municipalities, in most countries should provide free services to the poor, lowincome population and other uninsured, but often does not fulfill the legal mandate. In two pioneer-high countries, however, the public sector is the only or most important provider and has universal coverage: Brazil integrated multiple social insurances with incomplete coverage into a national public system (SUS) that offers free provision; Cuba did not have sickness insurance until the reform of the 1960s that created a national public system (SNS) that legally covers all the population free. In the remaining countries of the pioneer-high group and the most advanced of the intermediate group, the public sector plays a secondary role vis-à-vis social insurance. Argentina’s few central government hospitals were transferred to the provinces and, together with the largest municipalities, provide free services to the population in need.

Uruguay’s central ministry, departments, and municipalities cover gratis low-income workers, pregnant women, children below 1 year, and pensioners older than 65 or disabled. In Colombia, still part of the population is temporarily ‘tied’ to public hospitals subsidized by the state, just as prior to the reform, because budgetary resources are insufficient to cover all the potential population in the subsidized regime. Mexican poor and low-income uninsured are eligible for health care in the public sector (federal and state ministries, federal district, wealthy municipalities, etc.); multiple public health programs have targeted the poor, the latest is popular health insurance (SPS). In the latecomer-low group and the least developed intermediate-group countries the public sector is in charge of the majority of the population but has grossly insufficient resources, the worst case being Haiti that virtually lacks social insurance and has the highest poverty incidence in the region. Coverage in the private sector is predominant by profit seekers such as prepaid plans, insurance, and independent facilities and professionals that normally care for the uppermiddle and high-income population mainly in the capital city and urban areas, but also the low-income population without public access who pays out-of-pocket. Private providers are often under contract with social insurance or the public sector in Argentina, Brazil, Chile (high-income insured may transfer from public-social insurance to private ISAPREs), Costa Rica, Dominican Republic, Nicaragua, Peru, and Uruguay. Large enterprises have primary-care facilities or contract health plans for their employees in the biggest countries. Not-for-profit providers (NGOs, churches, charities, foundations, mutual-aid societies) care for the poor and low-income population in rural and urban-marginal areas. Traditional medicine, which commonly charges for its services, is important in at least six countries of the Andes and Central America and cares for the poor and rural population especially indigenous, for instance most of Paraguay’s Guarani Indians and Guatemala’s rural population (see Section 8.3).

Health Care Reforms on Universal Coverage 173

8.1.2. Fulfillment of reform coverage targets Most reforms did not establish coverage targets for the total population (e.g. Chile, Nicaragua, and Paraguay) and others set general vague goals such as to extend coverage to those uninsured or without access (Argentina, Bolivia—basic package, Ecuador,3 Guatemala, Honduras, Paraguay, and Uruguay). Four reforms set up specific coverage targets with a timetable, but only one met them. Colombia projected to reach 100% coverage in 2000– 1 with 70% in the contributory regime and 30% in the subsidized, but achieved 29 and 24% respectively for a total of 53%; the gap between projected and real coverage actually expanded in 1999–2000 and universal coverage remained unfulfilled in 2006 (Martínez et al. 2003). Dominican Republic aimed at 100% coverage in 2011, but by 2006 only one of its three regimes was in operation and with limited coverage; the rapid increase of the basic package cost in the contributory regime due to rising inflation forced to scale down the target to 33% in 2002, but the package cost jumped 119% in 2004 making impossible to meet the target (Lizardo 2004). Mexico’s SPS set in 2002 a target of 31% coverage of the total population in 2010 but in 2004 the minister of health declared that won’t be met and only 11% was covered in 2006 (Nigenda 2005, 2006). Panama sought to increase coverage from 80 to 85% in 2000–4 and probably met that target.

8.2. Estimates of population coverage According to ILO (2003a), the average coverage of salaried workers contributing to all social security programs in eleven Latin American countries declined from 67 to 64% in 1990–2002 (during the reform period in most countries): it diminished in seven countries, increased in only two and the comparison was not feasible in other two. Based on household

surveys taken in seven countries, ILO estimated the specific coverage of the health program as percentage of the labor force in 1994–2002: 56% in Argentina and 87% in Chile (pioneer-high group); 15%, 24%, and 33% respectively in Bolivia, Peru, and Ecuador (intermediate group); and 16% in Nicaragua and 26% in Guatemala (latecomer-low group). Coverage increased in Chile, Ecuador and Nicaragua, but decreased in Argentina, Bolivia and Peru, and there was only one observation for Guatemala (last column of Table 8.2). PAHO calculated the population excluded in Latin America and the Caribbean in 1989– 96 as follows: 46% without social or private health insurance, 27% did not have economic access, and 22% lacked geographic access; between 22 and 55% was excluded from coverage (100 to 260 million people) whereas the covered population fluctuated between 45 and 78%, a hefty range. Despite these distressing figures, PAHO warns that ‘exclusion from social protection for health does not usually appear on the list of priorities on international and social policy agendas [and] the health agendas of the region countries. Health sector reforms have generally not dealt with this issue either and, when they have, it has been in an indirect, piecemeal manner’ (PAHO 2002a: 144). Comparative studies on thirteen Andean and Central American countries only give figures on the impact of the reform on population coverage in Colombia: an increase from 21 to 57% in health insurance in 1994–7 (but see below). Estimates for other three countries are restricted to coverage by the basic package: 90% in Costa Rica, 32% in Guatemala, and 66% in Nicaragua (PAHO 2002b, 2002c). Three international organizations, based on household surveys and secondary data, measured the ‘population without health insurance’ in 1996–2001 ranging from 73 to 83% (lower to higher) in Peru, Dominican Republic, Ecuador, Paraguay, Guatemala, and Honduras (PAHO/WHO/SAID 2003). This section identifies the problems to estimate recent coverage and to compare it with the situation prior to the reform; provides and

174 Reassembling Social Security

Table 8.2. Health care coverage/access as percentage of the total population before the reform and in 2002–4, and as percentage of the labor force in 1994–8 and 2000–3 Total population Labor forceb Countries

Years

Public

Social insurance

Argentina

1991 2001

36.4 37.4

57.6 54.4

4.6 7.9

Bolivia

1997

30.0

25.8

10.5

0

Brazil

1998 2003

75.5 75.5

0 0

24.5 24.5

n.a. n.a.

100.0q 100.0q

n.a.

n.a.

Chile

1984 2003

3.1c 16.3c

13.5c 11.6c

100.0c 100.0c

1996 2000

86.2 87.3

Colombia

1993 2002

23.7 53.3d

n.a. n.a.

n.a. n.a.

73.6 100.0d

n.a.

n.a.

Costa Rica

1994 2003

0 0

86.2 86.8

13.8m 13.2m

0 0

100.0 100.0

n.a.

n.a.

Dominican R.

2000

60.0

7.0

12.0

5.0

84.0r

n.a.

n.a.

Ecuador

1994 2004

28.0 41.6

18.0 16.5

20.0 1.7

7.0 1.0

73.0s 60.8

1994 1998

23.1 33.2

El Salvador

2001

40.0e

15.8e

1.5e

n.a.

57.3

n.a.

n.a.

Guatemala

1995 2000

30.0 26.0

16.6 16.6

12.0 30.0n

n.a. n.a.

58.6 72.6n

2000

26.0

Haiti

2000

21.0f

0.5f

38.0f

1.0

60.5f

n.a.

n.a.

t

83.4c 72.1c 50.0 46.7d

Private

Othersa

Total

Years

Coverage

1.4 0.3

100.0 100.0

1997 2001

63.9 56.2

66.3p

1997 2002

17.4 15.2

Honduras

2000

52.0

11.7

1.5

65.2

n.a.

n.a.

Mexico

1985 2002

47.7g 41.8g

41.8g 45.3g

n.a. n.a.

10.5g 12.9g

100.0g 100.0g

n.a.

n.a.

Nicaragua

1990 2001

n.a. 60.0h

18.3 7.9

n.a. n.a.

n.a. 0.5

n.a. 68.4h

1998 2001

14.8 16.6

Panama

1996 2004

39.9i 35.4i

61.1 64.6

i

0 0

100.0i 100.0i

n.a.

n.a.

Paraguay

1999 2001

42.0 33.3

12.4 12.4

6.3 7.0o

1.2 n.a.

62.0 52.7

n.a.

n.a.

Peru

2002

30.0j

26.0

12.0j

3.0

71.0j

1994 2000

28.4 24.2

Uruguay

1987 2003

27.2 38.7k

15.8k 15.9k

30.8k 34.7k

13.9k 7.9k

87.7k 97.2k

n.a.

n.a.

Venezuela

2000 2004

l

38.4 38.3

l

l

l

n.a.

n.a.

l

l

l

l

i

n.a.

Health Care Reforms on Universal Coverage 175 analyzes estimates on total coverage and in the three sectors in all countries and at the regional level; and compares coverage before and after the reform. Major findings are: in most countries the reform resulted either in

unchanged or declining coverage; the regional coverage of social insurance decreased after the reform and the majority of the population remains uninsured; and there was a fall in public-sector access, stagnation or decline

← (Continued) Sources: Total population coverage from PAHO (2005a); Argentina Appendix 1; Bolivia World Bank (2004); Brazil IBGE (2000, 2005); Chile Appendix 2; Colombia Appendix 3; Costa Rica Mesa-Lago (2000 f ), CCSS (2005a); Dominican Republic Lizardo (2004), Rathe (2004); Ecuador Ecuador (2005), IESS (2005); El Salvador DIGESTYC (2002), ISSS (2002); Guatemala Mesa-Lago et al. (1997a), Durán and Cercone (2001); Haiti IADB (2003); Honduras Durán (2003), IHSS (2003); Mexico Appendix 4; Nicaragua Mesa-Lago et al. (1997b), Rossman and Valladares (2003), INSS (2005a, 2005b); Panama Mesa-Lago (2005a), CSS (2005); Paraguay Ramírez (2004, 2005a); Peru MINSA (2002), EsSalud (2004), MINSA (2005b), SEPS (2005); Uruguay ISSA (2003a), INE (2004), Triunfo (2005); and Venezuela D’Ella (2002), IVSS (2006). Labor force from ILO (2003a). Cuba does not publish data on access to the public system but reports that is universal. For country analysis see Mesa-Lago (2006a). n.a. = nonavailable. a In some countries, armed forces, policemen, civil servants, and other separate schemes. b Salaried employees, large and small enterprises, domestic servants, self-employed and unemployed. c No statistics available for 1980; united public-social insurance, private are ISAPREs, ‘others’ are armed forces and police, private insurance and ‘no affiliation’ or direct payment for care, total may include an unprotected segment. d Of the 53.3% affiliated, 29% was in the contributory regime and 24.3% in the subsidized; the remaining 46.7% is ‘tied’ (vinculado) to public hospitals, total is questionable due to duplication in coverage and unknown effective access of vinculados. e Gross estimates based on household surveys for public and private and statistics for social insurance; public responsible for the uninsured (82.7%) but only 40% of the sick were public users. f All data on primary care; social insurance probably on maternity only; total based on 40% without effective access. g Public responsible for the uninsured; social insurance is by the principal program (IMSS) including peasant scheme; there is small but growing private insurance; others are social insurances of federal and state civil servants, armed forces, and oil workers; total is overestimated. h Public coverage is ‘presumed’ hence the total is uncertain. i Public responsible for all without social insurance; a small fraction covered by private; total assumes law enforcement but 20% may lack access. j Public based on access from surveys; private includes care at pharmacies, traditional medicine, etc., total is a gross estimate. k In 1997 social insurance included BPS and ‘collective’ IAMC contracted by the BPS (but IAMC are private entities), private included the individual-voluntary IAMC, partial insurance and private mobile emergencies, and others included part of private; in 2003 public excluded coverage by private mobile emergencies, socialinsurance coverage estimated on affiliation, private included the rest of IAMC and mobile emergencies, and others were the armed forces. l Lack or unreliable data on public, private and others, impeded an estimation of total. m Nonaffiliated to social insurance are covered by private, there are no other groups. n Contradictory estimates on 30% covered by the basic package mainly by NGOs, question the total. o Private was also given as 54% including self-medication, shaman care, etc. p Unreliable estimate because of unknown effective public access. q Wide consensus on universal coverage, but 15% may lack access and pay out of pocket. r Doubtful figure; public is not effective access and probably overestimated. s Doubtful, one source gave 35% without protection. t Author’s gross estimates based on contradictory figures in 1999–2002; public access probably overestimated.

176 Reassembling Social Security in social-insurance coverage, and increase in private-sector coverage (sources for this section are those in Table 8.2, and Appendices 1–4).

8.2.1. Problems to estimate coverage The region lacks historical statistical series on coverage. Based mainly on data provided by institutions and household surveys from some countries, Table 8.2 compiles for the first time estimates of the percentage of the total population covered in nineteen countries, comparing whenever it was possible the year prior to the reform and the most recent year available. The covered population is divided into the three sectors (public, social insurance, and private) and the category ‘others’ (separate schemes, like armed forces, policemen, civil servants, etc.). The accuracy of the estimates is afflicted by several problems summarized below. (a) Most recent estimates are from five- to eleven-years old in nine countries: Argentina, Bolivia, Dominican Republic, El Salvador, Guatemala, Haiti, Honduras, Nicaragua, and Paraguay. In all of them and in Peru and Venezuela only gross estimates are available and incomplete in some cases; Venezuelan data are limited to social-insurance coverage. Seven countries have fairly recent data (2003– 4): Brazil, Chile, Costa Rica, Ecuador, Panama, and Uruguay. Surveys commonly lack standardized criteria to categorize the population protected by the three sectors and ‘others’ in most countries. Best and recent data are often found in the pioneer-high group and the worst in the latecomer-low group. (b) Half of the countries lack estimates on effective access to the public sector: Argentina, Bolivia, Brazil,4 Cuba, Ecuador, Honduras, Nicaragua, Paraguay, Uruguay, and Venezuela (Cuba was excluded from the table because has never published data on access to its public system, lacks social and private insurance and does not report on ‘others’). Available figures on public access in other countries diverge significantly, for example, from 26 to 54% in Guatemala in 2000. Bolivia access to

the public sector reportedly increased from 30 to 65–70% in 1997–2000, but only a minority of the population receives the basic package. Because in twelve countries most of the population is the responsibility of the public sector, the statistical vacuum on public access is the most daunting impediment to estimate total coverage. (c) Social-insurance coverage is overestimated in at least five countries. In the least developed (Guatemala, Honduras, and Paraguay) such distortion often results from: the lack of a comprehensive system of affiliation and identification embracing all beneficiaries; failure to systematically ‘clean’ the registry of affiliates who have left the labor force, emigrated or died (for instance pensioners), and estimates of dependent family with a right to care based on doubtful ratios of dependents per insured. Mexico’s overestimation is an outcome of double counting the same affiliate in two schemes, and in Panama from using one single contribution annually. (d) Coverage by the private sector is impossible or very difficult to estimate in nine countries: Colombia, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Uruguay, and Venezuela. Figures in 2000–1 vary widely, for example, in Honduras from 1.5 to 10%; in Guatemala, from 0.2 to 5% (in private insurance plans); and in Paraguay from 7 to 54% (the latter including self-medication, traditional medicine and care at pharmacies). The estimation of private coverage in Uruguay is extremely complex.5 (e) Similar difficulties were found to estimate the category ‘others’ in eight countries: Brazil, Colombia, Chile, El Salvador, Guatemala, Honduras, Uruguay, and Venezuela. (f) A fundamental difficulty in the calculation of ‘total coverage’ is that it combines insurance (social or private) with access to the public sector that might be effective or not. The least credible data in half of the countries are on total coverage (usually overestimated) and the uninsured, unprotected, or excluded (normally underestimated). Remarkable disparities in figures exist even for the same year

Health Care Reforms on Universal Coverage 177 and sometimes in the same document, for example, total coverage from 73 to 81% in 2000 in Guatemala (the uninsured from 52 to 82%); those without coverage from 18 to 40% in 2001 in Honduras;6 those excluded from 30 to 47% in 2004 in Paraguay;7 and the uninsured from 59 to 73% in 2001 in Peru. Bolivia’s total coverage jumped from 66% in 1997 to 95%–100% in 2000, because public effective access was unknown and guesstimates in the other two sectors varied as much as 10 percentage points.8 Colombia’s total coverage is dubious because of unknown effective access of the population tied to public hospitals. Uruguay’s total coverage is overestimated due to double coverage and because until 1987 data embraced the entire population but since 1999 are restricted to urban areas and exclude rural areas that probably have lower coverage. Argentina illustrates the overwhelming difficulties to assess total coverage: (i) data are given separately for each of the three sectors, based on different sources, neither consolidated nor comparable through time; (ii) no estimate is available on public access, due to this sector segmentation between provinces and some municipalities, with diverse coverage levels; (iii) OS information is neither complete nor transparent because of double coverage with private prepaid plans and mutualaid providers, only 41% of OS affiliates were active contributors in 2004 and it is unknown if the remaining 59% had a right to care; (iv) the private sector has many heterogeneous providers that combine a high-income group with good coverage and another that pays out of pocket (estimates differ by as much as 9 percentage points in the same year); and (v) there is no information on coverage in ‘others.’ (g) It is extremely difficult to contrast total coverage before and after the reform due to the data flaws explained above and because: (i) only one observation or year is available (Bolivia, Dominican Republic, El Salvador, Haiti, Honduras, and Peru); (ii) there are two observations but not one before the reform (Brazil, Chile, and Paraguay); (iii) the reform was not implemented (Ecuador) or it was but

had little importance (Panama and Uruguay); and (iv) there was considerable multiplicity of social insurances prior to the reform in Chile and Colombia, changes in affiliation among sectors before and after the reform in Argentina and Colombia, and unconsolidated series of insured and uninsured and shifts between them in Mexico. The starting year of the Colombian reform is imprecise because the law was enacted in 1993, the contributory regime began in 1995 and the subsidized regime in 1996.

8.2.2. Analysis of the available data on coverage: currently and before the reform Considering the problems discussed above, Table 8.2 indicates that as older and more developed the social security system and larger the size of the formal sector, higher the coverage of both social insurance and the total population, and vice versa. In 2001–3, the pioneerhigh group had the highest total coverage ranging from 97% in Uruguay to 100% in Argentina, Brazil, Chile, and Costa Rica (Cuba lacks data but claims universal public access). Universal coverage reported in 2000–4 in the three most advanced intermediategroup countries is debatable. In Mexico and Panama9 part of the insured do not have effective public access and total coverage is overestimated due to double insurance affiliation; in Colombia coverage could either be 100% or 53% depending on whether the population supposedly ‘tied’ to public hospitals is included or excluded in the calculation. In the three least developed intermediate countries, Bolivia, Ecuador and Peru, total coverage ranged from 61 to 71%. Venezuela’s total coverage is unavailable. In the latecomer-low group coverage in 2000–1 stretched from 57 to 68% in five countries: El Salvador, Haiti,10 Paraguay, Honduras, and Nicaragua. Higher coverage in the other two countries is doubtful. Guatemala total coverage of 73% is based on contradictory

178 Reassembling Social Security figures of public access and the uninsured as already explained. Dominican Republic’s total coverage of 84% relies on an unrealistic 60% access to the public sector and a total uninsured of 24% overestimated due to double coverage between social insurance and private prepaid plans; merely 7% of the population (basically poor) in 2005 was covered by the subsidized regime, the only one that was in operation (Lizardo 2006). Data from nine countries indicates that in four there was no change in total population coverage before and after the reform: Argentina, Chile and Costa Rica, which had universal coverage prior to the reform, and Mexico and Panama that had high coverage before. Total coverage rose in Uruguay and Guatemala, the former had virtual universal coverage before the reform whereas the latter substantial increase is doubtful as explained; coverage decreased in Ecuador and Paraguay, and either increased or fell in Colombia depending on inclusion or exclusion of the population tied to public hospitals. Brazil lacks figures before the reform, but a comparison in a period of five years after the reform indicates that there was no change in coverage and that is virtually universal. Complexities and distortions in comparing coverage before and after the reform require explanation in three countries. In Chile some thirty-five social-insurance health schemes covered 71% of the population in 1973 (the year of the military coup), that fell to 62% in 1980 (Mesa-Lago 1989). The reform of 1981 unified all previous schemes into the public-social-insurance sector (FONASA, except for the armed forces and the police) and introduced a single ID for all affiliates, as well as a registry of the new private ISAPRES, hence coverage figures became more precise. FONASA coverage declined from 83 to 59% in 1984–97 but rose to 72% in 2003 (about the same level as in 1973), whereas coverage by ISAPREs increased from 3 to a peak of 26% in 1997 and thereafter descended to 16% in 2003 (see Appendix 2).11 In Colombia several problems impede an adequate comparison of coverage before and

after the reform: (a) affiliation in many socialinsurance health schemes prior to the reform are comparable to contributory-regime affiliation under the reform, but there were no consolidation of previous data and part was lost when the collection system was changed; (b) about half of the uninsured population was the responsibility of the public sector (nation, departments, and municipalities) but effective access was uncertain; after the reform that segment appears as affiliated, mostly in the subsidized regime, but the uninsured poor are tied to public hospitals as before the reform; and (c) surveys that measured affiliation and access before and after the reform were taken by two different agencies with diverse questionnaires that obscure comparison of results. Statistics after the reform are also flawed: there is not a consistent series on affiliates in the contributory and subsidized regimes in 1995–2002 and available data are afflicted by many contradictions (see Appendix 3 notes). The reported extension of coverage in both regimes is overestimated due to double counting in the number of affiliates: the superintendence estimated that 13% of contributoryregime affiliates had double affiliation in 1999, if applied to the subsidized regime such correction had decreased its coverage by 7 percentage points. Significantly different estimates of total affiliation as percentage of the total population are given in the same year, for example, from 45 to 57% by three sources in 1997. Finally, the rise of affiliation in the contributory regime does not go together with a similar increase in contributors, actually the breach between affiliation and contribution has expanded and consequently the overestimation of coverage (Jaramillo 1999; MINSALUD 2000; SNS 2000; Castaño et al. 2001; Sojo 2001a; Felizzola 2002; Martínez et al. 2003; Uthoff 2003). Although not technically comparable, total affiliation was higher in 1993 than 2002 (74% vs. 53%) whereas the opposite was true of total nonaffiliation (26% vs. 47%).12 Total coverage (100%) in 2002 (Table 8.2) is unreliable because effective coverage of vinculados was impossible to assess.

Health Care Reforms on Universal Coverage 179 In Mexico two separate (nonconsolidated) series on population coverage report the insured in five social insurances and the uninsured that legally should be protected by the federal ministry of health (SSA). A consolidated series for 1985–2002 estimated in Appendix 4 shows the insured disaggregated by the five social insurances and the uninsured that used SSA services. The total insured are overestimated because part of them had double coverage, for example, 5 to 23% of IMSS affiliates also had private insurance. Those covered by IMSS peasant program were reported as uninsured in 1999 (accounting for 11% of the total population) but later were included among the insured, thus explaining the jump in insured coverage between 1995 and 2000.13 The insured oscillated but increased from 52 to 58% in 1985–2002, both as a result of the extension in peasant coverage and the mere shift in their inclusion from uninsured to insured. The uninsured also fluctuated but decreased from 48 to 42% during the period. Uninsured users of SSA rose from 26 to 28% in 1999–2003. The introduction of a public-sector popular health insurance (SPS) in 2003 and its reported expansion to 11% of the population in 2006 would result in 69% of the population insured in that year (assuming no change in other insured and SSA users) visà-vis 52% in 1985.

8.2.3. Rough estimate of regional coverage Chapter 1 estimated average total coverage in the region before the reform but restricted to social insurance, excluding coverage by the public and private sectors, hence can’t be compared with estimates on total population coverage by the three sectors and ‘others’ in Table 8.2. The comparison is further obstructed by uncertain effective access to the public sector and lack of data on ‘other’ programs in onethird of the countries. It is viable, however, to estimate a regional average (weighted by the population of each country) by social insurance before and after the reform (excluding

Brazil and Cuba public systems): increase from 43 to 52% in 1980–90 but decrease to 41% c.2001 for about 136 million insured (socialinsurance coverage from Table 8.2 and total population from ECLAC 2003c, 2005c). All these averages are below the ILO minimum norm of 75% coverage of the resident population, but an incalculable part of it has access to the public sector or is covered by the private sector and ‘others’.14 Roughly 15 million or 2.8% of the population (59 million and 11.5% including Brazil) was insured by the private sector c. 2001. Adding social and private insurance a total of 151 million had insurance for 45% of the total population leaving 181 million or 55% uninsured; including Brazil population and private insurance, the total number of insured ascended to 195 million leaving 325 million uninsured, higher than the previous estimate of 100 to 260 million uninsured made in 1989–96 (PAHO/WHO/SAID 2003).

8.2.4. Trends in coverage by sector Trends in coverage/access by sectors and countries before and after the reform (Table 8.2) can only be detected in a limited number of countries and with extreme caution; the comparison was unfeasible in Bolivia, Dominican Republic, El Salvador, Haiti, Honduras, and Peru because only one year was available. Public-sector access, the most uncertain (nine countries), appears to have fell in five countries (Colombia, Guatemala, Mexico, Panama, and Paraguay), increased in three (Argentina, Ecuador, and Uruguay) and stagnated in Brazil. Social-insurance coverage, the most certain (twelve countries), was stagnant in five of them: Costa Rica,15 Guatemala,16 Paraguay, Uruguay and Venezuela, the first with high coverage and the last four with relatively low coverage. Coverage declined in four countries (Argentina, Chile, Ecuador, and Nicaragua17 ) and increased in three (Colombia, Mexico— mainly by the peasant program, and Panama18 ). Additional data to that in Table 8.2

180 Reassembling Social Security show stagnation in El Salvador and decline in Peru.19 Private-sector coverage, relatively uncertain (eight countries), seemed to have risen in five countries where access to the public sector and social-insurance coverage contracted (Argentina, Chile, Guatemala, Paraguay, and Uruguay), to be stagnant in two (Brazil and Costa Rica) and to decrease in Ecuador but its figure is unreliable. Although data are needed from more countries, as well as better quality in most, the above figures suggest a fall in publicsector access, stagnation or decline in socialinsurance coverage, and increase in privatesector coverage.

8.3. Coverage by location, urban–rural zones, and ethnic groups Fifteen countries have data on differences in social-insurance coverage by geographic areas such as regions, states, departments, provinces, and municipalities; Cuba, Dominican Republic, El Salvador, Haiti, and Venezuela lack information, and figures are available only in urban and rural areas in Paraguay, and in the capital city and the rest of the country (interior) in Uruguay. Data refer to the public basic package in Brazil, all combined insurances (including private) in Ecuador and social insurance combined with IAMC in Uruguay. The resulting evidence summarized below unveils a strong positive relationship between degree of coverage and urbanization and development, and an inverse relationship between coverage and rural location, poverty and indigenous ethnicity.20 The five pioneer-high countries have nil or little geographical inequalities in coverage. Costa Rica’s social-insurance functions in the entire country and gave priority to peripheral areas in 1995–2002 increasing their coverage from 30 to 90%; Uruguay lacks data on the departments and rural areas but

urban coverage in the capital and the interior were equal at 97% in 2003. The extreme ratio between the best and worst covered geographic area was very small in the other three pioneer-high countries ranging from 1.2 times in Brazil to 1.4 in Argentina and Chile; the ratio gradually increased in the intermediate group and jumped in the latecomer group: 2 in Colombia, 3 in Mexico and 4 in Ecuador (in these two countries geographical inequalities are attenuated by peasant insurance), 5 in Panama and Paraguay, 155 times in Guatemala, 350 in Honduras (only 39% of departments are covered), and 400 in Nicaragua.21 The best covered geographic areas are the most developed, urbanized and wealthy, while the worst covered or not covered at all are the least developed, rural and with the highest poverty incidence. The segmentation and lack of coordination of the health care system in more than half of the countries intensifies geographical inequalities because of divergent coverage by the three sectors. Data for 2002–4 show that social insurance and private providers (particularly the latter) are greatly concentrated in capital cities and other highly urbanized areas, whereas rural and poor areas are mostly protected by public services: in Argentinean wealthiest provinces coverage by national social insurances (OS) and private providers was predominant whereas in the most rural and poorest provinces the public sector and provincial OS prevailed and coverage by private insurance was nil; in Chile 56% of the insured in ISAPREs were in the metropolitan region that had 41% of the total population; in Honduras coverage by private insurance ranged from 5% in the second city to 0.5% in rural areas; in Paraguay 31% of urban areas (mainly the capital) had social or private insurance but only 7% in the rural sector and the poorest departments lacked private insurance altogether; in Peru 61% of affiliates to private providers lived in the capital that had 29% of the total population; and in Uruguay 57% of urban coverage was by private IAMC and 54% of interior coverage by the public sector.

Health Care Reforms on Universal Coverage 181 The proportion of indigenous peoples relative to the total population varies significantly: 41%–62% in Bolivia and Guatemala, 25%–35% in Ecuador and Peru, 11%–15% in El Salvador and Honduras, 5%–8% in Chile and Mexico, and 1%–2% in Argentina, Costa Rica and Paraguay (data were not available on Colombia, Nicaragua, Panama, and Venezuela). Statistics on health care coverage by ethnicity are rare; no data were accessible from seven countries with sizable indigenous populations, but figures for 2000–3 indicate that they are concentrated in the poorest, most rural, and worst covered areas by social insurance and the public sector. For instance, 74% of Bolivian inhabitants in the Chuquisaca department was indigenous and suffered a poverty incidence of 70% but only 18% was covered, the lowest coverage in the country; and Mexico’s worst social-insurance coverage was in the poorest four states, which had the biggest concentration of indigenous population. Important exceptions are three programs that cover indigenous peoples, the first two predating the reform. Ecuador’s peasant program provides primary care to rural and poor departments where most indigenous people live and yet indigenous coverage was 69% of nonindigenous coverage and the quality of care of the peasant program was lower. In Mexico 35% of those covered by IMSSOportunidades are indigenous people living mainly in rural areas and two-thirds endure very high marginality; the program emphasizes prevention and primary care adapting to local patterns of morbidity and mortality (community volunteers articulate the program services with indigenous traditions and customs), and IMSS hospitals at second and third levels provide care to affiliates for highrisk pregnancy and other procedures. In Chile 80% of the indigenous population is covered by public-social insurance and 47% of them are indigents. Finally, countries that target the basic package on the poor and rural and urban-marginal areas where indigenous populations are concentrated should cover part of them (e.g. Guatemala and Peru).

An evaluation of the impact of the reform on geographical inequalities feasible only in Colombia showed an impressive improvement albeit with some problems: the extreme ratio between the best and worst covered departments decreased from 2.8 to 2 times in 1993–2001, while the ratio of urban–rural inequality declined from 4.4 to 1.3 times.22 There were increments in affiliation in all departments and areas, but they were stronger and faster in the least developed and rural than in the most developed and urban (recall that 1993 data exclude public-sector protection of the poor). Despite the increase in coverage, the gap between urban and rural district capitals remained the same in 1997–2003;23 coverage by the subsidized regime in municipalities was inversely correlated with unsatisfied basic needs (e.g. 17 and 89% in Bogotá vs. 24 and 44% in Pereira), and the subsidizedregime basic package for poor and low-income affiliates is half the contributory-regime package.

8.4. Coverage by income level Statistics on social-insurance coverage by income quintiles, available in thirteen countries for diverse years in 1996–2003, demonstrate that coverage (particularly private) increases with income and vice versa. Information from a few countries also indicates that coverage rises with education and decreases with age in the private sector but increases with age in the public sector. The impact of the reform on coverage could be assessed only in two countries with inconclusive results. The 1996 series from IADB reproduced in Table 8.3 for seven countries (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Paraguay, and Peru) is flawed; more recent statistics for 1997–2003 in eight countries are more accurate and allow comparison with the 1996 series in Bolivia and Colombia.24 The 1996 series overestimated coverage by income quintiles (particularly the lowest) in

182 Reassembling Social Security

Table 8.3. Total population covered by health social insurance distributed by income quintiles, between 1996 and 2003 Quintiles (from lowest to highest income) (%) Countriesa

Years

1

2

3

4

5

Argentina

1997

35.1

62.1

73.0

83.7

91.0

2.59

Bolivia

1996 2000

19.6 5.8

44.8 9.9

67.7 15.0

87.9 23.6

97.9 31.0

4.94 5.34

Brazil

1998

71.6

88.7

96.7

97.7

98.6

1.37

Colombia

1993 1996a 1997 2000

6.3 60.6 43.2 35.0

13.5 85.2 46.0 43.9

24.0 92.8 54.0 48.9

30.0 98.9 64.0 59.4

43.8 98.1 78.8 74.7

6.95 1.62 1.82 2.13

Dominican R.

2002

Ratio (5/1)

2.1

3.7

6.6

7.7

7.1

3.38

b

11.8

17.8

25.3

27.6

35.7

3.02

Ecuador

2003

Guatemala

1996

9.3

16.1

31.1

62.8

91.5

9.84

Haiti

1996c

24.0

37.3

47.4

60.7

78.2

3.26

Honduras

1999

0.5

4.0

10.5

16.4

27.1

54.2

Nicaragua

1996

32.9

58.8

79.8

86.0

92.3

2.80

Paraguay

1996

41.2

49.9

69.0

87.9

98.1

2.38

Peru

1996

14.3

49.6

75.4

87.2

96.7

6.76

24.3

56.4

75.5

89.6

96.5

3.97

Uruguay

d

1999

Sources: 1996 from IADB (2004b). Other years: Argentina Centrángolo and Devoto (2002); Bolivia World Bank (2004); Colombia 1993 Céspedes et al. (2001), and 1997 and 2000 Salud Colombia No. 58, (2001); Dominican Republic Lizardo (2004); Ecuador Quevedo (2005); Honduras Durán (2003); and Uruguay ISSA (2003b). Data were not available on Chile, Costa Rica, Cuba, El Salvador, Mexico, Panama, and Venezuela. a IADB (2004b) series, inconsistent with those for other years. b Includes also private insurance. c Sickness-maternity insurance (the latter initiated in 1999) has not been fully implemented and has few beneficiaries, hence data are unreliable. d Includes IAMC and other social insurances.

most countries for four reasons: (a) The series reported coverage in the poorest quintile is considerably higher in five countries than their respective total population overall coverage by the principal social-insurance program: Haiti 24 and 0.5%; Guatemala 9 and 3%; Nicaragua 33 and 13%; Paraguay 41 and 12%; and Peru 14 and 26%. (b) Reported 1996 coverage is also higher than survey data on access

in two countries: in Bolivia only 3% of the poorest quintile received care in social insurance in 2000 compared with reported coverage of 20% in 1996; in Guatemala merely 1% in the poorest quintile used social-insurance health services in 1999 versus reported coverage of 9% in 1996. (c) Reported 1996 coverage in the poorest quintile go against very high poverty incidences afflicting four countries in

Health Care Reforms on Universal Coverage 183 1997–8: 9% coverage vis-à-vis 61% incidence in Guatemala; 33 and 70% in Nicaragua; 41 and 61% in Paraguay; and 14 and 48% in Peru. (d) More recent surveys show considerably lower coverage in the poorest quintile in two countries: Bolivia 20% in 1996 versus 6% in 2000; and Colombia 61% in 1996 versus 43% in 1997 (see below). Once the flawed 1996 series is removed, Table 8.3 shows that the lowest coverage in the poorest quintile is found in the latecomerlow group (0.5% in Honduras and 2% in Dominican Republic), followed by those in the intermediate group (6% in Bolivia and 12% in Ecuador—including private insurance), and the highest coverage in Uruguay’s social insurance (24%) and Colombia’s system (35%). The highest extreme ratios of coverage between the wealthiest and the poorest quintiles are similarly found in the latecomer group (fiftyfour times in Honduras and ten times in Guatemala), declining in the intermediate group (seven times in Peru and five times in Bolivia) and the lowest ratio in the pioneerhigh group (twice in Argentina and Colombia and four times in Uruguay). Brazil is the only country that provides data on public-sector protection, and it has both the highest coverage in all quintiles (72% in the poorest and 99% in the wealthiest) and the lowest extreme ratio (1.37) among all countries. In all thirteen countries and in all years (even in the flawed 1996 series) coverage steadily increases with income—a trend more accentuated in the private sector of five countries. Argentina’s richest quintile was 30% covered in the private sector and only 8% in the public (61% in OS), whereas the poorest quintile was 4% covered in the private and 63% in the public (31% in OS). Brazil’s two wealthiest quintiles had 80% coverage in the private sector and only 2% of the wealthiest decile used the public sector. In Chile 65% of the population covered by ISAPREs was in the two wealthiest quintiles but 5% in the poorest quintile, whereas 60% of the population covered by public-social insurance was in the poorest two quintiles but 5% in the wealthiest quintile. In Dominican Republic a quintile

distribution for 2002 that includes privateinsurance compared with that in Table 8.3 restricted to social insurance shows significant coverage increases in the wealthiest two quintiles: from 8 to 29% in the fourth and from 7 to 44% in the fifth. Ecuador’s coverage by private insurance notably rose with income: from 0 to 0.6% in the poorest three quintiles to 8% in the wealthiest quintile. Data from two countries also exhibit a positive relationship between education and both degree of coverage and affiliation to private insurance/use of private services. Among those with incomplete primary education in Brazil (the poorest) in 2001, 54% used exclusively the public system (SUS), 34% resorted to public and private provision combined, and 12% were not users mainly for lack of access; conversely only 5% of those with higher education exclusively used the SUS, 47% resorted to public and private provision combined, and 48% were not users because they had private protection. Ecuador’s total insurance coverage increased with education from 14% among the uneducated to 49% among those with higher education in 2004, the use of private services increased with education from 1 to 19% respectively. Data for 2004 from Chile and Ecuador indicate a decreasing coverage in the private sector as affiliates’ age increases while the opposite occurs in the public sector. This was found among all income quintiles in Chile (even in the wealthiest), because health risks rise with age (particularly in the population age 65 and over) and the private-sector practice of cream skimming that pushes out the elderly to the public sector. Although data from other countries are needed, those summarized above question the capacity of the private sector to extend coverage among the lowest income, worst educated and older segments of the population because of risk-selection. An analysis of the impact of the reform on coverage by income levels was possible in only two countries and with uncertain results. Bolivia exhibited a significant decline of social-insurance coverage in all quintiles between 1996 (in the midst of the reform) and

184 Reassembling Social Security 2000, from 20 to 6% in the lowest quintile, albeit might be partly a result of the explained overestimation of coverage in the 1996 series (Table 8.3). It is widely asserted that Colombia 1993 reform reduced inequalities in coverage by income quintiles, and yet if the 1996 flawed series is used, Table 8.3 exhibits a steady decline of coverage in all quintiles in both 1997 and 2000. On the other hand, the assertion of expanded coverage holds when the years 1993 and 1997 are compared, but still a significant decrease occurred between 1997 and 2000. The population affiliated to social insurance in 1993 rose from 6% in the poorest quintile to 44% in the wealthiest quintile but those figures excluded the poor population with public access part of which became affiliated to the subsidized regime after 1993. Total affiliation in 1997 (adding up the contributory and subsidized regimes) increased in all quintiles vis-à-vis that in 1993, an outcome of coverage extension to dependent family in the contributory regime (that favored the wealthiest quintiles) and the creation of the subsidized regime (that helped the poorest quintiles). Conversely, total affiliation in 2000 declined in all quintiles relative to 1997, from 43 to 35% in the poorest quintile and from 79 to 75% in the wealthiest quintile, due to problems confronted by the reform. Apparently coverage in 2000 was higher than in 1993 but recall that the latter excludes publicsector protection of the poor. Data on the subsidized regime alone indicate that coverage in the two poorest quintiles fell between 1997 and 2000, whereas it increased in the two wealthiest quintiles.

8.5. Coverage of the self-employed and other groups difficult to incorporate This section compiles and compares all available data on social-insurance coverage of labor groups difficult to affiliate such as

self-employed, domestic servants, ruralagricultural workers, and employees of microenterprises, identifies major problems that obstruct their incorporation and a few successful approaches to overcome them, and explores the role of labor market composition vis-à-vis policies to extend coverage. Data on effective coverage of the selfemployed and other vulnerable groups are scarce and nonstandardized. The ILO (2003a) has calculated the level of social protection in the Latin American labor market based in household surveys taken between 1994–8 and 2000–2 in seven countries: Argentina and Chile in the pioneer-high group, Bolivia, Ecuador and Peru in the intermediate group, and Guatemala and Nicaragua in the latecomer-low group. Surveys were also taken in Brazil and Mexico but their data are not comparable with the rest because only salaried workers were included. Social protection levels were estimated according to the degree of vulnerability set by workers’ labor condition typified by three variables: kind of occupation (employer, salaried worker, selfemployed, nonsalaried worker, and unemployed), type of enterprise or sector (large, small, and public), and labor skills (professional or nonprofessional). Upon this classification seven groups were identified and ranked in social protection (from high to low): four with low vulnerability and high social protection (employer, salaried worker in large enterprise, salaried employee in the public sector, and self-employed professional) and three groups with high vulnerability and low social protection (salaried worker in small enterprise including domestic servant, unskilled self-employed, nonsalaried worker, and unemployed). Statistics on social-insurance health care coverage were thwarted because survey questionnaires were not always standardized and only the surveys of seven countries included questions that supported the following estimates in 2000–2: the highest coverage was among salaried employees in the public sector (68–93%), followed by salaried workers in large enterprises (39–94%), but significantly

Health Care Reforms on Universal Coverage 185 lower among salaried workers in small enterprises (5–42%), and the lowest coverage was in the three groups with high vulnerability. In the period of observation, coverage of unskilled self-employed decreased in three countries and increased in three; there was a single observation in one country (ILO 2003a; Gasparini and Betranou 2005). As already explained (Section 3.1.2), most of the labor force in the region is in the informal sector or in rural areas, with unstable employment, low salary and limited capacity to contribute, all significant barriers for incorporation. The informal sector steadily grew since the early 1980s to 47% of urban employment in 2002, and is much larger in the latecomer-low group and the least developed countries of the intermediate group, which have the lowest social-insurance coverage. The self-employed comprised most of the informal sector and they are very difficult to cover by social insurance because they have voluntary legal coverage in thirteen countries, are excluded in three and have mandatory coverage only in Colombia and Costa Rica (see 8.1 above). Furthermore, when affiliated, the self-employed pay a percentage contribution equal to the sum of the percentages paid by the salaried worker and the employer they lack, and it is very complex to detect and affiliate them and collect their contributions. Self-employment accounts for a significant percentage of the urban labor force (particularly in the least developed countries) but their coverage as a proportion of total affiliates in the main social-insurance program is extremely low, as demonstrated in six countries with legal voluntary affiliation in 2000– 3: 40 and 2% in Peru, 38 and 1% in Honduras, 35 and 0.4% in Nicaragua, 34 and 5% in Ecuador, 32 and 0.2% in Paraguay, and 23 and 0.1% in México (labor force percentages from Table 3.3, coverage from ISSA 2003b; Durán 2003; INSS 2003; Saldaín 2003; EsSalud 2004; IESS 2005). Additional data from four pioneer-high countries, similarly with voluntary legal coverage, also show a very small percentage of the self-employed covered. In Chile they accounted for 1.7% of total insured

in the private sector in 2002; most of them were covered by public-social insurance that provides free care to the poor and fiscal subsidies to those with low income (SI 2003). Argentinean self-employed accounted for 15% of total affiliates in OS in 2001–4 and most of them were covered by the public sector25 (Bertranou and Vázquez 2006). In Uruguay 92% of the self-employed was covered in 2000 but only 14% by social insurance versus 45% by the public sector, and 33% by private providers (ISSA 2003a). Costa Rica also had voluntary coverage until 2006 but 45% of the self-employed are affiliated to social insurance because those poor receive free care and the state provides a subsidy (in lieu of the employer’s contribution) to those with low income (Programa Estado 2002, 2004). Colombia has legal mandatory coverage of the self-employed and the reform goal was to incorporate 80% of them by 2000 but estimates of actual affiliation in that year ranged from 4 to 29% (Martínez et al. 2003; Ramírez 2004), despite that poor and low-income selfemployed receive fiscal subsidies. Estimates of nonprofessional self-employed coverage in three countries with optional affiliation to social insurance in 2000–1 were 0.8% in Nicaragua, 3% in Peru, and 35% in Chile; but coverage was from two to four times higher when considering all sources of protection (including the public sector): 3% in Nicaragua, 12% in Peru, and 74% in Chile (ILO 2003a). Inferences on the effect of social insurance legal voluntary versus mandatory coverage on actual protection of the self-employed are difficult because it is only compulsory in Colombia (and Costa Rica since 2006) and reported coverage in 2000 is contradictory. And yet, data on countries with voluntary affiliation show consistent low coverage (0.4–1% of total affiliation) in the latecomerlow and intermediate groups, which have high proportion of self-employment in their labor force. Coverage is considerably higher in pioneer-high countries that have a small proportion of self-employed: 14 to 34% in Argentina, Chile and Uruguay, where the selfemployed are mostly care for in the public or

186 Reassembling Social Security public-social-insurance sector; coverage peaks at 45% in Costa Rica due to free care to the poor and fiscal subsidies to low-income self-employed. Statistics are not available in Brazil and Cuba but their universal and largely free public systems probably cover all the self-employed (Cuba’s percentage of selfemployed is the lowest in the region). We are left with the question on whether the degree of coverage is a function of the composition of the labor force or specific policies to stimulate affiliation such as legal mandatory coverage or a combination of free and subsidized care either in social-insurance or public systems. In any case, social insurance in the least developed countries face an insurmountable task to cover the self-employed because of their relative large size, low income, and scarce fiscal resources to subsidize their incorporation; a better alternative to these countries would be to protect them by public systems. More developed countries could either follow the Costa Rican approach or that of Brazil and Cuba, the answer to whether the Colombian reform model constitutes another feasible alternative will have to wait until better data are available. Legal mandatory coverage of domestic servants in thirteen countries is applied in only ten (they have voluntary coverage in one country and are excluded in six) and does not appear to be effective as data on the percentage of those workers insured indicate: 3% in Paraguay, 11% in Panama, 13% in Ecuador, and 26%–27% in Colombia and Uruguay.26 Employees of microenterprises are covered by 51% in Argentina but only by 4% in Mexico; in Honduras coverage according to the size of enterprises in 2000 decreased from 46% among those with more than ten employees to 4% among those with only three employees27 (Flood 1997; Felizzola 2002; Bucheli 2003; ISSA 2003b; Saldaín 2003; MEF 2004; IESS 2005). In latecomer-low countries and the least developed in the intermediate group, the rural population and the agricultural labor force range from 29 to 55% and this segment is also difficult to incorporate because of the nature of its labor relations, dispersion, and very low income. Half of the countries legally exclude

agricultural workers and is not surprising that few figures are available on their coverage: 1.5% in Ecuador (but 18% of the rural population covered by peasant insurance), 2% in Honduras, and 3%–6% in Mexico (but 29% of the rural population covered by peasant insurance) (Durán 2003; IMSS 2004; IESS 2005a; Munguía 2006). Conversely, Argentinean agricultural workers and their dependents, affiliated in the rural OS, were 35% of the total rural population in 2005 (Paulucci 2006). There are virtually no statistics on coverage of dependent family of informal workers. Mexico’s principal social insurance (IMSS) created in 1997 a family health scheme (SSF) geared to the extension of coverage of informal workers and their families with the incentive of a 56% cut in workers contributions and state supplements, but SSF as percentage of total IMSS insured was only 1.3% in 2000 (ISSA 2003b). The major obstacle faced by this program is that half of the labor force earns less than three minimum wages; despite the incentives provided, still these workers’ contribution is very high and, because they lack an employer contribution, they pay six times what a salaried worker with the same income pays. SSF is more affordable for selfemployed earning higher wages such as taxi drivers and middle-size merchants (BrachetMárquez 2001).

8.6. Access and utilization According to ILO-PAHO (2005) estimates 110 to 140 million people or 20%–25% of the total population of the region lack access to health care. PAHO study on six Andean countries found ‘no evidence in any of them that the reform has had incidence on either . . . the five indicators selected to evaluate access, nor in the three chosen to measure the use of health facilities’ (PAHO 2002c: 23). This section examines several facets of access and use of services: percentage of the sick population that did not seek care and their reasons; where care was received (in the three sectors, selfmedication and traditional medicine); access

Health Care Reforms on Universal Coverage 187 disparities between rural and urban areas, income quintiles and indigenous and nonindigenous population, and potential reform impact on access.28 Data on access from households surveys in sixteen countries in 2000–4 (inaccessible for Cuba, Haiti, Panama, and Venezuela) show that from 4 to 62% of the population who was sick or suffered an accident did not seek health care; the lowest percentage was in the pioneerhigh group and the highest in the latecomerlow group: 4 and 7% in Uruguay (capital and interior respectively), 24% in Chile, 30 and 37% in Mexico (insured and uninsured respectively), 40% in El Salvador, 48% in Bolivia and Ecuador, 51% in Paraguay, 56% in Peru, and 62% in Dominican Republic. The most frequent reasons for not seeking health care identified by the surveys (in Brazil, Colombia, Chile, Dominican Republic, Ecuador, El Salvador, Paraguay, Peru, and Uruguay) were: lack of money or too expensive care, sickness perceived as minor, selfmedication or use of home remedies, too far away health facility (particularly in rural areas), long waiting list, deficient care or poor confidence in the personnel, lack of medicines or physicians, complex and excessive paperwork, and absence of information. Lack of money was the main reason given in Colombia in 1997–2003 (58% among nonaffiliates and 33%–67% among affiliates) and by 52% of respondents in Ecuador in 2004. A significant proportion of the sick that didn’t seek institutional treatment either selfmedicated or used home remedies, particularly among the poor, rural, and indigenous populations: 62% in Dominican Republic, 40%–48% in Ecuador and El Salvador, and 38% in Chile. Most of the sick sought care in the public sector: 63% in Chile, 53%–59% in Brazil, Honduras and Dominican Republic, 42% in El Salvador and Ecuador, and 27%– 30% in Bolivia, Paraguay, Peru and Mexico.29 Another part sought care in the private sector because of lack of public or social-insurance access: 54% in Paraguay, 41% in Brazil and Honduras, 30%–32% in Dominican Republic30 and Chile, 22% in Mexico, and 8%–11%

in Bolivia, Ecuador, El Salvador, and Peru. The smallest proportion was cared at social insurance: 5%–13% in Dominican Republic, Ecuador, El Salvador, Mexico, Paraguay, and Peru, but 55% among those insured in Mexico’s five schemes. Traditional medicine was used by 48% in Bolivia and 40% in El Salvador. Surveys also revealed four important access disparities: (a) access in rural areas was considerably more restricted than in the capital city and urban areas in Bolivia, Colombia, Guatemala, and Haiti. For instance, Bolivian medical personnel assisted 86% of urban deliveries but only 39% of rural deliveries (the remaining 61% resorted to midwives and shamans); 12% of the urban sick was cared by social insurance but only 1% of the rural sick whose access was restricted to the public sector and the basic package. (b) Access to or use of services is highly correlated with income, regardless of the level of development, for example in Bolivia, Chile, Dominican Republic, Ecuador, Honduras, Peru, and Uruguay. Medical personnel in Bolivia assisted 23% of deliveries in the poorest quintile but 89% in the wealthiest quintile. Among Chilean sick, the poorest quintile had systematically less access than the wealthiest quintile, and a higher proportion used home remedies or were impeded by facilities that were too far or failed to get an appointment. (c) Access by dependent family is considerably lower than access by direct insured, often due to legal coverage restrictions imposed to spouses and children: Nicaraguan access was 3% among spouses and 27% among children but 70% among direct insured. (d) Access to socialinsurance facilities measured by surveys was smaller than reported coverage by institutional statistics, for example, in Bolivia 7% versus 26%, and in Ecuador 5% versus 16%; Brazil access to the public system was 60% visà-vis 74% statistically reported. Indigenous population access is worse than in the rest of the population in at least four countries: In Guatemala half of indigenous peoples lack access to basic services and in Ecuador they endure half of the access to

188 Reassembling Social Security social insurance than the nonindigenous population. In Bolivia, 78% of indigenous households lacked potable water and 72% sanitary services, 30% used traditional medicine or self-medication and only 15% had access to conventional medicine; half of indigenous women deliveries were not assisted by health personnel but by midwives or shamans, and 31% of those speaking indigenous languages were sick and 44% received care contrasted to 18 and 56% respectively among Spanish speakers. In Paraguay’s four departments with the highest indigenous concentration, 90% of them recourse to traditional medicine, 68% don’t go to the health post, and 78% lack physician care. The impact of the reform on access is very difficult to assess because surveys were not systematically taken before and some years after the reform or data were not comparable among such surveys or there was contradictory information. Evidence from several countries indicates, however, that the basic health care package has been a key factor in facilitating and extending access, but confronting several problems (see Section 9.3.1). Bolivia’s access to the basic package reportedly rose from 33 to 52% in 1996–2000 (in the midst of the reform and two years latter, respectively), but it was not specified if relative to the total or target population; furthermore care was reported to be ‘deficient’ and ‘the information on coverage is not reliable’ (PAHO 2005a: 13, 16, 21). Mexican public and social-insurance programs targeted on the poor allegedly cut the population without access to institutional services from 13 to 0.5% in 1985–2000, and yet the target of the popular health insurance (SPS) initiated in 2003 is to provide access to a basic package for 31% of the population by 2010. A more detailed analysis is made of the reform potential impact on access in Brazil, Chile, Colombia, and Costa Rica. There is consensus that Brazil’s reform of 1990 achieved universal coverage, but access and quality of public services vary significantly: ‘We lack an exact measure of the population access to the health system, which

makes impossible to calculate the needed resources to expand such access’ (Biasoto 2004a: 35). A 1998 survey estimated that 15% of the population didn’t have effective access to SUS due to lack of services in the neighborhood or resources to pay for transportation or user fees or to long waiting lists or absence of a physician, and at least 7% of the poor were not users (Medici 2002a). This group confronts scarcity of services for pathologies, public hospitals with insufficient resources, and difficulties to access basic services in some areas. Access has been expanded by the basic package and the family health program, but in 2003 one-third of the population lacked access to medicine; the introduction of generic medication at lower prices albeit positive has not solved the problem of the poor because they lack resources. According to Chile’s survey of 2003, 7% of the population was not affiliated, 8% among the nonpoor, who probably buys private insurance or direct care, and 6% among the poor who may not have access. The publicsocial-insurance network reaches all the population, except in some towns in the north or deep in the south; in rural areas, however, geographic, climatic, transportation, and communication conditions are important barriers for access to emergency care and opportune medium-complex care, and curative care is not guaranteed for all; at the national level there are long waiting lists for surgery and highcomplex care. At least until 2005, ISAPREs neither ensured maintaining coverage for catastrophic illnesses or hospitalization or intensive care, nor coverage of the elderly, as the private firms could annually change coverage conditions, exclusions and co-payments, as well as impose other access restrictions. For example, periods without right or waiting periods for preexisting diseases and pregnancy (which were either excluded or charged higher co-payments for treatment); exclusions of certain pathologies, and ceilings for reimbursement or maximum costs, over which only the legal minimum benefits were granted. Democratic government policies have improved access and reduced ISAPRE restrictions, such

Health Care Reforms on Universal Coverage 189 as free vaccination against influenza and telephone calls for primary-care appointments; new emergency first-level services, a program to reduce hospital waiting lists, outpatient hospital rooms, and loans to pay health expenditures. Since 2004 a guaranteed minimum package of benefits is granted to all the population, and since 2005 exclusions are banned for both the accidental omission of a preexisting disease and when care for preexisting pathologies has not be used for five years (see Section 9.3.1). Colombian departments did not experience a significant change in the use of health services in 1993–7, except for a decline in Bogotá and an increase in San Andrés, but if coverage was indeed extended it was logical to expect a boost in use. At the start of the reform in 1993, 77% of urban inhabitants had access versus 58% among rural inhabitants; in 1997 the gap continued despite slight access increases to 79 and 60% respectively, explainable by geographic-cultural barriers, lack of information and right awareness, and lower rural income. In a 1997 survey of subsidizedregime affiliates, 88% said to be enrolled but 40% didn’t know what that regime was, and 30% of the poor were not enrolled. Only half of those identified as eligible in that regime were affiliated, 31% ignored their rights and 20% didn’t know how to claim them; 33% confronted geographic or cultural barriers of access and 16% said they lacked money to pay for services that were actually free. In a 1999 survey of the entire health care system, 69% said to be affiliated and 31% not to have any affiliation. Five additional access problems faced by affiliates were: 47% didn’t get prescribed medicines and those in the subsidized-regime got less than in the contributory regime; only 46% had an annual preventive consultation, 20% left the consultation ignoring what was their sickness and 50% without information to improve their health; 42% was denied or delayed admission in emergencies; 41% endured a delay of eight or more days for outpatient consultation and 39% for surgery beyond the legal maximum of fifteen days.

In Costa Rica a comparative study in the 1990s indicated that the reform had improved geographic access at the first level of care through the local integrated units (EBAIS): access (in terms of average distance to the closest facility) was in the same day and virtually equal in urban and rural areas, hours of physician availability increased, and coverage in prenatal and child care improved. In 2001, however, only 28% of EBAIS monitored and followed up references to secondary and tertiary levels, which despite extended hours of operation and subcontracting with private providers had not reduced previous long waiting lists (see Section 10.7). Two evaluations of first-care by EBAIS, conducted in 2000–1, showed that the least developed region was ordered last and without change: quality of care was ranked low both to first-level and to comprehensive services to adolescents and elder patients.

8.7. Causes of poor coverage and nonfulfillment of targets The low coverage/access in half of the countries, the seven in the latecomer-low group and Bolivia, Ecuador, and Peru in the intermediate group, is due to external factors to the system but also to the system itself and fiscal policies. External factors are: underdevelopment, economic crisis or low growth, high poverty incidence, large informal sector, substantial unemployment and underemployment, sizable rural population, regions poorly developed, scarce fiscal resources, poor taxing capacity, political instability or crisis, lack of government commitment, different languages and cultures, and ethnic and gender inequality. The health care system itself rather than been neutral can determine the degree of exclusion: a segmented or highly segmented system without coordination, weak regulation, and lack of solidarity are typical in all these countries (PAHO 2002a; PAHO/WHO/SAID 2003).

190 Reassembling Social Security In the latecomer-low group social insurance was introduced late, some countries has not expanded coverage to all geographic areas, and several of them have large rural and indigenous populations difficult to incorporate. Mechanisms for allocating and distributing financial resources have played a crucial negative role in coverage, such as higher expenditures of the social-insurance and private sectors relative to the public sector that is in charge of the protection of most of the population but have grossly insufficient resources, user fees charged by the public sector, as well as significant out-of-pocket expenditures that heavily burden low-income groups. Conversely, countries of the pioneer-high group developed social insurance early, extended its scope, better coordinated or integrated the system and infused elements of solidarity. Health care systems with different organization and financing in this group achieved virtual universal coverage, either with integrated systems (Costa Rica and Cuba) or dual systems with some coordination (Brazil and Chile), with two exceptions regarding integrationcoordination (Argentina and Uruguay). All these systems are either fully financed by the state or receive substantial fiscal transfers and taxes to protect the poor, and two fully cover the self-employed or grant fiscal transfers to those poor and with low income. External positive factors have facilitated coverage, like relatively high level of development, low poverty incidence, large formal sector, political stability and government commitment to extension, and absence of significant indigenous population in most of them.31 To test the facilitating and obstructing factors identified above, eight case studies were selected with diverse health models: three that achieved universal coverage in the pioneerhigh group (Chile, Costa Rica, and Cuba); four that has failed so far in that task (Ecuador and Peru in the intermediate group, and El Salvador and Honduras in the latecomer-low group); and Colombia that requires detailed analysis on the role such factors played in

the failure to meet the reform target of universal coverage. The three successful countries were chosen based on fairly common characteristics: relatively small territorial size, comparable level of development, good communication structure, high urbanization rate (except Costa Rica), low poverty incidence, high literacy, similar stage of epidemiological transition, and lack of significant physical or ethno-cultural barriers in most of them32 (see Mesa-Lago 2000h). In the four countries selected from the intermediate and latecomerlow groups, each pair is fairly comparable in size, economic and social security development, high informal sector and poverty incidence, sizable indigenous population and other factors (general sources in this section are Table 3.3; ECLAC 2004a, 2005a, 2005b; PAHO 2005a). Chile has the second oldest social security system in the continent and had achieved close to universal coverage prior to the 1981 reform that introduced the dual coordinated health care model: public-social insurance that covers most of the population (free for the poor and subsidized for low-income groups) and private that covers less than onefifth. Coverage diminished during the military regime due to two economic crises and severe cuts in the public health budget, but heavy investment in public-social-insurance health care since 1990 led to a recovery of virtual universal coverage. External favorable factors have been an average 4% GDP per capita growth rate in 1990–2004 (contrasted with a regional average of 1% in the period), combined with financial and political stability, as well as positive social indicators in 2003–5, such as a rate of urbanization of 87% (the third highest in the region); poverty incidence 19% and extreme poverty 4.7% (second lowest in the region); the informal sector accounts for only 29% of the labor force and self-employment for 15% (both lowest in the region); and open unemployment 8% (below the regional average of 9% and there is unemployment insurance). Proper policies have compensated significant physical and ethnic barriers: 80% of the indigenous population,

Health Care Reforms on Universal Coverage 191 mostly poor, is freely covered by public-social insurance with fiscal subsidies. Costa Rica’s system is younger that Chile’s but rapidly extended coverage since the early 1970s with unified social insurance (CCSS) that reached virtual universal coverage prior to the reforms of the 1990s and has maintained it for the past twenty-five years; the small private sector is largely confined to primary care through contracts with CCSS. Successive governments have stuck to the commitment to extend coverage and grant free care to the poor and subsidies to low-income self-employed with fiscal transfers. Favorable external factors have been political stability for sixty years and economic stability (except for the 1980–2 crisis); a poverty incidence of 20% in 200233 (third lowest in the region); a relatively large formal sector with 31% informal workers and 18% self-employment (the lowest after Chile); and unemployment of 7% in 2005 (the third lowest in the region). Although the rate of urbanization is low (52% in 2005), rural living standards are among the best in the region, and there are no significant physical or ethno-cultural barriers to obstruct coverage and access. To provide care to small indigenous groups with poor access, the CCSS trains people from the community for health activities, and incorporates positive aspects of traditional medicine. Cuba has a totally unified and gratuitous public system (there is no private sector) since the start of the 1960s that also has reached virtual universal coverage. There has been a strong politico-financial commitment first to extend and later to maintain universality notwithstanding the severe economic and social crisis since the start of the 1990s; political stability (without democracy) has reigned for forty-seven years despite drastic economic oscillations and crises. A family doctor program began in 1984 to facilitate neighborhood access at the primary level; 46% of all Cuban doctors were in that program and cared for 99% of the population in 2001 (but lacked basic medicine), however, one-fourth of them were sent to Venezuela and other countries thus reducing previous domestic

access. External favorable factors have been: a rate of urbanization of 82%; poverty incidence estimated at 20% in 1999–2002 (similar to Costa Rica), the largest state-formal sector in the region (although there has been increasing informality since the early 1990s), and open unemployment officially reported as 2% in 2005 (a questionable figure). There are no significant physical or ethno-cultural barriers to obstruct coverage (MCI 2005b; Mesa-Lago 2005b). Ecuador has suffered severe and prolonged politico-economic instability (GDP per capita averaged 0.6% in 1995–2003), as well as adverse external social factors: 49%–64% urban poverty incidence,34 50% informal labor, 65% underemployment and 14% unemployment (three-fourths of the urban labor force), and 35% of the population is rural, largely consisting of indigenous groups and has the greatest poverty incidence (88% in 1999), and 60% of the rural labor force is selfemployed. The health system is segmented without coordination, there is no basic package, social-insurance legal coverage of dependents is among the worst in the region, only a small fraction of domestic servants and selfemployed are covered, and the public sector charges user fees that increase barriers to access for the poorest. The most effective tool to extend social-insurance coverage has been peasant insurance (SSC) but affiliates fell from 21 to 18% of the rural population in 2000– 4. Selection of SSC beneficiaries is not based on poverty and low health standards, but on the community size and minimum organization (absent in those of small size and lowmobilization capacity), existing facilities, and feasibility; cultural differences have not been taken into account establishing SSC in diverse communities. Main reasons given in a 2004 survey for not been affiliated to social insurance were: lacking the proper age (an average of 28%, but 63%–74% within the 0–14 age bracket because only 9% of children younger than 5 are insured); self-employment work (23%, higher among indigenous people); lack of work (22%, higher among women and the elderly); and parents not affiliated (10%, 22%

192 Reassembling Social Security among indigenous persons) (Ecuador 2005; IESS 2005; González Jijón 2006). Peru’s GDP per capita averaged −0.8% in 1998–2001, the climatic phenomenon of El Niño caused vast damage and export prices fell, and the country faced a serious political crisis in 2000. Poverty incidence grew from 48 to 55% in 1997–2001; the informal sector accounted for 60% of the urban labor force and 42% was self-employed (both second largest in the region); unemployment was 9% and underemployment 51% in 2000. A constant feature has been frequent changes of health ministers and policies (including coverage programs), which have provoked instability and prevented policies to bear fruit. User fees introduced by the reform and suspended in 2001 were a financial barrier for public access of the poor. Dispersion of the population in the highlands and the jungle makes more difficult and expensive to extend health services, and facilities are often located far from communities. There are significant ethno-cultural barriers: indigenous communities distrust health facilities personnel, their modern childbirth techniques, and usual rejection of traditional medicine. The change of government in 2001 led to new policies and improvement in economic growth but the fundamental problems continue unabated (Arroyo 2000; Sanabria 2001; MINSA 2002; EsSalud 2004). El Salvador’s rural population was 42% of the total population in 2005 (the sixth largest in the region); poverty incidence was 49% in 2001 (62% in rural areas); half of the urban labor force was informal and 32% selfemployed, whereas 41% of the rural labor force were self-employed farmers, peasants, or cooperative members; urban unemployment was 7% in 2005 (below the regional average) but underemployment 29%. Both the informal sector and the rural population are excluded from social-insurance legal coverage, dependent female spouses lack sickness coverage and dependent children above 12 are excluded. The health care system is segmented and lacks coordination, health facilities and resources are concentrated in urban areas,

there is not a basic package, and the public sector that legally should protect the poor and low income charges user fees. Health care was also harmed by the prolonged and costly civil war and thereafter by the priority given to reconstruction, GDP per capita still averaged 0.6% in 1996–2004. Two earthquakes in 2001 damaged 25% of public health facilities in 79% of the departments, as well as 54% of socialinsurance hospitals and 87% of medical units; the earthquakes increased poverty and made more difficult to extend coverage (Mesa-Lago 2001d; ISSS 2002). Honduras’ GDP per capita averaged 0.5% in 1990–2004, half of the regional average. The rural population accounted for 48% of the total population in 2005 (the third largest in the region); poverty incidence was 77% in 2002 (86% in rural areas) and extreme poverty was 54% (70% in rural areas), both the highest in the region. The informal sector represented 54% of the urban labor force (the fourth largest), self-employment 37%, and subsistence microenterprises 25%; in the 1990s half of the jobs created consisted of self-employed women or unpaid family members, and salaried workers without contract jumped from 53 to 74%; and 63% of the rural labor force is comprised by self-employed farmers or peasants. All these sizable segments of the population are excluded from socialinsurance coverage, which does not grant sickness care to dependent female spouses and excludes dependent children older than age 11. The health care system is segmented without coordination, there is no basic package, and the public sector probably charges user fees (Durán 2003). The Colombian reform demands a deeper analysis because of its importance and for setting a target of 100% coverage that was not met. Adverse external factors that contributed to that failure were: the severe and long economic recession (GDP per capita averaged −0.6% in 1996–2002) partly caused by the civil war and partly by the macroeconomic adjustment that reduced fiscal transfers to the subsidized regime; a poverty incidence ranging from 51 to 55%, an informal sector

Health Care Reforms on Universal Coverage 193 tantamount to 44% of the labor force and 39% self-employment, and unemployment stretching from 15 to 18%. The most important negative factors related to the health care system and reform policies were: (a) the battle among policymakers over granting demand subsidies to the subsidized regime, combined with hesitation and reverses at national and departmental level, which postponed the start of that regime until 1996 and delayed and reduced resource allocation to it (funds in constant prices shrank since 1998 and 36% of the projected resources had not materialized by 2000); (b) substantial and rising evasion, payment delays and under-declaration of wages in the contributory regime, aggravated by transgressors who received free care in public hospitals, depleting resources intended for coverage expansion to the poor (with the lost resources the subsidized regime could have increased its affiliation by 31%); (c) part of the funds that should have been assigned to extend coverage of the poor were used in boosting salaries and fringe benefits of medical personnel; (d) public hospitals continued to receive supply subsidies, preventing their transformation into demand subsidies targeted to the population in need; (e) serious flaws in the information system (SISBEN) to measure both income and need, and its improper function in the countryside due to low education of the population and in urban areas because the poverty definition is linked to housing and public services—even if inadequate—that these groups have; ( f ) fast application of SISBEN in the most developed and capable municipalities that affiliated most of their eligible poor (90% in Bogotá), whereas the least developed municipalities only identified and affiliated 40% of their eligible poor, overall 30% of the poor were disqualified while 31% of nonpoor were affiliated; (g) an important segment of the population neither qualified for the subsidized regime nor had enough income to join the contributory regime, which is increasing exclusive due to restrictions and rejection of persons with low payment capacity and/or high risk; and (h) the reality of the labor market and a growing informal sector

ignored in a projection of very high selfemployed affiliation that failed to materialize (Jaramillo 1999; Málaga et al. 2000; Sojo 2001a; Ginneken 2003; Martínez et al. 2003).

8.8. Impact of the reforms on coverage and fulfillment of targets This section summarizes the chapter major findings on the impact of the reforms on legal coverage, groups difficult to incorporate, fulfillment of coverage targets, population covered before and after the reform, inequalities in coverage by income, location and ethnic groups, access and utilization, and causes of the low coverage in most countries and nonfulfillment of reform targets.

8.8.1. Legal coverage, groups difficult to incorporate and fulfillment of targets Political constitutions and social security/health laws in virtually all countries establish the right to health care coverage, often universal and free for the uninsured, a mandate not been enforced in the seven latecomer-low countries and the four least developed in the intermediate group. The public sector should legally protect the uninsured population in need in most countries, but in practice a significant part lacks effective access. The private sector, which has grown during the reform, mostly consists of profit seekers (prepaid plans, insurance firms, and providers) that take care of the upper-middle and high-income strata concentrated in urban areas. As prior to the reforms, the poor and low-income population, including indigenous persons, lacks effective access to the public sector or social insurance and continues to be inadequately covered by not-for-profit providers and traditional

194 Reassembling Social Security medicine, and paying out-of-pocket (see Section 11.1.1). In the past twenty-five years health reforms have extended legal coverage in only two countries: Brazil shifted from social-insurance coverage of workers to a public system with universal coverage; Colombia extended mandatory coverage to the self-employed, the insured dependent family and the poor, most of whom where not previously legally covered, and Costa Rica introduced compulsory coverage for the self-employed. The Dominican Republic’s reform law stipulates an extension of coverage but only one of its three regimes was operational and with limited coverage in 2006. Cuba initially lacked national sickness insurance but had maternity coverage; the reform of the 1960s, with a different orientation than the rest, created a universal free public health system. In most countries the reform has not significantly extended legal coverage, except for the basic package of benefits (see Section 9.1.1). Compulsory coverage of dependent female spouses predates the reform and previous restrictions mostly imposed in the latecomer-low group continue: seven countries grant maternity coverage to the female spouse but not sickness care or vice versa; and seven countries keep the exclusion of children above ages ranging from 1 to 12 (two of them recently increased those ages slightly). Pensioners remain covered in all countries albeit with limitations in some of them, which have not been lifted. Legal coverage was significantly expanded in Guatemala’s departments but not in Honduras. As before the reform, all salaried workers have social insurance mandatory legal coverage in the pioneer-high group and the most developed countries in the intermediate group, whereas they are only partly covered in the latecomer-low and the least developed of the intermediate group, because of the legal or practical exclusion of self-employed and agricultural workers. The majority of the labor force in the region is in the informal sector and/or in rural areas, with unstable employment, low salary and reduced capacity to contribute, negative features that obstruct its

social-insurance coverage. The informal sector has steadily grown in the last two decades reaching 47% of urban employment, and is much larger in the latecomer-low group and the least developed countries of the intermediate group, which have the lowest insurance coverage. The self-employed are very difficult to incorporate in social insurance because they are either excluded or have voluntary legal coverage in sixteen countries and mandatory coverage in only two, unskilled self-employed’s income is very low, their percentage contribution equals the sum of the salaried worker’s and employer’s percentages, and it is very complex for them to affiliate and pay contributions. In six countries of the latecomer-low and intermediate groups, all with legal voluntary coverage, self-employment stretches from 23 to 40% of the labor force but its percentage of total social-insurance affiliation ranges from 0.1 to 5%. Pioneer-high countries with a small proportion of self-employed in the labor force have higher coverage: 14%– 34% in Argentina, Chile and Uruguay, but most self-employed are actually cared by the public or social-insurance sectors (only 2% of these workers are affiliated in Chile’s private system after twenty-five years of reform). The highest social-insurance coverage of 45% in Costa Rica, despite voluntary affiliation until 2006, resulted from free care to the poor and fiscal subsidies to low-income selfemployed. Brazil and Cuba’s universal and largely free public systems probably cover all self-employed. Colombia’s reform stipulated mandatory legal coverage of the selfemployed by its two regimes, the subsidized one providing free care to the poor and low income, but affiliation was either 4% or 29% in 2000, failing to meet the reform target of 80%. Five reform laws or legal drafts stipulate mandatory coverage but had not been implemented in 2006. The impact of social insurance legal mandatory versus voluntary affiliation on actual coverage of the self-employed is inconclusive, as it is unclear whether the degree of coverage is an outcome of their share in the labor force or policies to

Health Care Reforms on Universal Coverage 195 incorporate/stimulate affiliation or a combination of both. The legal mandate of social-insurance affiliation of domestic servants in thirteen countries has not been enforced as their actual coverage ranges from 3% in Paraguay to 27%– 31% in Colombia and Uruguay. Employees of microenterprises, unpaid family workers, home, seasonal and part-time workers, those without a labor contract, the unemployed, and housewives are excluded from socialinsurance coverage or can join voluntarily or under certain conditions in a few countries. Coverage of employees of microenterprises stretches from 4% in Mexico to 51% in Argentina. The reform has not improved coverage of these workers. In the least developed countries, the rural population is still significant and half of all countries legally exclude agricultural workers (in the other half coverage is restricted to wage earners in plantations and co-op members) resulting in very low coverage: from 2 to 6% in three countries; coverage jumps to 18%–29% of the rural population in the peasants schemes of Ecuador and Mexico that predated the reform and weakened under it; but Brazil’s public system covers half of the rural labor force. Most reforms did not set coverage goals; only eight established them but four in a vague manner; of the four reforms that set up specific targets on coverage with a timetable, three did not meet them: Colombia’s 1993 reform target was 100% coverage of the total population in 2001 but it was not met in that year and remains unfulfilled; Dominican Republic’s target of universal coverage in 2011 was considered unfeasible in 2004; and Mexico’s SPS target of 31% coverage of the total population in 2010 is difficult to accomplish as only 11% was covered in 2006.

8.8.2. Population coverage and comparisons before and after the reform As older and more developed the social security system and larger the size of the formal

sector are, higher the coverage of both social insurance and the total population, and vice versa. In 2000–4 the countries with the highest total coverage (97%–100%) were five in the pioneer-high group. Reported universal coverage in the three most advanced intermediategroup countries was debatable because part of them lacked data on effective public access, total coverage was overestimated in Mexico and Panama, and access of the population tied to public hospitals was unknown in Colombia. In the three least developed intermediate countries, Bolivia, Ecuador, and Peru, total coverage was 61%–71%. In the latecomer-low group coverage was 57%–68% in five countries (El Salvador, Haiti, Paraguay, Honduras, and Nicaragua) whereas it was higher but unreliable in Guatemala and Dominican Republic. No data were available for Cuba and Venezuela. It was unfeasible to estimate total population coverage in the region as a whole and in eleven countries, both currently and relative to the situation before the reform, because of poor data reliability: (a) lack of historical series on statistical coverage, absence of standardized survey criteria to categorize the population covered by the three sectors, and only gross estimates available in half of the countries; (b) obsolete estimates (five- to eleven-year old) in nine countries, and fairly recent estimates in only six countries; (c) remarkable contradictions among divergent estimates, sometimes in the same document and for the same year; (d) unknown effective public access in nine countries; (e) overestimation in social-insurance coverage in at least five countries; ( f ) impossibility or high difficulty to estimate coverage by the private sector and ‘others’ in eight and seven countries respectively; and (g) unreliable estimates of total coverage and total uninsured in half of the countries because of the previous reasons. Four additional problems obstructed the comparison of coverage before and after the reform: (a) a single observation available in six countries; (b) two observations but not one before the reform in three countries; (c) twenty years elapsed between Chile’s reform

196 Reassembling Social Security and most of the others; and (d) considerable social-insurance multiplicity prior to the reform or changes in affiliation among sectors before and after the reform or unconsolidated series of insured and uninsured. A rough comparison of total coverage in nine countries approximately in the year before the reform and in 2000–4 indicated: (a) no change in four: Argentina, Costa Rica, Mexico, and Panama (the first two had virtual universal coverage before the reform and the last two high coverage); (b) increment in three: Chile, Uruguay (both with near universal coverage before the reform), and Guatemala (data are questionable); and (c) decline in two: Ecuador and Paraguay. Brazil lacks data prior to the reform but had virtual universal coverage in 1998–2003, and Colombia had either an increase or decline in coverage depending on whether the population tied to public hospitals was included or excluded. The estimated regional weighted average of the total population covered by social insurance before and after the reform (excluding Brazil and Cuba public systems) increased from 43 to 52% in 1980–90 but decreased to 41% c.2001. These averages are below the ILO minimum norm of 75% coverage of the resident population, but an undetermined part of it has access to the public sector and to less extend is covered by the private sector (3.8% of the population) and ‘others’. Social and private insurance combined accounted to 45% of the total population (excluding Brazil and Cuba); Brazil has one-third of the regional population and when its public-sector and private insured were added to the calculation, the total climbed to 62%, still below the ILO minimum norm. Trends in coverage/access by sectors and countries could only be detected in about half of the countries: public-sector access, the most uncertain (nine countries), fell in five of them, increased in three and stagnated in one; social-insurance coverage, the most certain (twelve countries), stagnated in five countries, declined in four and increased in three; and coverage in the private sector, relatively uncertain (eight countries) probably rose in

five countries (where coverage of public and social insurance contracted), stagnated in two and shrunk in one. Summarizing, total population coverage by all three sectors could not be estimated, but average regional coverage of social insurance decreased after the reform and most of the population remains uninsured; in most countries with available reliable data there was a fall in public-sector access, stagnation or decline in social-insurance coverage and increase in private-sector coverage.

8.8.3. Geographical, ethnic and income disparities in coverage Data from 14 countries on geographic differences in social-insurance coverage show that inequality is lowest in the pioneer-high group and highest in the latecomer-low group: the ratio between the best and worst covered areas were: virtually nil in Costa Rica and Uruguay; 1.2 times in Brazil (public system), 1.4 in Argentina and Chile, 2 in Colombia, 3 in Mexico and 4 in Ecuador (both with peasant insurance), 5 in Panama, 155 in Guatemala, 350 in Honduras, and 400 in Nicaragua. The bestcovered geographic areas are the most developed, urban, and wealthy, while the worst covered or not covered at all are the least developed, rural, and poor. The segmentation and lack of coordination of the health care system in more than half of the countries intensifies geographical inequalities because of divergent coverage through public, social insurance, and private providers: the last two (particularly private) are greatly concentrated in capital cities and other highly urbanized areas, whereas rural and poor areas are mostly protected by public services. An evaluation of the impact of the reforms on geographical differences in coverage was only feasible in Colombia where such inequalities have been significantly reduced: the ratio between the best and worst covered department decreased from 2.8 to 2 times in 1993– 2003, while the ratio of urban–rural inequality declined from 4.4 to 1.3 times. Nevertheless,

Health Care Reforms on Universal Coverage 197 the gap between urban and rural district capitals remained unchanged by the reform, coverage by the subsidized regime in municipalities was inversely correlated with unsatisfied basic needs, and the basic package in the subsidized regime that covers poor and low-income affiliates is half the size of the contributoryregime package. Information on coverage by ethnicity is very scarce (no data were available from seven countries with sizable indigenous populations) but the empirical evidence offered shows that indigenous peoples are concentrated in the poorest, most rural, and worst covered areas. Three important exceptions are Ecuador and Mexico peasant programs that provide primary care targeted on rural areas where most indigenous people live (both programs predated the reforms), and Chile where 80% of the indigenous population, half of which is poor, is covered by public-social insurance, an outcome of the reform particularly in the last decade. But most reforms have not improved coverage of indigenous people, except in a few countries that provide the basic package to poor and rural/urbanmarginal areas where most indigenous persons live. Data on the population covered by social insurance by income quintiles in thirteen countries in 1996–2003 demonstrate that such coverage increases with income and vice versa; the ratio of coverage between the wealthiest and the poorest quintile ranged from 2 to 54 times (the lowest was in Brazil public system: 1.4 times). In five countries that relationship is accentuated in private-sector coverage: a positive relationship was found between the level of education and private insurance or use of private services in Brazil and Ecuador, whereas an inverse relationship was noted between coverage and rising age in the private sector in Chile and Ecuador (the opposite in the public sector). Although statistics from other countries are needed, the available figures question the capacity of the private sector (expanded under the reform) to extend coverage among the lowest income, worst educated and older segments of the pop-

ulation because of risk-selection practiced by private providers, except when a basic package is mandated and guaranteed to all the population regardless of income. The impact of the reform on coverage by income quintiles was assessed in two countries: in Bolivia fell in all quintiles in 1996–2000 (but 1996 data probably overestimated coverage); and in Colombia it increased significantly in all quintiles in 1993–7, but decreased in all quintiles in 1997– 2000 (comparisons with 1993 were obstructed by changes in the groups covered).

8.8.4. Access and utilization Data on health care access in sixteen countries are quite diverse but recent surveys in several countries show that from 4 to 62% of the population who was sick did not seek health care; such percentage was lowest in the pioneer-high group and highest in the latecomer-low group (from Uruguay to the Dominican Republic). The most frequent reasons given for not seeking medical care in nine countries were: lack of money or too expensive care, sickness perceived as been minor, self-medication or use of home remedies, too far away health facility or long waiting list, deficient care or poor confidence in the personnel, lack of medicines or physicians, too complex procedures and lack of information. Of those who sought care, the majority went to the public sector but from 30 to 63% had to resort to the private sector (in half of the countries) because of lack of access to the public sector; the lowest proportion was taken care at a social-insurance facility (5%–13% in five less developed countries), and 40%–48% used traditional medicine in two countries. Access in rural areas is considerably more restricted than in urban areas and the capital city; access of those with resources is immediate but it takes a long wait for those who lack resources; access and use of services are highly correlated with income regardless of the level of development; access by dependent family is considerably lower than access by direct insured (often due to restrictions imposed

198 Reassembling Social Security to spouses and children); access measured by surveys is smaller than coverage reported by social-insurance statistics, and access by the indigenous population is worse than the rest of the population in four countries: they resort more to traditional medicine or selfmedication and have less access to basic services. The evaluation of the potential impact of the reform on access was inconclusive because surveys were not systematically taken before and some years after the reform or survey data were not comparable or there was contradictory information. Evidence from several countries indicates that the basic health care package has been a key factor in facilitating and extending access, but with some serious problems (see Section 9.3.1) and unclear outcomes in Bolivia and Mexico. About 7% of the population in Chile, 15% in Brazil, and 31% in Colombia may not have access; there are persistent geographical and ethnic barriers in Bolivia and other countries; there has been significant improvement at the first level of care but difficulties in referrals (Costa Rica); the population have virtual access to the first level of care by a physician but part of it lacks basic medication (Argentina, Colombia, and Cuba), and there are long waiting lists for surgery and other complex procedures (Brazil, Chile, Colombia, and Costa Rica).

8.8.5. Causes for low coverage and nonfulfillment of targets The low health care coverage in half of the countries (the seven in the latecomer-low group and three in the intermediate group) has been influenced by external factors to the system but also by the system itself. Eight case studies were chosen to explore the role of such factors: Chile, Costa Rica, and Cuba in the pioneer group with universal coverage; Ecuador and Peru in the intermediate group, and El Salvador and Honduras in the latecomer group, the last four with low coverage; and Colombia in the intermediate group concerning the failure to meet the reform target

of 100% coverage. An attempt was made to pair countries that share some important common features making them fairly comparable. Although a method is needed to evaluate the impact of these factors (isolating external from internal ones) and more countries included in the comparison, some preliminary conclusions can be derived from the analysis of the eight cases. Countries with universal coverage benefited from external factors such as fair economic growth and stability, low poverty incidence, relative small informal sector and selfemployment, high rate of urbanization, low unemployment, political stability, and the absence of physical and ethnic-cultural barriers for coverage. But the system itself (integrated or well coordinated) played a key role in coverage as well as specific policies such as: a strong public, social insurance or publicsocial-insurance sector that provides free coverage to the poor and either free or subsidized coverage to low-income groups, including the rural labor force as well as self-employed in most countries (and 80% of Chile’s indigenous population); fiscal transfers targeted on the poor; and continuous commitment to expand and preserve coverage. Conversely, low coverage in the other four countries can be explained by several external factors often combined: economic crisis and persistent negative or low growth, high poverty incidence, a majority of the labor force informal and sizable self-employment, high unemployment and underemployment, a large and dispersed rural population with significant proportion of self-employed small farmers and peasants; political instability (civil war in two countries); and a considerable indigenous population. Concerning internal factors, two of the countries introduced social security late; the four countries have segmented or highly segmented health care systems in bad need of coordination and most of them also lack targeted fiscal transfers; social insurance excludes the informal sector and the agricultural labor force and rural population (except Ecuador’s peasants but with declining coverage and lacking adequate targeting on the poor) and in

Health Care Reforms on Universal Coverage 199 three of them also restricts coverage of dependent female spouse and children above certain age; health facilities and personnel are concentrated in the capital city and urban areas; a public system poorly funded charges user fees (with one exception) and does not effectively protect most of the indigenous and rural populations and the poor, and three countries lack the basic package. Peru has frequently changed health ministers and policies, El Salvador suffered two earthquakes that damaged a significant proportion of facilities, and Honduras’ social-insurance coverage has not been extended to most departments. Later it will be shown that these four countries have a regressive distribution of health funds, allocating more to social-insurance and private sectors than to the public sector that is legally in charge of the protection of the majority of the population, as well as significant outof-pocket expenses that burden the poor and low-income groups. The Colombian reform failure to meet the target of 100% coverage of the total

population in 2001 can be mainly explained by internal flaws: lack of policy consensus; insufficient public health expenditures; inability to shift supply subsidies toward demand subsidies; postponement of the start of the subsidized regime for the poor and low income and delay and cut of funds designed for it; inadequate information system to identify the poor population qualified to receive subsidies; resources initially assigned to the poor that were shifted to health personnel salaries; and a contributory regime increasingly exclusive due to restrictions and afflicted by growing evasion, payment delays, and under-declaration of wages. External factors that contributed to the failure were: civil war, prolonged economic recession and political instability, which reduced fiscal transfers to the subsidized regime and increased unemployment, as well as persistent poverty incidence and a large informal sector and self-employment not realistically taken into account in setting the coverage target.

Notes 1. The universal right to health care is enthroned in the constitutions of at least sixteen countries and some of them specify its free nature: Bolivia, Brazil, Colombia, Costa Rica (free for the poor in social insurance), Cuba (free for all in the public system), Ecuador and El Salvador (free in the public sector), Guatemala, Haiti, Honduras, Mexico, Nicaragua, Paraguay, Peru, Uruguay (free for the poor in the public sector), and Venezuela (free for all). Laws of health care or/and social security grant that right in Dominican Republic and Panama (legislation; Madies, Chiarvetti, and Chorny 2000). 2. According to a presidential decree in 2006 all persons above age 60 affiliated to FONASA must be cared freely in all public hospitals presenting their identification to check their age, a measure expected to increase coverage mostly to pensioners and women (El Mercurio, March 16, 2006). 3. Ecuador’s universal health care insurance was launched in 2005 to extend coverage to dependents of insured and the self-employed; another project in 2006 aimed to expand coverage to 1.4 million people in the poorest two quintiles of the population by 2009, as well as in selected regions and municipalities; no data on these programs progress were available by mid-2006 (IESS 2005, 2006b; World Bank 2005e). 4. There is wide consensus about universal coverage in Brazil but the two coverage figures (in 1998 and 2003) are identical in Table 8.2 and restricted to the private sector; public access has been estimated as a residue, there is not social insurance, and no information on ‘others’. 5. It is impossible to separate IAMC affiliates in social and private insurance (IAMC are private entities thus complicating the allocation of their affiliates), there is double coverage in the public-sector and private partial insurances, as well as in IAMC and private mobile emergencies, affiliates in ‘others’ are given separately or jointly with private partial insurance and mobile emergencies, and recent surveys give overlapping coverage among programs.

200 Reassembling Social Security 6. There are two different sets of coverage estimates c.2001 in Honduras: 60 and 52% public sector, 12 and 11% social insurance, 10% and unknown private, and 82 and 63% total population covered, leaving 40 and 18% unprotected (IHSS 2003; PAHO 2005a). 7. In 1999–2001 there was an increase from 38 to 47% of Paraguay unprotected defined as ‘deficit in real coverage’ (Ramírez 2004, 2005a). 8. Two sets of disparate figures are given in Bolivia for 2000: 65 and 70% public sector, 17 and 20% social insurance, 4 and 5% private, and 0.1 and 5%–10% ‘others’ (World Bank 2004: 19–20, 92). Despite blurring the comparison with 1997, data indicate a jump in the uninsured from 64 to 79% by 2000. 9. Panama’s national development plan 2000–4 asserted that ‘close to 80% of the population has access to health services . . . the poorest families are excluded, due to cost and distance’ (MEF 1999: 41). 10. Haiti’s total coverage of 61% in 2000 is overestimated because it is limited to the first level of care and would be smaller if higher levels had been taken into account. 11. The combined total of the two sectors and ‘others’ suggests that all the population is covered, and yet since ‘others’ include those nonaffiliated to either FONASA or ISAPREs and directly paying for care, there might be a segment of the population without protection. 12. At the start of the reform in 1993, 50% of the total population had public access and 24% was affiliated to social insurances, a total of 74% was covered or with access and 26% was either unprotected or with private insurance/direct care. In 1997–2002 coverage fell from 33 to 29% in the contributory regime, increased from 17.6 to 24.3% in the subsidized, rose from 50.6 to 53.3% in total affiliation, and diminished from 49.4 to 46.7% in nonaffiliation albeit could be tied to public hospitals (Appendix 3). 13. Total coverage by IMSS (the largest insurer) rose from 36% in 1980 to 46% in 1990, fell to 38% in 1995 due to the economic crisis, peaked at 47% in 2000 and declined to 45% in 2002 below the 1990 and 2000 levels. Coverage of the peasant program stagnated in 2001–4 (IMSS 2005a, 2005b) and that by the other four social insurances virtually stagnated in 1995–2002 (Appendix 4). Private insurance coverage was 2.3% in 1997 and ranged from 0.5 to 3% in 2000 (Dávila and Guijarro 2000; ISSA 2003b; INEGI 2005; PAHO 2005a). 14. Brazil has one-third of the total population in the region and if access to its public system were added to the calculation, regional averages would be 61% in 1980, 64% in 1990 and 62% c.2001, still below the ILO minimum norm but leaving out the public sector and ‘others’ in the remaining countries. 15. Costa Rica’s coverage rose from 61 to 86% in 1976–94 (it fell in 1980–3 due to the economic crisis) and peaked at 89% in 1999 but decreased to 86% in 2003, at the same level as in 1994 when the reform started. 16. Guatemala’s coverage, instead of stagnant in 1995–2000, is also given as falling from 20% in 1994 (before the reform) to 18% in 2004 (IGSS 2005). 17. Nicaragua’s coverage shrunk from 24% in 1987 to 18% in 1990 (the year before the reform) and 13% in 1995 (due to civil war, economic crisis and merging of the public sector with social insurance); the restoration of social insurance and improvement of its services expanded coverage from 8 to 25% in 2001–4, above the pre-reform level but similar to 1987. 18. Panama’s coverage augmented from 53% in 1992 to 61% in 1996 (year of the reform in one hospital), peaked at 68% in 2000 and fell to 65% in 2004. 19. El Salvador’s coverage rose from 14.5% in 1997 (before the reform) to 16% in 1999 (at the end of the reform) but stagnated at 15.8% in 2001 (ISSS 2002). Peru’s coverage fell from 34% in 1990 to 20% in 1997 (at the end of the reform), but rose to 26% in 2003 still below the pre-reform level (Mesa-Lago 1992; EsSalud 2004). 20. Sources for this section besides Table 8.2 are CAR (2005); Bolivia Guerrero (2000); Brazil IPEA (1998), Medici (2002a); Chile Valenzuela (2003); Colombia Málaga et al. (2000), Castaño et al. (2001), Salud Colombia No. 63 (2002), AXESNET (2004), Ramírez (2004); Costa Rica Martínez (2005); Ecuador Carpio (2001), Quevedo (2005); Guatemala IGSS (2005), Herrera (2006); Mexico Munguía (2006); Nicaragua INSS (2005a, 2005b).

Health Care Reforms on Universal Coverage 201 21. Percentages of the worst and best geographic areas covered by social insurance in 2000– 4 (with exceptions noted) were respectively: 87 and 99% in Brazil (public-sector basic package), 54 and 78% in Chile, 45 and 61% in Argentina, 31 and 70% in Colombia, 22 and 71% in Mexico, 14 and 66% in Panama, 9.5 and 35% in Ecuador (including the peasant scheme and private insurance), 6 and 31% in Paraguay (rural and urban areas), 0.4 and 62% in Guatemala, 0.1 and 35% in Honduras and 0.1 and 40% in Nicaragua. Venezuela’s social insurance lacks health facilities in three poorest states and only has ambulatories in seven relatively poor states; 61% of such facilities are concentrated in the four most urbanized states. 22. Social-insurance affiliation in departments ranged from 16% in Atlántica to 39% in Bogotá in 1993 and increased to 36 and 70% respectively in 2001, whereas affiliation in rural areas was 7% versus 31% in urban areas, and rose to 48 and 61% respectively. 23. Urban capitals had 61% of their population affiliated in both regimes in 1997 (mostly in the contributory) while rural districts had 47% (mostly in the subsidized) with a gap of 14 percentage points between them; in 2003 the percentages had risen to 66 and 52% respectively, but the gap remained unchanged. 24. Sources for this section, in addition to those in Table 8.3 are: general from PAHO (2005a); Bolivia Narváez (2002); Brazil Medici (2002a); Chile Pollack (2002), MIDEPLAN (2004), Urriola (2005), Bravo et al. (2006); Dominican Republic Rathe (2005); Ecuador Ecuador (2005), Quevedo (2005); Guatemala Durán and Cercone (2001); and Uruguay ISSA (2003b), World Bank (2005d). Data on social-insurance coverage are from Table 8.2 and Section 8.2; poverty incidence from ECLAC (2004a). 25. The fixed minimum contribution in the OS (US$7 monthly) is relatively high for low-income self-employed and a barrier to their affiliation, a reason why most of them are covered by the public sector (Torres 2004). 26. Uruguay’s domestic servants accounted for 10% of the urban employed labor force in 2002 but only 3.5% of total insured in BPS; 74% of homes with domestic servants in 2004 did not have them insured (INE 2005). Domestic servants comprised 4.6% of Mexico’s urban employed labor force in 2002 but accounted for 0.004% of total insured in IMSS (ISSA 2003b). 27. Coverage of employees in microenterprises by all sources (not only social insurance) in 2000–2 was: 5% in Nicaragua, 7% in Bolivia, 12% in Guatemala and Peru, 20% in Ecuador, 42% in Argentina, and 88% in Chile. Between 1994 and 2002 coverage of these workers increased in half of the countries and decreased in the other half (ILO 2003a). 28. Sources for this section are general from PAHO (2005a); Bolivia Narváez (2002), World Bank (2004), Valenzuela (2004); Brazil Biasoto (2004b), IBGE (2005); Chile Barrientos (2000), Sánchez and Zuleta (2000), MIDEPLAN (2004), Urriola (2005); Colombia Jaramillo (1999), Málaga et al. (2000); Castaño et al. (2001), Salud Colombia No. 71 (2003); Costa Rica Programa Estado (2002); Dominican Republic Lizardo (2004), Rathe (2004); Ecuador Ecuador (2005); Guatemala Durán and Cercone (2001), O’Neil, Bartlett, and Mignone (2005); El Salvador DIGESTYC (2002); Honduras Durán (2003); Mexico ENE-SS (2001), SSA (2004a), Nigenda (2005); Nicaragua Quintanilla (2001); Paraguay Ramírez (2004, 2005a), World Bank (2005c); Peru Manrique (1999); and Uruguay Quijano (2002), World Bank (2005d). 29. According to Mexico’s 2000 survey, among the 63% uninsured that sought care, 27% did it in the public sector and 10% in social insurances, 22% paid for private services and 4% resorted to traditional medicine and self-medication. 30. In Dominican Republic 7% of the poorest quintile that was sick sought care paying out of pocket in small private clinics with inadequate equipment, lack of specialized personnel, and low quality of service. 31. A composite index of exclusion based on multiple factors and their interaction has determined that internal factors explained between 53 and 59% of the exclusion in Ecuador, Honduras and Paraguay, whereas external factors explained 54% in Peru (PAHO/WHO/SAID 2003).

202 Reassembling Social Security 32. Other countries that have achieved universal coverage could have been selected, like Argentina, Brazil and Uruguay, but they lack several characteristics common to the three chosen, such as very large territory and population in the first two and segmented systems without coordination in all. 33. The domestic estimate of poverty incidence is lower: 14% in 2002 and 12% in 2003 (Programa Estado 2004). 34. According to PAHO (2005a) the incidence of urban poverty rose from 19 to 55% in 1995–9, but ECLAC estimated an increase from 57 to 64% in the same period, and later reported a fall to 56% in 1997 and 49% in 2002 (ECLAC 2002a, 2004a).

9

Effects on Equal Treatment, Solidarity, and Comprehensiveness/ Sufficiency

9.1. Equal treatment Equal treatment in health care is closely related with equity, which has diverse meanings. In the first generation of reforms, IFO targeted health care on the poor (who suffer higher incidence of disease and mortality than middle- and high-income groups), but since the 1990s with the second generation of reforms they have focused on equity understood as the guarantee to the most vulnerable groups of minimum levels and access to health services (Vergara 2000). Equity is also conceived as the policy reduction of differences in health conditions of the population, once standardized by factors like age and gender (Lloyd-Sherlock 2003) or the elimination of unfair and avoidable disparities in access and quality of benefits (ILO-PAHO 2005). Four dimensions of equity are: (a) in coverage and access (analyzed in Chapter 8), which demands the elimination of inequalities by income, age, gender, ethnic group, or geographic location; (b) in financing that operates when individuals or households contribute to the system according to their payment capacity, not considering their health risks; (c) in benefits/services (discussed in Section 9.3), achieved when individuals or households receive care according to their needs, regardless of their payment capacity and risk, and (d) in health levels, which requires the eradication of important differences in health indicators (ILO 2000; Castaño et al. 2001; Bossert et al. 2003).

An equitable distribution of health financing among geographic areas or regions can be achieved through three alternative approaches: (a) redistribution from rich to poor regions in order to even resources between all regions (Chile); (b) allocation of any increase in public health funds to poor regions in order to close their gap with rich regions without harming the latter (Colombia); and (c) letting poor regions benefit from the development of bordering rich regions in pursuit of convergence on their level of resources. The third approach applies ‘trickledown’ economic theory to health and usually leads to inertia but it’s argued that, if its effects were superior to that of the first approach, the implementation of redistributive policies would hamper the development of rich regions and general growth. An evaluation of the impact of the reforms on equity based on indicators related to coverage, access, use, and financial distribution of resources (but not benefits and health levels) concluded that: only nine of the twenty countries had incorporated some definition of equity; in less than half of them the reforms were contributing to reduce disparities in coverage, and only a minority was lessening disparities in resource distribution—lack of data impeded to assess the reform impact on access and use of resources (Infante, de la Mata, and López-Acuña 2000; PAHO 2002a). WHO (2000a) has appraised and ranked health systems in 191 countries (including all in Latin America) based on seven indicators,

204 Reassembling Social Security a key one being financial equity, for example, financing proportional to individual income regardless of risk. A more advanced approach demands financing to be progressive: the heaviest financial burden should be supported by those with highest income or be proportionally inverse to individual or family income (Dominguez and Soares 2005). Health reforms in the region have stimulated competition and the expansion of the private sector, deepening the previous segmentation of the population with a negative impact on equity in several countries: (a) private insurance firms and providers practice risk selection (cream-skimming), excluding or charging higher premium to those who suffer chronic diseases or are old; (b) competition is centered on the most attractive groups (young, healthy, with low health risk, without a record of chronic diseases, and sufficient income to cover needs derived from their epidemiological risk profiles) and avoids the least attractive groups (old, ill, with high risk, chronic diseases, and insufficient income to finance needs derived from their epidemiological profiles) who are left to the public sector (‘the market specializes on the rich and the state on the poor’); (c) groups excluded by the private sector become attractive only when either benefits or their quality are reduced, often through a basic package of minimum benefits, and fiscal subsidies granted; (d) segmentation increases in countries where social insurance is not integrated with the public sector and covers the middle-income formal labor stratum, an effect aggravated when separate schemes cover several occupational groups with superior benefits and fiscal transfers; and (e) segmentation makes very difficult to ensure equivalent health care for affiliates, generates double or triple coverage within the same household, foments inefficiency in resource allocation and has adverse effects on equity (Giordano and Colina 2000: 36; PAHO 2002a; see also Sections 10.4 and 10.5). This section analyzes five aspects of equal treatment and equity: the standardization of benefits and subsistence of inequalities; financial equity (an analysis of WHO rankings);

inequalities in the distribution of human and physical resources, as well as in health levels by geographic areas, health care sectors, and ethnic groups, and gender equity.

9.1.1. Standardization of benefits and subsisting disparities The reform infused certain degree of standardization of health benefits in a minority of countries, but in the majority different levels and quality of care subsist between the three sectors: the public covers the uninsured particularly the poor, is generally in charge of promotion and prevention and offers the worst curative services; social insurance (several schemes in some countries) concentrates on curative medicine with better services; and the private sector (with remarkable differences between private insurance firms and providers) usually have the best services and they are exclusively curative. Countries with a federal structure add differences between state, provincial/departmental and municipal services; even those with unitary governments have disparities in benefits and quality care between the capital and the rest of the country (PAHO 2005a). The most standardized care was achieved in three countries with unified or relatively unified systems prior to the reforms that started in the 1980s. Cuba public system created in the 1960s legally offers homogeneous services to all the population, there are neither social insurance nor private services but a few separate schemes subsist with superior quality. Costa Rica unification of social insurance (CCSS) services occurred in the 1970s and also provides equal care to all (including an identical package of benefits mandatory even to contracted private providers), regardless whether affiliates are compulsory or voluntary, contributors or poor financed by the state, but illegal-informal practices weaken the legal rule. Panama relatively unified social insurance dating also from the 1970s is the principal provider and offers similar services to most of the population; the ministry offers

Equal Treatment, Solidarity, and Comprehensiveness lower-quality services to vulnerable groups in areas not covered by social insurance; there are no separated schemes and very few covered by private insurance. Costa Rica’s illegal or informal practices (the biombo or folding screen) embrace about 15% of the population: patients avoid the long waiting list for surgery paying a bribe; procedures not available at a CCSS facility are delivered by an employed doctor during his work schedule charging a fee; CCSS physicians send patients to their private office with the enticement of providing more personal treatment or refer the patient to a private provider. Patients’ denunciations to CCSS are often stopped due to the physician union esprit de corps and to avoid sanctions. Mainly because of the CCSS long waiting list in surgery, an increasing number of middle- and high-income insured opt out, especially at the primary level of care but also at the other two levels, and buy private services that are faster and/or better (Martinez and Mesa-Lago 2003; World Bank 2003; Martinez 2005; similar practices are reported in other countries by Medici 2002a; MesaLago 2005b). In other three countries the reform fused multiple social insurance health schemes with diverse benefits and fairly standardized them, but introducing significant differences through the new insurance firms and providers. Brazil integrated five social insurances and the ministry into its unified public health system (SUS), but the autonomy of all states and part of the municipalities generates differences, albeit cushioned by the basic package and a compensation fund. Chile eliminated several sickness-maternity schemes with diverse benefits and established two different sectors: public-social insurance (FONASA) covers the poor and low-income population with greater risks and provides basically uniform services, there are no restrictions and a co-payment is charged only to higher-income affiliates; and private ISAPREs cover the population with higher income and low risks with benefits that vary remarkably, thousands of health plans charge copayments according to the insured financial

205

capacity and risks, and services are restricted by deductibles, waiting periods, and exclusions (Bitrán and Almarza 1997). Colombia unified more than 1,000 highly diverse health schemes and set standardized conditions and benefits within each of two regimes, contributory and subsidized, with important disparities: the contributory regime basic package is far superior to that in the subsidized regime and so is the unit of capitation paid; the biggest public insurance firm concentrates 90% of the population with high risk, while private insurance firms have captured the population with highest income and lowest risks; and half of the public hospitals continue operating as before the reform, with significant differences in services (Giordano and Colina 2000; ISSA 2004). Argentina’s reform fell behind the previous six concerning standardization; it reduced the largest inequalities between social insurances (OS) but remarkable differences still subsist: the best OS are those of the private sector (specially in the capital and main cities in provinces) and direction personnel, followed by OS organized by unions and in provinces that have considerable variety in their benefits; low-income OS affiliates who transfer to OS with greater resources are only entitled to the basic package, whereas affiliates with more payment capacity can finance superior packages; and the public sector provides the lowest level of care and with considerable stratification because benefits in provinces vary according to their development level. Despite the reforms, twelve countries maintain the three sectors and substantial disparities in their services. Probably Bolivia is the country with the greatest inequality, because numerous social insurances have divergent benefits and services (Guerrero 2000). Dominican Republic’s reform law (not fully in force yet) stipulates three regimes that will cover various groups of the population with diverse benefits; currently the three health sectors operate with significantly disparate benefits; the private sector has the greatest diversity, for example, NGOs that only provide promotion services and prepaid

206 Reassembling Social Security plans that offer divergent packages according to the premium charged. In Ecuador and Mexico the principal social insurance program (IESS and IMSS respectively) grants comprehensive care, but the peasant program provides primary care; Mexico schemes for civil servants (ISSSTE) and oil workers have better benefits than IMSS. Nicaragua social insurance pensioners are covered with a ‘basic mini-package’ inferior to services provided for active insured; there are remarkable differences in services offered by providers, the big ones (especially in the capital) offer more and better benefits than those in small cities; the ministry provides care to women in their entire life and to all children regardless their age whereas social insurance restricts the spouse care to maternity and sets a maximum age to children of insured; and the ministry under contract with private providers (EMP) must offer better services to their insured than to the uninsured (Rossman and Valladares 2003; La Forgia 2005). Paraguay principal social insurance entitles the insured and dependent relatives to integral care, but female domestic servants only for maternity (IPS 2005). Uruguay divergent services in the three sectors are compounded by notfor-profit private providers (IAMC) and private insurance plans that offer different packages of benefits; additional disparities exist between the capital and the rest of the country (interior). Venezuela exhibits remarkable differences in the amount and quality of benefits provided by several social insurances and those of the ministry and private sector; the legal draft of reform, still pending in 2006, stipulates the standardization of services in the public sector and social insurances (RBV 2004). Excluding Costa Rica and Panama, the reforms left untouched the armed forces’ separate schemes and own facilities, which usually provide the best services after the private sector. In at least seven countries policemen and civil servants also have social or private insurance with better quality of services than in the principal social insurance program: Brazilian civil servants at the three

levels have private plans without losing their right to care in the public system; the same is true in Dominican Republic, Haiti, Paraguay, Uruguay (covered by IAMC and paid by the state), and Venezuela (insurance policy paid by the state). In eleven countries other groups have separated schemes with superior benefits and quality of care, for example congressmen and judges in Argentina, teachers and petroleum workers in Colombia, tourists and foreign visitors in Cuba, the national university, teachers, physicians, and nurses in Dominican Republic, and teachers, university professors, and oil workers in Venezuela. In a national survey in Mexico, users ranked the quality of private services as the best (both at the primary level and hospitals), followed by social insurances (IMSS and ISSSTE), and federal public services were ranked worst; armed forces, states, and other insurance services were excluded from the survey (Dávila and Guijarro 2000).

9.1.2. Financial equity by income levels and geographic areas WHO (2000a) evaluation of health systems in the world (including all in Latin America), with data from 1997, was based mainly on two indicators relevant for equity: the health general level and equality in its distribution, and fairness in financing or financial equity. These two and other indicators were merged into an index that ranked the countries on their overall health system performance that often mismatched the financial equity-ranking. Colombia ranked first in the region and the world in financial equity (also first in performance within the region); other countries ordered in high positions were Cuba, Uruguay, and Costa Rica. Surprisingly Haiti ranked 12th in financial equity (but last in performance), Chile tied with Bolivia in 14th place (but 2nd and 17th in performance respectively), and Brazil ranked last in financial equity but 16th in performance. The analysis below based on data on distribution of health financing by income levels and by geographical areas confirms the WHO ranking

Equal Treatment, Solidarity, and Comprehensiveness of Costa Rica and Colombia (during the first stage of the reform but not thereafter), and largely of Brazil and Uruguay, but disputes the rankings of Bolivia, Chile, Cuba, and Mexico. A major obstacle in the analysis was the lack of standardized statistics, for instance, financial distribution by income level was available by quintiles or deciles, referred to total or public health expenditures or fiscal subsidies (in percentages or per capita); the Gini coefficient was accessible in one country only; data on geographical distribution of health expenditures (in percentages or per capita, either total, public or social insurance) came from states, departments, provinces, and/or municipalities (capital and interior in Uruguay). An intriguing question is how the WHO ranking was done based on such scarce, incomplete and disparate data. Colombia’s ranking as best in the world on financial equity has been attributed to the reform changing previous allocation of more resources to wealthy than to poor municipalities, toward an increase in total financial resources and concentration of the additional funds on the poor municipalities, thus reducing their gap with the wealthy ones that grew at a lower rate. This argument was supported by financial distribution indicators in 1994–7: the ratio in the allocation of total health funds between the poorest and the wealthiest deciles diminished from 6.1 to 1.2 times; the allocation of funds per capita to the poor population cared by the subsidized regime in the departments increased 122%; the ratio between the departments that received most and least funds shrank from 15 to 5.5 times and between municipalities from 70 to 12 times; household expenditures on medicines decreased in all income quintiles; and outof-pocket expenses of households declined (see Section 11.1.1). Conversely, other data show deterioration in financial equity: outof-pocket expenses significantly rose in the poorest quintile and the two least developed departments and were stagnant in the countryside in 1994–7; the ratio of funds allocated between the wealthiest and the poorest deciles rose in 1997–2001, and department

207

funds per capita assigned to the poor in the subsidized regime shrank 28%; expenditure in medicines dropped but their prices jumped since 1999; the Gini index of concentration of health resources was virtually stagnant in 1984–99; allocation of resources did not take into account the quality of services and diverse health needs in municipalities, and the distribution of municipal financing was legally based on equity priorities (unsatisfied basic needs, poverty incidence, and living standards) but the obligation to assign 25% of central government revenue to health care was not met (Málaga et al. 2000; Castaño et al. 2001; Homedes and Ugalde 2003, 2005c). Cuba was ranked second in financial equity, justified because its health system is universal and gratuitous, and out-of-pocket expenses are the second lowest in the region (see Section 11.1.1). On the other hand, data have never been published on the distribution of health funds by income level and on the Gini coefficient, and the economic crisis caused a severe medicine shortage, forcing families to buy them in hard-currency shops at very high prices unaffordable to most of the population, still true at the time of WHO evaluation (MesaLago 2005b, 2006d). Uruguay was ranked third in financial equity, which is confirmed by four indicators: the poorest quintile received 12.5% of the public health expenditure in 2003, the second quintile 48%, and the richest quintile 3%, an improvement in the distribution in 1994–5 when it was concentrated in the third quintile; public health expenditure had a progressive impact as it reduced the Gini coefficient in 1999–2003; and outof-pocket expenses were the third lowest in the region and exhibited a declining trend (RISSSCS 2002; Presidencia de la Republica 2004; Section 11.1.1). But other data for 2002– 3 indicated a regressive effect: total payments to private IAMC increased in tandem with income, in the capital they took 30% of the poorest-quintile income but only 4% of the richest-quintile income, whereas in the interior the proportions were 40 and 7% respectively; the average cost per IAMC affiliate increased

208 Reassembling Social Security in 1998–2003, more in the interior than in the capital; and expenditure per capita in private-partial insurance was 4 times that of the ministry and in private IAMC was 3 times that of the ministry (MSP/WB 2004a). Costa Rica was ranked fourth in financial equity, which is ratified by several indicators: social insurance coverage is universal and the poor are gratuitously covered by fiscal transfers; the poorest quintile contributed only 3% of the total social insurance contribution in 1998; the poorest decile received 20 times more in health services than its contribution, whereas the wealthiest decile received one-fifth relative to its contribution; out-ofpocket expenses were the third lowest in the region and exhibited a decreasing trend; and public-social insurance expenditures were the most progressive among all social expenditures (Trejos 2004; Section 11.1.1). But data on geographical distribution were not as positive: in 2001 per capita primary care expenditure in 29% of the areas was higher than the national average and lower than such average in the remaining 71%, the area with the highest expenditure received 50 times what the area with the lowest expenditure received, and the poorest region had the lowest per capita expenditure (World Bank 2003). Mexico was ranked ninth in financial equity (well above Chile), but lacked data on distribution of health financing by income levels. Public health expenditure as percentage of total government expenditures in the states show that the reform of the 1990s did not significantly reduce regional inequalities: the national average was 16% in 2003, but a rich state received 24%, while the three poorest states received 12%–14%. Federal transfers were unrelated with states’ needs, and the allocation ratio to families between the richest and the poorest state was 100 times. The federal government assigned 3 times more per capita to 60% of the population affiliated to the five social insurances in 2000 than to 40% of the uninsured population left to the federal ministry (SSA 2004b, 2005). Such gap declined to two times in 2005, while the gap between states with the highest and lowest federal per capita decreased from five to four

times, largely due to SPS.1 And yet this happened after the WHO ranking and still in 2004 the three states with the highest allocation (twice the national average) were among the wealthiest (including the Federal District) while three poor states were below or slightly above the national average (Frenk 2005; Nigenda 2005). Chile’s ranking in 14th place in financial equity is supported by several indicators (financial distribution by income levels was not available): part of those classified as poor really were not and, together with others with payment capacity, used public services and received regressive fiscal subsidies; part of affiliates in private ISAPREs also used public services (in emergency, maternity, and catastrophic diseases), as they are better and cheaper, also benefiting from fiscal subsidies, and ISAPREs excluded or created affiliation barriers to old, chronically ill or terminal patients and pregnant women who were left to the public sector increasing its costs (Bitrán and Almarza 1997; Bitrán et al. 2000; Arteaga et al. 2002).2 But the ranking is contradicted by other data: the progressive effect of fiscal subsidies targeted on poor and low-income groups that received free care from publicsocial insurance services (conversely highincome affiliates paid more than the benefits they receive), and out-of-pocket expenses in Chile were considerably lower than in seven countries ranked higher in financial equity (Dominican Republic, Ecuador, Guatemala, Haiti, Mexico, Nicaragua, and Venezuela). Laws in 2004–5 improved financial equity by eliminating the discrimination against the elderly and women, through a universal basic package provided regardless of age and gender, and created a compensation fund to reduce inequalities (see Sections 9.1.3, 9.3.1, and 11.1.2). Brazil ranking last in financial equity is explained by several reasons: the fifth highest per capita health expenditure in the region but health level indicators well below the regional average, largely due to unequal distribution of health financing; the public system (SUS) lacks geographic, epidemiological, and social criteria to distribute its resources,

Equal Treatment, Solidarity, and Comprehensiveness hence health funds are geared to middle- and high-income groups, and poor families have high out-of-pocket expenses; the share of health expenditure in the family budget rose relative to both total expenditure and family consumption, thus augmenting its regressive impact (increased 52% in the poorest quintile in 1987–96 but only 15% in the wealthiest quintile); medicine expense took 60% from health expenditures of poor families and their hospitalization cost was proportionally greater than that of wealthy families, and the poorestdecile health expenses in the most developed states were higher than those of the wealthiest decile (Medici 2002a, 2002b). It is argued that such inequalities declined in 1995–2002 because of a significant increase in the allocation to primary care through the extension of the family health program and the basic package (expenditure in the latter actually stagnated). And yet recent data demonstrate continued regressivity: the ninth highest outof-pocket expenses in the region (the highest in the pioneer-high group); family out-ofpocket expenses took 7% of the poorestdecile income but only 3% of the wealthiest decile; 83% of private health expenditures went to purchase medicine in the poorest decile but only 42% in the wealthiest decile; public health expenditure per capita in the wealthy Southeast almost double that in the poor Northeast (six times between the Federal District and Maranhão), and the highest income families in developed São Paulo used more public free services than poor families in underdeveloped Recife (Piola and Biasoto 2001; Medici 2002a; Dominguez and Soares 2005; PAHO 2005a). Most countries in the latecomer-low group and the least developed in the intermediate group were ranked in the lowest places regarding financial equity (from worst to best: Peru,3 Honduras, Paraguay, El Salvador, Bolivia, and Nicaragua). Bolivia’s ranking in 14th place in a tie with Chile is questionable: total health expenditure per capita in the wealthiest quintile was 5 times the per capita in the poorest quintile in 1999; such ratio increased to 8 times in social insurance health expenditure

209

and to 12 times in private health expenditure; although public health expenditure was better distributed, it did not significantly reduce that remarkable gap because of its low proportion in total health system expenditure (World Bank 2004).

9.1.3. Inequalities in human and physical resources and in health indicators: geographic areas, health sectors, and ethnic groups Data on these types of inequalities are available in eight to ten countries, indicators are not standardized and years vary between 1998–9 and 2003–4, impeding the elaboration of comparative tables and obstructing the evaluation of the reform effects. The geographic distribution of human and physical health resources exhibits inequalities that usually parallel those of the financial distribution. The wealthiest and most urbanized areas (usually the capital city) have disproportionably higher resources than the poorest rural areas as the following indicators prove: (a) the ratio of physicians per 10,000 inhabitants was 1.6 higher in Costa Rica, twice in Cuba, 5 times in Brazil, Mexico, and Panama, 7 times in Peru, and 15 times in Guatemala; (b) the ratio of hospital beds per 1,000 inhabitants was 2 times in Cuba, 3 times in Costa Rica, 5 times in Uruguay, 7 times in Guatemala, and 10 times in Mexico;4 (c) the percentage of childbirths at a health facility was 1.5 times in Cuba, 2 times in El Salvador, 4 times in Bolivia, and 12 times in Ecuador; (d) the percentage of prenatal care was virtually identical in Cuba, 2 times in El Salvador and 5 times in Ecuador. Similar differences existed regarding the percentages of population with access to potable water and sanitation, outpatient consultation, ambulatory visits, and hospitalization (MINSA 2003; World Bank, 2003, 2004, 2005d; ONE 2005; Mesa-Lago 2006a). Information from eight countries on the distribution of human and physical resources

210 Reassembling Social Security by the three health sectors indicates that the public one has ratios of physicians and hospital beds considerably lower than the percentage of the population that legally should protect whereas the opposite occurs in social insurance, particularly in the separate schemes of the armed forces; data on the private sector from four countries show a similar or superior situation than in social insurance (inequalities in financial allocation by sector are discussed in Section 11.1.1). Ecuador public sector should protect 80% of the population with 29% of the physicians, social insurance 16% both the population and physicians, the armed forces 2% with 5%, and the private for-profit sector 2% with 4%. El Salvador’s social insurance ratio of hospital beds per affiliate was 1.6 times greater than in the public sector and the ratio of physicians was triple (Nicaragua’s ratio was 4 times). Haiti’s public sector is in charge of 60% of the population with 36% of the facilities and 28% of the physicians (social insurance is minimal). Mexico’s federal ministry was responsible for 42% of the population with 38% of the physicians, whereas the social insurance peasant program covered 11% with 0.2%, and the respective proportions in separate schemes were: civil servants 10 and 20%, oil workers 0.5 and 3%, and armed forces 0.5 and 1% (the rest were covered by IMSS main program). Paraguay’s public sector had 80% of the population and 39% of hospital beds, and the other sectors proportions were: social insurance, 13 and 21%; private, 6 and 17%; and separate schemes, 1 and 3%. Peru’s public sector was assigned 59% of the population and provided 44% of ambulatory actions, whereas the private sector covered 12% and provided 36%. Venezuela’s public sector jointly with social insurance should cover most of the population with 47% of the physicians, but 13% of them served the armed forces, teachers and other schemes, and 40% the private sector (MSPAS 2001a; MINSA 2002; Rossman and Valladares 2003; SSA 2004a; INEGI 2005; PAHO 2005a). Despite improvement in national averages in health indicators in most countries (see Section 11.2), data from nine of them show

that significant inequalities persist in such indicators between the wealthiest and poorest geographic area: (a) infant mortality was one half in Argentina, Brazil, Cuba, Mexico, and Venezuela, one-third in Ecuador and Guatemala, and one-fifth in Peru; (b) neonatal mortality was one-fifth in Venezuela and onetenth in Peru; (c) maternal mortality was onetenth in Peru and infectious-intestinal diseases one-ninth in Mexico; and (d) life expectancy was three years longer in Argentina, ten years in Panama and Venezuela, fifteen years in Ecuador, and twenty years in Peru. There are ever less data to assess trends: extreme differences in infant mortality decreased in Brazil and Cuba in 1990–2001, countries with significantly diverse systems, whereas infant and maternal mortality differences diminished in Peru in 1996–2000, but concentrated in the wealthiest two quintiles and worsened in the poorest two quintiles (MesaLago 2006a). In terms of health resources and indicators, available data from eleven countries suggest that geographical inequalities are smaller in pioneer-high countries (Argentina, Costa Rica, Cuba, and Uruguay), followed by Brazil and the most advanced intermediate countries (Mexico and Panama), and greater in latecomer-low group countries and the least developed in the intermediate group (Bolivia, Ecuador, Guatemala, and Peru). There is a close relationship between indigenous people, their location in rural and underdeveloped areas, poverty incidence and the worst health resources and standards. Statistics from eight countries show that indigenous people have: (a) half of the national average per capita health expenditure in Ecuador; (b) one-fifth of the national average ratio of physicians in Panama (73% of ambulatories caring for indigenous persons don’t have a physician in Venezuela and 74% of indigenous communities lack a doctor in Paraguay); (c) half of institutional childbirth than the nonindigenous population in Bolivia and half of the national average in Guatemala; (d) half of the national average on access to potable water and sanitation in Venezuela; (e) 10% women access

Equal Treatment, Solidarity, and Comprehensiveness to family planning versus 43% among nonindigenous women in Paraguay; (f ) 3 times higher retardation in growth among children due to malnutrition and twice infection by contagious diseases than national averages in Mexico; (g) 58% higher infant mortality than the national average in Mexico, and 11 times such average in Paraguay; (h) 3 times higher maternal mortality than the national average in Mexico, and 3 times the average among nonindigenous women in Paraguay; and (i) fourteen to seventeen years less life expectancy than the national average in Colombia, and six years less in Mexico and Panama. Data were not available from Brazil, El Salvador, Honduras, Nicaragua, and Peru, countries with significant indigenous population (Mesa-Lago 2006a, updated with Ecuador 2005; O’ Neil, Bartlett, and Mignone 2005; Herrera 2006; Munguía 2006).

9.1.4. Gender inequalities in access and care Gender equity demands that health resources and services are allocated and received according to the needs of each sex and paid based on the economic capacity of the individuals and families regardless of differential risks by gender. Gender inequities are the outcome of labor market factors and those internal to the health care system. Among the former, women compared with men suffer from the following discriminations: more than 50% doesn’t participate in the salaried labor force; have a higher unemployment rate; are concentrated in low-paid jobs and overrepresented in occupations not covered by social insurance; receive a lower salary for equal work, and suffer a higher poverty incidence when heading households, hence have less capacity to finance co-payments in private insurance or out-of-pocket expenses. Factors related to the health system are: direct access to social insurance is affected by permanent exit from the labor force to raise children that result in loss of coverage in sickness-maternity insurance; coverage is often indirect as a

211

dependent spouse of an insured male worker and in some countries only partial (in maternity but not in sickness or vice versa); coverage is usually lost by abandonment, divorce, or death of the insured; private providers often exclude women in fertile age because of the higher care costs (on pregnancy and some peculiar pathologies) or charge higher fees to compensate for such costs; complications during pregnancy, childbirth, and postdelivery are more frequent among women with poor living conditions, and user fees in the public sector particularly affect poor women because they use such services for them and their children more often than men (Gómez-Gómez 2002; see Sections 4.1.2 and 8.1.1). In most countries of the region, the percentage of women affiliated to social insurance is considerably lower than the percentage of men, with two notable exceptions: Costa Rica and Uruguay where the opposite is true but even so with more subtle inequalities. Costa Rica’s coverage of woman is 84% versus 77% of men, which results from indirect insurance as a spouse dependent of the insured (women have only 24% of direct insurance versus 52% in men); and women are the main users specially in the 20–45-year range (Martinez and Mesa-Lago 2003; Martinez 2005). Uruguay’s women also have greater coverage than men but the gap is narrower (96% versus 95%), nevertheless the affiliation is different: 22% of men are covered through social insurance payment to IAMC but only 14% of women, whereas 29% of men have individual voluntary affiliation to the IAMC versus 38% women, hence the cost is greater for women (ISSA 2003a). PAHO has studied gender and equity in health services (in Brazil, Chile, Colombia, Ecuador, and Peru) to identify gender inequalities in access and financing, as well as the interaction between such inequalities and different modalities of provision and financing. Women have more need of health services than do men because of their reproductive function and greater longevity that results in higher probability of suffering chronic diseases; furthermore, the usual

212 Reassembling Social Security concentration of health systems on curative care underestimates preventive needs that are greater for women, such as family planning and pregnancy care. Health reforms that have privatized services or introduced financing systems lacking solidarity worsen preexisting gender inequalities: women require more services but are forced to pay more than men due to their higher risks, a problem aggravated by the lower economic capacity of women; private insurance often practices cream-skimming and excludes fertile women or charge them higher fees, and the responsibility of raising the children falls on the woman who increasingly lacks a companion to share such burden—30% of households in the region are headed by a woman (Gómez Gómez 2002). Gender inequities are illustrated by the Chilean system, albeit part of them has been corrected by recent legislation. In 2002, 66% of contributors to ISAPREs were men and only 34% women, whereas in FONASA the proportions were 57 and 43% respectively, hence the gender gap was smaller in public-social insurance than in the private sector; on the other hand, combining both sectors, the proportion of women covered as dependents of insured was 28% higher than the proportion of men; these two opposite factors virtually compensated each other so that the percentage of beneficiaries by gender was similar (SI 2003; FONASA 2004). This phenomenon was partly explained by labor market discriminations already discussed: 39% of women participated in the labor force compared with 73% men, women earned one-third of men average wages, 45% of women were employed in the informal sector, and access of woman to health coverage was mainly as dependent of an insured male. But the phenomenon was also the result of important gender inequities practiced by ISAPREs whose plans discriminate against women because their risk factor is twice as high as among men. The premium charged to fertile-age woman (ages 20– 40) was between 2.4 and 3.1 times greater than the premium charged to men of the same age in 2001; therefore maternity costs were

exclusively financed by women but, even health plans that excluded childbirth care charged a premium to woman higher than men of equal age. ISAPREs could adjust premium in the annual renovation of contracts according to women age and their number of dependents (Pollack 2002). Due to these inequities the immense majority of women in fertile age are covered by the public-social insurance sector that thus grants subsidies to the private sector.5 The 2004 law established a universal basic package of benefits of equal cost regardless of gender, which must be granted both in the public-social insurance and private sectors at the same price within each ISAPRE. The compensation fund created by the 2005 law reduces discrimination by age and gender between open ISAPREs6 concerning financing of the basic package, and the annual increase in premiums must fall within a band fixed by the law. President Bachelet has promised to eradicate remaining discrimination of ISAPREs against fertile-age woman (MINSAL 2004; Inostroza 2005a; Gobierno de Chile 2005; Ley 20,015 2005; see Section 9.3.1).

9.2. Solidarity and income distribution Solidarity diverges with the type of insurance. Before the reform and still in many Latin American countries, the social insurance sickness-maternity program had solidarity within itself because its average premium is based on collective risk, the contribution is not related to individual risk, there are transfers between generations and benefits are standardized; therefore, high-income, healthy, low-risk, and young insured subsidize low-income, sick, high-risk, and old insured. Nevertheless solidarity is much stronger where social insurance coverage is high and includes rural and nonsalaried workers than where coverage is low and excludes those workers. In the latter, the uninsured population legally should be cared by the public sector but normally that sector has insufficient

Equal Treatment, Solidarity, and Comprehensiveness resources, limited access, and low quality of services, hence a significant part of the population (including the poor and low-income groups) is left without effective protection. Furthermore, social insurance is concentrated on curative services, shifting health prevention, promotion, and education to the public sector that also ought to provide curative services to most of the population, with proportionally less financial resources than social insurance. A national public health system (as exists in most of the English-speaking Caribbean and Cuba, and partly in Brazil and Chile) provides both preventive and curative care, compensates all risks, and is largely financed by fiscal transfers and salary contributions in some countries, hence it has greater solidarity. Most countries have several social insurance schemes for occupational groups or unions or mutual-aid societies that are typically closed, fix premiums and benefits according to the group risk and exclude the rest; these schemes have internal but not external solidarity. Micro-insurance (rare in Latin America) is organized by groups usually excluded from formal protection and based on their common occupation, ethnicity, or gender, as well as location; being self-managed their members decide priorities, fix a collective premium, and adjust benefits to the health needs of the group, thus micro-insurance has solidarity but limited to its membership. In the private sector, insurance firms/providers normally set their premiums and co-payments based on risk (age, gender, chronic diseases) and they don’t subsidize the poor and lowincome patients (a ‘subsidiary’ role left to the state), hence they lack solidarity: those enduring greater risks (e.g. fertile-age women, the elderly, the chronically sick) must pay higher premiums that often cannot afford, private insurance therefore is adequate for those with relatively high income who desire supplementary services and more personalized care but it is unsuitable for the entire population (ILO 2000; ILO/ISSA 2001b; Ginneken 2003). Fiscal transfers to health care have the most progressive impact on income redistribution

213

among all social expenditures, providing they are geared to the public sector and prevention services, targeted on the poor and low-income groups, rural and urban-marginal areas and indigenous peoples that endure the greatest poverty incidence (Mesa-Lago 1992). Diverse financing sources may have progressive, neutral, or regressive effects: (a) direct general taxes (e.g. a graduated income tax) are progressive, whereas indirect taxes (on consumption, e.g. VAT and sale taxes that generate half of tax revenue in the region) are regressive; specific taxes (on alcohol, tobacco, etc.) are neutral but some have positive effects on health; (b) payroll contributions without a ceiling on salaries are progressive, but if they have a ceiling are regressive; furthermore, the employer contribution may be transferred to the worker or the consumer through prices and be regressive; (c) recovery or user fees, usually charged to patients in the public sector to control demand or discourage overuse especially in outpatient consultation, laboratory tests and medicine, restrict the access of poor households, because such fees place a heavy burden on their shoulders, hence have regressive effects;7 (d) co-payments, similar to user fees but usually charged by private insurance (often added to the premium arguably to control excessive-unneeded demand), are financed by the insured and have a regressive impact on those of low income (Chile publicsocial insurance imposes them progressively according to income); and (e) out-of-pocket expenses are the most regressive among all financing sources, because they afflict the poorest households and generate new poverty (ILO 2000; Castaño et al. 2001; CISS 2004; Baeza and Packard 2005; for financing sources see Section 11.1.1).

9.2.1. Solidarity vis-à-vis equivalence Health care reforms in the region emphasize the new principle of equivalence between individual contributions and health benefits/quality of care but not to the degree of structural pension reforms. It is argued that

214 Reassembling Social Security strict equivalence decreases coverage, because of the exclusion of those in most need that cannot pay contributions; health care protection adapted to groups in need consequently demands a partial or total disconnection of contributions from benefits as well as redistribution of resources (Roberts, Stafford, and Ashworth 2002). It is important, however, to distinguish between equivalence at the individual and collective levels because they have diverse effects on solidarity. Under individual equivalence (as in Chile initial reform, later changed by second generation reforms) there is no solidarity: the contribution depends on the affiliate risk and benefits received, hence private for-profit insurance firms practice cream-skimming and reject those exposed to high risk because they are costlier and less profitable (as often have low income and can’t afford co-payments), while concentrating on groups of smaller risk and greater income capable of financing co-payments, allowing such firms to capture a sizable share of total health funds. The only existing solidarity is external and limited: the public sector is left in charge of the least profitable groups because they endure greater risk-costs and lower income; therefore, this sector has to protect most of the population with a smaller share of funds. Conversely, there is solidarity under collective equivalence (as in Colombia and Chile after 2005) because affiliates contribute according to their payment capacity instead of their risk, all receive a compulsory basic package of benefits, diverse individual risks are compensated (adjusted by a formula) and financed with fiscal subsidies and/or solidarity contributions. The basic package infuses an important element of solidarity within the system due to the incorporation of poor and low-income groups with equal minimum benefits, albeit not eliminating all differences in access and quality of care because of the package basic level and additional benefits coverage made contingent upon income and risks (solidarity is therefore considerably less than in universal systems with integral benefits such as

Cuba’s public system and Costa Rica’s social insurance), but such inequality can be reduced if the package involves universal guaranteed benefits that are gradually expanded (Titelman, Uthoff, and Jiménez 2000; see Section 9.3.1).

9.2.2. Mechanisms in favor and against solidarity There is not a methodology to properly assess the net result of mechanisms in favor and against solidarity, but a detailed analysis of eleven countries (Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, Mexico, Nicaragua, Peru, and Uruguay) led to the identification of fourteen mechanisms favorable to solidarity and seventeen contrary to solidarity.8 The selected countries are all six in the pioneer-high group, four out of seven in the intermediate group and one out of seven in the latecomer-low group.

9.2.2.1. Mechanisms favorable to solidarity 1. Universal coverage/access (in the six pioneer-high countries) or close to universal (in the two advanced intermediate countries: Colombia and Mexico), achieved by effective provision by the public sector to the poor or low-income population and financed with taxes or fiscal transfers (Argentina, Brazil, and Cuba) or by social insurance, covering all dependent family, with benefits equal to those in the contributory regime and financed with state transfers (Costa Rica) or by a regime financed by solidarity contributions and fiscal transfers (Colombia subsidized regime); 2. Integration of all the system with similar benefits either in the public sector (Cuba) or in social insurance (Costa Rica); 3. Absence of separate schemes that provide better care and receive fiscal subsidies (Costa Rica); 4. Granting of a universal basic package in all sectors or regimes regardless of income,

Equal Treatment, Solidarity, and Comprehensiveness age, gender, location and risk, hence redistributing from affiliates of higher income and healthy to those of lower income and sick: Argentina, Bolivia (but see below), Brazil, Chile (among ISAPREs to counteract risks by age and gender9 ), Colombia (in the subsidized regime financed by capitation according to age, gender and location, but see below), Costa Rica and Cuba (both integral care), and Peru (within social insurance while the public package targets the poor and maternal-infant population); 5. Public health financing concentrated on the poor and low-income population, the public sector offers free care to the poor and subsidies to the uninsured according to their income (Uruguay) or public-social insurance provides subsidies according to income, the poorest don’t pay and higher-income groups are charged increasing co-payments (Chile) or fiscal transfers to social insurance regimes for the poor and low-income groups or solidarity programs in the public sector or social insurance targeted on poor peasants, low-income families and/or maternal-infant groups (Brazil and Mexico); 6. Compensation or solidarity funds that either provide high-complex care among geographic areas (Brazil, Colombia—the fund also finances prevention and promotion, Peru— reimburse services provided to social insurance affiliates in the public sector, and Uruguay) or supplements financing of the basic package in the entire system guaranteeing a prefixed minimum for each affiliate and dependent family (Argentina) or provides additional resources to finance the package in municipalities with insufficient resources (Bolivia); 7. Fair proportions between the health sectors regarding their population covered and their share of total health expenditures, as well as human and physical resources (Costa Rica and Cuba—public sector only); 8. Relatively small inequalities in the geographical distribution of public financial funds (Costa Rica and Cuba) or increase in public financing and targeting it to poor geographic areas (Colombia—in the first stage

215

of the reform) or distribution of fiscal transfers to geographic areas based on their needs resulting in transfers from wealthier to poorest areas (Chile) or transferring fiscal resources to provinces particularly poor (Bolivia); 9. Substantial allocation of health financing to the first level of care (proportionally reducing allocations to the two higher levels), which is the one that solves most health problems suffered by the poor and lowincome groups (Costa Rica, Cuba, Nicaragua, and Peru); 10. Distribution of public/social insurance financial resources so that the poorest and low-income strata receive proportionally more benefits than what they contribute (Costa Rica and Chile); 11. Low out-of-pocket expenses as percentage of total health expenditures (Colombia, Cuba, Costa Rica, and Uruguay) and no user fees in the public sector (Brazil, Costa Rica, Cuba, and Peru); 12. Transfer of a solidarity contribution paid by high-income insured and assigned to expand coverage and benefits to poor and lowincome groups (Colombia—transferred from the contributory to the subsidized regime) or retention of the entire social insurance contribution to affiliates who cop out to other insurance firms or providers (Costa Rica and Peru); 13. Fiscal incentives to promote the affiliation of low-income, unskilled self-employed and similar informal groups (Costa Rica) or fiscal subsidies to low-income pensioners to pay their premium to providers, as well as employers’ payment of the difference in premium to low-income employees and extension of health care to low-income unemployed (Uruguay)10 ; and 14. Control or proscription of adverse risk selection (cream-skimming) by age, gender, risk, or income practiced by private insurance firms and providers (Colombia, Chile since 2004–5, Peru and Uruguay—among providers contracted by social insurance, but see below) or forcing providers to accept all social insurance affiliates regardless of preexistent chronic conditions (Nicaragua).

216 Reassembling Social Security

9.2.2.2. Mechanisms contrary to solidarity 1. Low coverage of the total population, particularly the legally assigned to the public sector in countries with a high percentage of uninsured (Bolivia, Nicaragua, and Peru), and lack of social insurance coverage (either sickness or maternity) or restrictions imposed to dependent spouses and children, which force them to get care in the public sector or to buy private insurance (Nicaragua and Uruguay); 2. Segmentation of the population between the three sectors, without solidarity among them and mainly based on income: poor and low income in the public sector, middle income in social insurance, and high income in the private sector (Argentina, Bolivia, Mexico, Nicaragua, Peru, and Uruguay) or segmentation between contributory and subsidized regimes, and with surviving public and private sectors (Colombia) or segmentation between a public and/or social insurance sector and a private sector (Brazil and Chile); 3. Continuation of separate schemes with superior benefits whose affiliates don’t contribute to the principal social insurance and receive fiscal subsidies financed by taxes usually on consumption and paid by all the population including the uninsured poor (Argentina, Bolivia, Brazil, Chile, Colombia, Mexico, Peru, and Uruguay) or separate better health facilities for groups with relative high income and political power (Cuba) or payment of private insurance to some of those groups (civil servants in Brazil); 4. Different basic packages between insured groups (Colombia—the contributory regime package is twice the subsidized regime) or exemption from the basic package of some groups or providers (Argentina provincial OS) or existence of geographical, ethnic, or cultural barriers of access to the basic package (Bolivia) or granting of a lower package to pensioners (Nicaragua); 5. Public subsidies to affiliates of social insurance who use public hospitals and highcomplex services without paying (Argentina, Bolivia—also for primary care, Mexico, Nicaragua, Peru, and Uruguay—also for

outpatient consultation and medicines) or similar subsidies to private insurance firms/ providers (Brazil, Chile—better control in the 2000s, Nicaragua and Peru), hence the public sector operates as default insurance, reducing the cost of social insurance and private providers (and their affiliates) and depleting resources needed for care of the poor; 6. Exemption of some groups from the solidarity or compensation funds (Argentina provincial OS) or shift in orientation from an original progressive allocation to poorest geographic units (Bolivia); 7. Significant disproportions between the three sectors regarding the populations covered and their percentages of total health expenditures and human and physical resources that are higher in the private and social insurance sectors than in the public sector (Argentina, Bolivia, Colombia, Mexico, Nicaragua, Peru, and Uruguay); 8. Substantial differences in geographical allocation of public financing that doesn’t take into account local epidemiological profile and needs, assigning proportionally more to wealthy than poor areas (Bolivian municipalities, Brazilian and Mexican states) and cut or stagnation of fiscal transfers to geographic areas in need during economic crisis (Argentinean provinces); 9. Small share of total health financing to the first level of care and overwhelming shares of the two higher levels (Chile and Uruguay); 10. Proportionally less health care payments by high-income affiliates than by lowand medium-income affiliates (payments proportionally increase as income diminishes) the former receiving a regressive fiscal subsidy (Uruguay) or fiscal subsidies unequally distributed by income groups (Brazil and Peru); 11. Substantial out-of-pocket expenses by poor and low-income families that lack effective access to public sector or social insurance (Brazil, Mexico, Nicaragua, and Peru), as well as user fees in the public sector (Argentina, Bolivia, Mexico, Nicaragua, and Uruguay); 12. Transfer of social insurance affiliates of high and upper-middle income to the private sector (Chile) or to other social

Equal Treatment, Solidarity, and Comprehensiveness insurances with better benefits (OS in Argentina), in both cases transporting their entire contribution, hence deteriorating services of the major social insurance program that is left with low-income affiliates who endure the highest risks;11 13. Setting a percentage contribution to unskilled, low-income self-employed equal to the sum of the percentage contributions of employers and salaried workers, hence discriminating relative to the latter who only pay their own contribution (Colombia, Mexico, and Nicaragua) or a percentage contribution equal to that exclusively paid by the employer (Bolivia and Peru) or a fixed contribution to the self-employed regardless their income (Argentina); 14. Adverse selection by private insurance firms and/or providers that concentrate on the young, low-risk and high-income population, leaving to the public sector the old, female, high-risk and poor or low-income population, as well as increment of co-payments to fertile-age women and the elderly (Chile12 and Uruguay—providers with individual voluntary affiliates); 15. ‘Closed’ insurance firms that cover groups within a trade with relatively high income (Chile) or banning low-income affiliates to change to a private firm with higherincome affiliates or payment by the latter of only the basic package to those who move with lower income (Argentina provincial and other OS); 16. Tax exemption on contributions, premium, and co-payments to the private sector favoring high-income groups (Brazil and Chile); and 17. Nonpoor individuals with paying capacity (free riders) receive fiscal subsidies (Chile and Colombia13 ).

9.3. Comprehensiveness and sufficiency of benefits This section describes and analyzes three key elements of comprehensiveness and

217

sufficiency of benefits: the basic package, catastrophic illnesses and high-complex care, and perceived quality of care.

9.3.1. Basic package of benefits One of the World Bank’s (1993) major recommendations was the creation of a basic package of health benefits adjusted to epidemiological and local needs of users, and the targeting of public financing on such package. Since the Bank report and until 2005, fifteen Latin American countries established the package; seven offer it compulsory by insurance firms and/or providers to all affiliates, regardless of their risk and income, whereas the other eight only offer partial or limited packages. The basic package is positive because it guarantees minimum sufficiency in health benefits, allows providers to practice risk pooling, and reassigns health expenditures giving priority to essential needs and primary care, more cost effective in the long term than curative care (especially high complex). But the package faces intricate issues, for instance, the redistribution formula for risk adjustment and the lack of an optimal definition of what is the acceptable minimum of benefits; a study of twelve developed countries with obligatory basic packages found that their definitions were too ample and vague and exhibited remarkable differences on the benefits included. The determination of the inclusion of benefits with certain quality in the package generates a conflict between equity-sufficiency on the one hand and competition-efficiency on the other. If benefits are reduced, the package will be cheaper, stimulate entry of more providers and a greater choice among beneficiaries, but will increase their out-ofpocket expenses and diminish the redistributive effect: the poor won’t be able to pay for benefits excluded from the package and hence receive a low level of care, whereas high-income beneficiaries will be able to pay for those extra benefits and receive better care. Conversely, if the package includes many benefits, its cost will be higher, fewer providers

218 Reassembling Social Security will join the plan and users will have reduced choices, but the redistributive effect will be greater. In many countries, concern on the package cost predominates over its sufficiency (Bertranou 1999). Other problems that afflict the package will be discussed at the end of this section. Out of the fifteen countries that have introduced the basic package covering all or part of the population, in seven it is fully operational and compulsory for insurance firms and/or providers to all affiliates regardless of their risk and income (Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, and Cuba) whereas in seven it is limited or partial (Dominican Republic, Guatemala, Mexico, Nicaragua, Panama, Paraguay, and Peru); Uruguay has not defined the package but it largely functions in practice with some restrictions (Table 9.1).14 In virtually all the countries social insurance offers integral care to all its insured (including minimum benefits); in Costa Rica it provides integral care to 87% of the population and in Cuba the public system offers integral care to all the population.15 Conversely, in Nicaragua there are restrictions on coverage and benefits, the so-called ‘integral system’ actually provides insufficient services, with significant exclusions and only during part of the life of the insured and their dependents. Because Costa Rican and Cuban systems provide full benefits, they are excluded from the following comparative analysis of the basic package.16 Five countries lack the package (Ecuador, El Salvador, Haiti, Honduras, and Venezuela); two of them have recent projects albeit were not operational in 2006. Ecuador’s health law stipulated the creation of the package and gave until the end of 2003 to define its contents; a new program was drafted in 2005 to expand health insurance to the poorest two quintiles of the population and in selected regions and municipalities (AUS); it changed name in 2006 (PROAUS), but none of the three was operational in 2006. El Salvador’s ministry offers priority services that include preventive and curative benefits to children until age 11, woman in fertile age, adolescents,

adults, and the elderly. Haiti’s package (BPS) was interrupted by the crises; a new plan projects its expansion in 2004–8 to 4% of the population with geographically and demographically defined criteria, delivered by NGOs and other public–private providers that will receive a per capita from the ministry. Honduras’ government proposed to define a package in 2002 and deliver it to 4% of the population; a new program (NAS) offers basic benefits at the primary level, such as full maternal-infant care in communities and communal houses for childbirth. Venezuela’s package is provided only by social insurances, but not in the public sector at least until 2006. There has been a proliferation of basicpackage programs in several countries (overlapping, substituting, or adding) with different names and abbreviations that create confusion and require clarification. Argentina’s original package (Programa Médico Obligatorio: PMO) was replaced by a second package in 2001 currently in force (PMOE), there are no significant differences between the two, except for inclusion of some medicines and minor restrictions; providers (OS and EMP) can offer additional plans at higher prices. Bolivia introduced the first basic package in 1996, which was replaced by a second package in 1998 and by a third package in 2003 currently in force (Seguro Universal Materno Infantil: SUMI). Brazil has two major programs: the basic package (Piso da Atençao Básica: PAB) introduced in 1998 has fixed and variable components, the later was expanded in 2001 (PABA); in addition there is a family health program (Programa de Saùde de Familia: PSF) that started in 1994. Chile’s package brought a major difference in 2004 because its access and benefits are universal and guaranteed (Garantías Explícitas en Salud: AUGE). Colombia’s package is the most complex: the 1993 reform law established an ‘integral system’ with three packages: the most basic is obligatory for all the population and managed by the ministry, departments, and municipalities (Plan de Atención Básica: PAB); the second has additional benefits, is compulsory for all contributory-regime

Equal Treatment, Solidarity, and Comprehensiveness affiliates and administered by insurance firms (Plan Obligatorio de Salud: POS), and the third, with less benefits than POS, is for subsidized-regime affiliates (Plan Obligatorio de Salud del Régimen Subsidiado: POSS); voluntary supplementary plans (PAC) managed by insured firms offer contributory-regime affiliates additional benefits to those in POS. Mexico has had a plethora of basic packages targeted on the poor population with fiscal subsidies and most with external aid, managed by either the ministry, the principal social insurance, and other federal entities; the family health program (Seguro Popular de Salud: SPS) that began implementation in 2003 is the latest and most ambitious of all programs. Peru ministry established two packages (1997 and 1998), which respectively covered children in public schools, and pregnant women and children to age five; the two programs merged in 2001 into the current basic package (Seguro Integral de Salud: SIS).

9.3.1.1. Benefits included in the basic package Most countries with a functional package give priority to prevention (first level) over curative care but a few countries provide more extended packages that include even certain high-complex procedures. The package is often targeted on vulnerable groups (pregnant women, children, and the elderly) and/or on extreme poverty areas (Bolivia, Guatemala, Nicaragua, Paraguay, and Peru), but in other countries it has universal scope. The basic package normally provides first-level benefits such as: health education, promotion, prevention, immunization against infectious diseases (AIDS in some countries), drugs and tobacco control (in a few countries), family planning, nutrition, control of diarrhea and hydration, diagnosis, maternal-infant care (usually prenatal, childbirth and postdelivery care, young children care in some countries), basic dental care (in the majority), essential medicines, and some curative actions at the primary care and emergencies.

219

Argentina’s package (PMOE) guarantees a set of essential benefits prioritized according to socio-demographic features of the population and clinic criteria; it is targeted on vulnerable groups (pregnant women, children, and the elderly), and determines what specialized procedures are included; all national OS and providers (EMP) must grant the package, but provincial OS and part-time workers are excluded. Bolivia’s three successive packages increased benefits from thirty-two to ninety-two, originally all were targeted on pregnant women and young children, later some benefits (treatment of some infectious diseases) were extended to the entire population, the current package (SUMI) added new benefits including a few complex procedures. Brazil’s package provides seventy-five benefits; the fixed component (PAB) plans to gradually expand first-level benefits to all the population but also complex treatment of high-incidence diseases; the variable component (PABA) includes some medium and high-complex procedures, while the family health program (PSF) focuses on the first level but offers some medium-complexity interventions. Chile’s package (AUGE) will gradually extend from twenty-five to fifty-six pathologies and health conditions in 2005–7, that causes most mortality and incapacity. Colombia’s basic package for all the population (PAB) offers first-level benefits; the one for the contributory regime (POS) adds diagnosis and curative care at all levels, rehabilitation, and most medicines; the subsidized-regime package has only 60% of the benefits included in the contributory-regime package (excludes surgery, dental care, cancer treatment, and other procedures, which might be available at public hospitals) and their benefits are of lower quality; supplementary plans provide additional benefits exclusively to those entitled to POS, such as better hotel services, higher technology, and excluded medicines. According to the Dominican Republic law (not fully enforced in 2006), the package (PBS) should include ample and costly benefits in all three regimes such as surgery, some high-complex procedures and medicines.

220 Reassembling Social Security Mexico’s package (SPS) is planned to progressively increase from 78 to 91 pathologies, concentrated on the first level—largely medicines—that cover 85% of demand, and some at higher levels; the SPS is expected to eventually cover nearly all services except for catastrophic illnesses. Peru’s package (SIS) essentially provides first-level benefits, initially it was expected to offer higher level care (hospitalization, major surgery, intensive care, etc.) but all these procedures combined were less than 1% of the total benefits in 2005.

9.3.1.2. Coping with cream-skimming Profit-seeking by insurance firms and providers and their cream-skimming practices have adverse effects on the sufficiency of benefits during the insured life time. To cope with risk selection, providers of the package in several countries cannot reject affiliates or discriminate by age, gender, risk, or preexisting diseases. Colombia was the pioneer in banning cream-skimming, contracts are not annual, and a basic plan is guaranteed for the entire life of the insured; insurance firms in the contributory regime can neither reject any applicant affiliated to that regime nor take into account preexisting health conditions (but can set an initial waiting period); but the system discriminates by income because the poor are confined to the subsidized regime that offers a worse package. The Dominican Republic package should be equal for all the population affiliated to the three regimes stipulated in the law, regardless of age, gender, health status, and income, but only the subsidized regime was in operation by the end of 2006. Mexico’s package(SPS) does not discriminate by risk or preexistent diseases and is granted free to the poorest two income quintiles. Chile is an example of correction of deeply rooted cream-skimming for almost two decades. In theory all affiliates to the publicsocial insurance (FONASA) had free health care access to municipal services, while access to second- and third-level services went from free to subsidized according to income, but

because of financial limitations in practice FONASA rationed services through long waiting lists for specialized medicine. ISAPRE legally had to offer the same services of FONASA, but coverage was restricted by tariffs set by insurance firms and restrictions imposed on affiliates. ISAPREs offered contracts designed according to individual risk (age, gender, and chronic diseases) and could annually change premiums and benefits, thus they did not integrally covered all risks during the insured entire life. Affiliates older than 60 or who suffered chronic or catastrophic diseases were ‘captive’, that is, they lacked an option to accept or reject contract changes because it would not be accepted by other ISAPREs. That situation has changed in the past decade particularly since 2004: since 1996 the superintendence sets a price index and maximum rates for the plans of ‘captive’ affiliates; the 2004 law guarantees access to a set of specific benefits, revised every three years, applied and gradually extended to all system affiliates (FONASA and ISAPRE) regardless of age, gender, risk, and income, and the ministry of finance must annually assign the needed resources; the same law sets a maximum waiting period to various health procedures and improved the definition of preexisting sicknesses, and the 2005 law created a fund to compensate risks among open ISAPREs.

9.3.1.3. Population coverage Only six countries have some data on the total population covered by the package, ranging from 11 to 47% but the lack of statistical standardization impedes proper comparisons.17 Brazil’s PSF functioned in 75% of municipalities in all states in 2003 and population coverage rose from 0.7% to 33% in 1994–2003; Guatemala’s SIAS covered 27% of the total population in 2002 (the reform goal was to cover 44%) targeted on rural, poor indigenous populations who suffered the worst health indicators, as well on some relatively urbanized populations; Mexico’s SPS covered 11% of the total population in all the states in 2006

Equal Treatment, Solidarity, and Comprehensiveness (versus a target of 31% in 2010), except the federal district; Paraguay lacks an adequate estimate of the population covered but the number of persons treated under the package was 47% of the total population in 2005 (an overestimation of population coverage because the same person may have received care more than once); and Peru’s SIS covered 33% of the population in 2004, all packages combined (including those in social insurances) covered less than half of the population. The total population care for the package is usually broken into areas that range from 4,500 people in Brazil18 and Panama to 10,000 in Guatemala and Haiti.

9.3.1.4. Cost, adjustment and financing Data on the basic package annual cost per capita were available in only seven countries for various years between 2002 and 2006, and ranged from US$8 to US$200 (see Table 9.1): US$8–12 in Brazil (including fixed and variable components), Guatemala, and Haiti; US$87 in Dominican Republic (in subsidized regime in force, the contributory regime projects US$166); US$83 and US$150 in Colombia (in the subsidized and contributory regimes respectively, the gap was planned to disappear in 2000 but actually expanded); US$154 in Nicaragua, and either US$87 or US$200 in Mexico according to two different estimates. Information on the periodic adjustment of the package is accessible in only four countries: Guatemala didn’t adjust it in 1997–2002 because of lack of resources, Argentina failed to adjust it for some time but was done annually in 2003–2004, Brazil adjusts it annually, and the Dominican Republic law stipulates an annual adjustment or every six months in case of high inflation.19 In Brazil both the PAB and the PSF are free; in Chile AUGE is free for the poor, the government sets a universal premium or value per capita of the package, defines maximum copayments in a range of prices for users and sets ceilings according to family income; in Colombia the most basic package (PAB) is free for all the population, whereas the subsidized-

221

regime package (POSS) should be free for the poor, but the contributory-regime package (POS) requires contributions; in the Dominican Republic the package is free in the subsidized regime, but it has co-payments in the contributory-subsidized and contributory regimes; in Mexico the SPS is free for the lowest two income quintiles that make up 93% of total coverage, the rest has to pay a premium according to income; and in Bolivia the package is not free but charges an affiliation fee and users’ payments for lab tests, sonograms, medicines, etc. (40% of municipal hospitals charge fees for their services). The package is financed by a per capita adjusted by risk (Colombia) or age and gender (Chile), to avoid risk selection, but also by a per capita equal for all the population (Argentina, Bolivia, Brazil, Dominican Republic, Guatemala, Haiti, and Nicaragua) that may help to reduce regional inequalities but does not take diversity into account, hence obstructing flexibility required by communities to adapt the package to their diverse features and needs (for detailed financing data see Sections 11.1.2 and 11.1.5).

9.3.1.5. Evaluation of the package, positive outcomes, and problems The package performance and effects (coverage, quality of care, cost, efficiency, and impact on the beneficiary population) has not been evaluated in at least six countries: Argentina, Brazil, Guatemala, Mexico, Panama, and Peru. An example of the adverse effects of lack of evaluation is Guatemala whose package (SIAS) began in 1997 and its monitoring has been sporadic and informal instead of periodical and systematic; in the initial stage monitoring focused on financing and neglected coverage and quality; measures were taken later to improve SIAS but their efficacy had not been assessed at least until 2004. Evaluation mechanisms for providers’ performance in SIAS first four years were inadequate, confusing and often contradictory; later there was moderate improvement, because of the inclusion of indicators (not goals) but

222 Reassembling Social Security

Table 9.1. Sufficiency of health care benefits and perceived quality, 2003–5

Countries Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican R. Ecuador El Salvador Guatemala Haiti Honduras Mexico Nicaragua Panama Paraguay Peru Uruguay Venezuela

Basic package of benefits

Annual per capita (US$)

Yes (PMOE) Yes (SUMI) Yes (PAB)a Yes (AUGE) Yes (POS, POSS) Yes (integral benefits) Yes (integral benefits) Yes (PBS), partialb Project (2005)c No Yes (SIAS), limited Project (2004) No Yes (SPS), partial Yes, very limitedd Yes, limitede Yes (PCSB), partial Yes (SIS), partialf Not clearly definedg No

n.a. n.a. 8.30 a 10.90a n.a. 150 and 83 n.a. n.a. 87

8 10 87 or 200h 154 n.a. n.a. n.a.

Catastrophic illnesses

Periodic surveys among users

Yesi Noj Yes Yesk Yes Yes Yes Partial Nol Nol Nol No Nol Partialh Nol Partiale Nol Nol Yesg Nol

Yesm No Yes Yes Yes Yes Yesn Noo No Yesp No Noq Nor Nos Not Nou Nov Now Nox No

Sources: Legislation; see also text. n.a. = non available a There is a fixed and a variable component of the basic package that includes the PSF. b Implemented only in part of the subsidized regime. c The 2001 reform law stipulated the creation of a package, new programs were drafted in 2005–6, none was operational in 2006. d In the public sector, it is limited to pregnant women and small children; in social insurance, it is limited by gender and age, a reduced package is granted to old-age pensioners. e In the public sector, it is granted in two regions; social insurance provides integral care. f By the ministry and social insurance. g Not a defined package but minimum benefits provided by the ministry and social insurance; catastrophic illnesses are covered by the FNR. h Author’s per capita estimates based on family figures of US$269 or US$625 and a ratio of 3 dependents; the basic package includes some second-level actions, and a catastrophic fund covers selected diseases; social insurances integrally cover the basic package and catastrophic illnesses. i In the public sector; in social insurance (OS) and private sectors varies according to contracts. j SUMI grants complex benefits to the maternal-infant group. k In the public sector, in the private sector requires co-payment. l Granted only by social insurance with restrictions in some countries: Ecuador only in the principal program not in the peasant scheme; Nicaragua’s law of 2005 (delayed to 2007) stipulates it in social insurance; Peru’s social insurance is the default for those affiliated by providers (EPS) and enterprises; Venezuela’s legal draft of 2005 not approved by 2006 stipulates it. m Among OS in Buenos Aires. n One survey in 1997 and subsequent biannual surveys reported without giving results.

Equal Treatment, Solidarity, and Comprehensiveness performance continues to be reported by providers and there has not been an evaluation based on surveys; an assessment of the results of SIAS and its impact on health levels had not been done by 2003. Poor control and evaluation of contracted NGOs that provide the package forced cancellation of such contracts to a third of NGOs due to incompetence or nonfulfillment of duties, for example, covered a population lower than agreed and undeclared surpluses.20 Undoubtedly, the basic package is a key solidarity measure to extend primary care to poor and vulnerable populations and currently three countries grant it universally without discrimination by risk or income. In some countries (Brazil and Chile), the package functions as the entry door to the health system and refers patients to specialized units of middle and high complexity, helping to transform the traditional curative, hospitalcentered model and reducing the congestion in hospitals particularly in outpatient consultation and emergencies, which previously took care of the first level, an inefficient and costly practice. Chile AUGE established a new model of care that promotes an annual preventive check up, early diagnosis of certain diseases, and primary care with enough capacity to cope with most emergencies and with adequate referrals to the second level. Colombia compensation fund stimulates prevention and promotion services. Brazil standardized package eliminated previous inequalities that existed between regions of diverse development, although inequalities

223

based on different needs remain. Mexico SPS reportedly cut out-of-pocket expenses by 25% (particularly on medicines) that account for half of total health expenditure and had a devastating effect on the poorest families; 40% of SPS total expense is on medicines and 27% on preventive services. On the other hand, the basic package confronts important problems in virtually all countries (data were unavailable on Paraguay and Peru), some common and others of diverse nature: (a) the package coverage is usually very low, does not reach the poorest populations and rural areas due to geographical and cultural barriers (40% of Bolivian women in the highlands does not get it) or requires minimum infrastructure in the area covered (in Mexico it demands modernization of obsolete equipment in some states, labs, and ambulances in many rural areas and timely distribution of medicines); (b) the target population lacks adequate information on the package and some ignores its existence (Bolivia, Colombia, and Guatemala) or the information is confuse (Mexico); (c) there are exclusions (provincial OS and other providers in Argentina) or benefits are seriously restricted (Nicaragua’s ‘mini-package’ is granted only to old-age pensioners in the capital city and seven departments have fewer benefits that the standard package); (d) imperfect estimates of the package cost have reduced delivery of services related to the epidemiological profile of the low-income population (Colombia) or the package provision is not adapted to important socioeconomic and cultural differences in

← (continued) o One survey in public and private sectors (not in social insurance) without giving specific results. p Surveys in one and five SIBASIs, both with detailed results; a web page open to opinions on SIBASI; not in the rest of the public sector or in social insurance. q A survey on the basic package was planned for 2006. r Sporadic surveys reported without giving results. s National surveys of user satisfaction in 1994 and 2000. t In social insurance but not by surveys but a web page. u Only in San Miguelito Hospital in 1999. v In social insurance but suspended after 2000. w One survey in 1997. x One survey by social insurance in 2004.

224 Reassembling Social Security the population (Mexico) or the package was initially designed as a single inflexible model incapable of adaptation to prevailing local diseases and later infused with more flexibility but disparities in funds and technical capacity led to an uneven development (Guatemala);21 (e) the package was rapidly expanded before completing pilot tests and lacked an adequate legal framework to regulate contracts with providers, hence resulting in improvization, errors, and abuses (Guatemala); (f ) instead of one unified package at the national level there have been several packages created by successive administrations, some of them discontinued or their funds cut (Mexico); (g) insurance firms and/or providers lack incentives to prioritize prevention and promotion (Brazil, Colombia, and Chile—up to 200522 ) and the third level of care may take most available funds, favoring middle-income urban groups and leaving not enough resources to primary services essential for the poor (Bolivia); (h) the personnel in charge of the package implementation is insufficient (Bolivia) or it is difficult to hire due to excessively rigid labor legislation (Brazil) or was hired temporarily and changed later due to political shifts (Colombia) or it has been very slow in providing the benefits (Bolivia and Dominican Republic); (i) universities and other educational centers have not adjusted their curricula and internship practice to train professionals able to work in the program (Brazil); (j) the package needs criteria for cost effectiveness (Argentina); (k) social insurance does not have incentives to provide the package to the uninsured population because it can’t recover costs, particularly at the third level (Bolivia); (l) a cut in the per capita paid to providers forces them to impose waiting periods or other restricting techniques (Nicaragua); (m) lack of calculation of benefits cost and their financial viability based on stipulated revenue have generated deficit (Argentina) or funds have been cut (Colombia, Guatemala, and Mexico) or are insufficient (Bolivia and Peru); and (n) there was not a previous study of comparative costs between internal and externally contracted services and still a solid estimation of

external-provider costs is needed (Guatemala’s NGOs).

9.3.2. Catastrophic illnesses and high-complex procedures Catastrophic illnesses (CI) that require highly complex procedures (HCP) cause enormous health expenditures that may lead to financial insolvency according to the given financial capacity of a family. Such risk raises serious concern among the insured, particularly when they are old and don’t have enough resources. Only seven countries in the region entirely cover this risk (the six in the pioneer-high group and Colombia), three countries partially (Dominican Republic, Mexico, and Panama), and the remaining ten don’t offer coverage at all, albeit do it through social insurance23 (see Table 9.1). In most countries the armed forces and the police have their own hospitals that cover CI without restrictions (this section is largely based on legislation and Mesa-Lago 2006a). Among the seven countries that provide universal coverage against CIs, the sector in charge varies considerably; four countries have risk-pooling mechanisms, and data from two countries show huge increases in CI number and costs. In Argentina, they are covered in the public sector and the OS (but with diverse scope according to the contributions paid) and to less extent in the private sector; since 2000 there is public insurance for high-cost but low-incidence pathologies with benefits regulated to control overuse and expenditures. Brazil’s federal government and states share responsibilities on HCP, while the expanded basic package provides middle-complex procedures through municipalities, and a federal compensatory fund (FAED) reduces regional disparities to finance HCP; the number of these procedures increased 246% between 1990–5 and 2001–2 (Biasoto 2004b). Chile’s public-social insurance (FONASA) is the default underwriter for CI; only a minority of affiliates in private ISAPREs has CI coverage due to high

Equal Treatment, Solidarity, and Comprehensiveness co-payments charged as well as the possibility of noncovered affiliates to receive care at FONASA (whose CI coverage jumped 351% and its costs 449% in 1994–2004); since 2000 ISAPREs include CI coverage in their contracts paying only for those whose cost exceeds the deductible (Urriola 2005; Borzutzky 2006). Colombia’s private insurance firms must have pool-risk reinsurance to cover CI and avoid risk-selection, but firms that sell policies to providers in the subsidized regime offer discounts for low-risk incidence, thus inducing discrimination against high-risk affiliates. Costa Rica and Cuba cover all CI, the former through social insurance that provides integral care to 87% of the population (the rest have resources to buy coverage) and the latter through the public sector; both countries however report long waiting lists for surgery, and Cuba also reports scarcity of inputs and nonfunctional HCP equipment. Uruguay’s HCP is provided by specialized public or private entities (Institutos de Medicina Altamente Especializada: IMAE), financed by a compensation fund that covers all the population regardless of its affiliation. Three countries provide partial CI coverage either by the public sector or the basic package, whereas social insurance offers full coverage, and only Mexico has a compensation fund: the Dominican Republic’s reform law had to limit CI to a few HCP because high cost made it incompatible with the extension of the basic package coverage (some private insurances cover CI with a ceiling and a high premium); Mexico’s basic package (SPS) includes some complex procedures at the second level and a federal fund covers selected HCP but with differences among states; and Panama’s public sector covers CI in two out of its fourteen regions. The public sector in ten countries either exclude CI (Haiti) or provide very low coverage (Bolivia’s basic package delivers certain HCP to the maternal-infant group). Social insurance covers CI albeit with limitations in some countries: exclusion of peasants in Ecuador, exclusion of certain CI and HCP in Nicaragua (particularly among pensioners),

225

and unclear criteria for the selection of beneficiaries with a high degree of discretionary power in Nicaragua and Venezuela. None of these countries has a compensation fund for HCP. Ecuador’s 2002 health law, not enforced in 2006, stipulates coverage of CI including peasants but submitted to prior proper funding. Nicaragua’s public sector has very limited capacity and resources to cover CI; the social security law of 2005, postponed to 2007, stipulates coverage of HCP currently excluded in social insurance. Peru’s public sector virtually doesn’t provide HCP but social insurance does and functions as default insurance for providers (EPS) and enterprises health services. Venezuela currently doesn’t provide CI coverage in the public sector, but it is stipulated in the legal draft on health reform not yet approved in 2006, and a national register of CI patients is being elaborated (RBV 2005).

9.3.3. Perceived quality of benefits The quality of health services, part of the principle of sufficiency, has two dimensions: technical and perceived. Technical quality requires accreditation of health facilities and educational institutions, certification of professional skills and instruments for evaluation of technologies, protocols and quality of services. Perceived quality demands: freedom of choice of the insured or users concerning insurance firms and providers (see Section 10.6), public entities for the protection of user rights, and surveys measuring user satisfaction (Medici 2000). Due to space limitations we limit the analysis herein to perceived quality (for technical quality see Mesa-Lago 2006a). PAHO regional evaluation of the impact of health care reforms on perceived quality found that only four countries showed concrete experiences that promoted it. Obstacles to evaluate this issue are: thirteen countries don’t conduct periodic surveys to measure quality, others don’t provide information on assessments of reform results (Argentina) or omit improvement in perceived quality among the reform main goals (Guatemala)

226 Reassembling Social Security or have not considered specific actions to improve perceived quality (Paraguay) (Infante, de la Mata, and López-Acuña 2000; PAHO 2002a, 2002b). La Forgia (2006) states that most reforms had little to do with quality. Table 9.1 (last column) shows the seven countries that conduct periodic surveys either general or exclusively dealing with perceived quality or users’ satisfaction.24 In 1995–2004 Chile, Colombia, Costa Rica, Brazil, and Argentina (in that order) conducted most surveys and published their results. Eleven countries conducted occasional surveys normally without reporting specific results.25 Six countries don’t report any mechanism to evaluate quality: Bolivia, Ecuador, Guatemala, Haiti, Venezuela, and Panama (except in one hospital). Comparisons of country results are very difficult, as well as within one country longitudinally, because surveys usually don’t ask the same questions through time, have diverse scopes (the entire system, one sector or a specific program) and sometimes give divergent results in the same country. A summary of the most important survey results mostly in 2001–4 in half of the countries follows. In Argentina, 75% of OS affiliates ranked them from good to excellent on coverage and operation, and 53% said that services were better than expected, but the perception was worse in the countryside than in the metropolitan area. In Brazil, 64%–88% of interviewees were satisfied with public services, but later reported that such services had not improved and were probably worse than before the reform. In Colombia, a survey after ten years of the reform revealed a minimum level of overall satisfaction with services (6 in a 0–10 range); 62% of health professionals thought that the general service level had declined, whereas the ministry, universities, and civil society concurred that the primary-care level had worsened. In Costa Rica, 70% of social insurance affiliates were satisfied with its services and 88% with its treatment, whereas 47% was dissatisfied with Cuban public services and 55% with Peruvian public services. In El Salvador, 90–98% of users of the first-established and best local health

unit (SIBASI) said that they would return for care and recommend that service to others, but reported significantly lower levels of satisfaction on personnel and equipment and waiting period for appointments.26 In Mexico, the perceived need of fundamental changes in the health system decreased from 80 to 77% in 1994–2000 (still quite high), while the positive evaluation of such services rose from 13 to 19% (still very low). In Nicaragua, satisfaction with social insurance-contracted providers (private and public) increased slightly from 57 to 59% in 1998–2000 (41% was not satisfied) and childcare coverage reportedly fell from 60 to 45%. In Uruguay 75% of users of social insurance maternal-infant services appraised them as good or very good and 96–98% said to be satisfied with physicians and nurses.27 Chile conducted surveys in 1993–2001 that compare user satisfaction between public-social insurance (FONASA) and private ISAPREs. In 1995, fourteen years after the reform, 68% of ISAPRE affiliates preferred them and 23% favored FONASA, while among FONASA affiliates 66% preferred its services and 29% ISAPRE; the small gap exhibited a declining trend relative to the 1993 survey and would have been more favorable to FONASA if the survey had specified that its cost is substantially less than in ISAPRE (most interviewees wanted to change from FONASA to ISAPREs but couldn’t do it because of high co-payments). Furthermore, surveys show that many ISAPRE affiliates used FONASA third-level services, the latter provided 43% more lab tests and surgeries and 3 times the number of hospital beds per insured than ISAPREs. In 1998, 60–70% of FONASA users perceived an improvement in the quality of its services, with even higher percentages among affiliates in the lowest-income quintiles. In 2001, 52% of beneficiaries of FONASA and 44% of ISAPREs were dissatisfied with their respective services, the former due to insufficient resources and the latter for equity reasons; as lower the ISAPRE affiliate income, higher the dissatisfaction and 25% of said affiliates complained to be insufficiently covered. Twenty years after the reform, a

Equal Treatment, Solidarity, and Comprehensiveness Chilean supporter of further privatization in health services stated: ‘The country confronts a generalized discontent in this [private] sector, reflected in public surveys’ (Cifuentes 2000: 68). A comparison of survey results in six countries between public, social insurance, and private services suggests that the latter are preferred because they are faster and have better quality, but their access is obstructed by high premium and satisfaction decreases with income; public services are used by the majority (except in Costa Rica) because they are either free or very cheap but not because of their quality, and social insurance services are favored for their coverage and solidarity. Most affiliates to Argentinean OS preferred them to prepaid providers (EMP) and ranked OS coverage, trust and solidarity from good to excellent, but considered EMP services faster and better albeit with high co-payments; and yet half of OS affiliates who transferred to EMP became unhappy with their coverage and services. Chile’s ISAPRE affiliates were somewhat more satisfied than FONASA affiliates, but the gap between the two gradually decreased; in FONASA satisfaction increased as income diminished, whereas in ISAPRE dissatisfaction increased in tandem with income. In Costa Rica, most insured at CCSS were satisfied with its services but users of primary-care cooperatives said that their waiting time was shorter. The overwhelming majority of beneficiaries of public services in Dominican Republic used them because were free or very cheap (only 7% because of adequate quality) and considered private services of better quality but very expensive and with poor availability of physicians; most affiliates to private providers were pleased with their services but complained about their high premium. In Ecuador, 75% of social insurance affiliates were not satisfied with their services but even so preferred them to public services. Conversely, in Peru public services were judged better than those of social insurance and the highest contentment was with private services. Surveys in seven countries have identified the four main causes of dissatisfaction of ser-

227

vices and the author has ranked them as follows: (a) bad quality or inadequate facilities or services or lack/deficient equipment; (b) difficulties in access such as delays to get an appointment or to receive care, and long waiting list for surgery and specialists; (c) deficient treatment or maltreatment by personnel; and (d) lack of medicines or they are inadequate or difficult to get or their costs are high. Only five countries protect users and their rights by an independent agency (‘defender of the people’ in Colombia, Costa Rica, and Dominican Republic), arbitrage commissions for medical claims in Mexico or users associations in Uruguay, which facilitate processing of complaints and claims, particularly among lowincome people that cannot afford litigation.

9.4. Impact of the reforms on equal treatment, solidarity, and comprehensiveness/ sufficiency 9.4.1. Equal treatment and equity With very few exceptions, health care reforms have not standardized benefits and in most countries have maintained the previous tripartite system and separate schemes, preserving or expanding unjustified differences that erode equal treatment. The most standardized systems predate the reforms of the 1980s: Cuba is totally public and largely standardized (two separate schemes have superior benefits), Costa Rica’s social insurance is fully standardized, and Panama’s social insurance is also standardized but still there is a small public sector, in the last two countries there is a tiny private sector. Brazil, Colombia, and Chile reforms unified a large number of social insurances that had diverse benefits and established relatively homogeneous norms (under a public system in Brazil) but expanded the private sector and its inequalities and preserved some separate schemes with superior benefits, and Colombia’s two regimes have significant differences in benefits. Argentina’s reform

228 Reassembling Social Security reduced but not eliminated inequalities in benefits among OS, maintained disparities between the three sectors and created new private providers with divergent benefits. Despite the reforms, the three sectors subsist in twelve countries, with notable differences in their services (the worst probably in Bolivia). Inequalities in Argentina, Brazil, and Mexico are accentuated with the fragmentation of services between federal government, states/provinces and municipalities, and private providers. The armed forces in all countries (except in Costa Rica and Panama) have separate schemes that usually provide better care than that offered by the public and social-insurance sectors, and there are separate schemes in eighteen countries for powerful groups also with superior benefits. The analysis of WHO’s ranking based on financial equity confronted the problem of scarce, incomplete, and nonstandardized statistics and was evaluated in eight countries. Colombia’s best financial equity in the region might have been true in an early stage of its reform due to rising coverage in all income quintiles as well as increase of public expenditures and good targeting that reduced the gap between wealthy and poor municipalities, but not later when the ratio in expenses between the richest and poorest deciles rose, out-ofpocket expenses decreased in an uneven manner, and medicine outlays diminished but their prices significantly rose. Cuba’s ranking was questioned based on the lack of crucial data and huge increase in price of medicines. Uruguay’s ranking was supported by an improvement in the progressive allocation of public resources, a Gini coefficient reduction and the third lowest out-of-pocket expenses, but payments to IAMC were regressive and cost per affiliate in IAMC and private insurance were 3–4 times higher than in the public sector. Costa Rica’s ranking was confirmed by its universal coverage, free for the poor with fiscal transfers, and the poorest quintile receiving 20 times more in services than its contribution, opposite to the richest quintile. Mexico’s ranking, well above Chile, was doubtful due to its highly unequal distribution of pub-

lic health expenditure between states and the three sectors; such inequalities were reduced (after the WHO’s ranking) by the reform introduction of the SPS but still are significant among states. Chile’s low ranking was justified by fiscal subsidies received by free riders and groups with paying capacity that utilize public services, as well as private providers practicing risk selection, but contradicted by fiscal subsidies targeted on the poor and low-income affiliates that enable them to receive services at much lower costs than their payments, opposite to high-income affiliates, and outof-pocket expenses lower than in other seven countries ranked higher by WHO (inequalities have been reduced by recent laws). Brazil ranked last due to clear indicators of inequality, allegedly reduced since 1995 due to the extension of the basic package and the family health program, but the basic package stagnated, out-of-pocket expenses grew, and the poorest decile proportionally paid more and received less than the richest decile. Most countries in the latecomer-low group and the least developed in the intermediate group were ranked in the lowest places; Bolivia’s tie with Chile was contested based on data that show worsening distribution of health expenditure per capita between the three sectors. Despite improvement in national averages in health standards in most countries, available data show that significant inequalities persist in the distribution between geographic areas of physicians, hospital beds, and health indicators. Statistics from eight countries on the distribution of human and physical resources between the three health sectors indicate that the public one continues to receive a proportion of such resources considerably lower relative to the population that legally should protect, while the opposite occurs in social insurance, aggravated when there are separate schemes that get more proportional resources than the population they cover; scarce information on the private sector suggest a similar or superior situation than in social insurance. There is a close relationship between the indigenous people, its location

Equal Treatment, Solidarity, and Comprehensiveness in rural and underdeveloped areas, its poverty incidence and the worst health resources and standards: statistics from eight countries on that ethnic group demonstrate that its proportional indicators on physicians, hospital beds, institutional delivery, infant and maternal mortality, and access to potable water and sanitation are considerably worse than national averages and those of nonindigenous populations. Reforms don’t seem to have reduced these inequality gaps but data on half of the countries are needed to confirm that conclusion. Gender equity demands that health resources and services are allocated and received according to the needs of each sex and paid based on the economic capacity of the individuals and families regardless of differential risks by gender. In all countries of the region, the percentage of women affiliated to social insurance is considerably lower than the men percentage; exceptions are Costa Rica and Uruguay where the opposite is true. Gender inequities are the outcome of labor market factors but also of the health care system itself: women direct access to social insurance is lost by permanent exit from the labor force to raise children; indirect affiliation as a dependent spouse of a male insured in some countries is granted only in either maternity or in sickness and is lost by abandonment, divorce or death of the insured. Health care reforms that privatized services have aggravated gender inequalities because private providers often exclude women in fertile age due to higher-care costs or charge them superior premium; and user fees in the public sector particularly affect poor women because they use such services for them and their children more often than men. Recent laws in Chile have corrected some but not all those inequalities

9.4.2. Solidarity The principle of equivalence introduced by most reforms has partly replaced the principle of solidarity but not to the degree achieved

229

by structural pension reforms; there is no solidarity under individual equivalence (as originally in Chile), but some solidarity is infused under collective equivalence (Colombia, Chile now), albeit not at the level of a couple of integrated public or social insurance systems that provide universal comprehensive benefits. The main pro-and-con mechanisms identified in eleven countries were: high versus low population coverage; integration with similar benefits in all sectors versus segmentation with dissimilar benefits largely based on income; absence versus subsistence of separate schemes with privileged benefits; granting of a universal basic package of benefits in the entire system regardless of age, gender, risk, location and income, versus partial or differentiated packages between insured groups or sectors; compensation or solidarity funds that finance the basic package or complexhigh care reducing geographical inequalities versus absence of those funds or exemption of groups from them; similar versus disparate proportions between the population covered by each sector and the financial, human, and physical resources respectively allocated to such sectors; relative small versus important inequalities in the geographical distribution of financial, human, and physical resources; progressive fiscal transfers to the public sector versus regressive cross subsidies from the public to social insurance and private sectors and affiliates; significant versus small proportional allocation of financial resources to the first level of care vis-à-vis the other two levels; fiscal subsidies that finance free care to the poor and low-income groups and decrease as income increases versus fiscal subsidies that increase with income; solidarity contributions or retention of contributions to those who opt out to the private sector versus portability of those contributions for those who shift; fiscal incentives to promote the affiliation of low-income self-employed versus charging them the sum of the percentages paid by employers and salaried workers; practice of adverse risk selection by private insurance firms and providers versus control or banning of those practices; substantial

230 Reassembling Social Security versus small out of pocket expenses; and absence versus charge of user fees in the public sector. Lack of historical standardized statistics impedes a thorough evaluation of the impact of the reforms and equivalence on solidarity. However, data from the eleven countries indicate that mechanisms against solidarity are predominant (eight to eleven versus two to five pro solidarity mechanisms) in eight countries (Argentina, Bolivia, Brazil, Chile, Mexico, Nicaragua, Peru, and Uruguay), whereas mechanisms favorable to solidarity prevail in only two countries (Costa Rica and Cuba: eight to ten versus zero to one respectively), and there is a balance between such mechanisms in one country (Colombia: six). There seems to be a relationship between low solidarity and high degree of privatization (the latter as measured in Section 10.5), albeit anti-solidarity mechanisms are found in public and social insurance sectors. Among countries with predominant mechanisms against solidarity, Brazil, Chile, and Uruguay (if private not-for-profit providers are included) have the highest privatization (in Chile most mechanisms of solidarity are in public-social insurance), followed by Argentina and Nicaragua, but Bolivia, Mexico, and Peru have low degrees of privatization. The two countries with most solidarity have an entirely public system (Cuba) and a unified social insurance system with a relative small subcontracted private sector (Costa Rica), and Colombia have a middlelevel degree of privatization and a balance in solidarity. Finally, countries with a significant private sector have specific mechanisms against solidarity such as adverse risk selection and tax exemptions. More data and a methodology to measure the net impact of opposite mechanisms are needed to confirm our conclusions. Countries with information on solidarity are all six in the pioneer-high group, four out of the seven in the intermediate group and only one out of seven in the latecomer-low group. Although we lack data on solidarity from most of these two

groups, available statistics demonstrate that they endure the lowest population coverage/access, the worst inequalities in the distribution of financial, human, and physical resources (by income level, health sector, and geographical areas) and several of them have a significant indigenous population that suffers significant exclusion and bear the worst health resources and standards, all of which allow us to conclude that mechanisms against solidarity are predominant in that group.

9.4.3. Comprehensiveness/sufficiency The reforms introduced basic health packages for all the population or for specific groups in fifteen countries but only seven of them have fully implemented the package (the pioneerhigh group—except Uruguay, Bolivia, and Colombia); in eight countries it is limited or partial, and five countries either lack the package or have projects not implemented yet. In contrast, prior to the reform and today, social insurance normally provides integral care to all their insured beyond the basic package. Vital statistics on the package are scarce, available from few countries and largely not comparable. Most packages are restricted to health actions at the primary level; only in five countries the package includes middle- and highcomplex actions but limited either to part of the system or a few actions or a small proportion of total expenditure. Package coverage of the total population available only in three countries ranges from 11 to 33%; figures from other three countries are not comparable; the package does not reach the poorest population and rural areas in some countries, and excludes groups in others. Only three countries prevent providers to practice risk selection in granting the package. The annual per capita cost of the package in seven countries ranges from US$8 to US$150; in Colombia the subsidized-regime package is about half the level of that in the contributory regime. Incomplete data on annual adjustment of the package offered by three countries impede any

Equal Treatment, Solidarity, and Comprehensiveness conclusion. The package is free for all only in Brazil, whereas in Bolivia even the poor pays affiliation and user fees; the other countries grant the package free to the poor and charges co-payments or premium to the rest of the population according to income. The package is an important measure to extend primarylevel care to poor and vulnerable populations, and to reduce out-of-pocket expenses, but its lack of evaluation in at least six countries hinders an adequate assessment of its results and impact on health standards. In addition, the package has suffered cuts in funds or deficit; the target population lacks adequate information on the program or ignores its existence; the personnel in charge often doesn’t have the needed skills and training or is insufficient or hired temporarily or has been slow in proving the benefits; providers lack incentives to prioritize prevention and promotion, and it is difficult to determine the beneficiaries’ profiles or provision has not been adapted to important local differences. Only seven countries provide universal and full coverage of catastrophic illnesses (CI) and high-complex procedures (the pioneer-high group and Colombia); three cover them partly and ten lack it in the public sector but protect it with significant restrictions in social insurance. Even in countries that provide full benefits there are long waiting lists for complex surgery and similar procedures. Four countries have risk-pooling mechanisms like a public compensation fund or ad hoc insurance or reinsurance. The private sector offers CI coverage to a minority, charging high co-payments in Chile, and receives cross-subsidies because affiliates who lack coverage resort to the public sector; in Colombia reinsurance firms discriminate against affiliates of high risk. The public burden of CI protection (significantly expanding in Brazil and Chile) makes this program difficult to finance in countries with scarce resources. Only two advanced countries in the intermediate group provide partial CI coverage, whereas the least-developed countries in the latecomer-low group don’t cover CI in the public sector and do it in social insurance but with significant restrictions.

231

In the least-developed countries, a trade-off exists between extending coverage, particularly of the basic package, to all the population and providing costly high-complex procedures (which are usually confined to a minority protected by social insurance and some separate schemes). In order to maximize social welfare and enforce solidarity and equity, the rational alternative in those countries is to achieve basic universal coverage and, when resources are available, gradually incorporate protection of the most common and devastating CI. Only seven countries conduct periodic surveys on perceived quality or users’ satisfaction and they have diverse scope and nonstandardized questions among countries and across time, but their results confirm that the reforms have not had a significant positive impact on perceived quality and, in at least one country, might have had the opposite effect. In Chile (after 14 years of reform), the percentages that preferred the public-social insurance and the private sector were similar and, although many wanted to change to private ISAPREs, they couldn’t do it because of high co-payments; in Colombia (after ten years of reform) there were minimum levels of satisfaction with quality and possibly a decline; in Mexico (after six years of reform, but before the SPS) a slight improvement of services was perceived by only 19% of interviewees; in Nicaragua (after nine years of reform) satisfaction with social insurance-contracted providers (EMP) was 59% and coverage of some services had declined; results were contradictory in Brazil. Affiliates to social insurances in Argentina, Costa Rica, and Uruguay are satisfied with their services from 70 to 98%, while users of public services are satisfied by 64–88% in Brazil, but 45–47% in Cuba and Peru. Public services are normally used by the majority not for their quality but because they are free or very cheap, where social insurance is preferred for its coverage and solidarity, and private services are favored because they are faster and have better quality, but their high premium is a strong barrier and satisfaction decreases with income.

232 Reassembling Social Security

Notes 1. Popular health insurance (SPS) has played a role in reducing inequalities: assigning an equal per capita to poor families (who receive a free basic package), allocating more federal transfers to poor states than to richer states, establishing an interstate compensation fund, publishing SPS information in indigenous languages, and cutting the gap in the federal allocation between the ministry and the principal social insurance (IMSS), but not to other schemes that continue receiving the highest fiscal subsidies per capita (Nigenda 2005). 2. An ISAPRE affiliate in the 60–5-year-old range contributes twice as much than an affiliate in the 18–44-year-old range, and if older than 75 contributes 4.5 times more, hence there is an elderly exodus to FONASA, and as the population ages such inequity will increase. 3. Peru allocated 20% of fiscal subsidies to the poorest quintile and 22% to the wealthiest in 2001, and Lima received 52% of social insurance health expenditure (with 29% of the population) in 2003 whereas twenty departments received 29% (MINSA 2002; EsSalud 2004). 4. In Venezuela the ratio of hospital beds in social insurance ranged from 6.5 to 16.5 in the three most developed urbanized states, but there were no facilities in the six poorest states (D’ Ella 2002). 5. Chile’s maternity paid-leave was initially paid by ISAPREs, further increasing the premium of fertile-age women; later the state decided to finance it but is highly regressive because it is paid in proportion to wage, therefore half of the population gets 80% of the total fiscal maternity subsidy that is financed by all the population mostly through consumption taxes imposed even on the poor (SPS 2002; Sojo 2006). 6. Closed ISAPREs avoid gender inequality because there is solidarity within the insured group regardless of age, gender, and risk, but in the long run this will lead to an increase in the premium or cut in benefits. 7. The World Bank (1993) first recommended user fees but changed course when evidence showed they generated inequalities and regressive effects (Homedes and Ugalde 2006). 8. The original analysis of individual countries was made in Mesa-Lago (2006a) based on legislation, PAHO (2005a), and other sources: Argentina: Centrángolo and Devoto (2002), Medici (2002c), Lloyd-Sherlock (2003), Torres (2004), and SSS (2005); Bolivia: World Bank (2004); Brazil: Medici (2002a); Chile: Barrientos (2000), Titelman (2000), SI (2004), Homedes and Ugalde (2005a), and Urriola (2005); Colombia: Jaramillo (1999), Málaga et al. (2000), Yepes (2000), and Salud Colombia (2004); Costa Rica: Herrero and Durán (2001), Martínez and Mesa-Lago (2003), Martínez (2005), and Sojo (2006); Cuba: Mesa-Lago (2003b); Mexico: González Pier (2005); Nicaragua: La Forgia (2005); Peru: MINSA (2002) and EsSalud (2005); and Uruguay: Bucheli (2003), ISSA (2003a), Presidencia (2004), World Bank (2005d), and Lagomarsino (2006) (see also Sections 8.3–8.4, 9.1–9.3, and 11.1.1). 9. A compensation fund between open ISAPRES introduced in 2005 balances health risks by age and gender, neutralizing the effects of morbidity differences among affiliates respect the basic health plan: all insured within each ISAPRE contribute equally to the fund and get a benefit calculated according to their health expenditure, hence fertile-age women and the elderly receive transfers mostly from young men (Ley 20,015 2005). 10. Uruguay’s two legal drafts on tax and health reforms in 2006 reduce the VAT and make more progressive the income tax, and improve financial solidarity by progressive contributions on salary and better access and quality of services. 11. Costa Rican upper-middle- and high-income CCSS affiliates are shifting to private providers due to long waiting lists particularly for surgery, in search of faster and more personalized services, gradually leaving CCSS with low-income affiliates; cop-outs must keep paying their contribution to CCSS to maintain solidarity, but the situation creates incentives to evade or under-declare salaries (Martínez and Mesa-Lago 2003). 12. Chilean ISAPRE covered only 8% of the population 65 years and older in 2003 vis-à-vis 92% covered by public-social insurance; after retirement income falls and health expenditure

Equal Treatment, Solidarity, and Comprehensiveness

13.

14.

15.

16.

17. 18.

19.

20.

21.

22.

23.

24.

233

increases hence ISAPREs raised their premium to the elderly most of whom transferred to FONASA. In Colombia 12% of affiliates in the two wealthiest quintiles were classified as poor in 2000, whereas 11% of the poor in the subsidized regime had co-payments 2–4 times higher than in the contributory regime. Uruguay’s ministry determines the minimum benefits granted in the public sector and the IAMC, but insured in the latter with insufficient means to protect their families pick IAMC that offer fewer benefits with lower co-payments or enroll in partial insurance that does not cover minimum benefits. Excluded are medicines except for hospitalized patients, as well as ear, dental, and orthopedic prosthesis, wheelchairs, crutches, glasses, etc., which must be bought out-of-pocket mainly in hard-currency shops. This section’s general sources are legislation and Medici (2000, 2002a, 2002c); PAHO (2002b, 2005a); World Bank (2004, 2005e); La Forgia (2005); Mesa-Lago (2006a); and Sojo (2006). Additional country sources are: Argentina: Torres (2004, 2005a, 2005b), SSS (2005); Brazil Barros (2002), Biasoto (2004b), Coelho and Moreno (2004), and Piola (2005); Chile: Bitrán and Almarza (1997), Gómez (2003), and SPS (2002); Colombia: Jaramillo (1999), Málaga et al. (2000), Yepes (2000), Felizzola (2002), and Ginneken (2003); Dominican Republic: Lizardo (2004, 2005) and Rathe (2005); Ecuador: IESS (2006a, 2006b) and World Bank (2005e); Guatemala: UNDP (2003); Haiti: IADB (2003); Honduras: Durán (2003); Mexico: Nigenda (2005, 2006) and Homedes and Ugalde (2006); Nicaragua: INSS (2005a, 2005b); Panama: CSS (2004); Paraguay: Ramírez (2005b); Peru: EsSalud (2005) and MINSA (2005a); Uruguay: ISSA (2003b). Population from ECLAC (2004c, 2005c). Bolivia’s basic package coverage was reported as 52% in 2000, without specifying if relative to the total or target population, and data was considered not reliable (PAHO 2005a). In its first stage, Brazil’s PSF prioritized the extension to small municipalities with low coverage, targeting women, children, elderly, and certain diseases; in a second stage the PSF is expanding to municipalities with more than 100,000 inhabitants. The cost of the Dominican Republic package is very high and annually adjusted; if the contributory regime implements the 2005 estimated cost of US$166 when it starts functioning, it would be more than double the amount in the subsidized regime, violating the constitution. A comparative evaluation of Guatemala’s three models of basic package provision (traditional, by NGOs and mixed) done by the World Bank found ‘limitations that impeded to assert in a conclusive manner the superiority of one model over the rest’ (La Forgia 2005: 95). The package major success has been in Escuintla department due to more funds, social insurance participation, hiring of health promoters, qualified personnel, and international aid than in other places. ISAPREs offered three free services: an annual prevention exam, control of pregnancy and postdelivery, and control of the newly born, but the annual check up was not obligatory and only one-fourth of the affiliates received it; ISAPREs lacked incentives to allocate resources to prevention because its returns are in the long-run and affiliates could change firms. Nicaragua’s law of 2005 (postponed to 2007) stipulates coverage of HCP in social insurance. In Venezuela when the main social insurance (IVSS) does not have specialized care for its insured they can seek care in the private sector and their expenses reimbursed by IVSS; there are no set criteria for the election to such practice, but in 2004 only 44 out of 9 million insured benefited from it at a cost of US$3,000 each (RBV 2005). This section integrates and summarizes the separate analysis of the twenty countries done by Mesa-Lago (2006a), based on PAHO (2005a) and the following sources: Argentina: Torres (2004, 2005a) and SSS (2004); Chile: Bitrán and Almarza (1997), Titelman (2000), and Pollack (2002); Colombia: Málaga et al. (2000), Felizzola (2002), Salud Colombia (2003) No.71, and Homedes and Ugalde (2005a); Costa Rica: World Bank (2003) and Martínez (2005); Dominican Republic: Lizardo (2004); Ecuador: Roldós (2003); El Salvador: APSAL

234 Reassembling Social Security (2002) and MSPAS/GTZ (2002); Honduras: Durán (2003) and IHSS (2003); Mexico: ISSA (2003b) and Arredondo (2005); Nicaragua: La Forgia (2005); Panama: Mesa-Lago (2000b) and CSS (2004); Paraguay: IPS (2005); Peru: Manrique (1999) and MINSA (2002); Uruguay: MSP/WB (2004b) and BPS (2005b). 25. Cuba and Honduras report periodic surveys without giving results (except Cuba in 1997); two surveys were conducted in El Salvador (limited to one program), Mexico and Nicaragua; one survey in Dominican Republic, Peru and Uruguay (limited to social insurance); three surveys in Nicaragua social insurance. El Salvador and Paraguay have a web page for users to express their opinions. 26. In a 2002 survey (including the best SIBASI), leaders and management reported improvement in access and preventive and curative care, but as much as half of the members of the consulting committee comprised by community representatives rejected those claims and only a minority of users reported to receive essential services that legally should be provided by SIBASI. 27. Paraguay has no surveys but social insurance has a web page with questions that didn’t require ID registration in 2005; 77% of those who answered reported improvement in ambulatory clinics but 71% acknowledged paying bribes and 55% said to be maltreated by the personnel.

10

Effects on Unity, State Responsibility, Efficiency, Costs, Social Participation, and Reform New Goals

Most reforms pursued key changes in the state role on health systems: (a) separation of functions of direction-regulation, supervision, insurance, financing, and provision; (b) yield financing and provision functions and concentrate on technical direction, coordination, regulation, supervision, prevention, improvement of equity, targeting fiscal subsidies on the needy, and evaluation of performance; (c) support an active private participation in provision and financing; (d) grant freedom of choice to the insured for selecting the best providers and thus encourage competition and efficiency and reduce costs; (e) foster hospital autonomy allowing their operation under private law and competition to ultimately attain self-sufficiency; ( f ) decentralize the system from the state to regional and/or municipal units, which are closer to users, know their needs better and can facilitate community involvement; and (g) promote social participation and control (World Bank 1993; Medici 2000). Health experts in Latin America and elsewhere have questioned the feasibility and actual results of the reform changes: (a) the state has failed before hence it is unlikely that will be capable to regulate and supervise diverse functions of numerous insurers, providers, and financers in a more complex system than that preceding the reform; (b) the health ministry, entrusted with direction, coordination, and supervision of the system, typically lacks enough material and human resources to exercise these functions; (c) the

public sector has been reduced substantially but in many countries continues to be responsible for the poor and low-income population; (d) system segmentation is accentuated by the reform making its integration or coordination even more essential; (e) decentralization should theoretically encourage responsibility, social participation, and adequate answer to local needs, but the transfer of authority is done in practice without assigning the necessary resources to local units that also lack capable personnel; ( f ) private and/or local initiative presumably will solve the system problems but rules for the transfer of responsibility and control of the decentralized units are weak and unclear; (g) the private sector has become one of the most important insurers and/or providers in some countries and hence empowered to resist regulation; (h) hospitals in most countries still are financed by central budgets or receive fiscal subsidies, without true competition between them; (i) some reforms neither have improved equity nor efficiency; and ( j) most reforms were implemented from the top down, with little or nil social participation in their design, execution, and supervision (Giordano and Colina 2000; Lloyd-Sherlock 2000a, 2003; Londoño and Frenk 2000; Homedes and Ugalde 2002, 2005a). This chapter analyzes the changes introduced by the reform and their effects on ten issues: (a) unity or integration of the system, (b) separation of functions, (c) state regulation and supervision, (d) decentralization,

236 Reassembling Social Security (e) competition, ( f ) privatization, (g) freedom of choice, (h) efficiency, (i) administrative costs, and ( j) social participation. Five are conventional social security principles (a, c, and h–j) and the reform impact on them is assessed, whereas five are reform goals (b and d–g) and we test if they have materialized.

10.1. Unity or integration of the system The degree of unity/integration or at least coordination of the health system in the twenty countries is ranked from higher to lower based on Table 10.1: (a) totally unified public with neither social insurance nor private sector in Cuba; (b) totally unified social insurance coordinated with a small private sector (public sector limited to policyregulation) in Costa Rica; (c) segmented with high coordination between public-social insurance and large private sector in Chile; (d) unified social insurance, functionally integrated with the public sector, and small private sector in Panama; (e) segmented but coordinated between contributory and subsidized regimes, with remnants of public sector in Colombia; ( f ) segmented with some coordination between a public system at three geographic levels and large private sector (no social insurance) in Brazil; (g) highly segmented with low coordination between public sector at three geographical levels, multiple social insurances, and diverse private providers in Argentina; (h) segmented without coordination of public and private sectors (virtually no social insurance) in Haiti; (i) segmented without coordination between the three sectors in Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, and Peru (in the last three, social insurance services are totally or partially channeled through external providers); and ( j) highly segmented without coordination between the three sectors in Bolivia, Dominican Republic (not-fully-enforced law orders coordination between three new regimes), Mexico,

Paraguay, Uruguay, and Venezuela (legal draft pending approval in 2006 stipulates full integration). In summary, fifteen systems are segmented or highly segmented (thirteen without coordination and two with low or some coordination), two are functionally integrated or highly coordinated, and only two are totally unified.1

10.1.1. Integration and coordination in public sector and social insurance Five reforms promoted integration and coordination, three of them before the reforms initiated by Chile in 1980 and spread in the 1990s. Cuba collectivized all private providers, cooperatives, and mutual-aid societies in the 1960s and integrated them, together with public facilities, into a national public system (SNS), banning private medicine. Costa Rica unified social insurance (CCSS) essential curative services in the 1970s, leaving the ministry with direction-regulation and prevention-promotion functions; the latter were transferred to the CCSS by the 1990s reform. Panama unified all social insurances in the 1970s and began a process of integration with the public sector that was later halted. Chile’s system was fragmented into thirtyfive autonomous social insurances (bluecollars including self-employed and poor, white collars, armed forces, policemen, main university, oil corporation, banking, etc.), the private sector was a marginal provider and insurer; the reform unified all social insurances in 1979, except the armed forces and police, into public-social insurance, and in 1981 created private insurance firms (ISAPRE), leading to a more integrated system but still segmented albeit well coordinated. Colombia endured the widest multiplicity: a public sector fragmented into nation, departments, and municipalities; more than 1,000 social insurances: private-sector workers (ISS), civil servants (Cajanal), communications (Caprecom),

Effects on Unity, State Responsibility, Efficiency

237

Table 10.1. Degree of unity/coordination or segmentation, number of sectors and separate health-care programs, c.2005

Countries

Degree of unity/coordination or segmentation

Argentina

Highly segmented with low coordination

Bolivia

Highly segmented without coordination

Brazil

Segmented with some coordination

Chile

Segmented with high coordination Segmented but well coordinated

Colombia

Costa Rica

Unifieda

Cuba Unified Dominican R. Highly segmented without coordination (reform not fully in force)

Ecuador

Segmented without coordination

El Salvador

Segmented without coordination Segmented without coordination Segmented without coordination Segmented without coordination

Guatemala Haiti Honduras

Number of sectors and division within each 3: Public (provinces, municipalities), social insurance and private 3: Public (provinces, departments, municipalities), social insurance (various) and private 2: Public (federal, states, municipalities) and private 2: Public-social insurance and private 4: Contributory, subsidized, private and ‘tied’ to public 2. Social insurance and private (small but growing) 1: Public 3. Current: public, social insurance and private. Under the Reform: subsidized, contributory and contributorysubsidized 3. Public, social insurance (peasants separated; external providers) and private 3. Public, social insurance and private 3: Public, social insurance and private 2: Public and private 3: Public, social insurance (part by private and public) and private (small)

Separate programs Armed forces

Others

Yes

Congress, judiciary, national universities

Yes

14 health social insurances

Yes

Yes

Civil servants have private plans, policemen Policemen

Yes

Teachers, oil workers

No

No

Yesc Yes

Tourists, top leaders Various groups of civil servants, teachers, physicians, etc.

Yes

Policemen

Yes

Teachers

Yes

Policemen, ministry of interior Civil servants have insurance plans National university

n.a. Yes

(cont.)

238 Reassembling Social Security

Table 10.1. (Countinued)

Countries Mexico

Degree of unity/coordination or segmentation Highly segmented without coordination

Nicaragua Segmented without coordination Panama

Functionally Integrated

Paraguay

Highly segmented without coordination Segmented without coordination Highly segmented without coordination

Peru Uruguay

Number of sectors and division within each Public (several), social insurance (five) and private 3: Public, social insurance (through prepaid) and private 2. Social insurance and public (private very small) 3. Public, social insurance (various) and private 3. Public, social insurance (various) and private 3. Public, social insurance (various) and private

3: Public, social insurance Venezuela Highly segmented and private without coordinationb

Separate programs Armed forces

Others

Yes (two) Federal-state civil servants, oil workers Yes

No

Yes Yes Yes

Yes

Ministry of Interior (employees, policemen) No

Civil servants,d policemen and others Policemen Civil servants, port, policemen, notary public, banking Policemen, teachers, universities, oil workersd

Sources: General from legislation; La Forgia (2005); PAHO (2005a); Sojo (2006); see also text. a

Some segmentation has been created and expanded with prívate provision.

b

The legal draft of 2005 (not yet approved in 2006) integrates social insurance and separate groups (except armed forces) to the national public health system under the coordination of the ministry.

c

Includes internal security personnel and the top political leadership.

d

Have a right to private health insurance with fiscal subsidy.

oil workers, armed forces, public teachers, congressmen, etc.; and private prepaid plans, facilities, and professionals. The 1994 reform replaced the public sector and social insurances with two new regimes, coordinated but without integration, and left out important separate schemes and part of the population tied to public hospitals. Segmentation in the public sector is compounded in countries with federal systems. Brazil had six separate sectors-schemes: public federal, state and municipal, social insurance, civil servants and armed forces; in 1993 all

social insurance health facilities, personnel, and users were transferred to and integrated with those of the public sector into a national public system (SUS) but fragmented into federal, state, and municipal levels. The federal ministry exerts some coordination in negotiation with states and municipalities (through a national health council), designs policies, provides most public financing, and supervises and evaluates the system; two public agencies regulate and supervise the private sector. The states also have functions of regulation, financing (part by federal transfers and their

Effects on Unity, State Responsibility, Efficiency own), provision and coordination of highcomplex actions; municipalities get federal and state funds according to their provision responsibilities at the local level, specially the basic package and the family health program (PSF). Little or no progress toward integration or high coordination has been achieved by other reforms. Argentina’s was not comprehensive but divided into public sector and social insurances (OS), increasing previous segmentation between them and within each, and lacking coordination. The central ministry legally has the direction and assists the government in health policies, but provincial ministries enjoy total autonomy in policy and are the providers of services (most central-government hospitals were transferred to the provinces), functions also performed by large municipalities. The national health council comprises the health ministries of all provinces and the capital but has limited coordinating power; the health superintendence approves, regulates and oversees OS and enforces the basic package, but excludes public provincial and municipal services. Mexico’s project in the 1980s to create a national health system did not materialize and currently there is high segmentation among the three sectors and within each of them. Five separate social insurances continue, the principal (IMSS) has separate programs for peasants (IMSS-Oportunidades)2 and family health insurance (SSF). The federal health secretariat (SSA) exerts policy, direction-regulation and supervision but with limited power over the states that have their own services (as well as the federal district), and provides direct care at the three levels. The ministry (alone or jointly with social insurance and the states) has developed several programs targeted on the poor, including popular health insurance (SPS); each administration usually creates a new program, some have continued after the presidential term ends (adding to the existing mosaic) whereas others have disappeared. A national health council weakly coordinates relations between the federal government, states, and large municipalities.

239

Despite its small size, Uruguay has one of the most segmented systems without coordination in the region. In the public sector the ministry ineffectively exerts policy-regulation and supervision and also provides services in all departments; separate facilities are operated by the university and several separate schemes, as well as municipal primary services. The principal social insurance (BPS) has its own facilities for maternity care in the capital but contracts with private not-for-profit collective providers (IAMC) for sickness care at the national level and for maternity care in the interior. Nicaragua’s social insurance health facilities were transferred to the ministry in 1979 that managed the unified health system (SNUS) for the insured and the uninsured population; in the 1980s, SNUS suffered a severe deterioration due to civil war, economic crisis, coverage extension without suitable infrastructure, administrative flaws, and widespread noncompliance; in 1992 all former facilities of social insurance were returned and INSS became independent; the SNUS was terminated in 2003 but a public sector subsists. Despite the health law of 2002, the system was still segmented and without coordination between the three sectors in 2006. In Bolivia after the creation of unified social insurance, fourteen autonomous schemes were approved against the social security code mandate for unity and the obligation of the national health institute to coordinate all schemes. Paraguay’s reform stipulated the creation of a national system coordinating the three sectors and eliminating duplication of services, but by 2006 the high segmentation without coordination continued and the three sectors had scarce relations among them. Peru’s three sectors continue as before, afflicted by many problems: they are watertight compartments without an institutional framework to coordinate inter-sector relations, segmentation between providers prevents the articulation of their services and creates overlapping, and health actions are fragmented into thirteen programs some of them paralleling external-cooperation

240 Reassembling Social Security programs. Guatemala’s reform pursued coordination between the ministry and social insurance; in 1999 both agreed to harmonize basic-package provision in one department but that model has not been extended to other departments. Honduras’ system is ‘fragmented and disintegrated with duplication of efforts, actions and expenditures . . . each of the three sectors has its own laws, without coordination, that blocks any integration process’ and few joint efforts exist between the sectors (PAHO 2005a: 22). Reform laws or legal drafts in three countries aim at integration but had not been implemented in 2006. Dominican Republic’s three sectors (including several social insurances) basically continue with very few relations between them; the 2001 reform law stipulated the creation of three regimes: contributory, contributory-subsidized, and subsidized but only the latter is partly operational. Ecuador’s constitutional mandate for a national health system was embodied in the reform law of 2002 that entrusted the ministry with that task, but is still without regulations. Venezuela’s constitution of 1999 and the social security organic law of 2002 mandated an integrated and coordinated system (SPNS) and the creation of three institutions: a national public health system to merge all services receiving fiscal funds, the superintendence of social security and the social security treasury; the legal draft of 2005 stipulates the integration of the principal social insurance (IVSS), separate schemes and public facilities into the SPNS that should cover the entire population at all levels of care; none of three institutions have been established, and both the regulations of the organic law and the legal draft of 2005 are still pending approval. Reforms failed to integrate most separate schemes in social insurance and in Brazil and Cuba public sectors (see Table 10.1): armed forces in all countries (except Costa Rica and Panama); policemen in nine countries; civil servants and teachers-universities in eight countries each, oil workers in three, and other groups in six countries. Argentina has schemes for congress and judiciary; Bolivia still has

fourteen schemes; Colombian schemes were given two years by the reform to transform into insurance firms or dissolve but subsisted in 2006; Cuba has separate facilities for internal-security personnel and political leaders, and for ‘health-care tourists’ paid in hard currency; Dominican Republic has a scheme for physicians; Paraguay principal social insurance has special regimes for domestic servants, teachers and war veterans; Uruguay has schemes for port public enterprises, banking and notary public.

10.1.2. Multiplicity in the private sector The region has a significant variety of private entities and plans, strengthened by most reforms by delegation of social insurance services (public in Brazil). Predominant are profit seekers such as insurance firms and providers, prepaid plans and independent hospitals, clinics and health professionals (either individually or associated or in cooperatives), concentrated in the capital city and most urbanized areas but much less in the countryside. For-profit insurance firms are important in Chile and Colombia (in Dominican Republic law not fully operational); but only a few in El Salvador and Honduras. Private providers that totally or partially deliver services to social-insurance affiliates exist in several countries. In Nicaragua the INSS virtually delegates all services to private or public providers albeit not authorized by the social security law and considered illegal. Peru’s social insurance health program was transferred to an autonomous entity (EsSalud) that kept its own facilities but the insured collectively can change to private, public, or mixed providers that operate with their own infrastructure or contract external services. Costa Rica’s reform introduced and promoted for-profit providers that contract with the CCSS; primary-care contracts were reportedly eliminated in 2004–6 and restricted to a few

Effects on Unity, State Responsibility, Efficiency high-tech services. Uruguay’s sickness insurance is partly provided by private for-profit partial and emergency insurances. Honduras’ social insurance has organized local entities that deliver primary care but delegate highercare levels to the private sector. Prepaid medicine is significant in Argentina (they contract with OS but also have individual affiliates); Brazil (supplementary to public care and with the largest coverage among health group private plans); Dominican Republic (igualas to whom half of affiliates in social insurance are members); Ecuador (also duplicate coverage with social insurance); Guatemala, Mexico, Nicaragua, Paraguay, and Venezuela. Insurance policies are common in Brazil, Mexico, Peru, and Venezuela. Large enterprises offer medical plans for their workers in Brazil, Mexico, Uruguay, and Venezuela, and there are enterprise physicians in Costa Rica and Honduras. Peruvian employers can affiliate their workers in social insurance or in providers or organize their own health services that typically provide first level care with higher-level care delivered by social insurance. Independent private hospitals, clinics, doctor and dental offices, laboratories, pharmaceutical enterprises, high-complex equipment, and so forth are found in all countries. For-profit traditional medicine, used by poor, rural, and indigenous populations, is important in at least six countries of the Andes and Central America (see Section 9.1.3; on privatization see Section 10.5). Not-for-profit providers are predominant in rural and marginal-urban areas and among the poor. Most common are NGOs, important in Dominican Republic (200 offer primary care), El Salvador, Guatemala (600 deliver primary care and 100 also the basic package through contracts with the ministry), Haiti and Nicaragua. Others are charities, foundations, religious organizations, mutual-aid societies, cooperatives, and the Red Cross. Costa Rica has innovated in this area with three programs associated with social insurance that existed prior to the reform and have been expanded thereafter: mixed medicine where

241

the insured select and pay a private physician for external consultation and the CCSS provides laboratory exams and medicines; enterprise medicine where the employer hires and pays a physician for external consultation of his workers and the CCSS offers diagnosis and medicines; and three cooperatives organized by health professionals that receive financial support from the ministry and the CCSS in exchange for care of an assigned population. Uruguay IAMC, mostly private but also public or quasi-public, are providers through contracts with social insurance, whereas highcomplex care is delivered by providers (IMAE), mainly private but also public.

10.2. Reform goal: separation of functions An important reform goal is the separation of functions: the public sector concentrates policy, direction-regulation and supervision, insurance could be split, and finance and provision transferred to social insurance and mainly to private providers. Table 10.2 summarizes the state of separation of functions in the twenty countries c.2005 showing that progress has been very modest. Only two countries have separated the functions entirely or very highly: Colombia and Chile respectively (Dominican Republic in the not fully-implemented law). Four countries have separated functions partially: Costa Rica, Nicaragua, Peru, and Uruguay. Six countries have either low separation (Argentina, El Salvador, and Guatemala) or very low separation (Bolivia, Brazil, and Honduras). In seven countries there is no separation at all as each of the three sectors combines functions of insurance, financing, and provision: Cuba, Ecuador, Haiti, Mexico, Panama, Paraguay, and Venezuela (the legal draft pending approval in 2006 integrates the system but does not separate its functions). Colombia’s reform achieved total separation of functions: The national council of social security in health, subordinated to the

Table 10.2. Separation of functions in health-care systems, c.2005 Separation of functions

Countries

Overall Degree

Direction/ Regulation

Argentina

Low

Bolivia Brazil

Very Low Very Low

Chile

Almost Total

Ministry: nation & provinces (weak) Ministry (weak) Ministry: federal & states (weak) Ministry

Colombia

Total

Ministry, CNSSS

Costa Rica

Partial

Ministry (weak)

Cuba Dominican Republic

No Total (but in the reform)a

Ministry, CNS (weak)

Ecuador

No

Ministry (weak)b

El Salvador Guatemala

Low Low

Ministry (weak) Ministry (weak)

Haiti Honduras Mexico

No Very Low No

Ministry (weak) Ministry (weak) Ministry (weak)

Nicaragua

Partial

Ministry (weak)

Financing

Insurance

Provision

Public: Nation, Fund (FSR), provinces, some municipalities. Social insurance: OS. Private: EMP, etc. Public (ministry), social insurance, and private Public (Federal, states, municipalities) and private; basic package by multiple providers Public-social insurance (FONASA) and Public-soc. ins (SNSS) Private (ISAPREs) & private (ISAPREs) Fund (FOSYGA) Insurance Firms EPS, Providers IPS, ESS, ARS ESE Social insurance (internal division) Social insurance and private (growing) State TSS and SNS Insurance Firms ARS Providers PSS

Public (ministry), social insurance (direct provision and few external providers) and private Public (ministryc ), social insurance, and private Public (ministry), social insurance, and private; separation of three functions in the basic package Public (ministry) and privated Public (ministry), social insurance, and private Public (federal, states,e municipalities), social insurances, and private Public (ministry), social insurance and private

Public, EMP, private

Supervision Superintendence (only OS) Ministry (weak) Divided between federal units (weak)g Superintendence (all the system) Superintendence (all the system) Ministry (weak)h

Superintendence (all the system, recently created) Ministry (weak) Auditing Court Ministry (very weak) Ministry (very weak)d Ministry (weak)i Ministry (weak and partial)e Ministry (weak)j

Social insurance; public (ministry)f Public (ministry), social insurance, and private

Panama Paraguay

No No

Ministry (weak) Ministry, CNS (weak)

Peru

Partial

Ministry (weak)

Public (ministry); social insurance (EsSalud), and private

Uruguay

Partial

Ministry (weak)

Venezuela

No

Ministry (weak)

Public (ministry); Fund (FNR); social insurance Public: ASSEBPS, (BPS), and private IMAC, IMAE Public (ministry), social insurance (various) and private (the legal draft integrates the first two in the SPNS, but without separating their functions)

Public, EsSalud, EPS, private

Ministry Superintendence not autonomous (weak) Ministry, Superintendence (weak)k Ministry (weak) Superintendence of social security (not implemented)l

Sources: Same as in Table 10.1. a Only one of three reform regimes had been implemented in 2006; collection of contributions will be by the Social Security Treasury (TSS) and financial allocations by the TSS and the National Health Insurance (SNS). b The 2002 health law not implemented by mid-2006 strengthens the ministry but does not separate system functions. c The finance ministry allocates budget funds to the health ministry and financially autonomous hospitals. d The 2004 basic package strategy separates its functions: direction and financing to the ministry, provision to district private and public entities, and financial supervision to the Comptrollers. e Some separation of functions between the federal government and the states; concerning supervision, the states are independent; social insurances are autonomous; the private sector is neither regulated nor effectively supervised by the ministry, and private insurances have a separate supervisor. f In one integrated hospital, insurance and financing (shared by public and social insurance sectors) are separated from provision through contracts with private providers. g A National Council and a National Agency of Supplementary Health Care respectively regulate and supervise the private sector. h The General Comptroller audits health expenditures, but social insurance has its own auditing services. i The ministry supervises social insurance, but not private insurances that are regulated-supervised by the National Commission of Bank and Insurances. j The law of 2005 (postponed to 2007), creates the superintendence of social security to oversee all social insurances. k The ministry supervises the public sector (not social insurance), the superintendence of EPS oversees private providers, and the superintendence of bank and insurances watch over private insurances. l The legal draft of 2005 (not approved in 2006) entrusts the social security superintendence with overseeing the system, but grants other supervising functions to the General Comptroller and a new auditing office in the ministry; the 2002 social security law not regulated in 2006 creates a superintendence of insurances to watch over private health insurances and plans.

244 Reassembling Social Security ministry, leads the system, defines the two basic packages, determines the capitation unit to finance services and manages the solidarity fund. The ministry is responsible for policy, technical direction, general norms, and evaluation. The superintendence regulates and oversees finances of insurance firms/providers, ensures quality of services and sets the framework for the transformation of public hospitals into autonomous entities. The solidarity fund (FOSYGA) assigns capitation units to insurance firms, receives their surpluses and covers their deficits. Insurance firms are of two types: EPS can operate in the contributory and subsidized regimes whereas ARS are limited to the subsidized regime; both types affiliate, collect contributions, manage finances, receive the capitation unit, transfer surpluses and get subsidies, and contract providers. Providers are of three types: IPS in the contributory regime, ESS in the subsidized regime, and ESE that are the former public hospitals becoming decentralized, autonomous, and self-sufficient (half of them continues to be dependent on the state and receives fiscal subsidies); IPS are contracted by EPS, whereas ESS are contracted by ARS, and both insurance firms and providers contract ESE. Chile’s initial reform, modified by subsequent legislation, has achieved very high separation of functions. The ministry is responsible for policy design, coordination, directionregulation and evaluation, standardization of the basic package (AUGE), and accreditation of its providers. The health superintendence, dependent on the ministry, oversees the whole system. Insurance, financing, and provision are split into the two sectors. In public-social insurance, that covers most of the population, the insurer is the public national health fund (FONASA), contributions are collected by the institute that manages the closed public pensions (INP) and by a private firm (Previred) and transferred to FONASA, which finances public goods and the three health-care levels, and assigns subsidies to the poor and low-income population; the provider (the most important in the country) is the national system of health services

(SNSS), under the responsibility of the ministry but decentralized in regional services for second and third level, and municipalities for the first level, affiliates also have an option to choose private services. In the private sector, ISAPREs insure, receive contributions from Previred, collect premium and co-payments, and finance services (a compensation fund among all open ISAPREs neutralizes differences resulting from divergent age and gender risks); the large majority of ISAPREs contracts services with private providers, a minority has its own infrastructure. Under the Dominican Republic law the ministry exerts direction whereas policy advice is by the national health council; the superintendence regulates and oversees the system, including the basic package by the family health program (SFS); insurance firms (ARS) are mandatory for all affiliates in the contributory-subsidized and subsidized regimes as well as for civil servants in the contributory regime, one ARS is public (SENASA), another is by social insurance and the rest are private; provision is to be done by providers (PSS) of multiple nature; the social security treasury collects all contributions and transfers funds to insurance firms. In 2006 only the subsidized regime was operational in some regions, the superintendence, the treasury, the council, and SENASA had been created and the latter started to provide services. Partial separation of functions varies in four countries: the ministry is in charge of policy and regulation in all of them whereas supervision can be by the ministry or another agency; functions of insurance, financing, and/or provision are partly separated. Costa Rica’s ministry exerts overall supervision and the general comptroller oversees expenditures; the three CCSS functions are divided into internal administrative units: insurance, financing, and buying of services (mostly from CCSS own facilities but also from private and mixed providers). Nicaragua’s ministry supervises the entire system, finances public facilities, and provides services; the INSS has insurance and financing functions but buy services

Effects on Unity, State Responsibility, Efficiency from providers (EMP); the law of 2005 postponed to 2007 stipulates a superintendence of social security to oversee INSS and regulate EMP. Peru’s ministry supervises the public sector and provides the basic package; social insurance (EsSalud) handles financing whereas insurance and provision are through its own services or by external providers (EPS), and supervision is divided into two entities (for providers and private insurance). Uruguay’s ministry performs supervision and insurance, whereas provision is by a ministry agency (ASSE) through its own facilities but can transfer some services to municipal ambulatories and private providers; social insurance (BPS) is insurer and financier in its sickness program and IAMC are providers, whereas in its maternity program the BPS is also the provider; the public national resource fund (FNR) is the insurer and financier of highcomplex care but buys all services from mixed providers (IMAE). In six countries policy design and direction are by the ministry, as well as regulation and supervision in most, whereas each of the three sectors combines the functions of insurance, financing, and provision, albeit with certain separation in provision (some delegation to external providers or in the basic package); the distinction between low and very low separation is very tenuous in these countries. Argentinean provinces and some large municipalities enjoy full autonomy in policy and provision; regulation, and supervision of OS is by a ministry agency; financing is by each sector but the solidarity redistribution fund (FSR) guarantees financing for the basic package; and provision is also by each sector: provinces and municipalities, OS, private prepaid providers (EMP), etc. El Salvador’s supervision is exerted by the auditing court; provision at the first and second level is partly done by integrated local units (SIBASI) that combine the ministry, social insurance, NGOs, private enterprises, and traditional healers. In Guatemala only the basic package separates regulation-supervision (ministry), financing (state), and provision by three providers: the ministry with its own

245

services, direct providers (PSS), and administrators (ASS) that pay the personnel, buy inputs and finance ministry providers; PSS and ASS are NGOs and to less extend social insurance, cooperatives, municipalities, and churches. Bolivia’s basic package is offered by multiple providers: public, social insurances, NGOs, churches, and private for-profit; a national solidarity fund should assign funds to municipalities with deficit to meet the demand for services. Brazil’s federal ministry has coordination and supervision functions, and with the states shares regulation and financing (there is a regional compensation fund: FAEC); provision is by states and municipalities with supplementary private plans. Honduras’ social insurance delegates part of provision to public municipal or private providers. Finally in seven countries the ministry concentrates the functions of policy design, direction-regulation and supervision (the latter with one exception), and each of the three sectors combines the remaining three functions with minute delegation of provision in a couple. Ecuador’s health reform law of 2002 (not implemented by 2006) and social security law of 2001 (partially declared unconstitutional) entrust the ministry with design and overseeing the basic package; social insurance buys services through contracts from its own facilities and a few external private providers (not in force in 2006). Mexico’s ministry has the supervision but the Comptroller oversees public expenditures. In Panama only three health regions and one hospital have separated financing (jointly by the ministry and social insurance) and provision through contracts with external providers; the suspended law of 2005 stipulated some separation of functions. Paraguay’s superintendence oversees public and private insurers. Haiti’s ministry concentrates financing (partly subsidizing NGO and other providers) and provision; a new strategy (not yet implemented) keeps financing in the ministry while delegating basic-package provision to municipal districts through contracts with external providers. Venezuela’s ministry has provision

246 Reassembling Social Security at the central level and through states and municipalities; the three levels may sign provision agreements with social insurance, and the organic law authorizes delegation of collection and investment to the private sector; under the legal draft pending approval in 2006 all functions are concentrated into the integrated public system (SPNS) through central or autonomous bodies (treasury, ministry of finance, health agency, etc.), and the supervision is by four entities whose respective functions are not specified. Cuba’s ministry exerts all functions except financing done by the state central budget. Usually countries with high or very high segmentation (lacking integration and coordination) don’t have separation of functions either, but integration and separation of functions not always go in tandem and may be opposite. For instance, Cuba has a totally integrated system but without separation of functions at all; Costa Rica also have an integrated system with separated functions of direction and provision (partly) but not separated in insurance and financing; and Colombia has a coordinated system without integration but complete separation of functions.

10.3. Role of the state: regulation and supervision A decentralized market with multiple providers engaged in competition generates flaws that require strong state regulation, but this task is complex and the reformers took for granted that governments had the required capacity. Nevertheless, before the reform in most of Latin America, the health ministry was weak and incapable to regulate the segmented systems, particularly social insurance that was uncoordinated with the public sector, and to execute the law and exercise an effective supervision over the private system (Mesa-Lago 1992). For example, Colombia’s health superintendence created years before the reform was

notoriously ineffective in coordinating and regulating public and social insurance services. Mexican health authorities lacked a culture of evaluation and monitoring, as well as accountability tools, problems that became accentuated with the decentralization (Arredondo 2005). Domestic and foreign experts warned in Argentina that free market forces combined with lack of proper regulation of the private sector, as well as ineffective coordination and enforcement of norms by the superintendence, expanded the for-profit private providers space of action and weakened equity and solidarity, without ensuring that potential efficiency gains would lead to better coverage and quality of services (Flood 1997; Tafani 1997; Centrángolo and Devoto 2002; Torres 2005b). Reformers ignored the serious difficulties to regulate the health market faced even in industrialized nations with vast experience and strong agencies, like the United States (Homedes and Ugalde 2005a). PAHO’s regional evaluation of the reforms found that only five countries had created new regulatory structures/institutions, and even those had not fortified accountability or restricted it to financial auditing (Infante, de la Mata, and López-Acuña 2000). The World Bank (1993) recommended strengthening the government capacity to regulate and control the entire health system, including the private sector, to guarantee security and quality of services. This section examines the current regulatory and supervisory powers in the region showing that the reform has achieved little progress in creating a single agency (ministry or superintendence) effectively empowered with those functions over the entire system and that in most countries both social insurance and the private sector are unregulated or weekly regulated and overseen by a separate agency if at all.3 In nineteen countries the function of regulation is entrusted to the ministry but is weak in all but two. In half of the countries the ministry neither can regulate nor has control over autonomous social insurance and the private sector, a situation aggravated in countries

Effects on Unity, State Responsibility, Efficiency that have several social insurances like Bolivia and Mexico. The Cuban state regulates and supervises the public system and there is no private sector. In virtually all countries, the armed forces, the police, and other separate schemes are normally independent from the ministry regulatory power. In Argentina, Brazil, and Mexico the regulatory power of the federal/central ministry is diluted because it ought to be shared with states/provinces. Argentina’s central ministry (through the superintendence) regulates OS and mandate prepaid providers (EMP) to deliver the same mandatory benefits than the OS but many norms are essentially indicative, and public state and provincial services are excluded. Brazilian federal government negotiates public-sector regulations with states and municipalities; it also establishes rules for private individual plans and insurance (through the national council of supplementary health) but excluding health group enterprises that are the majority and with inadequate evaluation and monitoring of private insurance. Mexico federal ministry is even weaker because it excludes the five social insurances, has limited power over states, the private sector is not regulated by the general health law, commercial insurance is regulated and supervised by a national commission of insurance and bail, and prepaid plans have separate rules. The function of supervision is exercised by the ministry in ten countries and in all is weak or very weak. Brazil’s federal government shares it with the states and a federal agency of supplementary health care (ANSS) supervises the private sector; Honduras’ private sector has its own law and is supervised by the national commission of banking and insurance; Peru’s ministry jurisdiction is circumscribed to the public sector, lacks power over social insurance that has its own internal auditing, and over providers (EPS) that are overseen by a separate superintendence, and has scarce control over the private sector watched by the superintendence of banking and insurance. Only five countries have a health superintendence often autonomous

247

but with diverse scope: in Colombia and Chile it oversees all the system; in Argentina it watches over OS (excluded are provinces and municipalities, prepaid EMP supervised by the ministry of economics, and other private providers); in Paraguay it is not autonomous and very weak to supervise the private sector; and in Peru it is restricted to providers (the ministry supervises the public sector). El Salvador’s supervision of the entire system is done by the auditing court. Dominican Republic reform law envisages a superintendence with regulatory-supervisory powers over all the system but incipient and weak. The cases of five countries (three in the pioneer-high group, and one each in the intermediate and latecomer-low group) are discussed in more detail below: Chile successfully fortified the regulation-supervision functions, Costa Rica’s ministry regulationsupervision is weak, Uruguay lacks a single regulatory-supervisory agency, Venezuela not yet enforced law contemplates four different supervisory agencies, and Nicaragua social insurance and private providers require proper regulation-supervision. Chile’s private sector virtually lacked regulation in the 1980s, which led to ISAPRE abuses, risk-selection, and cream-skimming. In 1990–2004 three laws enacted by democratic governments strengthened the ministry direction and normative powers, created the superintendence4 and regulated ISAPREs. The superintendence controls the entire health system (FONASA, ISAPREs, and other private providers), keeps national registries of all providers, oversees fulfillment of norms on health contracts, their exclusions and treatment of preexisting diseases, fixes premiums for elderly affiliates, guarantees minimum benefits in the entire system (including ISAPREs), standardizes ISAPRE information provided to users, publicizes a chart comparing different health plans, arbitrates conflicts between ISAPREs and their affiliates, and imposes sanctions on transgressors. Costa Rica’s law charges the ministry with the system regulation, a weak function for

248 Reassembling Social Security several reasons: cannot impose norms on social insurance (CCSS) or even agree with it to strengthen the evaluation of results; private insurance, and prepaid plans are not regulated and accreditation is voluntary thus exposing the population to significant risks (a prepaid provider bankruptcy left 15,000 affiliates without services in 2000); the information gathered is fragmented and incomplete; part of the ministry personnel has not learned the new functions of direction, regulation and monitoring; and decentralization and delegation of services make the ministry vulnerable to pressures from private providers. The CCSS regulates and audits its direct and contracted services with private providers but rules are unclear and imprecise; nevertheless, private providers must deposit a guarantee to meet at least 85% of CCSS goals, and transgressors are charged up to 5% of the value of their contracts. The general comptroller watches over health expenditures in the entire system. Uruguay lacks a single agency in charge of adequate regulation and supervision of the entire system, the ministry exerts general overseeing but the public provider (ASSE) lacks an explicit regulatory framework for the operation, structure, and quality of its services; social insurances are independent entities; the IAMC are also autonomous but the law stipulates the minimum requirements for their operation, regulates their benefits and investment, and the state sets a ceiling to their premium (but with poor effective control on the quality of their services and evaluation criteria); the ministries of health and finances have supervision over the IAMC but not integrated; partial insurances have a separate, less strict regulation. Venezuela’s legal draft not yet approved in 2006 makes the ministry responsible for direction-regulation of the future integrated public system, under the supervision of four entities, which augurs duplication and confusion: the superintendence of social security over nonfiscal funds; the general comptroller over fiscal resources; the health ministry over health facilities, industries, equipment, and

goods; and the superintendence over funds for health education in insurance companies. Nicaragua law of 2002–3 entrusted the ministry with the regulation of the system, but leaving out social insurance (INSS) that must be regulated by the law not yet enforced in 2006; the ministry should also audit the private sector (including NGOs) and the armed forces and the police but lacks information to make that task effective. The INSS neither demands a report on providers (EMP) services nor collects the necessary data from them on the use of contracted services, payment mechanisms, and the sums paid. The law of 2005 postponed until 2007 stipulates the creation of a superintendence to supervise the INSS and EMP and regulate the latter.

10.4. Reform goal: decentralization There is no clear agreement in the literature on the meaning of decentralization because the term has different connotations and degrees, involving diverse management, functions and geographical levels. Decentralization could be done exclusively with public entities, a mix of public and private, or only private either forprofit or not-for-profit. Functions susceptible of decentralization are decision-making, management, financing, human resources, provision, maintenance, buying of inputs, and so forth, and each one involves various activities. Geographical units of decentralization could be intermediate (regions, states, provinces, and departments) and local (municipalities, districts, communities, and neighborhoods); each unit may embrace various levels of care or types of decentralized facilities, for instance only primary level, basic package, hospitals, etc. Before deciding to decentralize, therefore, its management, functions, geographical units, and care levels should be precisely determined. Long before the reform there were several processes of decentralization in the region (Colombia and Mexico) not always successful.

Effects on Unity, State Responsibility, Efficiency The World Bank (1993) strongly supported decentralization and hospital autonomy and self-sufficiency, arguing that they have positive effects vis-à-vis centralized agencies: those who make decisions at the local level know better the needs of their communities than central bureaucrats (top-down-decisionmakers) hence can adapt services to divergent needs (epidemiological, demographic, socioeconomic, etc.); community participation monitors quantity and quality of services, demands accountability from health workers, facilitates cultural acceptance of services, and improves democracy; decisions are simplified, the infrastructure and resources are used more efficiently, responses to urgent needs are more agile and services are delivered faster; costs are reduced, and efficiency, quality, and users’ satisfaction are enhanced (for the evaluation of actual results see Section 10.4.3).

10.4.1. Degree of decentralization achieved There is a semantic discussion on various types of decentralization based on diverse functions, mainly decision-making, management-administration, financing and provision. In all types, the central government keeps the policy-design function but transfers one function or more to lower geographical units, for instance, decision-making in ‘deconcentration;’ decision-making and management in ‘delegation;’ and all four key functions in ‘devolution’ (Ugalde and Homedes 2002; ECLAC 2006; Homedes and Ugalde 2006). The comparative analysis of the twenty countries unveiled considerably more combinations of functions than in the above semantic differentiation hence a new approach is introduced herein. Table 10.3 (first column) ranks the countries on the degree of decentralization achieved in their public sectors (the entire system in Colombia, social insurance in Costa Rica, and public-social insurance in Chile); plans or projects were not taken into account. The ranking was roughly based on eight variables:

249

degree of actual implementation/progress, basic-package management by local units, existence of management agreements, hospital autonomy reached, and decentralized functions in financing, control of human resources, management and provision. A problem encountered was the lack of information on all variables in eighteen countries, particularly on the degree of implementation (only available in eight countries).5 The resulting ranking was: very high in Argentina and Brazil; high in Colombia and Mexico; middle in Bolivia, Chile, and Costa Rica; low in Cuba, Nicaragua, Peru, and Uruguay; and very low in Dominican Republic, El Salvador, Guatemala, Ecuador, Panama, Haiti, Honduras, Paraguay, and Venezuela. In thirteen countries the decentralization of the public sector is either low or very low. Social insurance has resisted decentralization with few exceptions: Costa Rica’s CCSS decentralized all functions at the first level except financing; Mexico’s IMSS delegated managerial, financial, and provision to health-care areas with relative autonomy but responding to national directives (the other four social insurances remain centralized); Nicaragua’s INSS delivers all services through external mixed providers; Peru’s EsSalud partly cares the insured with its own facilities and for the rest delegates to mixed providers; and Uruguay’s BPS transfers sickness-maternity care to private providers (IAMC) and high-complex actions to mixed providers (IMAE). Virtually all countries have transferred some responsibilities to intermediate units but few to local units. Table 10.3 (second column) confirms that decentralization in the public sector is from the central government to intermediate units in six countries (Dominican Republic, Guatemala, Honduras, Mexico, Panama, and Paraguay) and to a combination of both units in twelve countries but predominantly intermediate in most of them (Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, El Salvador, Nicaragua, Peru, Uruguay, and Venezuela). No country has decentralized mainly to local units, except for

250 Reassembling Social Security Table 10.3. Decentralization of the public health sector,a between 1998 and 2005

Countries

Degree of decentralization

Argentina

Very High

Bolivia

Middle

Brazil

Very High

Chile

Middle

Colombia

High

Costa Rica

Middle

Cuba

Low

Dominican R. Very Low Ecuador El Salvador

Very Low Very Low

Guatemala

Very Low

Haiti

Very Low

Honduras

Very Low

Mexico

High

Nicaragua

Low

Panama

Very Low

Paraguay

Very Low

Peru

Low

Uruguay

Low

Venezuela

Very Low

Decentralization of the central government toward intermediate and local geographic units 24 provinces (manage 65% of hospitals and 69% of beds), and some municipalities (manage 35% and 29% respectively)b Central government, departments, municipalities (very low autonomy), districts (DILOS)b Federal government to 27 states and 5,507 municipalities (90% control primary care and 10% all levels) 28 regions and 342 municipalities; hospitals to become decentralized under 2004 law 32 departments and 524 municipalities (not completed); half of the hospitals are decentralized and autonomousb Social insurance: local EBAIS in all health areas provide fist level care to 85% of the population; all hospitals are autonomousb Central government, 14 provinces, 170 municipalities; family doctor in neighborhoods for 99% of total population; no autonomous hospitals Central government, the principal, regions, provinces (little progress in districts) Central government, the principal, to health areas in municipalities Central government, departments and municipalities to 27 SIBASI (embracing various municipalities each)b Central government, the principal; regions and departments (few municipalities) Central government, the principal, to districts of municipalities (UCS only operates in 2% of districts) Central government, the principal; regions, little at the local level (SILOSS); some hospitals partly decentralized Federal government to all 31 states and federal district, also urban/wealthy municipalitiesb,c Central government, the principal, to municipalities (local SILAIS incipient in 21% of them); a minority of hospitals are decentralizedb,d Central government, the principal, provinces, regions (only three decentralized), one hospital decentralizede Central government, the principal, to regions and departments (only in 8% of municipalities); incipient hospital decentralization Central gov’t, depart’s; local units (CLAS) in 34% of primary facilities for 25% of population; since 2005 shift to regional decentralizationb Central government to 19 departments and municipalities (only in 15% of the facilities); a minority of hospitals are decentralizedb Central government to states and municipalities (reversed)f

Effects on Unity, State Responsibility, Efficiency a minority in the first level. Seven countries have organized ad hoc local managing healthcare units: Bolivia (DILOS, with little autonomy), Costa Rica (EBAIS, the most extended in the region), El Salvador (SIBASI, extended but with low decentralization), Haiti (UCS, interrupted by the crises), Honduras (SILOSS) and Nicaragua (SILAIS) both little extended, and Peru (CLAS, stagnant). The basic package is administered by local special entities in five countries (Bolivia, Brazil, Colombia, Costa Rica, and Dominican Republic—in the law), and municipalities manage the program in three countries (Brazil, Chile, and Cuba). Hospital autonomy has reached diverse degree of scope and progress in eleven countries: Costa Rica (all hospitals), Argentina (significant but progress halted), Colombia (half), Honduras (some and partial), Nicaragua and Uruguay (small minority), Panama (only one), Chile (stipulated by 2004 law), Paraguay (planned or incipient), Peru (obstructed), and Venezuela (reversed) (see Section 11.1.4). In twelve countries the central authority signs managing agreements (compromisos de gestión) delegating certain functions to intermediate and/or local units (Table 10.3): Argentina, Bolivia, Colombia, Costa Rica, Guatemala, Mexico, Nicaragua, Panama, Paraguay (planned), Peru (in only 20% of departments), Uruguay, and Venezuela (reversed, see below). Such agreements usually are from ‘top-down’ and most countries have not evaluated their results. Costa Rica

251

is a notable exception: the CCSS has signed fairly standardized agreements with all health areas and hospitals (by negotiating commissions appointed by both parties), setting targets on coverage, quality and productivity in the delivery of a set of benefits emphasizing health needs of those protected; health areas receive a per capita related to demographic variables whereas hospitals are paid according to their performance; 81% of total resources to health areas and hospitals were allocated through such agreements in 2001, which are also used to negotiate among the various levels within CCSS and with private and mixed providers; renewal of annual agreement requires CCSS evaluation of target fulfillment. There are not standardized statistics on the percentage of decentralization actually achieved, only very rough indicators not technically comparable in ten countries: 100% of states (but only in urban wealthy municipalities) own/manage their facilities in Mexico; 90% of municipalities own/manage first level facilities in Brazil and 10% at all levels (the municipal share increased from 69 to 92% of all public facilities in the 1990s); 99% of the total population receives primary care by neighborhood family doctors in Cuba; 85% of the total population gets primary care through local EBAIS in Costa Rica; 69% of hospital beds belong to provinces and 29% to municipalities in Argentina (the central government lacks hospitals and OS have 2% of beds); 25% of the total population is covered by and 34%

← (Continued) Sources: General from legislation; PAHO (2002b, 2002c, 2005a); also Brazil: Medici (2002a); Costa Rica: López (2005); Dominican Republic (2006); El Salvador: MSPAS/GTZ (2002); Haiti Republic (2004); Mexico: Homedes and Ugalde (2006); Panama: La Forgia (2005); Paraguay: Ramírez (2004, 2005a); Peru: Díaz (2001, 2007), Manrique (2005), Portocarrero (2006); Venezuela: RBV (2004), González (2005). a Entire system in Colombia, public-social insurance in Chile and social insurance in Costa Rica. b Management agreements signed between higher and lower units. c Social insurance (IMSS) has delegated managerial, financial, and provision to health-care areas with relative autonomy. d Social insurance has decentralized services buying them from mixed providers (EMP). e Social insurance decentralization limited to one hospital jointly with the ministry. f The legal draft of 2005 not approved in 2006 reverses the decentralization process recentralizing in the ministry previous functions transferred to states and municipalities that had signed decentralization agreements.

252 Reassembling Social Security of primary facilities have CLAS in Peru; 21% of municipalities have SILAIS in Nicaragua; 15% of health facilities are decentralized in Uruguay; 8% of the municipalities are decentralized in Paraguay, and 2% of districts have UCS in Haiti (see Table 10.3).

10.4.2. Decentralization of functions Decentralized functions are quite diverse but the central ministry or public autonomous entity retains functions of directionregulation-supervision in all countries, as well as those of financing and human resources in most of them, and transfers management and provision to intermediate and/or local units in part of the countries and in diverse degree—primary-level units do referrals to higher levels (sources of this section are in Table 10.3) Financing is centralized in seventeen countries, the thirteen ranked with low and very low decentralization, and Chile (FONASA), Colombia (solidarity fund: FOSYGA), Costa Rica (CCSS) and Cuba (central budget). Exceptions are countries with cofinancing: between the central government and provinces/departments in Argentina and Bolivia, and between the federal government and states in Brazil and Mexico; El Salvador central government, departments, and municipalities transfer financial resources to local teams (SIBASI). Human resources remain also centralized in the thirteen countries ranked with low or very low decentralization. Bolivian human resources are practically managed by the ministry regional offices (two attempts to transfer such function to the local level failed); Costa Rica’s CCSS centrally manages human resources, and Cuban personnel is centrally appointed and assigned by the state. Conversely, human resources are managed by provinces/municipalities in Argentina, municipalities in Brazil (authorized by the states), regions and municipalities in Chile, municipalities in Colombia (uncompleted process) and states in Mexico.

In Cuba the ministry determines functions and services of local health authorities, controls university hospitals, research centers, the pharmaceutical industry, import and export of health goods, and domestic distribution of medicines by state pharmacies; the ministry provincial offices manage all facilities in their territories whereas municipal offices are in charge of all local facilities; most decentralized is the family doctor program in neighborhoods. Conversely, Costa Rica CCSS transfers management and provision of a standardized set of first level services to local integrated teams (EBAIS) in the ninety health areas that the country is divided (3,500–4,000 inhabitants each); EBAIS work with their communities to identify high-risk persons and families, and do referrals to higher levels. El Salvador central government, departments, municipalities, and communities transfer authority and resources to local SIBASI that are responsible for first- and second-level care through mixed providers (public, social insurance, private including NGOs); each SIBASI covers 80,000–500,000 inhabitants embracing several municipalities, coordinates a health center with one referral hospital that counterrefers to specialty hospitals. Guatemala has a similar combination of central management and provision at the second and third levels with those functions decentralized at the first level that provides the basic package, and in Bolivia combining departments and local units; Nicaragua’s ministry is transferring authority to local integrated units (SILAIS) but they lack enough autonomy and resources; and Haiti’s ministry new strategy envisages a transfer of the basic-package provision to district health units (UCS), through contracts with mixed providers (for-profit private, NGOs, and public).6 Ecuador’s reform law of 2001 (not enforced in 2006) stipulates decentralization at the three levels; municipal health councils would design and evaluate the basic package. The Dominican Republic’s reform law mandated the transformation of nine regional services into autonomous entities to articulate the basic package network of providers and

Effects on Unity, State Responsibility, Efficiency deliver it to affiliates but they had not materialized in 2006. Panama’s decentralized management and provision of services is restricted to one autonomous integrated hospital: an independent organization receives budgetary transfers from the ministry and social insurance to contract and buy a comprehensive package of first- and second-level services for the hospital that takes care of area residents affiliated to social insurance or the ministry, subcontracting all services with private enterprises. Central government management and provision are shared with states/departments and municipalities in Mexico and Uruguay, and regions in Honduras and Paraguay (in process). The two functions have been completely transferred to: states/provinces and municipalities (Argentina, Brazil, and Mexico); regions and municipalities (Chile); regions and departments (Peru, in process), and insurance firms, providers and public hospitals (Colombia). Decentralization policies have changed in some countries as the following three examples illustrate. Brazil’s federal ministry norms gave prominence in 1993–6 to municipalities, depleting federal and state functions and creating several problems; new norms in 2000–1 restored some functions to federal and states governments to correct said problems and achieve better balance. The states are in charge of regional policy, referral and high-complex provision (cofinanced with federal transfers), identify regions according to their epidemiological characteristics, number of users and capacity for care, and determine the functions and levels of complexity granted to municipalities within each region. Municipalities receive federal funds to finance and manage the basic package and the health family program (PSF), both at the first level, as well as state funding according to the authorized level of middle-complex actions; the PSF and the budget are approved by states that empower municipalities to hire and train their personnel and organize community agents. Peru’s decentralization policy changed three times. Local committees for health administration (CLAS) organized in 1994 were

253

entrusted with some first level facilities and signed agreements with the ministry that transferred them funds, infrastructure, equipment and personnel; CLAS managed those resources, collected user fees, provided care to the assigned population, met the agreed targets and was accountable to the ministry. In 2002 decentralization shifted course from the central government to regional and departmental health offices, charging the latter with direction, control and provision of services in specific zones, submitted to the fulfillment of agreed targets; public facilities were to be autonomous and based on the articulation of provincial and local first level facilities with the corresponding hospital. In 2005 more emphasis was placed on transfers from the central government to regions that became the principal unit of decentralization and decide whether to keep, modify, or discontinue CLAS that became stagnant.7 Venezuela’s central government transferred to the states in the 1990s the functions of policy implementation, financing (cosharing central and state funds) and provision, whereas municipalities became responsible for management of services based on agreements with the other two levels. But the legal draft pending approval in 2006 reversed the previous process recentralizing in a public entity the services transferred to states and municipalities even in the 70% that had signed decentralization agreements.

10.4.3. Evaluation of decentralization results Several experts affirm that decentralization was done without a previous study of the most proper and viable forms according to health, socioeconomic and political features of the countries, failing to identify potential obstacles and seek solutions to them, and lacking adequate transfer of authority and resources from the central government to the decentralized units, as well as trained personnel needed for the new functions (Homedes and Ugalde 2006; ECLAC 2006).

254 Reassembling Social Security We lack rigorous studies on the effectiveness of decentralization and evaluation of its results relative to policy design, implementation, and goal fulfillment. Some countries show scarce progress in establishing indicators based on results and costs, whereas others have an excessive variety and complexity of indicators.8 For instance, the Bolivian ministry should evaluate the performance of agreements based on a set of indicators, but there is neither precise criteria to determine targets, nor a clear methodology to measure indicators and the agreements have not been evaluated periodically. Most Central American countries have not conducted a professional quantitative assessment of the impact of decentralization. An exception is Costa Rica’s thorough evaluation in 2001 with mixed outcomes that revealed minimum progress in agreements establishing indicators based on results and directly related to financial resources; the increasing variety of indicators and targets in the agreements led to excessive complexity and difficult assessment. El Salvador conducted an appraisal of the oldest SIBASI in 2001 and a survey in one-fifth of them in 2002. Some positive decentralization effects were found in four countries: (a) improvement in coverage of children and adolescents (but decline in coverage of the elderly) in Costa Rica, and in vaccinations in Bolivia; (b) municipal poles in Brazil that take care of their own population and that of neighboring municipalities with benefits of scale and expansion of access; (c) financing of health areas related to demographic variables and of hospitals according to their performance leading to increase in output in most hospitals (Costa Rica, but third-level hospitals generalperformance declined); (d) response to most problems at the local level, improvement in the opportunity and personal treatment of users, better coordination with the second level, increment in the ratio of consultations and reduction in external consultation costs, more references to the third level and cut in hospital mortality (El Salvador oldest SIBASI).

But data gathered on fifteen countries demonstrate that the World Bank expected positive effects have not materialized in most of them and the opposite have occurred in some: (a) Coordination and regulatory framework Decentralization was often done without a nationally coherent regulatory frame, coordinating mechanisms among various levels and precise determination of their responsibilities, actually there were complex and unclear norms, fragmentation of responsibilities between government levels, duplication, and confusion. Bolivia’s division of responsibilities into four levels provoked lack of definition on priorities and responsibilities, overlapping of functions due to ambiguity, uncoordinated investment between departments and municipalities (the latter are responsible for the first level and cannot do referrals at higher levels). Colombia had multiple territorial reforms with diverse progress and little interchange (instead of a national reform) and lack of clear regulations to achieve adequate vertical and horizontal integration of all care levels, which eroded the national health authority and system coordination. El Salvador doesn’t have a coherent legal frame for SIBASI, their strategies and interaction between the public and private sectors, as well as their financing, payments, and priorities. Coordination between Mexican federal and states governments entrusted to a national health council has been very difficult at least until 2005 and also provoked overlapping and gaps in services. Paraguay failed to design coordination mechanisms between the three public sector levels, and to previously specify the functions of the new health departments, whose functions duplicated those of the regional directors subordinated to the ministry, creating confusion and conflicts.

Effects on Unity, State Responsibility, Efficiency Peru’s CLAS were not integrated with the ministry team in charge of policy; the first level was not always the entry to the system and lacked adequate mechanisms of referral to higher levels; facilities were not articulated into a network and didn’t share resources, and funds were channeled from the central ministry or departments hampering coordination. (b) Local autonomy Decision-making on financing and human resources continues to be centralized and too strict regulations impede adaptation to community needs; local units have received responsibilities and resources but without proper managerial capacity. Brazil federal regulations restrain the ability of municipalities to adapt to local health conditions often obstructing positive innovations, and some municipalities are unable to meet federal and state conditions. Chilean decentralization was not profound because directors of municipal health services didn’t get real autonomy. Colombian municipalities have not taken full responsibility due to absence of incentives to decentralize, poor capacity and experience to take up complex functions, and lack of previous experience with capitation and the basic package. Costa Rican EBAIS enjoy little flexibility to subcontract services with private providers; most data are historical rather than on demographics, morbidity, and projections needed to estimate the real population needs, define policies on basic services, plan their buying and make investment decisions. Salvadorian SIBASI have not functioned in a decentralized manner due to vertical handling from the central level; most administrators and management committees have ranked very low the level of decentralization achieved (even in the oldest and most successful SIBASI), whereas the committees of social consultation

255

ranked it very low or middle or did not know. (c) Hospitals autonomy, self-financing, and coordination Most hospitals neither has become autonomous and self-sufficient (with Costa Rica’s exception) nor function in a network connecting various care levels nor share resources, and lack adequate referral and counter-referral mechanisms. Argentina’s reform aimed to transform provincial/municipal hospitals into self-financed, self-managed entities able to control their operations and budget, flexibilize hiring, promotion, and transfer of employees, and contract nonmedical services with private providers. But hospitals lacked standardized national regulations with minimum criteria for contracts and quality of services in provinces/municipalities (positive to facilitate adaptation to local conditions but actually leading to atomization); new self-managed hospitals staff was not properly trained, and countryside hospitals lacked the capacity to become self-managed specially in poor areas. Colombia’s public hospitals are far from the planned shift to autonomous entities. Costa Rica faces problems with referrals and counterreferrals because only 28% of EBAIS monitor and follow up their patients; hospitals don’t function as an integrated network at all levels and the existing coordination is more a result of individual initiative than general policy. Panama initially planned to extend the decentralized model in one hospital to five new hospitals, but it was blocked by strong opposition from unions and some political parties; laws in 1999– 2000 stipulated greater management autonomy to two hospitals expected to become models for the rest, but there was no progress. Some Peruvian hospitals began to organize according to self-management norms in 2000 with the goal to finance them based on their performance and yet the lack of

256 Reassembling Social Security an explicit hospital policy obstructed the process. The ten hospitals that had signed decentralization agreements in Uruguay accounted for only 15% of total public facilities in 2002 and an attempt to grant more autonomy to them failed. (d) Financing Decentralization often lacked previous precise definition of financing and cost studies and has been harmed by oscillation in central transfers, cut in resources, payment delays and uncoordinated investment between intermediate and local units. Argentina reform goal that hospitals sell services to OS, providers and users with paying capacity didn’t materialize: resources from paying users didn’t come through whereas reimbursement delays by OS and providers occurred, and decentralized units became indebted because they didn’t receive sufficient central funds or suffered a crisis or lacked fiscal discipline. Paraguay reform failed to define financing of the decentralized system and ways to transfer funds. Peru cut CLAS budgets by 21% in 2003–5, failed to honor agreements signed with them and stopped new ones since 2003. Brazilian oscillations in transfers of federal funds and lack of state and municipal own resources in the 1990s have been alleviated by minima set at each level since 2000. (e) Efficiency/cost reduction In four countries, local politicians did not make the best decisions for their communities and inefficiently assigned centralgovernment health funds: in Bolivia such funds were used to construct roads and in Colombia to build unneeded municipal hospitals; in Peru a primary care center increased user fees to finance the construction of a hospital and another center discontinued the generic medicine program and sold brand medicines; and some Nicaraguan families

in rural areas sold the latrines they had received to satisfy other necessities. Chilean decentralization stimulated primary care doctors to unnecessary references to higher levels to reduce their work load and expenses; similar behavior was displayed by Costa Rican cooperatives that signed agreements with social insurance. Decentralization also caused a loss of economies of scale in purchase of medicines and other health inputs (as a result prices rose in Paraguay); many provinces and municipalities lacked the technical capacity to purchase by bidding with international suppliers hence bought products to local wholesale and even retail sellers, and they couldn’t offer high-complex costly services (see Section 10.8.2). (f) Personnel The transfer of responsibility and resources was not usually accompanied by adequate managerial knowledge; personnel was appointed without technical criteria, not properly trained to perform the new functions, often on a temporary basis, and without proper incentives to increase productivity and ensure the quality of services. Poor provinces and municipalities have encountered difficulty to recruit technical personnel locally and more so to attract it from abroad. Hiring of personnel in decentralized levels not always has reduced the central bureaucracy and its remuneration, because of its union power and the new regulationsupervision functions assigned to such bureaucracy. Bolivian personnel were appointed not relying on skills and experience, with the same salary in different levels of authority and without productivity incentives. Chilean directors of municipal services lacked flexibility (at least until 2004) to hire new personnel and existing employees could not be removed and kept their salary because of trade union resistance and opposition to part of the salary being linked to

Effects on Unity, State Responsibility, Efficiency performance. Colombian health professionals were not trained for the reform and many resisted change. El Salvador’s SIBASI needed permanent personnel and suffered from frequent absence of employees who were attending other activities. Peru’s chiefs of health centers and posts, physicians and community leaders lacked skills to manage facilities and funds, and personnel hired under private-sector labor rules lost their benefits under the publicsector labor law creating discontent (see Section 10.8.2). (g) Equity Bolivia’s and Mexico’s decentralization expanded disparities between regions, municipalities, and urban–rural zones because high- and middle-income inhabitants in urban and developed regions have greater power to pressure and obtain central/federal resources and services than poor and low-income inhabitants in the countryside and underdeveloped regions (Mexico’s SPS, however, has reduced such inequalities). In Chile, the transference of municipal funds from wealthy to poor municipalities has reportedly improved equity, but it is necessary to know who pay taxes in the wealthy municipalities and who receive benefits in the poor municipalities. Due to inadequate allocation of local funds, equity has not improved in Bolivia and Peru (see Section 9.3.3). (h) Quality Methodological difficulties have obstructed an adequate assessment on whether quality has improved as a result of decentralization, and users, particularly the poor, lack skills to evaluate the quality of services received (see Section 9.3.3). (i) Community participation Decentralization cannot promote democracy by itself, local elites are not willing to share decision-making with the population, thus the impact on community participation has been minor and uneven (see Section 10.10).

257

10.5. Reform goal: competition The monopoly of the public and social insurance sectors has been criticized for lacking incentives to improve efficiency. Reformers claimed that opening these sectors to competition, entry and diversification of private providers, and freedom of choice would make health markets more efficient, reduce costs, improve service quality and increase user’s satisfaction (World Bank 1993). But competition in health markets is obstructed by several causes: (a) strong restrictions to freedom of choice (see Section 10.7); (b) insured remain affiliated to a firm/provider despite the option to change and improve benefits and premium, either because of inertia, desire to keep the same doctor, or due to difficulties to get comparable data on health plans and understand them; (c) efficient insurers are pushed out of the market by inefficient insurers that successfully practice cream-skimming, the latter win but transfer their costs to other insurers and the global effect can be null or negative; and (d) competition works in cities or urbanized areas but not in rural zones with scarce population (Bertranou 1999; ILO/ISSA 2001b). Other flaws of competition are discussed below. Albeit abundant evidence show that competition does not function properly in reformed health systems, it is still superior than competition in the same countries relative to structurally reformed pensions: the number of health insurance firms compared with the number of pension administrators in 2004 was 309 and 12 in Argentina, 97 and 6 in Colombia, 17 and 6 in Chile, and 95 and 4 in Uruguay, and this without counting hundreds or thousands of providers.9 An explanation for this difference is that pension administration is more concentrated than health care that requires considerably more facilities and wide dispersion of services. Colombia’s reform was unique in the region for its features to maximize competition: (a) insurance firms and providers are separated, the latter competing to offer services

258 Reassembling Social Security to the former; (b) insurance firms/providers can be public, private, and mixed, either forprofit or not-for-profit; (c) social insurances can become insurance firms, whereas familyallowance funds, prepaid and health insurance companies may associate with insurance firms, thus facilitating entry; (d) affiliates have freedom of choice between insurance firms and between providers with a contract with the firm to which they are affiliated; (e) insurance firms must have at least two providers and municipalities at least one; ( f ) affiliates can change insurance firms annually; and (g) the basic package and capitation payment are not competitive but insurance firms can offer plans with additional benefits in the contributory regime. These features, however, were weakened by policies that hampered competition (see below). Despite the competition relevance, scarce statistics are available making quantitative comparisons difficult. Table 10.4 compiles all accessible data c.2000–5 in the twenty countries on the number of insurance firms, their nature and degree of concentration, and freedom of choice status. Insurance firms and providers are private for-profit in most countries, but of multiple nature (public, private, social insurance, etc.) in Colombia and Dominican Republic; in Brazil they are all private but of different types (group medicine, cooperatives, enterprise plans and insurance companies); in Nicaragua can be public or private, and in Uruguay IAMC are private not-for-profit mutual-aid societies, cooperatives or enterprises, whereas IMAE may be public or private. Colombia’s private insurance firms (EPS), both in total number and affiliates, increased from 28% to 73% of total firms in 1997–2002 while public firms (EPS) fell from 72% to 27%; in Nicaragua 75% of the EMP were private and 25% public in 2002, and in Uruguay private not-for-profit IAMC have decreased both in number and in affiliates while private partial insurance and mobile emergencies have risen. Conversely, Chile’s percentage of affiliates in public-social insurance has expanded while the percentage in private ISAPREs has shrunk (sources for this

section are those in Table 10.4 and Mesa-Lago 2006a). Due to the lack of standardized statistics on the number of private insurance firms and estimates on providers in most countries, only an approximate ranking of countries by their degree of competition was feasible: Brazil 1,587 mostly private; Argentina 309 social insurances (OS) and 270 prepaid providers (EMP); Colombia 97 insurance firms (EPS and ARS) and unknown number of providers, all of multiple nature; Chile 17 private insurance firms (ISAPREs) and possibly 3,500 providers; Uruguay 40 providers (IAMC private not-forprofit), 55 highly complex medicine providers (IMAE: multiple nature) and 68 private partial insurance; Dominican Republic 22 insured firms (ARS) and 40 providers (PSS) of multiple nature, 30 prepaid (igualas) and 12 private insurances; Nicaragua 48 providers of multiple nature (EMP); Paraguay 28 prepaid private providers; Honduras 10 private insurance firms, and Peru 2 private insurance firms (EPS). There are few insurance firms or providers in Costa Rica, Ecuador, El Salvador, Guatemala, Panama, and Venezuela (in Mexico prepaid providers were banned until 2001). There are no insurance firms but providers in Haiti, and neither private insurance nor providers in Cuba. It was impossible to find data of any kind from Bolivia and on competition in Honduras and Paraguay. The degree of concentration of affiliates in the three largest insurance firms could be measured only in six countries: 28% in Argentina OS, 38% in Colombia EPS, 47% in Dominican Republic ARS, 61% in Chile open ISAPREs, and 100% in Peru EPS. Based on the largest twenty firms concentration was 67% in Argentina, 73% in Nicaragua, and 96% in Dominican Republic. Both in small and big countries there is also concentration in urban areas, for instance, proportions of total providers in the capital city were 61%, 58%, and 42% in Peru, Argentina, and Nicaragua respectively. The number of insurance firms has diminished in most countries while concentration has increased. Argentina’s OS fell 21% in 1994– 2004 and concentration in the largest twenty

Effects on Unity, State Responsibility, Efficiency

259

Table 10.4. Indicators of competition and freedom of choice in health systems, between 2000 and 2005 Concentration (%) in largest firms/providers Countries

Number and nature of insurance firms/providers

3

10

20

Freedom of choice

28.0

51.5

67.1

Bolivia

309 (OS: social insurance), 270 EMP (private)a n.a.

n.a.

n.a.

n.a.

Brazil

1,587 (private)

n.a.

n.a.

n.a.

Chile

17 (ISAPREs)b (private)

61.4

n.a.

n.a.

Colombia

97 (EPS, ARS)c (multiple)

38.0

n.a.

n.a.

n.a.

n.a.

n.a.

47.0

82.0

96.0

Yes, but many OS are closed, affiliates cannot select an EMP Yes, within social insurance since 2000 Yes, affiliates can leave the public sector and choose a private provider Yes, between FONASA and ISAPREs (limits on income and other obstacles) Yes, more within the contributory and between insured than within the subsidized and between providers Noi No Yes, in various public hospitals and part of the contributory (not in force) No; yes in 2001 law (not in force) No No No No No, an attempt in the IMSS failed Yes, in social insurance but only within the area of residence No, except in one hospital No Yes, between social insurance (EsSalud) and the providers (EPS) by majority vote Yes, affiliates in social insurance (BPS) can select an IAMC but cannot shift later No

Argentina

Costa Rica Few (private)d Cuba No, only state Dominican R. 22 (ARS) and 40 PSSe (multiple) Ecuador

Very few (private)

n.a.

n.a.

n.a.

El Salvador Guatemala Haiti Honduras Mexico

Few (private) Very fewf Nog 10 (private) Few insurance firmsh

n.a. n.a.

n.a. n.a.

n.a. n.a.

n.a. n.a.

n.a. n.a.

n.a. n.a.

Nicaragua

48 (EMP) (public and private) Very few (private) 28 (prepaid) (private) 2 (EPS) (private)

n.a.

n.a.

73.0

Panama Paraguay Peru

Uruguay

Venezuela

40 (IAMC) and 55 (IMAE) (multiple); 68 (insurance, private) Very few (private)

n.a. n.a. n.a. n.a. 100.0 100.0

n.a. n.a. 100.0

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

(cont.)

260 Reassembling Social Security rose from 42% to 67% (the largest ten had 52%) whereas EMP fell 10% and the largest ten concentrated 78%. Colombia’s total insurance firms (EPS and ARS) decreased by 24% in 1997–2002; EPS in the contributory regime diminished 15% and concentration rose to 38%, and ARS in the subsidized regime shrank by 35% but concentration was considerably lower. Chile’s ISAPREs declined 53% in 1994– 2004 (affiliation fell from 27 to 18% of the population)10 and concentration in the largest three augmented from 48 to 61%. Peruvian providers (EPS) were cut by half in 2002–4 from four to two and concentration jumped to 100%. Uruguay’s IAMC decreased 25% in 1998–2005 and concentration rose (46% of IAMC had less than 20,000 affiliates each but 8% had more than 50,000 affiliates). On the other hand, Nicaraguan providers (EMP) increased more than twice in 1996–2002 and affiliates in social insurance who received their services jumped from 70 to 100%, and there has been an increase in Dominican Republic’s insurance firms (ARS) too. Multiple causes explain the decline in insurance firms and growing concentration in most countries. Argentina’s freedom of choice promoted competition and led to bankruptcy and merging of inefficient or financially unsustain-

able OS and prepaid EMP. Chilean ISAPREs and affiliates decline was caused by the elimination of fiscal subsidies granted to workers with insufficient contribution to joint an ISAPRE, market saturation due to the limited number of affiliates who can afford high premium and co-payments, ISAPREs reluctance to expand their plans at lower rates, transfers of retired affiliates from ISAPREs to FONASA, and bankruptcy of small ISAPREs unable to compete against the largest and most profitable. Uruguay IAMC growing co-payments and shrinking services provoked bankruptcies and rising concentration; large IAMC survived taking advantage of economies of scale and having lower average cost than small ones. Significant obstructions to competition exist in five countries. There are closed insurance firms in Argentina and Chile; in the former the entry of prepaid EMP was postponed so that OS, with state grants, could adapt to the competition; there is not competition between OS and EMP because an affiliate cannot choose an EMP directly but only through his own OS. Colombia’s competition, one of the strongest in the region, is not exempted from flaws: to allow the conversion of three social insurances in the contributory r