A W O R L D B A N K C O U N T R Y S T U D Y 50357 Argentina Income Support Policies toward the Bicentennial THE WORLD BANK A W O R L D B A N K C O U N T R Y S T U D Y Argentina Income Support Policies toward the Bicentennial Copyright © 2009 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First Printing: June 2009 Printed on recycled paper 12345 12 11 10 09 World Bank Working Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally-edited texts. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, Tel: 978-750-8400, Fax: 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-13: 978-0-8213-8051-2 eISBN: 978-0-8213-8052-9 ISSN: 0253-2123 DOI: 10.1596/978-0-8213-8051-2 Library of Congress Cataloging-in-Publication Data Argentina : income support policies toward the bicentennial / The World Bank ----p. cm. --ISBN 978-0-8213-8051-2 (pbk.) -- ISBN 978-0-8213-8052-9 (ebook) ­1. Income maintenance programs--Argentina. 2. Economic assistance, Domestic--Argentina. 3. Social security--Argentina. 4. Argentina--Economic policy. 5. Argentina--Social policy. I. World Bank. Argentina, Chile, Paraguay and Uruguay Country Management Dept. --HC180.P63A74 2009 --362.5'820982--dc22 -- 2009019209 Abbreviations and Acronyms AAA Analytical and advisory activities ALMP Active labor market program ANSES Administración Nacional de la Seguridad Social CCT Conditional cash transfer CEDLAS Centro de Estudios Distributivos, Laborales y Sociales / Center for Distributive, Labor and Social Studies CIC Integrated Community Center EPH Continuous Permanent Household Survey EPPS Encuesta de Percepción de Planes Sociales en la Argentina / National Survey of Perceptions of Social Policy in Argentina FNE Fondo Nacional de Empleo INDEC Instituto Nacional de Estadísticas y Censos ITP Income Transfer Program MDS Ministry of Social Development MTESS Ministry of Labor, Employment and Social Security NCP Noncontributory pensions NEA Argentinian North East region NOA Argentinian North West region OECD Organisation for Economic Co-operation and Development PIP Plan de Inclusión Previsional (Pension Inclusion Plan, or Moratoria) UTDT Universidad Torcuato di Tella Currency Equivalents Currency Unit = Argentine Peso (Arg$) 1 U.S. Dollar = Arg$3.0202 1 Peso = US$0.331 (as of June 23, 2006) FISCAL YEAR January 1 to December 31 iii Contents Acknowledgments ................................................................................................................ viii Preface ........................................................................................................................................ ix Executive Summary.................................................................................................................. xi Why Does Argentina Need Social Protection?.............................................................. xii What Is the Status of Social Protection in Argentina?.................................................xiii What Happened to Income Support during the Crisis? .............................................. xv What Has Happened since the Crisis? ........................................................................... xv What Do Argentines Think about Social Protection?..................................................xvi What Does the Future Hold for Income Support in Argentina? .............................xviii What Is the Potential of Income Support to Reduce Structural Poverty in Argentina?............................................................................................................. xix Final Reflections: What Are the Preconditions for Integrated Social Protection? ... xxi The Report .......................................................................................................................xxii 1. Income Transfer Programs and Social Protection .......................................................... 1 Introduction ......................................................................................................................... 1 Definitions: Social Protection and Income Support ....................................................... 2 The Size of Income Transfer Programs in Argentina ..................................................... 5 2. Income Support Policies in Argentina: Origins and Development ........................... 7 The Origins of Income Transfer Programs in Argentina............................................... 7 Emergency Response and Current Income Support Policy ........................................ 14 Current Income Support Policy ...................................................................................... 25 3. Opinions and Preferences Toward Income Support.................................................... 32 Data and Approach .......................................................................................................... 33 Familiarity and Experience with Programs................................................................... 37 Views about Programs ..................................................................................................... 39 Perceptions of Beneficiaries ............................................................................................. 41 Program Design ................................................................................................................ 42 Conclusions ....................................................................................................................... 47 4. Income Support and the Role of the Provinces............................................................. 50 The Federal Model in Argentina..................................................................................... 51 Social Spending at the Subnational Level in Argentina .............................................. 54 Income Support Policies at the Provincial and Municipal Levels .............................. 57 5. Policy Options and Future Challenges........................................................................... 60 The Role of Income Support for the Poor ...................................................................... 61 A Typology of Options..................................................................................................... 61 iv Contents v The Potential Impact of Income Transfers..................................................................... 63 The Simulations................................................................................................................. 65 Introducing Behavioral Responses ................................................................................. 70 Implementation Challenges............................................................................................. 73 6. Reflections, Challenges, and Policy Considerations ................................................... 75 7. References............................................................................................................................. 78 Appendixes............................................................................................................................... 81 Appendix A: Methodology of the EPPS............................................................................. 83 Appendix B: Provincial Consultations: Process and Methodology ............................. 85 Selection of Provinces and Key Informants................................................................... 85 Methodology ..................................................................................................................... 86 Appendix C: Methodology of the Behavioral Simulation .............................................. 87 Tables Table 1. Simulated Policy Impacts ............................................................................................. xx Table 1.1. Public Expenditure and Other Subitems in Constant Nominal Terms and as % of GDP, 2007 ................................................................................................................... 5 Table 2.1. Income Transfer Programs in Argentina, 2007 ....................................................... 25 Table 2.2. Income Transfer Programs in Argentina--Basic Elements as of 2007 ................. 30 Table 5.1. Income Transfers Typology and the Three Visions................................................ 62 Table 5.2. Basic Descriptive Statistics for the Urban Population, Second semester 2006.... 63 Table 5.3. Adjusted Baseline Data Used for the Simulations.................................................. 65 Table 5.4. Simulated Scenarios.................................................................................................... 65 Table 5.5. Simulated Effect on School Attendance--Transition Matrix, Children Ages 12­18 ....................................................................................................................................... 71 Table A1.1 Distribution of Homes by Region and Surveys Carried Out .............................. 83 Table A1.2 Distribution of Homes by Strata and Surveys Carried Out ................................ 84 Table A3.1 Log Earnings Regression for Reported Earnings of Children Ages 12­18........ 88 Figures Figure 1. Trends in Informality and Poverty, 1992­2006 ........................................................ xii Figure 2. Evolution of Public Expenditures on Income Transfer Programs, 1995­2007 .............................................................................................................................. xiii Figure 3. Social Protection Spending at the Provincial Level, 1980­2006 (% of GDP) ....... xiv Figure 4. Plan Jefes Exit Paths of Jefes Beneficiaries, 2003­2007 ............................................. xvi Figure 5. Evolution of Social Security Benefits.......................................................................xvii Figure 1.1. Evolution of Income Transfer Spending, 1994­2006 .............................................. 6 Figure 2.1. Benefits and Beneficiaries of the Pension System, 1950­2007 ............................... 8 Figure 2.2. Beneficiaries of the Pension System by Administrator, 1950­2007 ...................... 9 Figure 2.3. Contributory Resources as a Percentage of Social Security Spending, 1950­2007 ............................................................................................................................... 10 Figure 2.4. Beneficiaries of the Income Support Programs, 1992­2006 ................................. 11 vi A World Bank Country Study Figure 2.5. Pensions and Survivors' Pensions--Beneficiaries and Average Benefits, 1970­2007 ............................................................................................................................... 12 Figure 2.6. Family Allowances and Unemployment Insurance--Number of Beneficiaries, 1992­2007 ....................................................................................................... 13 Figure 2.7. GDP, Poverty, and Inequality, 1980­2007.............................................................. 15 Figure 2.8. GDP, Unemployment, and Informality, 1980­2007.............................................. 16 Figure 2.9. Cash Transfer Programs (CCTs and Workfare) from the Federal Government--Number and Amount of Benefits, 1993­2007 ......................................... 18 Figure 2.10 Plan Jefes: Exit paths of Jefes' beneficiaries, 2003­2007 ......................................... 19 Figure 2.11. NCP--Number and Amount of Benefits, 1989­2007 ......................................... 22 Figure 2.12. Average and Minimum Real Benefits, and Percentage of Beneficiaries Earning the Minimum Benefit, 1977­2007......................................................................... 24 Figure 2.13. Potential Poverty Rate Increase in the Absence of Income Transfer Programs, 1997­2007 ............................................................................................................ 26 Figure 2.14. Number of Income Transfers Received by Households--Total and First Quintile, 1997­2007............................................................................................................... 27 Figure 2.15. Percentage of Households Receiving Program Benefits by Poverty Condition, 1997­2007 ........................................................................................................... 28 Figure 2.16. Lorenz Curves for Per Capita Income and Income Transfer Programs, 1997­2007 ............................................................................................................................... 29 Figure 3.1. Objective of Social Programs: Universal Benefit and Family Allowances ........ 37 Figure 3.2. Familiarity with Social Programs (Spontaneous and Guided Responses) ........ 38 Figure 3.3. Interaction with Social Programs, 2002­2007 ........................................................ 39 Figure 3.4. Overall Views on Existing Social Programs (Leave or Replace Them).............. 40 Figure 3.5. General Evaluation of Programs (Agreement with the Statement "The Programs Work Well and Should Be Left as They Are")....................................... 41 Figure 3.6. Reasons for Exiting Benefit Programs .................................................................... 42 Figure 3.7. Types of Contraprestaciones....................................................................................... 43 Figure 3.8. Preferred Amount of Monthly Plan Jefes Benefit................................................... 45 Figure 3.9. Opinions on Form of Benefits.................................................................................. 46 Figure 3.10. The Choice between Jefes and Hypothetical Programs (similar to Familias/Seguro)...................................................................................................................... 46 Figure 3.11. Preferred Characteristics of Social Programs (Submodule) .............................. 48 Figure 4.1. Structure of Public Revenues and Expenditures, by Level, 1961­2004.............. 52 Figure 4.2. Total Social Spending, by Government Level, 2000­2007 ................................... 54 Figure 4.3. Participation of Government Levels in Spending on Social Sectors, 2007 (Percent of GDP) ................................................................................................................... 55 Figure 4.4. Provincial Spending on Social Protection, by Sector, 1980­2007 (Percent of GDP) ................................................................................................................... 56 Figure 4.5. Number of Beneficiaries of Provincial Employment Programs, 2000­2006...... 58 Figure 5.1. Simulation Results for a Transfer to Unemployed People .................................. 66 Figure 5.2. Simulation Results for a Transfer to Poor Households........................................ 67 Figure 5.3. Simulation Results for a Transfer to Households with Children under 18 ....... 68 Figure 5.4. Simulation Results for Family Allowances and a Transfer to the Elderly......... 69 Figure 5.5. No School Attendance of Beneficiaries, by Age.................................................... 71 Contents vii Boxes Box 1. The Preference for Employment ..................................................................................xviii Box 1.1. Integration of Social Protection...................................................................................... 4 Box 2.1. The Transition of Plan Jefes, Familias, and Seguro ....................................................... 21 Box 2.2. The Moratoria................................................................................................................. 23 Box 3.1. Do Opinion Surveys Matter?........................................................................................ 33 Box 4.1. The Case of Ciudadania Porteña ................................................................................. 58 Box 5.1. Family Allowances in Uruguay ................................................................................... 72 Acknowledgments T his report summarizes the work of a team led by Rafael Rofman (Co-TTL, FY08) and Dena Ringold (Co-TTL, FY07-08), and consisting of David Seth Warren (Co-TTL, FY07), Juan Martin Moreno, Gaston Mariano Blanco, Polly Jones, Marcela Salvador, Evelyn Vezza, Sarah Ruth Bailey, and Febe Mackey. The report itself was written by Dena Ringold, Rafael Rofman, Juan Martin Moreno, and Evelyn Vezza, with support from Sarah Ruth Bailey. The work was undertaken with support and guidance from Jesko Hentschel (Human Development Country Sector Leader until August 2007), Hermann von Gersdorff (Human Development Country Sector Leader from November 2007), Helena Ribe (Social Protection Sector Manager), Axel van Trotsenburg (Country Director until June 2007), and Pedro Alba (Country Director from July 2007). Peer reviewers were Andrew Norton, Margaret Grosh, and Omar Arias. This report synthesizes the findings of a two-year program of analytical and advisory activities (AAA), which benefited from the outstanding collaboration and participation of a wide range of organizations and individuals. The World Bank team is grateful to these partners for their support and contributions. The team would especially like to thank all of the individuals who gave their time to be interviewed in the provincial consultations and who participated in the opinion survey. Consultations at the provincial level were led by a team from the Universidad Torcuato di Tella (UTDT), led by Catalina Smulovitz and Alejandro Bonvecchi, assisted by a team of field researchers. The opinion survey (EPPS) was undertaken by a joint team from the Centro de Estudios Distributivos, Laborales y Sociales at the Universidad Nacional de La Plata (CEDLAS) and Equipos MORI Argentina, led by Guillermo Cruces (CEDLAS) and Helena Rovner (Equipos MORI) and assisted by a team of fieldworkers and researchers. Francisca Pereyra contributed a paper reviewing existing policy proposals circulating in Argentina. Alejandro Bonvecchi undertook a study of the determinants of subnational social policy in Argentina and other countries in the region. Ariela Goldschmit prepared a background paper on provincial financing of social policy. Earlier versions of the work included in this report were presented at two major conferences on "Social Policy in Argentina: Towards the Bicentenary," I and II, held in Buenos Aires in June 2007 and April 2008. These events were jointly organized with the Ministry of Labor, Employment and Social Security (MTESS); the Universidad de Buenos Aires; the Asociación Argentina de Políticas Sociales; CEDLAS; and the British embassy. The World Bank team thanks these partners for their excellent collaboration, as well as all of the speakers and participants in the two events. This program would not have been possible without the involvement of Minister of Labor, Employment and Social Security Carlos Tomada and his team and Minister of Social Development Alicia Kirchner and the officials in her ministry. viii Preface T his report synthesizes the findings of an intensive program of analysis, dialogue, and partnerships focused on the future of social protection and income support in Argentina. The work was carried out between July 2006 and June 2008 by a World Bank team in collaboration with a wide range of partners in Argentina, particularly the MTESS and the Ministry of Social Development (MDS). The scope of the work went beyond the analysis presented in this document and incorporated an extensive process of dialogue, consultations, outreach, and stakeholder involvement across the country. The starting point for the work was the recognition that Argentina has an important opportunity to develop a long-term strategy for social protection. Following the economic crisis of 2001­02, the government introduced a set of emergency measures to mitigate the effects of escalating poverty and unemployment. The centerpiece of this strategy, Plan Jefes y Jefas (Jefes, or the "Heads of Household" Program), succeeded in reaching over 2 million beneficiaries by 2003. Now that growth has returned, and poverty and unemployment have returned to precrisis levels and are declining further, there is space to develop a vision for social protection beyond the emergency programs. Plan Jefes aimed to contribute to and inform the debate about the future of income support policies in Argentina, taking as starting points the views, values, and preferences of stakeholders and the population. The work involved two innovative efforts to collect and understand the landscape of ideas regarding options for social protection circulating in Argentina. The first effort involved an extensive set of consultations with policy makers and practitioners in social policy, mainly at the provincial level. The second effort was a nationally representative opinion survey on the views and perceptions of the population regarding social policy and income support programs in particular. These findings were used to identify questions and issues for further analysis and discussion. The consultative process uncovered a rich set of views and perspectives regarding social policy. In general, the results found broad support for the future role of income transfer programs in Argentina. Argentina has a rich history in this area. More recently, income transfer programs and community development programs have played an important role in overcoming the crisis and improving the living conditions of Argentina's poorest and most vulnerable populations. The challenge ahead is to develop an integrated approach to social protection that extends coverage, avoids gaps and duplication across programs for both formal and informal workers, integrates contributory and noncontributory financing mechanisms, coordinates national and provincial efforts, and promotes work incentives. The issues raised were also developed further in two major conferences: "Social Policy in Argentina: Towards the Bicentenary," I and II, held in Buenos Aires in June ix x A World Bank Country Study 2007 and April 2008.1 The conferences allowed for further formulation of ideas as well as the discussion of evaluations of existing experience and international examples. It is hoped that this report and the conversations initiated during the dialogue throughout this program will continue to contribute to the development of a long-term strategy for social protection in Argentina. Notes 1 The events were jointly organized with the MTESS, the Universidad de Buenos Aires, the Asociación Argentina de Políticas Sociales, CEDLAS, and the British embassy. Vice President Pamela Cox Country Director Pedro Alba Sector Director Evangeline Javier Social Protection Sector Manager Helena Ribe Human Development Sector Leader Hermann von Gersdorff Co-Team Leaders Dena Ringold/Rafael Rofman Executive Summary A s Argentina approaches its bicentennial as an independent republic, it has a window of opportunity in social protection policy. Following the most serious economic crisis in its history during 2001­02, the country mobilized an unprecedented effort to provide income support to the population in need. Now, as growth has returned and social indicators have recovered to precrisis levels, there is an opening to move from emergency income support programs to a more comprehensive, long-term, and sustainable strategy for social protection. The emergency response was effective, as it helped the country to overcome the worst of the crisis. The centerpiece of the strategy, Plan Jefes y Jefas, provided benefits to nearly 2 million households during a period when poverty affected more than half the population and unemployment reached record levels. The number of beneficiaries slowly declined beginning in 2003, and was at nearly one-third of its maximum value by early 2008. This reduction was achieved by the reentry of beneficiaries into the formal labor market, the loss of eligibility (in most cases, when children turned 18), and the shift of beneficiaries to Familias and Seguro de Capacitación y Empleo (Seguro), the successor programs to Jefes. Now that the crisis has passed, the policy debate has shifted toward the future of social protection over the longer term. The improvement in overall economic conditions since 2003 has resulted in a decline in unemployment, poverty, and inequality, and a recovery of formal employment and real salaries to precrisis levels. These positive trends have generated opportunities to consider longer-term and structural issues, including a debate over the future of whether this new type of noncontributory social policies, based on income transfers to households and individuals, should continue. There is a strong preference in Argentina for social programs that will expand employment opportunities. National opinion research and stakeholder consultations that were conducted for this report found that people favor programs that yield tangible labor market outcomes. Potential beneficiaries prefer benefits accompanied by training, access to employment services, or job placements, even if it would mean a smaller monetary benefit. The challenges lie in program design, and equally, if not more important, ensuring effective implementation. Policies will need to be coordinated across multiple agencies and government levels, respond to concrete demands from beneficiaries, and ensure that an adequate set of incentives and opportunities are available for beneficiaries to become self-sufficient within the labor market. The challenge is to design and fully implement a social protection system that has adequate coverage and benefits and is integrated and sustainable. The model of income transfer programs currently under implementation in Argentina has the potential to fill this niche. The new approach combines two transfer programs, Seguro and Familias, combined with more direct interventions at the community level--including the Integrated Community Centers (CICs). However, challenges include covering all vulnerable groups, including those in the informal and formal sectors; ensuring that benefits are adequate to meet the programs' objectives; avoiding gaps and overlaps between programs, institutions, and levels of government; and ensuring that the overall social protection system is fiscally and politically sustainable. Social protection is defined here as the set of programs that help individuals and households manage social risks, cope with their impacts, and overcome structural xi xii A World Bank Country Study poverty. Risks and shocks can include those such as job loss, health, childbirth, old age, droughts, and floods. The objectives of social protection are poverty relief, employment opportunities, and consumption smoothing. Social protection programs range from contributory social insurance benefits, including pensions and other benefits linked to formal employment, as well as noncontributory social assistance and social safety net benefits and services targeted at the poor. Income support refers to the set of social protection programs that, through cash and in-kind transfers (contributory and noncontributory), help households manage income-related risks and shocks. Why Does Argentina Need Social Protection? The 2001­02 crisis marked one of the most significant shocks to incomes and employment in Argentina's history. However, owing to the emergency response and resumption of growth, social indicators have been improving during the past six years. Although the poverty rate leaped 28 percentage points to 57.5 percent in 2002, by the end of 2006 it had declined to 26 percent. Between 2002 and 2006, approximately 11.2 million Argentines emerged from poverty. Similarly, unemployment reached a historical peak in 2002, at 20 percent, and has since declined to under 10 percent. Informality also spiked to 48.5 percent during the crisis but has declined more slowly than poverty and unemployment. Recent estimates suggest that the rate of informal employment persists at 43 percent (second half of 2006, figure 1). Despite the recovery, social needs persist and coverage of current social protection programs is limited. Poverty remains high at about 20 percent, with rates significantly higher in poorer regions, indicating the need for a safety net. Similarly, there are gaps in coverage of social insurance programs. Close to 3.3 million salaried workers and a very high number of the self-employed continue to be excluded from the social security system, and about 9 million Argentines lack access to health care coverage through social security. Figure 1. Trends in Informality and Poverty, 1992­2006 a. Informality and social security coverage. b. Poverty. 100 80 80 Coverage 65+ 60 % of population % of population 60 40 Informal salaried workers 40 20 20 0 0 1992 1994 1996 1998 2000 2002 2004 2006 1992 1994 1996 1998 2000 2002 2004 2006 Sources: Rofman and Lucchetti 2006; World Bank staff compilation using CEDLAS and Instituto Nacional de Estadísticas y Censos (INDEC) data sources. Note: The discontinuities in the series are due to methodological changes in the Permanent Household Survey (from episodic to continuous). Executive Summary xiii What Is the Status of Social Protection in Argentina? Argentina has a rich history to build on in the area of income support. From the outset, programs mainly have been tied to formal employment. Formal sector social insurance programs, including pensions, family benefits, and unemployment, evolved from early efforts by trade unions to cover their members. Noncontributory social assistance programs have evolved more recently. The country's response to the 2001­02 crisis included the implementation of a widespread emergency workfare program--Plan Jefes y Jefas--which covered nearly 2 million beneficiaries during its peak in 2002. Jefes was effective in reducing the immediate effect of the crisis. This experience, along with the set of existing social protection programs, provides an important starting point for a future strategy for social protection. In 2007, Argentina spent 23.6 percent of GDP on social policies--including education, health, and social protection--and 11 percent of GDP on income transfer programs (figure 2). Within income transfers, contributory pensions represented the largest share of spending (almost 80 percent), and the rest was allocated to Plan Jefes y Jefas, family allowances for formal sector workers, noncontributory pensions (NCP), Familias, and the contributory unemployment insurance program. With respect to social assistance, a transition is currently under way, from the emergency support program, Jefes, to the government's new programs of income support--an employment insurance and training program (Seguro), run by the MTESS, and a conditional cash transfer (CCT) program, Familias, administered by the MDS. In addition, changes in the social security system have allowed many older individuals to receive a pension regardless of their contributory history. Many workers are moving into jobs or into the new programs. Questions remain about the future of those in Jefes and the pace of transition, given the phased rollout of Seguro. The challenge is not just to move people from one program to another but to contribute to the development of an integrated and equitable system that meets the needs of those coming from the formal and informal sectors. Coverage is also a concern, as to date, eligibility for the programs has been restricted to former Jefes beneficiaries. Figure 2. Evolution of Public Expenditures on Income Transfer Programs, 1995­2007 a. Spending as % of GDP. b. Breakdown of spending. 10% 100% Trabajar, Jefes, Familias, Seguro 9% Unemployment 80% insurance 8% Family 60% allowances 7% Noncontributory 40% pensions Pensions 6% (moratoria) 20% Pensions (old 5% age and 0% 1995 1997 1999 2001 2003 2005 2007 survivor) 1995 1997 1999 2001 2003 2005 2007 Source: World Bank staff compilation using Ministry of the Economy and Administración Nacional de la Seguridad Social (ANSES) data. xiv A World Bank Country Study Argentina also has a strong social insurance system. Its main programs are pensions (including old age, survivors, and disability), family allowances, and a contributory unemployment insurance. To be eligible for a contributory old-age pension, men need to be older than 65 years, and women older than 60, both with at least 30 years of contributions. Family allowances are restricted to children of formal salaried workers with low- and medium-wage levels. Unemployment insurance can be claimed only by formal workers with at least six months of contributions, and it has a limit of one year of benefits. Federal social protection policies in Argentina go beyond income transfers. Other programs include in-kind transfers and, more importantly, efforts to promote stronger community development. Among those is a large and growing program to build CICs in poor neighborhoods around the country. The MDS has actively supported this program, resulting in more than 250 centers built by 2007. The centers are managed by local community councils, with participation of municipal authorities, and provide multiple services, such as basic health controls, training, social activities, and child care. The National Council of Social Policies has recently developed a plan to integrate social policies at the local level. The new program, called Plan Ahi, will focus on coordinating sector policies, including housing, roads, communication, education, health, community development, job market inclusion, NCP, culture, and others. The objective of Ahi is not to replace the existing policies and programs, but to contribute to a better coordination among them. Figure 3. Social Protection Spending at the Provincial Level, 1980­2006 (% of GDP) Source: MECON 2007. Dirección de Análisis de Gasto Público y Programas Sociales (DAGPPS) 2007. Note: Other SP = Family allowances, unemployment insurance, employment programs Executive Summary xv Provincial governments play an important role in social protection. Provincial authorities in Argentina have the power to implement their own social protection programs, including income transfers. Until the early 1990s, most of their social protection spending was on pensions, but since 1990 the resources spent on other transfers have grown consistently. By 1998 nonpension spending at the provincial level reached 0.5 percent of GDP, and in 2007 it reached almost 1 percent. Most programs at the subnational level focus on basic social assistance, nutrition, and in-kind transfers, but employment programs also have grown and nowadays make up 0.1 percent of GDP. The pension schemes, on the other hand, had a stable situation after the reforms and transfers of the mid-1990s, with spending at about 1.5 percent of GDP (figure 3). The combination of social insurance and social assistance, executed both at the federal and provincial levels, results in a very complex social protection policy, with multiple layers and instruments. Table 5.1 in chapter 2 offers a simplified overview of the main characteristics of these programs, including the rules for participation, financing sources, implementing agencies, and duration of benefits. What Happened to Income Support during the Crisis? Jefes was the government's emergency response to the crisis, which reached more than 2 million households at its peak. The program grants a benefit to households with children whose head is unemployed (or at least not formally employed) for a value that at the beginning of the plan was close to the poverty line for a male adult.1 The program combines a workfare program and a CCT program, in which beneficiaries are required to work and ensure that their children go to school and get medical checkups. Jefes succeeded in targeting the poor population and mitigating the impact of the social and economic crisis. About 80 percent of the benefits were received by the poorest 40 percent of the population, a performance that in terms of targeting is very good compared with other programs in the region. In the absence of the program, nearly 10 percent of the beneficiaries would have fallen below the poverty line. The coverage was also quite efficient, since the program reached a third of the population in the lowest quintile. The program was likewise fundamental in improving the income distribution during the first stage of economic recovery in Argentina (May 2002­May 2003). What Has Happened since the Crisis? More recently, the government has launched two new programs to gradually replace Jefes. Familias offers a basic income to households with three or more children.2 At the same time, Seguro is a noncontributory unemployment benefit accompanied by training and employment services. Many Jefes beneficiaries have transitioned out of the program because they got a formal job (close to 550,000) or because they moved into new programs (326,000) (figure 4). As a result, the total number of Jefes beneficiaries, which peaked at 2 million in 2003, was about 800,000 at the end of 2007. A core challenge is expanding the coverage of the new noncontributory income support programs. Both Seguro and Familias have operated as transition strategies for the beneficiaries of Jefes; they have not been opened to new beneficiaries who were not participants in Jefes. This represents an important coverage gap, as nearly 1.8 million households in the country meet the eligibility criteria of Jefes but are not receiving it because they did not apply in 2003. xvi A World Bank Country Study Figure 4. Plan Jefes Exit Paths of Jefes Beneficiaries, 2003­2007 7/03 11/03 3/04 7/04 11/04 3/05 7/05 11/05 3/06 7/06 11/06 3/07 0 25 % of initial stock of beneficiaries 50 75 100 Other Formal labor market Seguro de Capacitación y Empleo Familias Jefas y Jefes de Hogar Source: Staff calculations, based on data from the MTESS. The government recently implemented a strategy to expand pension coverage among the elderly. The Pension Inclusion Plan (PIP), a one-time opportunity that closed in 2007, aimed at including most of the elderly in the pension system. The plan allowed individuals who met the minimum retirement age, but not the minimum contribution requirements, to receive a pension. This provided for the one-off inclusion of approximately 1.7 million new beneficiaries in the contributory system, although it is not clear how many of them were already receiving other benefits. Figure 5 shows that, by the end of 2007, the national system was paying slightly over 4 million individual benefits, to an unknown number of beneficiaries. What Do Argentines Think about Social Protection? The general consensus in Argentina is that the current income support programs are important.3 Less than 30 percent of the population would eliminate income support programs if they could not be improved. A majority (61 percent) agreed that "programs are always needed because there are people who need them," whereas 28 percent agreed that "programs are only needed during economic crises." Similarly, 41 percent believe that poverty would be worse in Argentina without the programs, 38 percent think that unemployment would be worse, and 37 percent think that inequality would be negatively affected. Executive Summary xvii Figure 5. Evolution of Social Security Benefits Source: Moreno 2007. The surveys showed general agreement among stakeholders that the emergency programs (that is, Jefes) were appropriate for the postcrisis period and had served their purpose. However, there was also agreement on the need for a longer-term vision and strategy. Respondents did not express any particular disagreement with the transition strategy to the new Seguro and Familias programs but showed concerns about whether the programs would be enough to satisfy the existing demand for social policies. A major theme emerged surrounding the link between income support programs and the labor market. While there is debate about the level to which social programs should be linked to the labor market, both the consultations and the survey sent strong messages that employment should be an objective of social policy, and that labor market incentives and opportunities should be considered in policy design. Related to this sentiment, there was nearly universal support (93 percent in the EPPS) for contraprestaciones, or for some form of conditionality to be included in the design of social programs. xviii A World Bank Country Study Box 1. The Preference for Employment The survey of preferences on social programs, prepared for the World Bank by CEDLAS-MORI-BM in 2007, showed that the most vulnerable populations have a clear preference for employment as a source of economic security. Indeed, the survey detected that most of the respondents prefer better labor opportunities to receiving higher transfers. Out of three options of combinations between transfers and access to the labor market, the respondents (identified as vulnerable based on their employment or income status) clearly preferred the options with greater employment opportunities. Options proposed Chosen by (%) * A plan that pays Arg$400 without job training 16 * A plan that pays Arg$250 and offers job training 35 * A plan that pays Arg$150 and offers job training that 47 grants, upon completion, a certificate for a trade Source: EPPS; CEDLAS-MORI-BM 2007. Note: The percentages do not add to 100% because of nonresponses (2%). The potential beneficiary population expressed a strong preference for programs with tangible labor market outcomes. Beneficiaries prefer programs that would provide training, employment services, and ideally a guaranteed job upon completion, even if it would mean a lower benefit amount or duration of eligibility (see box 1). This finding indicates support for the type of program the government is currently introducing: Seguro. Also, it suggests that other programs, such as Familias, could be adapted to incorporate more of a labor market orientation, for example, through the introduction of training opportunities or work requirements. What Does the Future Hold for Income Support in Argentina? The future role of income transfer programs in social policy is unquestionable. Public income transfer programs for the most vulnerable, such as Jefes and Familias, as well as community development programs, have played important roles in overcoming the crisis, improving the living conditions of Argentina's poorest and most vulnerable populations, and aiding in households' accumulation of human capital in order to break out of the vicious cycle of poverty. However, a forward-looking approach is needed. Because economic growth has returned and poverty and unemployment have declined since the crisis, there is a window of opportunity to develop a long-term strategy. The consultations indicated support for programs to assist the poor, not only during times of crisis. The building blocks for a sustainable and integrated social protection system exist in Argentina, including administrative capacity and strong interest in greater involvement of the provincial level. The Government has demonstrated its capacity--through the experience of the Jefes program--to mobilize national income support programs. On the fiscal side, the unprecedented strong fiscal situation of the Executive Summary xix national government, with sustained surpluses and growing financial resources, also represents a great opportunity. An integrated social protection strategy needs to encompass three objectives: poverty relief, earning opportunities, and consumption smoothing (insurance). Though Argentina has a range of programs in place at the national and provincial levels, there are challenges ahead to ensure that each of these objectives is addressed in a coordinated, equitable, and efficient manner. In particular, there is an important need to reduce gaps in coverage--both between formal and informal workers and between regions--as some programs are offered in a few areas but not in others. This is particularly the case for poverty-targeted benefits. Although some provinces are starting to offer programs to assist the poorest (for example, Buenos Aires province), these are not part of a national strategy. The number of people who have become poor since entry into the Jefes program was closed in 2003 suggests an urgent need for a national approach to income support for the poor. What Is the Potential of Income Support to Reduce Structural Poverty in Argentina? To understand what a national program aimed at poverty relief would entail, this study simulated the impacts of four different policies. The scenarios were developed based on the visions of social policy that emerged from the consultations. The exercise found that a benefit targeted to children in the bottom two quintiles would be a potentially effective source of income support for the poorest households. To test some of the policy ideas that emerged from the consultations, the exercise simulated four scenarios of noncontributory benefits. The first was a universal benefit of Arg$400 per month for the unemployed. The second scenario simulated the impact of an income transfer of Arg$400 per month for all poor households. The third scenario considered a family allowances program of Arg$155 per month for each child under 18 in the poorest households (the bottom two quintiles). Finally, the fourth scenario assessed the combined impact of family allowances and a transfer to elderly individuals not receiving pension benefits for Arg$270 (equivalent to the amount of the PIP, or Moratoria, benefits). The simulations considered effects on coverage, poverty, and inequality, as well as the simulated programs' overall costs. It is important to note that these simulations were simplified exercises that considered income effects only and did not incorporate other potential design aspects of programs, such as training, conditionalities, or community-level development. The results of the scenarios indicate that alternative income transfer programs could be cost-effective and have an important impact on welfare. As shown in table 1, several options with costs of about 0.5­0.6 percent of GDP could have significant impacts on welfare indicators. For example, a family allowances program aimed to children in the first and second quintiles could result in a dramatic reduction of extreme poverty levels, from 8 percent to below 3 percent of the population. Other alternatives, such as universal unemployment insurance or a transfer targeted to poor households, would also be effective. Instead, expanding coverage of the elderly in addition to the family allowances would have little effect, mostly because the Moratoria benefits have already reached near universal coverage for that age-group. xx A World Bank Country Study Table 1. Simulated Policy Impacts Family Poverty Family allowances Baseline Unemployment transfer allowances and old-age (%) transfer (%) (%) (%) transfer (%) Extreme poverty 7.8 6.5 2.1 2.7 2.6 Poverty 25.0 22.4 14.5 19.8 19.8 Poverty gap 10.2 8.9 4.1 5.7 5.7 Gini 0.4765 0.4643 0.4411 0.4542 0.454 Coverage (2 quintiles) 12.0 19.0 44.0 38.0 47.0 Total cost (% GDP) 9.11 9.66 9.77 9.58 9.73 Marginal cost (% GDP) 0.55 0.66 0.47 0.61 Source: World Bank staff calculations. Note: Values are percentages except where noted (Gini coefficients). Of course, the exercise using simulations masks the complexity of implementation and the political economy involved in any reform of existing programs. The simulations represent extremely simplified models of reality, where no implementation challenges, political economy considerations, behavioral responses from the population, or other obstacles are considered. Under these scenarios, assumptions of perfect targeting, no errors or coverage gaps due to operational or design problems, and high efficiency in management are freely made. The real world consists of many immediate and potential problems, which must be carefully considered before proposing any reform. The following are some of the most relevant: Institutions. Coherent institutional arrangements for program implementation are essential in any country but are particularly relevant in the federal context of Argentina. Institutional design needs to ensure coordination among actors at the national and provincial levels and provide sound arrangements for program delivery, financing, monitoring, and evaluation. Administrative cost of targeting systems. Targeting benefits to the poor entails administrative costs, especially those that involve an income test. Categorical benefits to specific population groups (for example, children) are less expensive and easier to implement, reducing costs and inclusion or exclusion errors. Limitations in reaching beneficiaries. Even in the best-designed programs, a number of potential beneficiaries never show up to claim their benefits. This may sometimes operate as a self-targeting effect, as individuals or households that were assumed to need the transfer are not interested on it, but in many other cases the problem is insufficient access, information, or resources to apply for the benefits. Information systems. Transfer programs are information intensive and require reliable beneficiary registries and coordination of databases to manage determination of eligibility, payments, monitoring, and controls. If conditionalities are to be included in the system, an efficient system to verify compliance would be necessary. Executive Summary xxi Labor market incentives. Labor market disincentive effects of transfer programs cannot be eliminated; however, they can be mitigated. Any transfer not related to labor effort may act as a disincentive to work. The challenge is not to eliminate these effects, but to reduce their relevance, in particular in comparison with potential positive effects of the program. Organisation for Economic Co- operation and Development (OECD) countries are increasingly incorporating labor market incentives into transfer programs to incentivize employment and ease the transition from welfare to work. Existing programs. Argentina has a strong basis of existing social protection programs on which to build. Any new programs will need to take into account the existing national and provincial programs to ensure coordination of incentives and institutional arrangements. Final Reflections: What Are the Preconditions for Integrated Social Protection? The question of coverage is probably the most urgent. Familias and Seguro, as defined today, do not cover all individuals and households in need of support. The constraints are short-term fiscal restrictions and the consequence of a demanding implementation process (Seguro is not only an income transfer, but also a set of employment services that need to be set up at the local level throughout the country). On the other hand, the pension system dramatically expanded its coverage within a year, but it is not clear whether this expansion has reached all the poor elderly or if further efforts will be needed. Adequacy of benefits is also critical. The benefit levels of Jefes have been fixed since the program's creation in 2002, resulting in a real decline of more than 50 percent. The new programs provide higher benefits (around Arg$225, on average), but benefits have not been increased since the programs' creation, and no adjustment mechanism is defined for them. This problem is similar among contributory benefits (pensions, family allowances, and unemployment insurance); although in these cases there have been discretional adjustments. This strategy does not seem consistent with a long-term view, as beneficiaries should have a clearer expectation regarding the purchasing power of their benefits, and definition of benefit levels should consider policy goals. Integration is a challenging issue. Ensuring coherence across programs, levels of government, and institutions requires particular attention. Social protection programs in Argentina seem to work, in most cases, as independent systems with little interaction among the programs. Overlap between programs occurs across jurisdictions, as some provinces have schemes that replicate or duplicate those at the national level, but also between national agencies, and even within specific programs. A clear example of insufficient integration in policy planning and implementation was the recent Moratoria, a program that granted retirement benefits to nearly 1.7 million older citizens without verifying whether they already had other benefits or transfers. Because resources are limited, the example points out the critical need to improve the coordination of policies and actions between the different actors at the national, provincial, and municipal levels. Sustainability has not been a critical concern in recent years, but it is potentially fragile. Pension expenditures in Argentina have been below 8 percent of GDP since the mid-1990s. The introduction of Jefes in 2002 and Familias and Seguro later on had a small impact, as their consolidated cost barely reached 1 percent of GDP in 2003 and then xxii A World Bank Country Study declined to half that figure in a few years. The simulations presented in chapter 5 show that an additional 0.5 percent of GDP could yield high impacts on welfare and coverage. On the other hand, the recent reforms in the pension system created a potential fiscal pressure, which might end up crowding out other programs. First, the Moratoria program has an estimated cost of 1.5 to 2 percent of GDP in 2008 and is expected to decline over time. Also, the February 2007 reform produced a one-time increase in fiscal revenues as a result of the transfer of contributors to the public second pillar pension scheme, but as these contributors begin to retire, the impact will be neutral or negative in fiscal terms. The 2002 crisis showed how important it is to have a social protection system able to respond effectively to short-term shocks. The response to the crisis was effective, but most policies and programs were not in place and had to be designed and implemented in a hurry, resulting in unnecessary delays and problems. This experience has provided an important lesson for policy makers in Argentina. Rapid response to critical situations is very difficult to implement, unless institutions, programs, and instruments are already in place, providing poverty alleviation to those in need, and can be scaled up if necessary. The combination of positive trends in macroeconomic and social indicators and the consensus across society about the need to advance structural policies has opened a window of opportunity for the social protection sector. The strong economic growth of recent years and the effects of several policy initiatives have relaxed pressures on the sector, as fiscal restrictions relaxed and demands for urgent interventions diminished. Within this context, policy makers, analysts, and the population in general show a growing awareness of the need to implement long-term policies. This combination has created a rare window of opportunity for authorities to discuss, design, and implement consistent reforms to address the issues of coverage, adequacy, integration, and sustainability discussed here. The Report The objective of this report is to contribute to the discussion of options for social protection. Specifically, it focuses on noncontributory income support in Argentina through discussion of the current set of programs--in terms of their coverage (focusing particularly on the truncated system for informal and formal workers), effectiveness, costs, and sustainability. It looks at the potential impact of the government's new social protection programs--Seguro and Familias--which are in the early stages of being rolled out. The report also looks specifically at the institutional setup for social protection and the potential role of provinces in Argentina's federal system. The report presents a synthesis of findings and analysis after an intensive two-year process of analysis, dialogue, and partnerships focused on income support in Argentina. The work was carried out between July 2006 and June 2008 by a World Bank team in collaboration with a wide range of partners in Argentina. The scope of the work went beyond the analysis presented in this document and incorporated an extensive process of dialogue, consultation, outreach, and stakeholder involvement across the country. In particular, the analytical program incorporated a strong emphasis on consultation and participation. This effort aimed to supplement quantitative and institutional analysis with a discussion of the values and preferences of different population groups. The aim was to understand the landscape of ideas regarding options Executive Summary xxiii for social protection held by opinion leaders and the population. The work developed two new data sources, which are presented in this report: first, an extensive set of consultations with policy makers and practitioners in social policy, mainly at the provincial level; and second, a nationally representative opinion survey on the views and perceptions of the population regarding social policy and income support programs in particular. The contents of the report include the following. The first chapter of this report introduces the concepts of social protection and income support and provides an overview of the current system. Chapter 2 discusses the origin of programs and analysis of the effectiveness of the current programs. Chapter 3 presents the findings from the consultation process with provincial stakeholders, as well as opinions and preferences of the population toward income support policy. Chapter 4 discusses the role of provinces in providing income support. Chapter 5 presents simulations of alternative policy scenarios, and chapter 6 presents conclusions. Notes 1 Because the amount was never updated, by the end of 2007 it was approximately equivalent to half the poverty line, or the extreme poverty line for an adult male. 2 Plan Familias was in fact launched in 2002, but it remained small until authorities began to promote the transfer of beneficiaries of Jefes to it in 2006. 3 Two instruments were used to assess the population views: first, over 300 consultations with stakeholders at the provincial level, and second, EPPS. Both are discussed in chapter 3 of this report. xxiv World Bank Working Paper CHAPTER 1 1. Income Transfer Programs and Social Protection Introduction In 2001­02 Argentina was hit by one of the most severe economic crises in its history. Subsequently, it has experienced one of the most sustained periods of economic growth. The crisis was a turning point for social protection and income support, as the government rapidly expanded social assistance to mitigate the effects of escalating poverty and unemployment. At the same time, coverage of social insurance programs declined, and fiscal sustainability was weakened, with increasing unemployment and informality in the labor market. The result was a fundamental shift in the content of social protection in Argentina. As growth has returned and social indicators have recovered to precrisis levels, there is an opening to move from emergency income support programs to a more comprehensive, long-term, and sustainable strategy for social protection. Because economic expansions are transitory, this period of growth is an appropriate time to reconsider approaches to social policy. The recent policy response to the crisis provides invaluable lessons from experience and considerations of income support policy for the future. This report aims to contribute to the discussion of future options for social protection and income support in Argentina. It focuses on understanding the origins of income support in Argentina and analyzing the current set of programs--their coverage (particularly the fragmented system for informal and formal workers), effectiveness, costs, and sustainability. The report also aims at supplementing the debate about future paths, with a discussion of values and preferences of stakeholders and population groups, based on two new data sources. The objective of this program of analysis was to understand the landscape of ideas regarding options for social protection held by opinion leaders and the population. The first new data source was an extensive set of consultations with policy makers and practitioners in social policy, mainly at the provincial level. The second research instrument involved a nationally representative opinion survey on the views and perceptions of the population regarding social policy and income support programs in particular. Using quantitative and qualitative discussions of current programs and future options, the report identifies a set of policy options and simulates their impact, both in terms of welfare effects and fiscal sustainability. Although the report reviews the totality of social protection in its discussion of the current system--including income transfers within social insurance and social assistance, in-kind transfers, and community development policies--the policy simulations focus on the potential impact of different noncontributory income transfer schemes. This choice was motivated by the need to focus 1 2 A World Bank Country Study the scope of the analysis, and because the future of existing programs such as Jefes y Jefas, Familias, and Seguro is currently a central debate in Argentina. This first chapter introduces the concepts of income support and social protection. Special attention is devoted to definitions and analytical approaches for classifying the wide spectrum of income transfer programs found in Argentina. The chapter concludes with an overview of the size of the social protection sector. Chapter 2 traces the origins of programs and analyzes the effectiveness of the current programs. Chapter 3 presents the findings from the consultation process with provincial stakeholders, as well as opinions and preferences of the population toward income support policies. Chapter 4 discusses the role of provinces in providing income support. Chapter 5 presents simulations of alternative policy scenarios, and chapter 6 presents conclusions. Definitions: Social Protection and Income Support Social protection refers to the set of programs that help individuals and households manage social risks and shocks, alleviate their impacts, and reduce the effect of structural poverty on the welfare of the population. The literature contains multiple definitions of the term. For example, Bertranou and Bonari (2006) define social protection as the set of public and private interventions that aim to reduce the impact of social risks on households and individuals. Ortiz (2001) describes social protection as "the set of policies and programs designed to combat poverty and vulnerability through the promotion of efficient labor markets, diminishing the population's exposure to risk and enhancing their capacity to protect themselves against hazards and interruptions/loss of income." Arriagada and Hall (2000) employ a similar definition, recognizing social protection as the combination of both contributory social insurance and noncontributory social assistance mechanisms. The objectives of social protection are poverty relief, employment opportunities, and consumption smoothing (Barr 2008). This report draws on these definitions and uses the term social protection to refer to contributory social insurance benefits, including pensions and other benefits linked to formal employment, as well as noncontributory social assistance and social safety net benefits and services targeted at the poor. Income support is defined in this report as the set of social protection benefits-- mainly cash transfers--intended to help households and individuals cope with risks and shocks to their income, as well as cope with structural poverty. Risks and shocks can include job loss, health, childbirth, and old age, as well as other broader events such as economic or social crises and climate-related events such as droughts and floods. Income support refers mainly to cash transfers; however, some programs--such as Argentina's Seguro program--may incorporate in-kind elements alongside cash transfers, such as training or employment services. The main programs include contributory benefits (old-age pensions, family benefits, and unemployment insurance), as well as noncontributory social assistance benefits for the poor or unemployed (including workfare programs and CCTs). Health insurance, although included as part of social protection in some accounting methodologies, is not included in this analysis. Income transfer programs can be classified along various dimensions. First, they can be either insurance or assistance. Programs can both protect individuals and households against the risks of certain shocks and provide assistance once the shocks occur. Second, income transfer programs can be financed out of contributions or out of general revenues. Third, programs can provide universal coverage of specific groups, or they can be Argentina 3 targeted more narrowly based on eligibility criteria. Fourth, transfers can be provided in cash or in-kind. Finally, administration and implementation of programs can be located at the national, provincial, or even municipal level. The first classification separates the programs into insurance and assistance schemes. Whereas insurance-based systems are designed to protect different population groups from inherent risks, assistance programs provide ex post relief. Health, disability pensions, and unemployment insurance are typical examples of the former, and noncontributory assistance programs such as CCTs are examples of the latter. Social insurance programs are normally financed on a contributory basis, whereas social assistance programs are financed through general revenues. However, even when systems are contributory, it is not uncommon for them to cover their deficits from general revenues. Because of their contributory basis, insurance programs tend to cover formal workers only and have led to a "truncated welfare state," in which an increasing number of informal workers are left out of social insurance coverage (de Ferranti et al. 2000; Fiszbein 2005) Most income transfer programs are targeted in some way. Price subsidies are the only truly universal program. Most income transfers are targeted to population groups. There are three main types of targeting. The first, income (or means) testing, directs support to individuals and households below a set income threshold and may or may not include some type of assets test. The second is indicator targeting, in which eligibility is based on one or more indicators that are correlated with poverty and are easier to observe than income, such as age, household size, and geographic location. Child allowances and pensions are examples of indicator targeting. In the third, self-targeting, individuals self- select to participate in programs based on their own decisions and choices. Social assistance benefits may be self-targeted if the opportunity costs (for example, time spent on paperwork and waiting) are high and wealthier households do not bother to apply. Alongside eligibility criteria, income transfer programs commonly include contraprestaciones, or conditionalities, that may affect their level of coverage. In Argentina, contraprestaciones refers to a requirement for participation in work activities (including community services or attendance at education or training courses) for a specific number of hours a week. The Jefes program requires participants to take part in various types of work or community service. Similarly, CCT programs incorporate conditions such as requiring that children in the household attend school or go to health clinics. The levels and types of conditions are shaped by the overall objectives of the program. For example, alongside their income assistance objectives, many CCT programs aim to invest in human capital by providing incentives for school attendance and utilization of health services. Finally, programs can be implemented and financed by different levels of government and by different types of agencies. Responsibilities for delivering and financing programs can be divided, or shared, across national, provincial, and municipal governments. Governments can also contract out provision of benefits and services to outside agencies such as nongovernmental organizations (NGOs) and other entities. Decisions about the distribution of roles and responsibilities for programs can be based on various factors, including institutional arrangements, capacity, political agreements, and economies of scale. Social insurance programs are generally national because they depend on the pooling of risks and require a sizable population base. Social assistance programs are sometimes managed at the local level, to respond to local needs and facilitate targeting, although that is not always the case. However, there is a risk that horizontal inequalities may emerge if financing of social assistance programs is decentralized and poorer localities are left less able to provide for a safety net. 4 A World Bank Country Study Box 1.1. Integration of Social Protection Social protection systems in Latin America are affected by a high level of fragmentation, affecting their coverage, design, management, and financing. This fragmentation is rarely intentional, but rather the consequence of an accumulation of "geological layers" of policies introduced with different objectives over time and lacking the necessary coherence. This fragmentation has reduced the systems' effectiveness, as many programs are inefficient, poorly targeted, or inadequately financed. The situation has led to a growing consensus about the need to integrate the policies, eliminating inequalities and inefficiencies and increasing their level of internal coherence. The problem of integration and coherence of social protection systems can be considered along a number of different dimensions. First, coverage and equity. Social protection programs in the region have been traditionally targeted to formal workers, excluding those who are unemployed or informally employed. This situation, in which some individuals receive full coverage and adequate services while others are completely excluded, regardless of their vulnerability to social risks, has led to the characterization of Latin American welfare states as "truncated." In recent years, several countries have begun to expand their noncontributory schemes as a way to provide basic protection to those who were excluded. However, in most cases the quality of these schemes--in terms of access, benefit levels, or quality of services--is clearly below that of the traditional contributory schemes, maintaining the truncation in the system. Second, program design and incentives. Because social protection systems aim at providing protection to as many workers as possible, they sometimes suffer from serious problems concerning the indirect impacts of their programs design. Programs that are too generous may generate negative incentives for the labor market, or very ambitious schemes may become too expensive and, if financed by contributions, end up pushing workers into informality. The existence of multiple, autonomous social protection schemes that are not complementary, or at a minimum, coherently designed, may also result in negative impacts on the labor market. Third, institutions. The existence of different institutions without sufficient coordination may result in negative effects. This includes two main aspects: Institutional fragmentation across levels of government. In a federal country such as Argentina, many programs are run at the federal level, while others are managed by provincial or municipal authorities. While this is reasonable and can be advantageous for the system, there is a need to ensure that the different institutions coordinate their actions and strategies, avoiding gaps and overlaps in coverage (in terms of risks and individuals), and ensuring that strategies are integrated in a coherent framework. Administrative coordination. Among institutions and agencies involved in the delivery of social protection programs, administrative coordination is needed to ensure efficiency and effectiveness. Shared databases for cross-checking information on eligibility are one key element. Administrative integration can also improve delivery of services and reduce costs imposed on beneficiaries. A number of OECD countries have moved toward the approach of "one-stop shops" where benefits under the auspices of different government agencies are provided at a single service center (for example, unemployment benefits and housing). Fourth, financing of social protection. This refers to several aspects. The division of programs into contributory and noncontributory programs, in a context where a significant part of revenues come from general taxation, creates unnecessary inefficiencies in financing strategies. Also, the role of the different levels of government in financing programs may create further inequalities. For example, in Argentina some provinces run their own pension systems for civil servants, which have more generous rules than the national program, but they then receive financial support from the federal government to finance their deficits. The question of integration and coherence of social protection systems is a serious challenge. There is no ex ante reason to argue that centralization of management of all programs into one single agency is more efficient, as it depends on the characteristics of the programs. But it is clear that policies need to be coherent, leaving minimum space for inequalities, inefficiencies, gaps, and overlaps across those policies. Argentina 5 Latin American countries have been characterized as "truncated welfare states," because of the uneven coverage of population and risks. The phrase refers to the division between contributory social insurance programs that cover formal workers and noncontributory social assistance programs for the poor. The latter programs tend to be small and fragmented (Lindert 2003; de Ferranti et al. 2005, 2006). The basic idea is that countries in the region have developed their social protection systems based on contributory regimes that are closely tied to the labor market. The increases in informality and unemployment have reduced the proportion of the population who are paying into and covered by these systems, limiting both coverage and the capacity of the government to react in the face of economic downturns. The Size of Income Transfer Programs in Argentina In 2007, Argentina spent 23.6 percent of GDP on social policy, of which 11 percentage points were spent on income transfer programs. Within income transfers, contributory pensions represented the largest share of spending, at approximately 80 percent of the total, while the rest was allocated to Jefes y Jefas, family allowances, NCP, Familias (a CCT program), and unemployment insurance. Most of these transfers were financed with the federal budget, but some provinces run some smaller pension schemes for their employees, which represented nearly 1.5 percent of GDP. In 2007, federal income support spending amounted to 8.2 percent of GDP, mostly allocated to the pension system. Consolidated public spending accounted for nearly 30 percent of GDP. Two-thirds of government expenditure was allocated to social spending, while the rest was allocated to state functions (administration, justice, and defense and security), debt service, and economic services (transport, communication, energy, and so on). Spending on income support has fluctuated pro-cyclically during the past 15 years. Real spending on social protection has increased alongside GDP (figure 1.1). Given that GDP levels are more volatile than expenditures, a reverse trend is observed when measured as a share of GDP. Spending on income transfers relative to GDP declines during periods of economic expansion and increases during recessions. As income support spending is a volatile component of social protection spending, trends are less sharp when measured as a percentage of GDP. Figure 1.1 shows that social protection and GDP fluctuate in tandem, as a result of variations in spending on income support. Table 1.1. Public Expenditure and Other Subitems in Constant Nominal Terms and as % of GDP, 2007 Aggregate % of GDP Consolidated public spending 36.9 Social spending 23.6 Spending on income support 11.1 Federal spending on income support 08.2 Source: Goldschmit 2008. 6 A World Bank Country Study Spending increased with expanded coverage of beneficiaries--particularly following the 2001­02 crisis--and contracted with reductions in the real value of benefits. Lack of indexation rules for pension benefits allowed for ad hoc nominal adjustments that significantly reduced the dispersion of benefits while they increased the real value of the average benefit from 2002 onward. Both unemployment insurance and family allowances were adjusted alongside nominal wage increases after the devaluation of the local currency in 2002. Finally, Jefes has had a flat benefit fixed at its initial 2002 level. This was most likely intended to increase the opportunity cost of not participating in the formal labor market, as well as to provide an incentive for beneficiaries to move to the new programs, Familias and Seguro, which have succeeded Jefes. Figure 1.1. Evolution of Income Transfer Spending, 1994­2006 50,000 350 300 GDP (billions of 1993 constant Arg$) 250 Millions of constant Arg$ (2001) 30,000 200 22,000 10,000 0 0 1994 1996 1998 2000 2002 2004 2006 1995 1997 1999 2001 2003 2005 Year Income transfer progs. (m. Arg$ de 2001) Labor market and unemployment (m. $ de 2001) Other social spending (m. Arg$ de 2001) GDP Sources: Ministry of Economy; DAGPPS (2007). CHAPTER 2 2. Income Support Policies in Argentina: Origins and Development T he 2002 crisis was a turning point for social protection and income support in Argentina. The rapid increase in poverty, unemployment, and labor market informality dramatically increased the need for social protection and challenged the role of the existing system. The government rapidly scaled up social assistance, most notably through the introduction of the emergency workfare program Plan Jefes y Jefas shifting the balance of social protection from social insurance to social assistance. Although the economy entered a high and continuous period of growth starting in 2003, social assistance has continued to expand as a share of total social protection. This chapter sets the scene for discussions of the origins and future of income support in Argentina and the government's response to the 2002 crisis. The chapter concludes with a description of current income transfer programs, focusing on targeting and coverage. The Origins of Income Transfer Programs in Argentina Argentina has one of the oldest traditions of income support policy in Latin America. Although some small NCP for war veterans can be traced back to the colonial period, the appearance of pension funds within certain sectors and guilds in the 1880s is considered the starting point of social security in Argentina. The programs began to expand through government-mandated schemes in the early 1900s, and by the 1950s most unionized workers were covered by pensions. Family allowances began to expand by that time, also linked to formal workers. These early pensions and family allowance funds were the roots of the state-run social security system. The unemployment insurance system was introduced as part of the contributory system in the early 1990s. On the social assistance side, a pension reform in 1967 introduced a noncontributory pension regime, which remained relatively small until recently. Since the mid-1990s, national noncontributory income transfer programs have been used for different labor and social development goals. Income Transfer Programs in Argentina through 1990 Although originally social security consisted of a small number of NCP programs, its development and expansion during the 1940s and 1950s was characterized by contributory schemes, designed with a Bismarckian influence. Analyzing the history of Argentina's social protection policies, Mesa Lago (1978) identified four phases of development. Two additional phases can be added to his list. These six phases are (i) from 7 8 A World Bank Country Study the colonial period to the First World War, (ii) the interwar period, (iii) from 1945 to the mid-1950s, (iv) from the late 1950s to the mid-1970s, (v) from the mid-1970s until 1994, and (vi) from the reforms of the 1990s to the present. First, from the colonial period to the First World War, social protection was limited to isolated NCP to military and civil servants. At the outset, only disability and old-age pensions were offered on a noncontributory basis. In 1785 pensions were granted to war survivors, first the navy, and then the army. In 1803 high-ranking civil servants began to receive these same benefits (Mesa Lago 1978). These mostly noncontributory benefits continued to expand to other groups of civil servants during the late 1800s, and by 1905 the first pension schemes for private sector workers were created, covering railroad workers and other utilities. The second period corresponds to the years between the First and Second World Wars. During this period, some organized unions that played a key role in supporting the export-led growth model negotiated the creation of pension funds for their members. The contributory design of these funds created conflicts among workers; some considered these programs unfair because they required workers to contribute. Isuani (1985) suggested that the implementation of these funds was the result of collective bargaining processes, rather than a bottom-up process. Pension funds were offered as a concession during wage negotiations. Finally, other benefits such as occupational health and safety insurance were introduced to complement the disability and old-age pension schemes. Figure 2.1. Benefits and Beneficiaries of the Pension System, 1950­2007 2000 4,000,000 00 000s Arg$ (2001) 1800 3,000,000 Number of benefits 1100 2,000,000 1,000,000 400 2001 Arg$ 200 0 0 1950 1960 1968 1980 1990 2000 1955 1965 1975 1985 1994 2007 Year Pensions (old age and survival) Pensions benefit (2001 Arg$) Pensions spending (00 000s Arg from 2001) Pensions benefit trend Sources: ANSES 2007; Schultess 1990; Ministry of Labor ("Historical series" from the Secretary of Social Security). Based on a compilation of series. Argentina 9 During the third period, from the late 1940s to the mid-1950s, Argentina entered a rapid process of import substitution development. Coverage of pension funds rapidly expanded during the Peronist years, when the government introduced industry-specific funds, and urban labor markets became increasingly salaried, with union-led contracts. Also, health insurance schemes were created and managed by unions, and family allowances were slowly introduced. Independent workers also were covered by pension schemes beginning in the 1950s. The fourth period, from the late 1950s to the mid-1970s, continued the expansion of coverage and included an important administrative reform. This 1967 reform clustered the universe of pension and family allowances funds into three funds: independent workers, private sector salaried workers, and public sector, including state-run firms (see figure 2.2). The fifth period, from the mid-1970s until 1994, marks a period when the economic development strategy abandoned the import substitution model, and marks the beginning of a process of slow deterioration of contributory social security through the labor market. Informality and unemployment slowly damaged the revenue side and pushed the system into a crisis. Deterioration of the purchasing power of benefits and concentration of beneficiaries on the minimum benefit level led a large number of beneficiaries to sue the state for higher benefits. By the end of the 1980s, social security had been running deficits for 10 years (see figure 2.3). Figure 2.2. Beneficiaries of the Pension System by Administrator, 1950­2007 4,000 Number of benefits (thousand) 3,000 2,000 1,000 0 1950 1960 1970 1980 1990 1994 2007 1955 1965 1975 1985 2000 Year Self-employed Private sector Public sector Defined contribution Transferred provincial funds Defined benefit Sources: ANSES 2007; Schultess 1990; Ministry of Labor ("Historical series" from the Secretary of Social Security). 10 A World Bank Country Study Figure 2.3. Contributory Resources as a Percentage of Social Security Spending, 1950­ 2007 350 300 Percent of spending financed by contributions 250 200 150 100 50 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Dieguez y Petrecolla (1977) Secretary of Social Securtiy Schultess (1990) ANSES (2007) Sources: Dieguez and Petrecolla 1974; Secretary of Social Security; Schultess 1990; ANSES 2007. The social security system in Argentina entered the 1990s immersed in a crisis. The response was a series of reforms of income support policies, including the introduction of a contributory unemployment insurance program, the incorporation of a defined contribution scheme within the pension system, and the reform of the targeting model of the family allowance program. However, these seem not to have been enough to respond to the crisis. In fact, the sustained decline of coverage among the poorest segments of the population was one of the problems that triggered the emergency response of 2002, which expanded social assistance. The sixth and final period, from 1994 to the present, is discussed in detail in the following section. The Reforms of the 1990s Both structural and parametric reforms were applied to the traditional social insurance programs in the 1990s. Aimed at ensuring financial sustainability (or reducing deficits), most of these reforms implied a decline in coverage. Figure 2.4 shows that the aggregate number of beneficiaries from income transfer programs grew by almost 10 percent in the early 1990s, but then became stable for almost a decade. Argentina 11 Several social protection programs were introduced or reformed during the 1990s. First, unemployment insurance and workfare programs were introduced in 1991. Second, reforms of pensions (old age and survivors) included both parametric and structural components in 1994. Third, in 1996, family allowances were parametrically reformed, targeting benefits to the bottom part of the distribution of formal salaried workers. Finally, NCP were left unchanged. Unemployment insurance was launched at the national level in 1991, jointly with a series of structural reforms including privatization of state-run firms and public sector administrative reform. The system was designed as a contributory system with strict eligibility conditions that resulted in low rates of coverage. A vesting period of one year was required to receive benefits for four months, a period that increased benefits up to one year in the case of workers with three or more years of continued employment. Unemployment insurance was the smallest program in terms of the number of beneficiaries. The number of beneficiaries depends, on the one hand, on the size of the eligible population (formal workers that satisfied the vesting requirement), and on the other hand, on the level of unemployment in the economy. Figure 2.4. Beneficiaries of the Income Support Programs, 1992­2006 10,000,000 Trabajar, Jefes, Familias, Seguro, etc. 8,000,000 Unemployment insurance Number of beneftis Family 6,000,000 allowances Noncontributory pensions 4,000,000 Survival pension 2,000,000 Pensions (old age) - 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Year Source: ANSES 2007. 12 A World Bank Country Study The pension system experienced a major reform in 1994, which consisted of two central components. First, a funded, defined-contribution regime was introduced to deepen the contributory feature of the pension system. Second, a parametric reform increased the vesting period and the retirement age and changed the calculation method for the replacement rate for all workers. The new benefit formula was more progressive, with a flat amount to be paid to all retirees, regardless of their preretirement wages, and a second component that would be proportional to past contributions, or the result of converting accumulated contributions into an annuity. The new law established an indexation rule for the unfunded benefits, but that rule was eliminated in 1995 and replaced by a discretionary mechanism. In parallel, an administrative reform that had started a couple of years earlier unified all national institutions managing social security programs into a single agency, which became the manager of pensions, family allowances, and unemployment insurance benefits. Figure 2.5. Pensions and Survivors' Pensions--Beneficiaries and Average Benefits, 1970­2007 Sources: ANSES 2007; "Historical series" from the Secretary of Social Security. Argentina 13 As expected, the coverage of the pension system among the elderly declined as a consequence of the parametric reform. This decline was exacerbated by the rise in informality and unemployment that had begun in the 1980s, which reduced the density of contributions to social security. The administrative reform also led to a better record- keeping system, resulting in stricter control that reduced the inflow of beneficiaries into the pension system. Figure 2.5 shows that after the reform, the fall in the total numbers of beneficiaries of the pension system resulted from a decline in beneficiaries of old-age pensions, rather than of survivors' benefits. At the same time, after 1994, average benefits for retirees increased faster than those for survivors, which remained practically unchanged. Family allowances were reformed twice during the 1990s. First, the 1992 consolidation of pension funds into a single agency also led to the centralization of the family allowance program. Second, a parametric reform in 1996 targeted the benefits to the lower part of the wage distribution among formal workers. Three levels of benefits were defined, such that workers with lower income received higher benefits. Figure 2.6. Family Allowances and Unemployment Insurance--Number of Beneficiaries, 1992­2007 Source: ANSES 2007. 14 A World Bank Country Study By the end of the century, the effectiveness of social protection (strongly based on social insurance) for poverty alleviation was reduced by low coverage resulting from informality and the parametric reforms introduced during the 1990s. This reduced effectiveness become evident when Argentina entered the worst phase of the economic crisis. When the economic recession turned into a deep economic crisis, the policy response came through social assistance, rather than social insurance. Emergency Response and Current Income Support Policy When Argentina was hit by one of the deepest crises in its history in 2002, apart from the fall in GDP, the social consequences were quite severe. Unemployment, poverty, and inequality reached all-time highs. Informality continued its two-decades-long increasing trend. Banking and political crises accompanied the process that ended in the biggest default of sovereign debt in history. In this context, a new income transfer program, Plan Jefes y Jefas, was launched and reached 2 million households, amounting to nearly 25 percent of the population. The aftermath of the crisis found Argentina recovering quickly from the recession and entering a path of solid and sustained growth. Although the emergency seems to have passed, nearly one-third of the beneficiaries remained in the emergency program as of early 2008. Among those who left the program, some reentered the formal labor market while others shifted to other programs. On the social insurance side, major innovations have been introduced in the past few years, resulting in an expansion of coverage. NCP experienced a twofold increase since 2002. The number of traditional pensions went up by 50 percent within a year through the PIP, which closely resembles noncontributory access. This process was also marked by the discretionary indexation of benefits, the contraction of the benefit distribution, and the increase of the average benefit. The 2002 Crisis Argentina's GDP proved to be highly volatile (figure 2.7). Declining during most of the "lost decade" (1980s), GDP began to rise at the beginning of the 1990s with the inflow of international financing. Some of these resources were foreign direct investment; the rest proved to be speculative. In 1995, after the default of Mexican sovereign debt, the capital flows reverted and hit the Argentine economy. The so-called Tequila Crisis resulted in a temporary disruption of the upward phase of the economic cycle, which returned to its previous path one year later. After 1999, the economy entered a declining phase of the cycle, which turned out to be quite pronounced during 2001. The abandonment of the fixed currency scheme that governed the economic logic of the 1990s was traumatic but represented the beginning of the economic recovery. Since 2003, GDP rose at a steady rate of 9 percent, one of the most sustained periods of economic growth in the history of Argentina, supported by a favorable international context and driven by high commodity prices. The economic recession resulted in record levels of poverty. Half of the population fell into poverty in 2002 and almost 25 percent were in extreme poverty. Poverty rates at the beginning of the 1990s were below 20 percent. Although GDP fluctuated during the mid-1990s, the poverty rate rose sharply until 1996. When the economy faced a downturn at the end of the 1990s, poverty rates began to rise again. The sustained decline in GDP accelerated the fall of large parts of the population into poverty. The economic recovery brought back the poverty rate to mid-1990s levels. Extreme poverty followed a similar path, although at a lower level. Argentina 15 Figure 2.7. GDP, Poverty, and Inequality, 1980­2007 Sources: Ministry of Economy; INDEC Permanent Household Survey. Income inequality has worsened since the beginning of the 1990s. The Gini coefficient experienced cyclical fluctuations along a rising trend. Although inequality seems to have mirrored the economic cycle until 1998, it continued its upward trend while the latter fluctuated. The latest figures seem to show a reversal of the trend of increasing inequality. Nevertheless, the time span is too short to assess the trend conclusively. Unemployment reached almost one-quarter of the active population in 2002. From a relatively stable low level in the 1980s, the unemployment rate began to rise during the early 1990s, with an abrupt and temporary increase in 1995 when the Tequila Crisis hit Argentina. The sustained upward trend drove the unemployment rate up from 5 percent to 20 percent between 1991 and 2002. By 2007 the percentage of people not working but looking for a job was below 10 percent. Until 1995, unemployment and GDP seemed to be positively correlated. However, since the second half of the 1990s the indicators were negatively correlated, although GDP seems to be slightly more volatile. Inflation may have cushioned this elasticity. 16 A World Bank Country Study Figure 2.8. GDP, Unemployment, and Informality, 1980­2007 Sources: Ministry of the Economy; INDEC Permanent Household Survey. Informality, a more structural phenomenon, continued its upward trend.1 Unlike unemployment that moved countercyclically for the past 12 years, informality has followed a sluggish upward trend that seems to be independent of the economic cycle (figure 2.8). Until 2002 the evolution of informality seemed to experience a nonlinear trend, converging toward 38 percent. Although informality seems to have radically increased at the beginning of the new century, this phenomenon was explained by two discontinuous jumps (between 2002 and 2004) rather than by a true increase. First, for 2002 two different values of the indicators for informality can be computed, including or excluding beneficiaries from the Jefes program. Second, a methodological change in the questionnaire of the household survey breaks the comparability of the data over time, reflected by the discontinuous jump in the data. Finally, from 2004 onward, informality began to decline, likely as a result of the combination between favorable labor market conditions (for example, growing economy and low labor costs) and the introduction of tighter enforcement mechanisms. Argentina 17 The impact of the crisis on welfare was ameliorated by the introduction of social assistance programs in 2002. Copying the structure of the workfare program Trabajar, the strategy to combat poverty was channeled through the welfare-to-work program Plan Jefes y Jefas, a fast-response emergency program that helped address the skyrocketing levels of extreme poverty nationwide. The Social Assistance Response Active labor market programs (ALMPs) in Argentina have been financed with employers' contributions. A series of important reforms took place at the beginning of the 1990s, including privatization and trade liberalization, which led to deep transformations in the labor market. The legislation seems to have anticipated the social effect of these reforms and modified the social insurance components of social protection. Law 24.013 (also known as the National Law on Employment) launched a new institution, the Fondo Nacional de Empleo (FNE; National Employment Fund). The latter was in charge of administering the new unemployment insurance, and was also responsible for designing and implementing a series of ALMPs. The main aim of these programs was to provide employment or training while transferring income to the beneficiaries. Although access to the unemployment insurance program was subject to past contributions, the eligibility requirements for participating in ALMPs were not conditioned on similar rules. To finance these programs, the contribution rate for family allowances was reduced from 9 percent to 7.5 percent, and the remaining 1.5 percent was allocated to finance the FNE. Although several small labor market programs were launched as part of the active employment policy, Plan Trabajar is the real ancestor of Plan Jefes. Several other workfare programs were designed and implemented in the 1990s, although their coverage and impact were very limited. Figure 2.9 shows that until 2001 the peak of the series of beneficiaries had only reached 130,000. Originally divided into three groups (private, public, and training), some of the federal programs received external financing from the Inter-American Development Bank (that is, Proyecto Joven). Plan Trabajar is one of the ALMPs that outlived the initial group of programs and was later financed by the World Bank. The impact of this program was evaluated several times between 1999 and 2001. The repeated conclusion was that the program was welfare enhancing, given that the forgone income of the beneficiaries was lower than what they actually received from participating (Ravallion, Galasso, and Salvia 2001; Jalan and Ravallion 1999; Ravallion et al. 2001). Jefes represented the national response to the crisis and was aimed at mitigating poverty. Officially launched in January 2002, this national strategy enrolled up to 2.2 million beneficiaries by May 2002, when entry was closed. Plan Jefes paid (and still pays) Arg$150 per month and required beneficiaries to either work or finish incomplete school cycles (see figure 2.9). Eligibility conditions intended to target the benefits to families with unemployed household heads with at least one child below 18 years old. However, in practice, any person that could prove to have a child of that age, and who was not formally employed or receiving a state benefit, was a potential beneficiary, since the cross- checks could not differentiate whether he or she was unemployed, inactive, or even informally employed. This particularity helps explain the larger share of women among the beneficiaries. 18 A World Bank Country Study Figure 2.9. Cash Transfer Programs (CCTs and Workfare) from the Federal Government--Number and Amount of Benefits, 1993­2007 Source: Ministry of Labor; labor statistics bulletin. Jefes introduced a series of changes with respect to Trabajar and previous employment programs. First, within the Jefes program, coverage is national and the geographical distribution is determined from the bottom up, as opposed to the more centralized allocation model of Trabajar. Second, within Jefes, payment of benefits is administered by ANSES and paid by the Banco Nación through a debit card (a model that was fully implemented after the program started), whereas Trabajar benefits were paid in cash at financial institutions. Third, within Jefes, local municipalities are responsible for arranging contraprestaciones and verifying compliance with education conditionalities, but they have limited or no authority or capacity to enforce compliance (they can request that federal authorities drop beneficiaries, but rarely do so).2 However, within Trabajar, municipalities had the authority to reject beneficiaries in case of noncompliance with conditions. Fourth, Plan Jefes was launched explicitly as a permanent right (of participants) rather than as a temporary program as Trabajar was. Finally, Jefes had no turnover of participants, since enrollment took place only at the beginning until the entrance to the program was virtually closed, while Trabajar beneficiaries could rotate when new projects were started. Argentina 19 Several evaluations highlight the positive impact of Jefes. MTESS (2002) states that participant households escaped extreme poverty and improved their income more often than households not participating in the program. Galasso and Ravallion (2004) focused on the welfare-enhancing features of the cash benefit, noting that targeting was better than national spending but worse than for Plan Trabajar, and they emphasized the importance of enforcing eligibility conditions. Other analyses of Jefes and its influence on the labor market and welfare in Argentina show interesting results. Among those, Monza and Giacometti (2003) found that a significant number of the women (the majority of beneficiaries) participating in the program were previously part of the economically inactive population. Also, Cetrángolo and Jimenez (2003) point out that Plan Jefes represented one of the most radical changes in terms of social protection in Argentina, as it did not depend on the availability of international financing, and they open the decentralization debate and discussion regarding the roles of provinces in the administration of social programs. Others in the literature present a more skeptical view of Plan Jefes. Barbeito, Giosa- Zuazua, and Rodriguez-Enriquez (2003) criticized the regressive feature of paying a flat benefit, regardless of the size of the household, highlighting that benefits are too low to combat poverty. Garcia and Langieri (2003) suggested that the veto power of municipal mayors may be counterproductive for the consejos consultivos, the local-level institutions that are in charge of allocating the beneficiaries to different contraprestaciones. Figure 2.10. Plan Jefes: Exit paths of Jefes' beneficiaries, 2003­2007 7/03 11/03 3/04 7/04 11/04 3/05 7/05 11/05 3/06 7/06 11/06 3/07 0 Percent of original stock of beneficiaries 25 50 75 100 Other Formal labor market Seguro de Capacitación y Empleo Familias Jefas y Jefes de Hogar Sources: Ministry of Labor; Monitoring and Evaluation bulletin of Plan Jefes. 20 A World Bank Country Study One of the most prominent debates in the literature revolves around the universal character of Plan Jefes. Pautassi (2003) claimed that Jefes was far from being a universal social right (as stated on the presidential decree that launched the program), but rather represented a short-term measure with political intentions. Stressing the fact that entrance to the program was closed in May 2003, Pautassi postulated that a truly universal right should not be conditional and that the benefits should guarantee a certain minimum level, for which a reform on the tax structure should be passed to ensure sufficient financing. Six years after the introduction of Plan Jefes, the program is still active and pays between two and three times more benefits than the maximum level of its predecessor, Plan Trabajar. Figure 2.10 shows that the main exit path from Jefes was the return of participants to the formal labor market, which accounts for a quarter of the initial stock of beneficiaries (2 million). A little more than 10 percent of beneficiaries have switched to the Familias program since April 2005. In May 2006, the first round of beneficiaries switched from Plan Jefes to Seguro, reaching almost 3 percent of the total stock of beneficiaries. The latter began transferring beneficiaries later than Familias, and the flow seems to be slower. Nearly 15 percent of Jefes beneficiaries left the program for a number of reasons, including graduation (the children of the beneficiaries are no longer are below 18 years old) and tighter controls. If the exit trend continues at the current level, Jefes might disappear in two or three years. The average reduction rate of beneficiaries is 22,600 beneficiaries per month, or 1.6 percent of beneficiaries per month. Nevertheless, a large part of this outflow is explained by the migration toward Familias and Seguro. Considering the set of three programs, these cash transfers would run out of beneficiaries in 2013 or 2015. If migration to other programs is not considered, the average reduction of beneficiaries was 16,600 persons per month, or 1.1 percent each month. However, these raw estimates are based only on past flows of beneficiaries, and labor market absorption rates they reflect would have taken place in a period of high and sustained economic growth. On the other hand, inflation has reduced the purchasing power of benefits. Purchasing power seems to have been nominally fixed as an incentive device to move people out of the Jefes, or toward the alternative programs. Figure 2.9 shows that, as of late 2007, the real value of benefits was equivalent to Arg$90 (in 1994 pesos), almost half of its original value at the beginning of 2002 when Jefes was launched. Familias, like Jefes, has not established a closing date, and the outflow of beneficiaries seems to depend on the graduation of beneficiaries (when their children reach 18 years old) or on the voluntarily opt-out offered by the program. The bias toward women with a large number of children is explicit, both in the actions taken by the MDS (in charge of the administration) and in the design of the program. On the other hand, Seguro, administered by the Ministry of Labor, requires beneficiaries to formally express that, by subscribing to the new program, they quit participation in Jefes and that they are aware that Seguro has a maximum duration of two years. Familias, set as a CCT, requires beneficiaries to send their children to school and to medical checks. Seguro, by comparison, is set up as a workfare program that asks participants to attend a training course as well as to enroll in public employment services. Argentina 21 Box 2.1. The Transition of Plan Jefes, Familias, and Seguro MTESS (2004) carried out an evaluation of Plan Jefes, concluding that the stock of beneficiaries was heterogeneous and that four groups could be identified. First, the group of beneficiaries that remained inactive (around 30 percent) were clearly differentiated from the other three groups of active beneficiaries. Those three groups included youngsters with low to median levels of formal education with labor experience (20 percent), young adults with similar levels of education but without labor experience (20 percent), and older workers with low education levels and no labor experience (30 percent). In 2005 the federal government launched a strategy to migrate beneficiaries from Plan Jefes to other programs with the objective of generating more homogeneous groups. In the case of the two programs Familias and Seguro, beneficiaries were invited to voluntarily move out of Jefes into the new program. The voluntary process reflects respect for both the beneficiary and the limitations of forcing participants out of a program that was formally stated as a social and universal right. Political activism on the part of some the beneficiaries (also known as piqueteros) may have represented a potential threat that induced authorities to carry out the migration process as smoothly as possible. Moreover, these legal and political factors reflect the difficulties and low probabilities of closing the program abruptly, as well as an explanation of the persistency of an institution like Plan Jefes. The two new programs have some important differences. Both offer beneficiaries higher benefits than what they would receive if they stayed with Jefes. Familias pays Arg$125 plus Arg$25 per child, and Seguro pays a flat benefit of Arg$225, which represents a 50 percent increase with respect to Jefes. The beneficiaries' transition to both programs is voluntary, and the switch is carried out by each municipality according to a schedule that is planned to cover the whole territory progressively. This territorial approach is another explanation for the pace of the transition. In the 1990s, provincial governments started developing their own income transfer programs. This process resulted in a spectrum of programs including complementary programs targeted to households and individuals already receiving Jefes benefits. For example, in 2003 San Luis launched a workfare program that included more beneficiaries than the number of Jefes beneficiaries in the province. At the beginning of the program, the benefit was Arg$300, double the Arg$150 paid by Jefes, and was later increased to Arg$450. In the province of Buenos Aires, three employment programs, which covered as many as 55,000 beneficiaries by 2000, provided training and basic income support. More recently, the beneficiaries of a food program called Mas Vida were switched to the Plan Tarjeta Alimentaria, which provides them with a debit card that can only be spent on food. Similarly, the program Ciudadania Porteña, from the city of Buenos Aires, provides a debit card that can only be used in a predetermined network of shops. This program pays varying amounts that depend on household size and composition, and ranges from Arg$80 up to Arg$800. The 70,000 beneficiaries receive on average Arg$250. A more detailed description of the subnational social policy is included in chapter 4. Finally, the traditionally marginal NCP program has grown to nearly 500,000 beneficiaries in recent years. This program consists of three types of benefits. First, the so- called social protection types group old-age and disability beneficiaries, which traditionally accounted for one-third of all NCPs (figure 2.11). The MDS is in charge of determining eligibility conditions. This type of pension explains most of the increase in total NCPs since 2002. Second, the so-called graciables are NCPs granted by the members of the parliament through discretionary mechanisms. The third and last group consists of pensions granted under special laws and comprises Nobel laureates, Olympic medalists, bishops, war veterans (from the Malvinas war), and others. 22 A World Bank Country Study Figure 2.11. NCP--Number and Amount of Benefits, 1989­2007 Source: Bertranou and Grushka 2002; Secretary of Social Security, Social Security statistical bulletin. Unlike other noncontributory cash transfers (such as Trabajar, Jefes, Familas, and Seguro), by 2007 the purchasing power of NCP benefits was nearly 50 percent higher than what it was during the 1990s. In fact, the nominal value of the average NCP is similar to the minimum contributory pension and survivors' benefits. The large expansion of the program from 2003 onward started a trend consistent with the PIP, which represented the most significant change in social protection following the introduction of Jefes. The Social Insurance Innovations The social insurance programs experienced two different types of reforms in recent years. The most important one was the PIP, also known as the Moratoria, which incorporated 1.7 million new beneficiaries into the pension system in less than a year (Box 2.2). The other change comprised a set of parametric reforms to the pension system, as well as modifications of the balance between the different pillars of the contributory pension program. Argentina 23 Box 2.2. The Moratoria The PIP is a set of legislative acts aimed at making the contributory requisites to access a pension more flexible thereby increase coverage. Law 24.476 (from 1995) allowed independent workers with incomplete contribution records to "buy" the missing contributions. This debt (maximum of 10 years) could be paid in several installments (up to 60), which could be deducted from the amount of the pension, resulting in a lower net benefit. This law allowed beneficiaries to buy the missing contributions corresponding to the previous pension law (18.036 and 18.037), which was valid from 1967 until 1994. Law 25.865, passed in 2003, set a similar moratorium system for several taxes, including pension contributions, which expired in 2005. Law 25.994, also known as the law of early retirement (jubilación anticipada), was passed in December 2004 and represents the real Moratoria. Copying the mechanisms of the previous laws, this time the potential beneficiary was allowed to buy years regardless of the period (both Law 18.038 from 1967 and 24.241 from 2004). Independent workers (everyone eligible, in practice) who were of pensionable age by November 30, 2004 (age 65 for men and age 60 for women), were eligible to buy up to 30 years of contributions (the whole vesting period). The enrollment period was first set to close in December 2006 but was later extended until April 2007. Since claimants needed to be registered as autonomous workers, and the debt (to be discounted) needed to be estimated while taking into account its category, the vast majority of beneficiaries enrolled themselves in the lowest one. Thus, they became eligible for the minimum pension, like most of the pensioners at that time. In practice, the payments of the Moratoria represent a wedge between the minimum (gross) pension and the net pension. A still determined share of new beneficiaries were already earning another type of pension. The Moratoria increased the number of pensioners by 60 percent within one year. Overall, the pension system went from 3 million beneficiaries to almost 5 million (considering the rise in NCP) in from 2005 to 2007. This size jump (see figures 2.1, 2.2, and 2.5) has never been seen in the history of the Argentine pension system and has opened the debate around the universalization of pensions and income transfers. The exact impact of the new program on coverage remains unknown. Authorities have declared that coverage among the elderly reached 95 percent (65 years and above). Nevertheless, this might be an overstatement, since a significant number of new benefits could have been granted to individuals who were already receiving survivors' pensions, information that remains unknown. The larger the overlap in beneficiaries of the two programs, the smaller the impact on coverage. As a result of the PIP, three different types of pensions are being paid under the umbrella of the same system. First, people that met the eligibility conditions receive a normal contributory pension. Second, people with incomplete contributory histories receive proportional contributory pensions, even though no such pension was stipulated on the last pension reform law (24.241). Third, people who are at retirement age but lack a contribution history receive pensions, although smaller amounts. The latter could easily be considered to be universal pension that is granted to people above the pensionable age, with no means test but as a self-targeting mechanism. Another reform of the pension system took place in early 2007, as the result of a law reform promoted by the government. This reform included several changes in the parameters and institutional structure of the contributory system. Among them, three modifications appear to have had the most significant impact. First, the flow of new beneficiaries to the defined contribution scheme has decreased drastically. Previously, new workers not declaring their preference between defined contributions and defined benefits were automatically enrolled in the defined contributions scheme. This default 24 A World Bank Country Study option was reversed, implying that 90 percent of new entrants to the labor market will now go to the defined benefits scheme every month. Second, workers contributing to the defined contributions scheme were allowed to switch back to defined contributions. Third, contributors older than 50 years of age with an accumulated fund below Arg$20,000 were automatically transferred to the defined benefits scheme. The absence of a pre-established indexation rule for the pension benefits resulted in some changes in the structure of benefits. First, the gap between the average and the minimum benefit has narrowed since 2003 because of discretionary increases to the minimum benefit applied since then. Within a relatively short period, pensioners received flat increases, proportional increases, proportional increases with caps, and increases of the minimum benefit. As a result, the benefit pyramid became flatter, with more beneficiaries at the lower benefit levels. By early 2008, nearly 70 percent of pensioners received the minimum benefit. Second, the real value of average benefits recovered from the decline produced by the 2001­02 crisis and surpassed the highest level of the benefit reached during the 1990s. Figure 2.12 shows that average benefits in September 2007 were higher than in December 2000 by 7 percent in real terms. The effect is even stronger for the minimum benefit, which has grown by 110 percent since 2000. Figure 2.12. Average and Minimum Real Benefits, and Percentage of Beneficiaries Earning the Minimum Benefit, 1977­2007 100% 350 90% 80% 300 70% 250 % of benefits at the minimum pension real AR$ 60% 200 50% 40% 150 30% 100 20% 50 10% 0% - Jul-94 Jan-95 Jul-95 Jan-96 Jul-96 Jan-97 Jul-97 Jan-98 Jul-98 Jan-99 Jul-99 Jan-00 Jul-00 Jan-01 Jul-01 Jan-02 Jul-02 Jan-03 Jul-03 Jan-04 Jul-04 Jan-05 Jul-05 Jan-06 Jul-06 Jan-07 Jul-07 1977 1983 % of benefits at the minimum Real minimum benefit Real mean benefit Sources: Schultess 1990; Secretary of Social Security, Social Security Statistical Bulletin 2007. Argentina 25 The inflationary process that started after 2002 demonstrated that the parameters and indexation rules of the social insurance components were designed for different macroeconomic contexts. Apart from the aforementioned innovations of the pension benefit structure, both the family allowance and the unemployment insurance needed to be parametrically updated. The family allowance system, after a reform in 1996, was targeted on the lower end of the wage distribution, when workers earning more than Arg$1,500 were declared not eligible for benefits. Although inflation since 2002 resulted in an increase of nominal wages, the eligibility parameters and the amount of the benefits were nominally fixed, and many workers became ineligible for the family allowance benefits. Two discretionary adjustments introduced since 2002, modifying both the amount of the benefit and the eligibility parameters, partially corrected this effect. Nevertheless, no indexation rule was established. The contributory unemployment insurance was also adjusted to include changes aimed at easing access to the benefits and extending the length of the eligibility period. Since its origins in the early 1990s, unemployment insurance paid a benefit equivalent to half the worker's wage for up to 12 months, with a minimum of Arg$300 and a maximum of Arg$600. The payment of the benefit decreased over time, and the worker was eligible to receive the benefit for 4, 8, or 12 months, depending on whether the worker had a contribution record of 12, 24, or 36 months continuously with the same employer. The simple parametric changes involved widening the wage range to Arg$250­400 and shortening the minimum vesting period to six months (entitling the beneficiary to two months pay). Workers older than 45 can automatically extend the payment period for six months, with a benefit equivalent to 70 percent of the original amount. Current Income Support Policy As of 2007, Argentina had five national income support programs: (i) pensions (both retirement and survivors'); (ii) family allowances; (iii) unemployment insurance and the work-related benefits (Jefes and Seguro); (iv) a CCT (Familias); and (v) NCP. The pension system is by far the largest program in terms of spending, at 8.5 percent of GDP, and covers 4.6 million beneficiaries. The family allowance program is as large in terms of beneficiaries, but represents a much smaller level of spending given the size of the benefit (Arg$111 on average). The last contributory benefit is unemployment insurance, which on average paid 93,000 beneficiaries in 2007 (see table 2.1). Table 2.1. Income Transfer Programs in Argentina, 2007 Income transfer Average benefit Spending Access and program Benefits (Arg$) (% GDP) financing Pensions 4,658,526 645 8.47 Family allowances 4,150,576 111 1.30 contributory Unemployment insurance 92,345 340 0.09 Familias and Seguro 1,280,781 180 0.65 noncontributory Noncontributory pension 571,415 468 0.75 Source: ANSES 2007. 26 A World Bank Country Study Noncontributory social assistance programs amounted to 1.4 percent of GDP in 2007. The traditional noncontributory pension program pays Arg$468 to 570,000 people. Jefes, Familias, and Seguro together are as large as the NCP, with larger coverage and lower benefits. In 2007, these three programs together paid an average benefit of Arg$180 to almost 1.3 million beneficiaries. Unfortunately, it is not possible to distinguish contributory pensions from NCP. Within the latter category, Jefes expenditures were 0.13 percent of GDP, down from as much as 0.96 percent in 2003. This decline is explained by the combined effect of a reduction in the number of beneficiaries and increases in nominal GDP. Whereas pensions are the most effective programs in terms of poverty reduction, family allowances are most efficient among the existing income transfer programs. Figure 2.13 shows that, without pensions, poverty rates would be 10 percentage points higher than their current levels. Similarly, family allowances would reduce the headcount ratio about 2 percentage points, and together, Jefes, Familias, and Seguro would reduce the poverty rate about 1 percentage point. Nevertheless, the pension program is six times larger than family allowances. The latter implies that, for pensions, the reduction of 1 percentage point of the poverty rate costs 1 percent of GDP, while family allowances achieve the same results with 0.62 percent of GDP. Jefes, Familias, and Seguro are the most efficient programs at reducing extreme poverty. This simple cost and benefit analysis raises the question of whether, from the perspective of poverty reduction, scaling up the pension program was the best option. Figure 2.13. Potential Poverty Rate Increase in the Absence of Income Transfer Programs, 1997­2007 percentage points of poverty rate 0 3 6 9 12 15 Unemployment insurance 1997 Family allowance 2001 Trabajar, Jefes, 2003 Familias, Seguro Pensions 2005 2007 All ITP Source: World Bank staff calculations based on INDEC EPH. Argentina 27 Gaps in coverage have been relatively constant over the past decade, while overlap of benefit programs remained at fairly low levels. Figure 2.14 reflects that nearly 40 percent of households did not receive a benefit from any of the income transfer programs between 1997 and 2007. This level was almost 10 percentage points higher for the first quintile of the per capita income distribution. During this decade, only 3 percent of households received more than one benefit from the programs. The introduction of Plan Jefes reversed the growth of the coverage gap in 2003, as it succeeded in reaching the bottom of the income distribution. In 2001, 60 percent of households in the first quintile of the distribution did not receive income support programs. The massive expansion of Jefes reduced this gap to 40 percent two years later. Existing income support programs are not adequately focused on the poor, and coverage has been declining in recent years. Although the share of the population not receiving benefits appears constant on aggregate, the share of the first quintile not receiving benefits increased between 2005 and 2007. Two main factors are driving this. First, because the social assistance programs--Jefes, Seguro, and Familias--are all closed to new entrants, the share of the poor left out of the programs is growing relative to the share receiving the benefits. Second, the radical scaling up of pension provided large transfers to parts of the population not at the bottom of the income distribution. Figure 2.14. Number of Income Transfers Received by Households--Total and First Quintile, 1997­2007 All households 1st quintile 100 75 50 25 0 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 Gap Coverage Overlap Source: World Bank staff calculations, based on EPH-INDEC. 28 A World Bank Country Study While coverage of family allowances has increased among the poor, coverage of the social assistance programs has declined because of exits from Jefes. Figure 21/2.15 indicates that 30 percent of households received benefits from family allowances, a level of coverage similar to pensions. In 2007 the coverage rate for Jefes, Familias, and Seguro declined to 7 percent, from its peak of 13 percent in 2003. Coverage of unemployment insurance is negligible. The factors that explain the different trends in coverage of family allowances are quite different from those behind the trends in the social assistance programs. While the percentage of poor households receiving family allowances declined, the same indicator increased for nonpoor households. In the case of Jefes, Familias, and Seguro, the aggregate coverage rate went down when poor households left Jefes. In other words, the percentage of poor families that do not participate in Jefes, Familias, and Seguro is increasing because of the combination of poverty transitions and closed entrance to the social assistance programs. Figure 2.15. Percentage of Households Receiving Program Benefits by Poverty Condition, 1997­2007 Trabajar, Jefes, Familias, Seguro Family allowances 0.400 0.400 0.300 0.300 0.200 0.200 0.100 0.100 - - 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 All households Poor Non Poor All households Poor Non Poor Pensions Unmeployment insurance 0.400 0.400 0.300 0.300 0.200 0.200 0.100 0.100 - - 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 1997 2001 2003 2005 2007 All households Poor Non Poor All households Poor Non Poor Source: World Bank staff calculations, based on EPH-INDEC. Argentina 29 Despite declining coverage, the social assistance programs are the most pro-poor of the income transfer programs in Argentina. The Lorenz curves in figure 2.16 show that although pensions were more equally distributed than per capita income, both family allowances and unemployment insurance were more targeted to the bottom of the distribution. 3 Because of the low level of coverage, the number of observations of unemployment insurance in the household survey is too low to be robust. The fact that family allowances are above the imaginary 45-degree line reflects that the lowest percentiles of the distribution receive a more than equitable fraction of the program. On the other hand, the concentration curve for pensions is below the equity line, reflecting that the lowest percentiles of the distribution receive a transfer that is less than proportional, concentrating the people that receive the benefits at the middle and upper parts of the distribution (probably owing to the receipt of the benefit and the size of those households). Figure 2.16. Lorenz Curves for Per Capita Income and Income Transfer Programs, 1997­ 2007 Source: World Bank staff calculations based on EPH-INDEC. 30 A World Bank Country Study Table 2.2. Income Transfer Programs in Argentina--Basic Elements as of 2007 Duration of Financing Implementing Program Description benefits Eligibility source agency Retirement age: 65 for men, 60 for Contributory National Multipillar old-age, women; vesting (employee and ANSES, Pension survivors, and Indefinite period: 30 years; for employer); deficit AFJPs System disability benefits disability and covered through survivors: vesting general taxes period of 18 months PAYG old-age, Retirement age and Contributory Provincial survivors, and Provincial vesting period: varied (employee and pension disability benefits Indefinite pension depending on the employer systems for provincial civil agencies province contribution) servants seven benefits For child ANSES (pregnancy, birth, benefits, Contributory (private Family adoption, per Formal salaried until they (employer sector), allowances children, disabled workers reach 18 contribution) government children, school years (public sector) support, marriage) Unemployment Formal salaried Unemploy- benefit, decreasing workers not in the Contributory 4 to 12 ment over time; agriculture or public (employer ANSES months insurance compatible with sector; vesting period: contribution) family allowances 6 months Beneficiary of special law (war veterans, Three benefits: old Nobel laureates, NCP age, special laws, Indefinite Olympic medalists, Noncontributory MDS-ANSES social protection etc.); for old-age pensions, older than 70 and poor Cash transfer program; While Unemployed head of counterpart work Jefes de children are household with MTESS- ranging from Noncontributory Hogar younger children younger than ANSES completion of than 18 18 schooling cycles to community work CCT conditional on Unemployed head of school attendance household with two or Familias Indefinite Noncontributory MDS and medical more children checks for children younger than 18 Transfer program, conditional on Unemployed head of Seguro de training and active household with Capacitaci Maximum labor search (public children younger than Noncontributory MTESS ón y two years employment 18 that participated in Empleo services the Jefes program enrollment) Cash transfers, Provincial linked to poverty Income Varies by Provincial alleviation, food Varies by program Noncontributory Transfer program governments distribution, and Programs workfare policies Source: World Bank staff based on legislation. Note: AFJPs = private pension funds; PAYG = Pay As You Go. Argentina 31 During the past decade the two largest income transfer programs have become more equally distributed, although for different reasons. While pensions approached the equity line from below, the family allowances curve did so from above. In the case of pensions, the expansion of coverage and, in particular, the flattening of the benefit structure, made the distribution of pensions more equitable. Family allowances, on the other hand, reflect that the reduction of informality expanded the coverage of the program, especially among the bottom of the distribution. The program's benefit is nearly flat. Jefes, Familias, and Seguro are still the most pro-poor programs, although the line becomes more unstable as a result of declining coverage. The accumulation of layers of income transfer programs throughout the years has resulted in a complex sector with many components. Table 2.2 presents a concise review of the basic characteristics of the most important programs, providing a description of the program, the duration of benefits, eligibility rules, financing source, and implementing agency for each program. Notes 1 The phenomenon of informality in Argentina is discussed in World Bank (2008a). 2 Contraprestaciones refer to the requirement for participation in work activities (including community services or attendance at education or training courses) for a specific number of hours a week. 3 Lorenz curves represent the cumulative frequency for each income transfer program by per capita income percentile. CHAPTER 3 3. Opinions and Preferences Toward Income Support P olicy outcomes are influenced by policy makers and the voters who elect them. Stakeholders contribute to the policy-making process by articulating their values, opinions, and preferences, as well as their experiences as beneficiaries, citizens, and participants. As part of this study, two major efforts were undertaken to listen to, understand, and document the views and experiences of stakeholders and the population at large with regard to social policies and income support, in particular. The first effort was a series of consultations, mainly at the provincial level, with key actors involved in the delivery of social policies. The second effort was a nationally representative opinion survey aimed at capturing the views of the general public, including beneficiaries of social programs, regarding social protection and income support. On its own, public opinion cannot determine policy choices and program design. However, it can be a useful input alongside considerations of other factors, including the effectiveness of existing programs and institutions, cost-benefit analysis, the availability of fiscal resources, objective analysis of needs (for example, poverty analysis), administrative feasibility, among other issues (see box 3.1). There are reasons why public opinion analysis of social policy in Argentina could be particularly informative. Policy makers and the population at large have recent experience with and awareness of the emergency social programs in Argentina, which means that the population is likely to have views about the effectiveness and future of these programs. These views could contribute both to discussion of the objectives of future policies and to perspectives about what works best in terms of implementation. Both the consultations and the opinion survey looked at the experiences that stakeholders in Argentina have had with social programs in the recent past. They also raised questions about priorities and options for future policies. The surveys provide a rich information base for considering different policy scenarios. Priorities and preferences can contribute to the policy dialogue and provide a channel for broader participation in the policy process. However, although perceptions and views expressed in the surveys are vital inputs for the policy-making process, they are only one factor to consider. This chapter summarizes the findings of the stakeholder consultations and the national opinion survey and discusses the implications for policy. The first section reviews the data sets; the second introduces a typology of three visions of social policy identified in the consultations; the third discusses the level of knowledge about existing programs within the population; the fourth reviews perspectives about programs; the fifth looks at views about program design and preferences for policy scenarios; and the last section presents conclusions. 32 Argentina 33 Box 3.1. Do Opinion Surveys Matter? There is much debate about the significance of public opinion in the policy process. Though some authors argue that public opinion is a social construct without intrinsic meaning, others interpret it as a map of society that has its own relevance, based on the interaction between the views of individuals and their interchange with the multiplicity of opinions that exist in the public space. Within this second point of view, public opinion is conceptualized as an actor that can influence political processes and deliver and communicate political messages. In this regard, public opinion has a permanent relationship with two other important actors: political elites and the media. These three actors aim to influence each other, and at the same time they are influenced and transformed by the opinions, ideas, and conflicts that circulate among them. Each can affect the others in distinct ways--through images (both superficial and profound, based on values or attitudes), provocative ideas, dynamic interchange of ideas, and so forth. Within this theoretical perspective, opinion surveys are considered to be one of a set of tools that can communicate public opinion. Surveys can play the role of transmitting public ideas and interests to the other actors involved in political communication (for example, political elites and the media), as well as reflect public reactions to the messages given by these actors. In this way, surveys can be legitimizing instruments for orienting political decisions. In the debate about how much influence these processes actually have on the policy process, opinions range from limited effects (Lazarsfeld) to a high level of impact (Noelle-Neumann), which find an insignificant impact of public opinion in influencing any aspect of the policy process. Even though some theorists are pessimistic about influences of public opinion on policy makers (as public opinion is seen to be too incoherent and volatile), other authors have thought of public opinion, expressed in surveys, as internally coherent, nonvolatile, and predictable and thus an important consideration in the decisions of policy makers. Empirical evidence on this latter view is not conclusive, but some moderate correlations have been demonstrated (see Erikson 1981; Page et al. 1983), showing that it is likely that constituents' opinions do affect the votes of elected representatives. Regardless of the theoretical debate, it is clear that survey quality--particularly sample selection, questionnaire design, data collection methods, and processing--is a determining factor that ensures that the results accurately reflect the views expressed by the respondents. Results of opinion surveys can have a dynamic of their own owing to measurement factors and underlying social processes, and can help to interpret and understand a complex social reality. Conversely, they can be artificially manipulated to define so-called majorities and minorities that do not necessarily exist outside of the page of processed data. Perhaps an even more important point to emphasize is that defining the ultimate views and desires of the majority within society is not a task of those who study public opinion: surveys can never attempt to replace the work of electoral processes and collective participation. Regardless, surveys can be an important tool for policy management and communication in that they do not attempt to illustrate preconceived prejudices, but rather offer a snapshot of the distribution of attitudes and ideas (and prejudices) of the public at a fixed point in time, with the recognition that this snapshot is under the influence of constant change and does not contain more or less truth or legitimacy than other forms of expression circulating in the society. Sources: Lazarsfeld 1948; Noelle-Neumann 1984; Converse 1987; Page and Shapiro 1992; Erikson 1981; Page et al. 1983. Data and Approach The findings in this chapter are based on two instruments for gauging perceptions and experiences with social policy and income support in Argentina. The first is a process of consultations at the national and provincial levels; the second is a national opinion 34 A World Bank Country Study survey. The consultations were conducted first, at the end of 2006, to map out the range of views among stakeholders at the provincial level. The consultations also included a number of selected interviews with key figures in social protection at the national level. The results of the consultations inspired the design of the national opinion survey, which was fielded between June and August 2007. These consultations found converging views on the intersection of employment and social policy, on views toward cash benefits and contraprestaciones, and on the role of provincial governments in social policy making. The questionnaire for the opinion survey was formulated around the issues that emerged from the consultations, with the aim of using the design of the survey to test and quantify the extent of these perceptions among the population. The stakeholder consultations were undertaken by a team led by the UTDT, in close collaboration with the World Bank.1 The consultations were held between September and December 2006 in 10 provinces.2 Over 300 individuals were interviewed, including provincial and municipal government officials; representatives from NGOs, trade unions, the church, and regional universities; and other key local officials. The interviews were open-ended following a common interview guide. More information on the approach of the consultations is included in appendix 2. The national opinion survey, the EPPS, was conducted between mid-June and mid- August 2007.3 It was undertaken by a joint team of Equipos MORI Argentina and the CEDLAS. The EPPS included a survey of 2,500 cases and was designed to be representative at the national and regional levels. In each case, either the head of the household or spouse was interviewed, depending on their availability. In addition to the national sample, the survey included a module of questions that was applied to a subsample of 600 potential beneficiaries of social programs, to deepen the analysis among the target populations. Selection of interviewees into the oversample was based on self-reported income. The survey questionnaire included a submodule for the oversample that included specific questions relevant to the beneficiary population on adequacy of benefit amounts, contraprestaciones, the duration of programs, and views regarding training and other forms of labor market insertion to accompany program design. Three Visions of Social Policy Though the provincial consultations yielded a diverse set of views, the spectrum of ideas can be synthesized into three visions of the future of social policy in Argentina. The visions are an idealized typology and are not mutually exclusive--some interviewees expressed views that overlap across the three types--however, they are a useful tool for characterizing the wide-ranging views that exist in the country toward social policies. The three visions can be summarized as follows: Rebuilding the social fabric. The first vision is grounded in the objective of rebuilding social capital and reintegrating those individuals and groups who were excluded following the severe crisis of 2001­02. The vision is family centered, community based, and aimed at targeting social policy to those in need. Under this vision, social policy should be planned and financed by the nation, with contributions to policy design and administration from the subnational level. Ideas in this area were most commonly expressed in the provinces of Chaco, Corrientes, Río Negro, and Tucumán and by representatives from the MDS, the Catholic Church, NGOs from the middle and upper classes, and civil society organizations (CSOs) with social-Christian values. Argentina 35 The creation of citizens' participation centers or CICs (Integrated Community Centers) is important for personal development in the "barrios." Even the community feeding programs that already exist, or NGOs in the local communities that have the infrastructure to become centers for people to participate, capacity to train people, projectors to watch movies, whiteboards, computers . . . This part of the neighborhood where people can come together, share opinions, learn; this is fundamental for us. (Undersecretary of public expenditures and municipal anticorruption in a city in the north of Argentina) Growth-oriented social policies. The second vision emphasizes economic policy as a primary objective and social policy as a secondary objective for public policy. Values expressed by this set of views are "shared growth" and "work first." Similar to the first vision, the role of the nation is envisaged to set economic policy and finance subnational interventions. Local (provincial and municipal) actors have a strong role in promoting local development and employment from both demand and supply sides. The regions where this view came out strongest overlap with those where the rebuilding vision was prevalent; they include Buenos Aires, Chubut, Mendoza, Misiones, and Santa Fe. The social actors included Ministry of Labor officials, private sector interests (especially small and medium enterprises), CSOs from the socialist tradition, and trade unions. The natural way out of poverty is the creation of employment . . . The natural way to solve problems is the creation of wealth. And the creation of wealth depends on a few basic things: the transfer of human capital to productive activities, incorporation of technology. (President of the Unión Industrial's regional headquarters) Universal, rights-based social policies. The third vision is a rights-based approach that prioritizes universal programs. This can be compared with rights-oriented views of social policies discussed in OECD countries. A primary objective of this vision is the introduction of a guaranteed universal minimum income, also referred to as a "citizens' benefit," that would provide income support to all households up to a set amount. The rights-based view emphasizes solidarity as a value with a strong role for the nation in financing and administering social policies. The focus of the subnational level is in addressing microsocial problems such as violence, addiction, and delinquency. This perspective was less commonly represented than the other two types. It was more common among unions from the social- democratic tradition, the political left, and leftist CSOs. If an income makes one a citizen and no longer a poor or homeless person where this status is important, then there is a need to change the model. It would no longer be necessary to have a troupe of social or technical workers who determine if the indicators show that a certain family should be included in the social plan. (President of an NGO in Cuyo) Although the three visions have distinctive aspects, there are areas of overlap and consensus that can be distilled. First, regarding the overall national policy approach, there was general agreement among stakeholders that the emergency programs (for example, Plan Jefes) were appropriate for the postcrisis period and had served their purpose. 36 A World Bank Country Study However, there is a need for a longer-term vision and strategy. Respondents did not express any particular disagreement with the transition strategy to the new Seguro and Familias programs. It may be that these programs were still too new to have engendered specific opinions among the interviewees. Employment emerged as a central focus for social policy across the three visions. However, there was debate about the utility of distinguishing between the "employable" and the "nonemployable." Although some felt that all social programs should aim at improving employability and prospects for labor market insertion, others felt that programs should distinguish between groups that would benefit from help in reentering the labor market (the employable) and others who would be unable to work, and hence need basic assistance. For employment-oriented transfers there was consensus on the value and need for contraprestaciones to be incorporated into the benefit. Across the board, respondents mentioned clientelism and politicization as influences in social policy. These imply the need for a gradual and incremental approach to reform. There was also a debate about the extent to which greater involvement of local actors leads to more or less clientelism and politicization. Finally, across all provincial and municipal actors, public and private, a desire was expressed for greater involvement and participation in social policy making and implementation. These issues are discussed further in chapter 4. The consultations also revealed notable areas of debate and disagreement among the actors, which derived from their different views about the objectives of social policies. In particular, areas of discord emerged surrounding the level of targeting, criteria for targeting, and strictness with which these criteria should be applied. Interviewees also expressed different views about policy approaches to dealing with informality, promotion of employment and local development, and the necessary prerequisites for reaching social objectives. One of the objectives of the EPPS was to look at the extent to which the three visions were prevalent at the national level. The survey included questions on values, as well as specific questions about the policy instruments advocated by the three perspectives (for example, family allowances, minimum benefits, training, and employment services). The results found that the principal elements of the three visions could be identified in national views, but--as the consultations revealed--there is overlap, and it was not possible to find a clear division of the views across specific population groups. Consistent with the second vision of growth-oriented policies, the EPPS found broad agreement that social programs should be linked to the labor market. A majority of the population (83 percent) agreed with the statement "programs should train people and help them get a job." The survey also indicated support for programs oriented to community development. A sizable share (44 percent) agreed with the statement "programs should only be directed at community programs that help neighborhoods, family groups, etc." Respondents indicated a preference for family allowances--a measure advocated by proponents of the first vision--and less support for a guaranteed minimum benefit-- advocated by the rights-based vision. More than half of the respondents (55 percent) indicated that they did not agree that the government should provide a monthly guaranteed income to all residents (figure 3.1). In contrast, 60 percent were in agreement that the government should pay benefits to families with children. Argentina 37 Figure 3.1. Objective of Social Programs: Universal Benefit and Family Allowances Agree 33% 60% 8% Doesn't agree or disagree 55% 7% Disagree 29% 4% 4% Doesn't know/No Government should pay family The state should ensure an answer acceptable living standard allowances to all families with for all, and to achieve this, children, not only to those it should pay a fixed employed allowance per month to all inhabitants, whatever their economic situation Source: Calculations, based on EPPS (CEDLAS-MORI-BM 2007). Familiarity and Experience with Programs Jefes is widely known among the population, whereas newer programs are less recognized. When asked to name social programs, 93 percent of those interviewed in the EPPS mentioned Jefes (figure 3.2).4 A large share of the population (61 percent) also mentioned Trabajar, the predecessor to Jefes, which was discontinued in 2001, suggesting recognition of the plans but less familiarity with recent developments. These findings illustrate the widespread familiarity with the emergency programs, especially Jefes, because of the high level of coverage within the population, as well as the significant media attention to the program. On the other hand, newer programs, particularly Seguro and Familias, were much less recognized. Only 2 percent of interviewees mentioned Seguro spontaneously. When prompted with a list of programs and asked which were familiar, the number of respondents indicating familiarity increased to 42 percent. Similarly, with Familias only 4 percent mentioned the program unprompted, whereas 55 percent recognized it from a list. These findings are not surprising, given that the programs are not yet fully implemented across the country. 38 A World Bank Country Study Figure 3.2. Familiarity with Social Programs (Spontaneous and Guided Responses) Plan Jefes y Jefas de Hogar Desocupados 93% 7% Plan Trabajar 61% 39% Plan Familias 55% 45% Seguro de Desempleo 42% 58% Seguro de Capacitación y Empleo 25% 75% Familiar Unfamiliar Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: total sample. Beyond basic name recognition, the survey also asked respondents about their knowledge of the eligibility criteria for participation in the programs. The findings indicated a relatively broad knowledge of basic characteristics of eligibility: 54 percent of interviewees recognized that unemployment was a criteria for receiving the program, 40 percent identified poverty, and 36 percent identified having children in the household as necessary. In addition to the eligibility criteria, 87 percent recognized that contraprestaciones were a requirement for receiving benefits. A significant share of the population has had direct experience with social programs, again reflecting the high coverage of Jefes. In the EPPS, 20 percent of heads of households interviewed indicated that either they themselves, or a household member, had received a benefit during the past five years (2002­07) (figure 3.3). This level of coverage is reflective of the peak level of coverage reached by Jefes in 2003. Of this 20 percent, 82 percent reported a monthly income below Arg$1,249--the level of median total family income at the time of the survey. On the other hand, 12 percent of respondents indicated that during the past five years, a family member had applied for a program but did not receive a benefit. Argentina 39 Figure 3.3. Interaction with Social Programs, 2002­2007 Have you ever had one of these plans which give 14% money monthly? Has your spouse had any of these plans which give 5% money monthly? And has any other member of your family had one 3% of plans which give money each month? Have you or any member of your household ever 2% been offered a plan and not accepted it? Have you or any member of your household sought 12% out or requested a plan and didn't receive it? Have you ever considered enrolling in a plan? (Base: Those interviewed who weren't part of a plan 14% and whose spouse wasn't either) Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: total sample. Views about Programs The survey indicated that the population generally supports the social programs, but it also signaled concerns about their implementation. When asked whether to "leave the programs as they are" or "eliminate them without replacing them," half of the respondents opted to leave the programs as they are, and 28 percent would opt to eliminate them. The remainder chose neither or did not respond. This suggests that there is a large share of the population that finds value in the programs, despite concerns about implementation (discussed further later), while at the same time, a significant group feels strongly enough to suggest that the programs should be eliminated entirely. As would be expected, beneficiaries were more enthusiastic about keeping the programs unchanged (71 percent), while the share who would opt to eliminate them was substantially higher (44 percent) among those who did not receive benefits. Approval of the programs appears to decline with age. Among those between 18 and 35 years old, 19 percent would eliminate programs; this increased to 27 percent for those between 36 and 55 years old, and to 34 percent for respondents older than 55 (figure 3.4). 40 A World Bank Country Study Figure 3.4. Overall Views on Existing Social Programs (Leave or Replace Them) Doesn't know, 6% None, 16% Eliminate the plans and don't replace them, 28% Leave the plans as they are now, 50% Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: total sample. The survey illustrated an overall view that the social programs play an important policy role and have an effect on economic and social issues in the country. A majority (61 percent) agreed that "programs are always needed because there are people who need them," while 28 percent agreed that "programs are only needed during economic crises." Support for permanent programs, rather than temporary programs during crises only, was strongest among people in the poorest regions, young people, and those with beneficiaries living in the household. Similarly, 41 percent believed that poverty would be worse in Argentina without the programs, 38 percent thought that unemployment would be worse, and 37 percent thought that inequality would be negatively affected.5 The survey results suggest deep concerns about the design and implementation of the programs. A large majority (58 percent) of respondents either disagreed, or disagreed strongly, that the programs should continue as they are, whereas only 21 percent agreed that the programs are fine without changes (figure 3.5). The main concerns with the programs indicated in the survey were issues with access, low benefit levels, and lack of information about eligibility. Argentina 41 Figure 3.5. General Evaluation of Programs (Agreement with the Statement "The Programs Work Well and Should Be Left as They Are") Doesn't know/No Agree + strongly answer agree 6% 21% Neither agrees or disagrees 15% Disagree + strongly disagree, 58% Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: total sample. Perceptions of Beneficiaries Within the EPPS sample, 17 percent of respondents, or their spouses, reported receiving some type of a social benefit between 2002 and 2007. Programs received included a mix of national programs including Jefes (60 percent), Familias (24 percent), and Seguro, as well as one from a range of provincial-level programs.6 In terms of the overall evaluation of the programs, 35 percent of beneficiaries reported that they believed that their quality of life had improved from participating in the program, and 26 percent maintained that their potential to get a job had improved. The majority reported that both their quality of life and employment prospects had stayed the same regardless of receiving the benefit (57 and 66 percent respectively). These results may not necessarily represent a negative assessment, if interviewees did not include the counterfactual in their assessment. In other words, in the absence of the benefit their quality of life and employment prospects might have worsened further. Of those respondents (or their spouses) who reported having received a benefit during the reference period, 66 percent were continuing to receive benefits at the time of the survey. For those who were no longer receiving benefits, the main reason cited for going off benefits was getting a job (36 percent; figure 3.6). Other reasons for discontinuing participation were that the program ended (8 percent), retirement (7 percent), or failure to meet the eligibility criteria of the program (6 percent). 42 A World Bank Country Study Figure 3.6. Reasons for Exiting Benefit Programs Found employment 36% I really don't know why, they just stopped giving it to me 16% It was suspended 16% The program ended 8% I didn't need it any more 8% I was asked to retire and they stopped giving it to me/I retired 7% I didn't fulfill the requirements because of the children's age 5% I couldn't complete the "contraprestacion" because of illness 2% The recipient died 1% I didn't carry out the contraprestacion 1% I stopped participating in "piquetes" 1% Move (the person who gave it out or the beneficiary) 1% Other 1% Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Respondents who had received (or whose spouse received) a benefit in the past. The survey also reflected the government's strategy of shifting Jefes beneficiaries to the new programs. One-quarter of the interviewees who reported receiving a benefit noted that they had transferred from Jefes to another program. In the majority of cases, this transition was one year or less from the time of the survey. The main reason cited for switching programs was that the new program offered a higher benefit level. Program Design Public opinion cannot determine the design of social policies; however, it can be a useful input for policy makers in understanding preferences and determining what design aspects may work better in practice. Because the recent experience of the emergency programs is fresh in the minds of the Argentine public, these perspectives can be particularly useful. The survey uncovered strong preferences and views regarding key design issues, including conditionality for receiving benefits, targeting, benefit levels, and the intersection between income support and employment policies. Argentines are nearly unanimous regarding the need for including contraprestaciones, or some form of conditionality, in the design of social programs. Contraprestaciones refer to the requirement for participation in work activities (including community services or attendance at education or training courses) for a specific number of hours a week. Argentina 43 Popular awareness is high because of the incorporation of contraprestaciones into the Jefes program. Proponents of all three visions in the consultations expressed unanimous support for contraprestaciones. Stakeholders indicated that a requirement to work should be obligatory because of the persistence of poverty and unemployment, and as a means to discourage welfare dependency (asistentialismo). Support for conditionalities attached to receipt of benefits was nearly the same within the EPPS survey. Of interviewees, 93 percent agreed that beneficiaries needed to "do something in exchange for receiving benefits." When asked about preferences regarding the type of conditionality, the majority (58 percent) supported a work requirement, while others supported participation in training courses (42 percent), a requirement that beneficiaries be actively seeking work (36 percent), or participation in a productive community project (20 percent). Among beneficiaries surveyed in the EPPS, an overwhelming share (89 percent) reported that they had actually participated in some form of contraprestacion (figure 3.7). The majority participated in a work activity (55 percent), while the second most common activities were compliance with school attendance (27 percent) and health checkups for children (26 percent). Figure 3.7. Types of Contraprestaciones Work in exchange for the plan (without receiving 55% another salary or wage) Check children's school attendance 27% Regular health checks for children 26% Attend employment training 14% Participate in some community organization (feeding 14% program) Finish primary or secondary school 10% Actively seek work 10% None of the above 10% Work for a political organization 4% Care for a disabled spouse 2% Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Respondents who had received a benefit and participated in some form of contraprestacion. 44 A World Bank Country Study There must be a contraprestación, because it's difficult to get back into a work culture after not having worked for a long time. I believe that this is of great value. This is why we tell the people in our organization that they have to also do something for us. Not like before, when they had to do everything at the same time. If an odd job comes up, do it. Then come back again and do your part. We understand that this helps to discipline time, production and it's difficult ... That's why we understand that social plans should be designed to regain all of that. (National coordinator of a piquetero group) Views regarding targeting did not converge as clearly as for contraprestaciones. Most stakeholders expressed a preference for universal programs during the consultations, but they recognized that some level of targeting was necessary because of resource constraints. Proponents of the rights-based vision supported a basic universal citizen's benefit, while supporters of the other two visions expressed the need for some form of targeting. The EPPS respondents also indicated support for some level of targeting. When asked which criteria should be applied in determining eligibility for support, the most common responses in the EPPS were disability and inability to work (56 percent), unemployment (47 percent), and poverty (42 percent).7 The survey indicated a strong preference for tightly targeted programs, rather than loosely applied eligibility criteria. The majority of respondents (60 percent) would prefer to have rigidly applied eligibility criteria for benefits, even if it would mean that some people in need would be left out. In contrast, 17 percent of the population would opt for more flexibly applied criteria that would potentially include beneficiaries who did not need support. Another key design aspect was the level of benefits. Despite the ambivalence about cash transfers indicated in some aspects of the survey, the general message on this issue was that current benefit amounts are too low and should cover basic household needs and not food costs alone; for example, they should cover clothing and transportation. More than 80 percent of interviewees agreed that benefits are not sufficient to meet monthly payment obligations. In households with a beneficiary, 67 percent of respondents agreed that the benefit should go beyond basic subsistence needs. Similarly, 78 percent of beneficiaries were in favor of increasing the benefit amount of Plan Jefes above the current Arg$150 per month (figure 3.8). Although most agreed that benefits should be increased, few said they are willing to pay the cost of doing it. Even though the survey indicated a strong interest in increasing benefit amounts, there was also a strong rejection of the idea that the increase should be financed through a reduction in the budget for education (90 percent against) or health (89 percent against), or through an increase in taxes (79 percent against). These sentiments signal an awareness of the trade-offs that income support policies entail. Respondents also indicated a preference for receiving benefits in cash, rather than in- kind. Of respondents, 65 percent expressed support for cash and 35 percent supported in- kind. In a related question on the logistics of benefit delivery, 48 percent expressed a preference that transfers should be provided only to the mother, or female head of household, whereas only 8 percent preferred the benefit to go to the male head of household. Argentina 45 Figure 3.8. Preferred Amount of Monthly Plan Jefes Benefit 31% 19% 18% 16% 12% 1% 3% Between Between Between Between Between $600 or Doesn't $151 and $200 and $300 and $400 and $400 and more know/no $199 $299 $399 $499 $599 answer Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Respondents who relieve that the amount of the Jefes benefit should be changed. Figure 3.9 presents preferences of the interviewees regarding other aspects of the delivery and administration of income support programs. There was a notable agreement that benefit amounts should be higher for larger families with more children (74 percent). There was also general agreement that benefits should be provided only for a limited period of time and that each household should receive only one benefit at a time. Finally, the survey found a divergence between those who thought foreign residents should be eligible for benefits (37 percent) and those who did not (44 percent). Beneficiaries and potential beneficiaries expressed an overwhelming preference for programs that would facilitate entry to employment. The module explored policy options in a number of ways. Respondents were first asked to choose among three different program options: Jefes, with a benefit of Arg$150 per month; and two hypothetical programs, one similar to Familias, described as a benefit of Arg$125 per month, with an additional Arg$25 per child and no work requirement, and the second corresponding to Seguro, a plan of Arg$225 per month that offers training and access to employment services for a maximum of two years. There was overwhelming support for the third option (72 percent), signaling that Seguro would be preferred among the potential beneficiary population. The results did not indicate any difference according to socioeconomic characteristics (for example, number of children or employment status). 46 A World Bank Country Study Figure 3.9. Opinions on Form of Benefits 37% 63% 76% 74% 15% 11% 44% 10% 9% 22% 13% 14% 1% 3% 4% 4% They should give more They should only give Families should receive Plans should also be money to families with one plan per family plans for a limited time given to foreign more children residents who need them Strongly agree + agree Neither agree or disagree Strongly disagree + disagree Doesn't know/No answer Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Total sample. Figure 3.10. The Choice between Jefes and Hypothetical Programs (similar to Familias/Seguro) A plan which pays $225 per month, offering training and access to employment agencies for a 72% maxiumum of two years A plan that pays $125 per month, plus $25 for each 16% child, with no obligatory work in the community Plan Jefes, with the same conditionalities as now ($150 per month for working in the community for 8% an as yet undetermined period) None 1% Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Submodule. Argentina 47 To explore this question of policy preference further, the submodule included additional questions around five hypothetical program options. They varied according to benefit amount (in a range of Arg$150, Arg$250, and Arg$400 per month), duration, extent of work requirement (contraprestacion), and relation to the labor market (for example, provision of training, employment at the end of the program). The results indicated a strong preference for higher monthly benefit amounts, even if this would imply greater commitment to meeting conditions and complying with contraprestaciones. On the other hand, respondents expressed a preference for lower benefits in exchange for training, access to employment services (bolsas de trabajo), or a concrete job offer. These results indicate a preference for employment-oriented social programs, similar to the design of Seguro (figure 3.10). Respondents preferred programs that provide tangible benefits in the labor market, even if it means receiving a lower transfer amount (figure 3.11). At the same time, the responses indicated a preference for higher benefits, even if those benefits are associated with a higher level of contraprestaciones such as school attendance or health checks for children. Potential beneficiaries also preferred to receive a higher level of benefits, rather than receive less over a longer period of time. Population groups that are more likely to be disconnected from the labor market are more likely to prefer employment-related benefits. Multivariate analysis found that respondents in the poorest regions of the country, young people, and those with low levels of education were more likely to choose the policy scenarios linked to labor market entry (training, employment services, or a job offer). On the other hand, respondents from households that had recently received benefits and residents in large cities were more likely to opt for program scenarios with a larger benefit amount. Finally, women respondents were more likely to choose programs promising training. Conclusions The findings from the consultations and the EPPS provide a snapshot of the views, perceptions, and opinions of stakeholders and the general public toward social programs in Argentina. In general, the findings suggest that stakeholders believe that the emergency programs were an appropriate response to the crisis but that they have concerns about implementation, which translate into views about the future of social programs in the country. A major theme throughout both the consultations and EPPS was the link between income support programs and the labor market. The level at which social programs should be linked to the labor market, and the extent to which programs should support those who are not able to work, are debated. However, the respondents sent strong messages that employment should be an objective of social policy and labor market incentives, and that opportunities should be considered in policy design. The EPPS found that there is widespread awareness of programs across the population. Jefes, in particular, was familiar, but respondents were less familiar with the new programs, Familias and Seguro, which are currently being rolled out. 48 A World Bank Country Study Figure 3.11. Preferred Characteristics of Social Programs (Submodule) 45% If a secure job will be offered when the plan ends 65% 12% If employment training is offered during the plan 38% 17% How long the plan lasts 32% If health and school certificates should be presented in 15% exchange 28% The amount of hours per day that should be worked in 8% exchange 23% 4% Doesn't know/No answer Total mentions First option Source: Calculations based on EPPS (CEDLAS-MORI-BM 2007). Note: Base: Submodule. The EPPS results indicate an important level of public support for existing social programs. A majority prefer not to eliminate the programs completely, whereas a minority (approximately one-third of the population) would opt to get rid of the programs completely. However, even those who would keep the programs expressed doubts about their effectiveness. Fewer than half of respondents believe that the programs have contributed to reducing poverty and employment. Similarly, an important share of beneficiaries noted that some changes should be made to the programs. A majority of the population would prefer a system of permanent social protection to help people who are not able to work, and a minority believe these programs should be in place only during crises. Among the potential beneficiary population, respondents expressed a strong preference for programs with tangible benefits in the labor market. Beneficiaries prefer programs that would provide training, employment services, and, ideally, guaranteed employment. These preferences were particularly strong among those living in poor regions, those with low education levels, and women. The preference for social programs that have the potential to facilitate labor market entry has a number of policy implications. First, it indicates support for the type of program the government is currently introducing--the Seguro. Second, it suggests that other programs, such as Familias, could be adapted to incorporate more of a labor market orientation, for example, through the introduction of training opportunities or work Argentina 49 requirements. This could be particularly relevant, as the majority of Familias beneficiaries are women. Notes 1 The findings from the consultations are synthesized in Bonvecchi and Smulovitz (2007). 2 The provinces are Buenos Aires, Corrientes, Chaco, Chubut, Mendoza, Misiones, Río Negro, Salta, Santa Fe, and Tucumán. 3 The findings from the EPPS are synthesized in Cruces, Rovner, and Schijman (2008). 4 Spontaneous response plus guided response (interviewee was asked to select programs from a list); 62 percent mentioned Jefes spontaneously. 5 Poverty, unemployment, and inequality were separate (not mutually exclusive) choices in the question. 6 Provincial programs mentioned included Plan Bonarense, Fondo de Desempleo, Plan Vida, Vale lo Nuestro, and Plan Primer Paso y Plan Medicamentos. 7 Multiple responses were allowed for this question. CHAPTER 4 4. Income Support and the Role of the Provinces A s in other federal countries, provincial and municipal governments in Argentina have a role in the definition and management of income support policies. The legal system in Argentina gives subnational authorities significant powers to operate in this area. However, not all provinces have the same level of activity, and though some limit themselves to complementing national policies with similar programs, others prefer to develop their own set of policies and programs. Fiscal considerations are central in defining the role of provinces, as provinces' resource availability depends on a complex tax-sharing system. Tax collection is mostly centralized by the federal government, which then distributes resources to provinces. Although legislation has been in place to govern this system since 1935, there have been many revisions, renegotiations, and exceptions. As a result, since the mid-1980s no coherent framework has been in place, resulting in a high degree of uncertainty about the fiscal capacity of the provinces in the medium term. However, greater involvement of subnational governments in social policies does not depend exclusively on the availability of resources. There is substantial variation across provinces regarding the level of involvement they have in social policies and in income support in particular. Though some provinces focus on implementing national programs, others have introduced their own supplementary or separate programs. In addition to resources, factors such as local demand for such policies, political alignments between the national and provincial governments, and political competition at the local level seem to be important determinants. Because of the federal structure and fiscal arrangements in Argentina, the design and implementation of income support policies should take into account both the federal and provincial levels. Provinces in Argentina have important roles in financing (they finance a share of social protection expenditures) and implementation (they execute a share of federal expenditures) of social protection. Although most of this report focuses on income transfers at the federal level, this chapter discusses the intergovernmental context as a contribution to a more forward-looking discussion of income transfer policies in Argentina. This chapter reviews the institutional and fiscal determinants of social protection at the provincial level in Argentina, and describes recent trends in subnational social spending. As discussed in the previous chapter, a strong message that emerged from the consultations with stakeholders at the provincial level was their strong interest in becoming more involved in the design and implementation of social policy. The chapter 50 Argentina 51 concludes with descriptions of some aspects of income support policies at the provincial and municipal levels.1 The Federal Model in Argentina Argentina, by its history and legal framework, is a federal country. The first article of the constitution establishes that Argentina is "representative, republican, and federal." The constitution grants the federal government exclusive authority over foreign relations, currency, commerce regulation, navigation, legal codes, and defense. All other policy areas not explicitly delegated to the federal government correspond to the provincial level. Provincial authorities are elected by citizens residing in the province, and each province has its own legislative and judiciary system. Although the system gives strong powers to the provinces on paper, the political model has tended toward centralization. The economic and demographic weight of Buenos Aires has created, since the origin of the country, a political strain that has never been fully resolved. A long civil war in the first half of the nineteenth century, followed by continuous economic and political conflicts between the porteños (that is, those who live in the port) and the rest of the country, established a difficult framework for collaboration and integration of policies across government levels. These conflicts have affected the way institutions operate, with periods in which federal authorities have been able to impose policy preferences on provinces thanks only to the control of fiscal resources, and periods when provincial authorities have greatly reduced the ability of the federal government to manage the country. The Fiscal Design The constitution in Argentina establishes a clear distribution of roles regarding tax collection. According to the constitution, provincial governments have the exclusive authority to collect direct taxes, the national government has exclusive rights to collect import and export taxes, and other indirect taxes can be collected by either level. This structure was designed explicitly to separate the roles of the national and provincial governments. However, in practice the tax system is less transparent. Provinces collect land, property, and some sales taxes, and municipalities collect fees for street lighting and cleaning, health safety checks, licenses, and so forth. The rest of taxes, including value added tax, fuel imports and exports, labor taxes, income taxes, and others, are collected by the national government. Some of the taxes are within the limits established by the constitution for federal taxes; others should be collected by the provinces or can be shared. For shared taxes, the federal government collects the payments and then transfers part of the revenues to the provinces under a scheme known as tax sharing (or coparticipación). The first tax-sharing agreement was enacted in 1935, when new direct taxes were imposed by the national government as a response to a crisis. In 1935 new sales taxes and income taxes were created to increase revenues at the federal level. As some provinces were already collecting these taxes, it was agreed that the provincial taxes would be eliminated and the federal government would share the revenue with the provinces. This model operated for decades, with some reviews and adjustments during the 1950s. Each tax was distributed in different proportions based on a set of indicators, and it was only in 1973 that the system was unified. The new scheme established a primary distribution 52 A World Bank Country Study between the federal and provincial governments, and a secondary distribution among the provinces. This model provided a balanced scheme of revenue and spending across the different government levels. The combination of growing fiscal pressures and a conflictive political climate brought the tax-sharing system to a critical situation in the mid-1980s. Because the system was supposed to be redesigned in 1984, the political climate (with most provincial governors coming from the opposition party) and the critical fiscal situation generated by macroeconomic conditions made it very difficult to reach a new agreement. In 1988 a new law was approved, establishing a higher share of the tax rate for provinces, and leading to a period of continuous conflicts, renegotiations, and so-called exceptional agreements. The new situation resulted in a larger role for the federal government in collecting and distributing revenues, but a reduced role in spending them. One clear result of the negotiations during the 1980s was that the percentage of shared taxes to be transferred to provinces grew from 53 percent at the beginning of the decade to 58 percent by the end. At the same time, the federal government increased its role as collector, and the proportion of taxes not included in the sharing agreement declined. In other words, the provincial governments ended up receiving a larger proportion of the national tax revenues. However, this effect was compensated for by an increase in the role of provinces in public spending. Although by 1984 the federal government was responsible for 65 percent of total spending, this figure declined to 52 percent in 1993. Figure 4.1. Structure of Public Revenues and Expenditures, by Level, 1961­2004 Revenues (%) Spending (%) 1961 6.4 24.7 69.3 72.1 21 6.8 1961 1970 4.5 28 67.4 66.6 27.5 5.8 1970 1984 4.3 27.6 68.1 64.5 30.9 4.5 1984 1993 5.3 19 75.7 52 37.2 10.8 1993 2000 5.5 18.5 76 52.6 37.9 9.5 2000 2002 5.9 19.7 74.4 51.4 38.6 10 2002 2004 3.9 17.8 78.4 50.5 40.2 9.3 2004 0 20 40 60 80 100 0 20 40 60 80 100 National Provinces Municip. Source: Cetrángolo and Jimenez 2004. Argentina 53 Some of the changes in the structure of spending were caused by the transfer of health and education services from the nation to the provinces, which started in 1978 and was completed in the early 1990s. This transfer explains a large part of the decline in national spending between 1984 and 1993, a period when revenues continued to concentrate at the national level. In 1978, most primary schools under the control of the national government were transferred to the provinces, and in 1991 the same process affected the secondary schools. Also, many public hospitals were transferred from the national level to the provinces in 1992. These changes created new fiscal pressures for the provinces, which triggered new policies and reforms in the social sectors and the tax-sharing system. Between 1994 and 1997, 10 provinces transferred their pension systems to the nation in an attempt to reduce the fiscal difficulties that were affecting them. Others passed civil service reforms, also trying to reduce or control spending trends. At the same time, the federal government began to provide one-time or programmatic transfers to provinces outside of the tax- sharing system. Many of these transfers were conditional, either because they were earmarked to finance specific expenditures, or because they were made in exchange for the implementation of certain policies. On the other hand, since 1992 a series of Federal Agreements revised the tax-sharing system, creating special extrabudgetary funds, directing part of the common resources to specific programs, and, in general, eliminating any explicit criteria to regulate the transfers. The recent evolution of the tax collection system has further debilitated the tax- sharing arrangements. Since 2001, the federal government created two important new taxes that are not shared with the provinces. These are a transaction tax, applied to all bank debits and credits, and a new export tax, which became a very important source of revenues for the federal government as quantities and prices of food and other commodities exports rapidly grew. The effect of this change in the revenue structure was to reduce the fiscal space of the provinces, and to make them more dependent on discretionary transfers from the national government. The Determinants of Autonomous Policies Whereas normative and fiscal determinants have affected most provinces in similar ways, not all provinces have responded in the same way. Some provinces have limited their role in social policies to manage the educational and health budgets, but others have taken a more aggressive approach to the social protection agenda by looking for policy options. In a background paper prepared for this report, Bonvecchi (2008) identified four groups of determinants to explain why some provinces have more independent (or autonomous) social policies than others. The first set of factors comprises the social conditions at the local level, which affect the demand for policies. The better the social environment, the lower the demand for interventions at the local level, and consequently, it will be less probable that an autonomous policy will be implemented. The second group includes the institutional conditions, defined by the ability of the provincial-level administration to adequately respond to and meet the needs of the population, as defined by normative and capacity aspects. Third, the fiscal determinants refer to the availability of funds to finance autonomous policies. The fourth group of determinants includes the political dimension, as defined by the incentives and interests of the local authorities to advance with an autonomous policy. 54 A World Bank Country Study The first two groups of determinants (social demand and institutional capacities) represent what could be considered structural conditions, and the other two groups (fiscal and political capacities) have a role in the short term. Only when demand for local services, and capacities to provide them, are present do the fiscal and political questions become relevant. In a province with no social needs, or with little institutional capacity to implement a program, neither the availability of funds nor the lack of political restrictions will, by themselves, motivate the development of an autonomous policy. The analysis using an empirical model with data from Argentina, Brazil, Mexico, and Chile shows that these concepts are empirically relevant. Autonomous social policies were found more often in countries with stronger federal systems (that is, higher institutional capacity at the subnational level) and stronger social demands, compared with countries with fiscal restrictions and reduced incentives for local politicians to differentiate themselves from the national authorities, either because of limited political competition at the local level or because the local authorities are politically aligned with the national government. Social Spending at the Subnational Level in Argentina Provinces and municipalities are responsible for nearly 50 percent of total public spending and social spending in Argentina. The distribution of social spending, including social protection programs, health, education, and other similar programs, has been very stable in recent years. Provincial governments execute 40 percent to 45 percent of all social expenditures, and municipal authorities cover 7 percent to 7.5 percent (figure 4.2). Figure 4.2. Total Social Spending, by Government Level, 2000­2007 100 7.5 7.6 7.1 6.9 7.4 7.2 7.6 6.7 80 41.5 41.0 40.0 41.4 41.8 42.9 45.0 44.9 60 % 40 51.0 49.6 51.9 53.0 51.2 51.5 47.9 47.6 20 0 2000 2001 2002 2003 2004 2005 2006 2007 Year Federal Provincial Municipal Source: Goldschmit 2008. Argentina 55 Of course, the distribution of spending is not the same across sectors. The federal government concentrates most expenditures on labor-related programs (including unemployment insurance and workfare programs), which make up more than 90 percent of the total, and on social security, which makes up more than 80 percent of the total (figure 4.3). The provinces spend three-quarters of all educational expenditures, and municipalities have their most important role in social assistance, where they cover almost 20 percent of all expenditures. This division of labor is the result of the different trends registered in each sector. In education, the transfers of schools to provinces in the 1970s and 1990s gave provinces a central role, as nearly all public schools in Argentina are now managed by provinces. The decentralization process in the health sector was less dramatic. Although some national hospitals were transferred to provinces in the early 1990s, others remained under the authority and budget of the federal government, along with other health programs. The pension system saw a reverse trend. After the 1993 pension reform, the federal government absorbed the civil servant schemes of 10 provinces, moving toward a more centralized system. Social assistance programs (not including workfare) have been mostly in the form of direct support to vulnerable populations, Figure 4.3. Participation of Government Levels in Spending on Social Sectors, 2007 (Percent of GDP) 100 80 60 40 20 0 Education Health Pensions Social Labor assistance Federal Provincial Municipal Source: Goldschmit 2008. 56 A World Bank Country Study providing temporary housing, food, or clothing in emergency situations. This is the area in which municipalities have had a more significant role. The close proximity of authorities to their constituents makes them more accessible and potentially allows for faster responses. Finally, spending on labor market programs includes unemployment insurance, family allowances (two programs that have always been run at the federal level), and workfare programs, which have been mostly managed at the federal level since the 1990s. Social protection programs at the provincial level rapidly expanded during the 1980s and early 1990s. After 1992, spending stabilized at about 2.0­2.5 percent of GDP, but the composition of this spending changed (figure 4.4). Provincial spending in social assistance has grown continuously since 1980, going from 0.1 percent of GDP in that year to almost 1 percent in 2007. This growth was possible partly because of a reduction in pension expenditures, caused by the transfer of 10 provincial schemes to the national system in the mid-1990s and the reduction of real expenditures in the early 2000s. Figure 4.4. Provincial Spending on Social Protection, by Sector, 1980­2007 (Percent of GDP) 3.0 Pensions 2.5 Other SP Total 2.0 1.5 1.0 0.5 0.0 80 82 84 86 88 90 92 94 96 98 00 02 04 06 Year Source: World Bank staff calculations, based on MECON (2007) and Goldschmit (2008). Note: Other SP = Family allowances, unemployment insurance, employment programs Argentina 57 Income Support Policies at the Provincial and Municipal Levels Provinces and municipalities in Argentina have three roles in financing and implementing social protection policies. First, local governments create and run their own programs to alleviate the impact of poverty or unemployment at the provincial level. As shown in the previous section, provinces spend approximately 1 percent of GDP on social protection programs, not including their pension schemes. However, much of this spending (as well as the more limited spending by municipal authorities) is done by in-kind transfers (mostly through food and nutrition programs) and through social and community development programs. The second role is linked to federal programs and their implementation. Several programs designed and financed by federal authorities, such as Jefes, Familias, and Seguro, require the participation of municipal governments in their implementation at the local level, either in identifying beneficiaries, organizing public works projects, providing employment services, or verifying compliance with conditions for eligibility. Finally, in the third role, which is of particular interest for this report, some provinces have set up their own income transfer programs. Income transfer policies at the provincial level have been implemented through employment programs, contributory pensions and NCP, and discretionary transfers in response to emergencies or social shocks. Among these, employment programs at the provincial level are particularly interesting, as they sometimes complement and in other cases overlap with national policies. Though their design, magnitude, impact, and performance vary notably, a review of their characteristics provides some useful insights on the policy-making process in Argentina. Several provinces began to introduce employment programs in the mid-1990s as a response to the growing unemployment and informality in the labor market. Between 1996 and 1999, 11 provinces created programs to provide emergency employment, training, or job placement services to their population. This trend accelerated with the crisis and has continued during recent years, and by 2006 all but three provinces had their own employment programs. Most of these programs covered unemployed workers, providing a basic income and health services for the beneficiary and their families, and required participation in public works. By 2000, the provincial programs had nearly 120,000 beneficiaries, almost 50 percent of whom were in Buenos Aires province, and in some cases coverage was extremely low. For example, in the city of Buenos Aires, which had approximately 165,000 unemployed workers in October 2000, the program New Labor Roles covered 2,000 individuals, less than 1.5 percent of the unemployed (figure 4.5). Overall, coverage grew dramatically in 2001, as the economic crisis produced a sharp increase in demand for these programs, reaching 205,000 beneficiaries. Once again, most of the cases were in Buenos Aires. Once Jefes was introduced in 2002, the number of beneficiaries of the provincial programs declined, as many were transferred to the national scheme. However, the figure recovered in 2003 and 2004, when other provinces, such as San Luis and Neuquen, introduced their own programs. 58 A World Bank Country Study Figure 4.5. Number of Beneficiaries of Provincial Employment Programs, 2000­2006 2500 2000 Buenos Aires Other provinces Beneficiaries, thousands 1500 1000 500 0 2000 2001 2002 2003 2004 2005 2006 Year Source: World Bank staff calculations, based on MECON (2007) and Goldschmit (2008). Box 4.1. The Case of Ciudadania Porteña Ciudadanía Porteña is a program, created in the city of Buenos Aires in 2005, that aims to provide a minimum income to poor families, helping them to access basic health and education services and promoting their integration in the labor market. The program began in mid-2006 to provide a monthly transfer set at 75 percent of the poverty line for households living in extreme poverty and 50 percent for those living in poverty. The amount is adjusted every three months, and by early 2008 the average benefit was Arg$250 (US$80). The transfer is made through a cash card that can be used in 1,400 stores within the city to buy food and some basic goods. As of early 2008, there were 65,000 households receiving benefits from the program. This implies that 75 percent of poor households are enrolled in Ciudadania. Benefits are paid to women, whenever possible, as a way to increase the impact on the household nutrition situation. Although the program is not incompatible with Jefes or other national schemes, the income received from them is considered to define whether applicants qualify to receive benefits. Thus, Ciudadania complements income from other sources, regardless of their origin. The program design included a detailed impact evaluation strategy, including a baseline; regular sampling of beneficiary and nonbeneficiary populations to assess the evolution of welfare, education, and other indicators; and a monitoring system. These components are being developed, but no results have been published yet. Argentina 59 Most provincial employment programs target the unemployed populations. One out of three programs tried to target the benefits to particular groups, generally by gender and age. With a few exceptions, these programs do not restrict the number of beneficiaries, but in practice, several programs have regulated the inflow of beneficiaries according to the budget available and their managing capacity. Again, in most cases the benefits are conditioned to participation in public works activities (although in some, private employment is accepted). Though there are several income transfer programs that are not linked to the labor market, data on these programs are limited. A few provinces and the city of Buenos Aires have implemented programs aimed at transferring income to poor families, regardless of the employment situation of their members. These programs provide basic income, access to food through vouchers or cash cards to be used in neighborhood stores, and other services. With a few exceptions, these programs are limited in coverage and budget. Notes 1 This paper is based on background papers by Bonvecchi (2008) and Goldschmit (2008). CHAPTER 5 5. Policy Options and Future Challenges T he views and preferences of stakeholders and the general population regarding income support policies suggest a range of possible policy options. The earlier discussion of the consultations and opinion survey presented the values, priorities, and preferences regarding the objectives of income support. The various actors also assessed the advantages and disadvantages of options based on expected policy outcomes, experiences with existing programs (as beneficiaries or providers of programs), feasibility of implementation, and required resources. This chapter explores the potential impact and feasibility of some of the main options identified for income transfers. A comprehensive approach toward social protection would require a range of measures beyond noncontributory income support, including social insurance, labor market policies, and community development programs. The exercises presented in this chapter simulate policy options for income support to the poor. The findings attempt to contribute to the discussion of possible future paths for income support, but they represent only one component of an integrated social protection system. To understand what a national program aimed at poverty relief would entail, the exercise simulated impacts of different policies. The exercise found that a benefit targeted to children in the bottom two quintiles would be a potentially effective source of income support for the poorest households. Different scenarios of noncontributory benefits were simulated to test some of the policy ideas that emerged from the consultations, The scenarios included (i) an unemployment benefit of Arg$400 per month for all unemployed workers (including those who were informally employed); (ii) a transfer of Arg$400 per month for all poor households; (iii) a family allowance of Arg$155 per month for each child under 18 in the poorest households--the bottom two quintiles--which results in an average transfer of Arg$401 per family; and (iv) a noncontributory benefit to the elderly not receiving pension benefits, amounting to Arg$270 (equivalent to the amount of the Moratoria benefit). The simulations considered effects on coverage, poverty, and inequality, as well as their overall costs. It is important to note that the simulations are simplified exercises that consider income effects only and do not incorporate other potential design aspects of programs, such as training, conditionalities, or community-level development. This chapter first presents the typology of policy options to be simulated. It is followed by a discussion of the approach used in the simulations and a presentation of the results. The chapter then discusses some of the implementation issues that would be 60 Argentina 61 involved in the introduction of the transfer programs. The chapter ends with a discussion of overall policy considerations for social protection. The Role of Income Support for the Poor During the past decade, many countries in Latin America have introduced noncontributory income support, or social assistance, to households as a pillar of social protection. Programs such as Progresa/Oportunidades in Mexico, Bolsa Escola/Familia in Brazil, Chile Solidario in Chile, Ingreso Ciudadano in Uruguay, and Jefes in Argentina are examples. Programs have been introduced with or without conditions and have varying objectives, design features, and implementation arrangements. Income support can have a wide range of objectives. As discussed in the first chapter of this report, programs can focus on (i) reducing poverty, (ii) reducing income inequality, (iii) promoting incentives for formal sector employment, and (iv) reaching and improving living conditions for specific vulnerable groups. Income support programs have a critical role to play for the poor, and have a potentially wider impact on welfare beyond income support. Targeted assistance can help low-income households meet expenses to send their children to school and pay for health care services, and can provide a springboard into employment as a temporary safety net for the unemployed and those isolated from the labor market. Programs also provide assistance and consumption smoothing for the working poor and households facing temporary income shocks such as illness and disability. The debate about the objectives and design of income support in Argentina is ongoing. The previous chapter presented the points of consensus and disagreement that emerged from the stakeholder consultations (Bonvecchi and Smulovitz 2007), the analysis of the political parties' proposals (Pereyra 2006), and the national opinion poll (Cruces, Rovner, and Schijman 2008). While some actors look to income support to play a structural role, others see a more circumscribed role for these programs. At the same time, views about the specific design and implementation arrangements for these income support policies varied widely. A central discussion in the Argentina debate on income support is the extent to which employment objectives should be linked to program design. Though all actors agreed with increasing employment as a social policy goal, not everyone considered income transfers to be the most appropriate tool. As a result, a set of policy options emerge that are supported by different groups to different extents. A stylized typology was developed to synthesize these perspectives. A Typology of Options Income transfer programs respond to at least five different policy goals. Five types of objectives were identified in the analysis of the stakeholder views and the opinion survey. They include (i) promoting universal employment, (ii) providing poverty relief, (iii) guaranteeing child protection, (iv) ensuring income protection of the elderly, and (v) guaranteeing a minimum income for all citizens. These different goals imply different program designs, and even within the same general objective, it is possible to find differences in the implementation strategy. Nonuniversal programs require the adoption of eligibility criteria, which can result in differences in coverage. Programs can be targeted based on an income threshold (for example, using a national poverty line or income 62 A World Bank Country Study quintiles), employment status, demographic characteristics (for example, age or household size), or some combination of factors. Beyond these aspects, some programs include additional requirements or conditions that may affect their coverage level. For example, CCTs can require either school attendance or health checkups or both as a prerequisite for receiving benefits. These five policy goals can be linked to the three visions of the stakeholder consultations that were discussed in chapter 4. Roughly speaking, a program of income transfer to the poor is in line with the vision of rebuilding the social fabric, while an income transfer to the unemployed can be considered as a tool for the growth-oriented program vision. Finally, a universal transfer to all children fits with the views of the rights-based approach to social policy (Bonvecchi and Smulovitz 2007). In reality, this typology simplifies the views of stakeholders in each vision and does not reflect overlapping views and goals. However, it is useful as an analytical device and a point of departure for simulating policy options. There are areas of consensus between the two predominant visions--the vision focused on rebuilding and the vision focused on growth--regarding the need to provide income support for those people who face difficulties in entering the labor market. Nevertheless, the two visions differ in approach. The first vision would allocate benefits to the unemployed, as well as to households in poverty, but for the second vision, these requirements are not enough. In other words, there is a belief that beneficiaries should also be provided with training as a way to increase qualifications and employment prospects. In contrast, the third vision prioritizes universal transfers without conditionalities instead of targeted programs. Table 5.1 presents the typology of transfers and the intersections among them. Table 5.1. Income Transfers Typology and the Three Visions Goal Beneficiaries Conditionality Vision: Growth oriented Employment Unemployed Training, work Children's school Poverty relief Poor households attendance, work None Vision: Rebuilding the social fabric Children--up to 18 years old, in households Children's school Child protection where head/spouse are not formal workers and attendance they have low income None Elderly people, 65 years old or above, without Elderly protection receiving pensions or retirement incomes and low None income Vision: Rights based Universal income guarantee Children, adults, and elderly None Source: Based on Bonvecchi and Smulovitz (2007). Argentina 63 Table 5.2. Basic Descriptive Statistics for the Urban Population, Second semester 2006 Indicators 1st quintile 2nd quintile Total Income Household per capita income $105 $242 $583 Poverty headcount ratio--official poverty line 100% 32% 27% Poverty headcount ratio--relative poverty line 100% 47% 29% Labor market and social protection Unemployment rate 22% 13% 9% Salaried informality rate 86% 37% 43% Pensions coverage rate 30% 53% 71% Sociodemographic Household members 6 5 4 Children in the household, up to 18 years old 2 1 1 Enrollment rate, children's age 5­18 88% 91% 92% Source: World Bank staff calculations based on Continuous Permanent Household Survey (EPH). Translating the typologies into concrete policy scenarios requires several assumptions and caveats. Table 5.2 summarizes the main variables that characterize potential beneficiaries in the first and second income quintiles. In general, the characteristics for eligibility are concentrated in the poorest quintiles: unemployment, families with many children, low school attainment, and insufficient income to reach a basic food and services basket. Some transfers with characteristics similar to those transfers identified in table 5.2 already exist. This is the case of Jefes, Familias, Seguro, and the new transfer policy for the elderly, PIP. Existing programs are focused on households in poverty with children, unemployed workers, or older people without pension coverage. The existing programs need to be considered in the formulation of the scenarios in order to avoid duplication of transfers. The Potential Impact of Income Transfers Methodologies for Ex Ante Evaluation The simulation exercise requires assumptions about the potential behavioral response of beneficiaries. Although arithmetic simulations assume that policy changes will have a direct impact on individuals and households that participate in the programs, behavioral models assume that some participants may change their behavior as a consequence of the policy. For example, a CCT program may require that families send their children to school. Whereas an arithmetic simulation considers the impact of the new transfer on the families that are already sending their children to school, a behavioral simulation would also consider the effect that the condition would have, as some families would now enroll their children in school as a response to the program's incentives. 64 A World Bank Country Study Modeling behavior is of course complicated, as it requires arbitrary judgments regarding participants' reactions to incentives. To advance in the analysis, the scenarios presented in this report are mostly arithmetic, except for one case in which a behavioral simulation is applied to try to illustrate the potential effects of introducing conditionalities. In the simulations the policy effects are assessed using a set of indicators. The indicators considered to assess welfare are (i) extreme poverty, (ii) moderate poverty, (iii) the poverty gap, and (iv) inequality, as measured by the Gini coefficient. Also, two indicators about the size of the program are included: (i) coverage in the first and second quintiles, that is, the percentage of households in these income strata receiving the transfer, and (ii) fiscal costs, in terms of annual expenditures as a percentage of GDP (net of Jefes). The Data The latest microdata available are from the EPH for the second half of 2006. The EPH is drawn from 33 urban centers with more than 100,000 citizens. Given that the EPH represents 71 percent of the urban population, the survey represents about 62 percent of the total Argentine population. The survey was conducted over the whole year, and INDEC publishes statistics on a quarterly basis. The EPH data set provides information on the variables to be included in the simulations and on the outcome indicators of welfare. The EPH collects a wide range of socioeconomic statistics regarding labor, income, education, and demographic variables. However, most of the existing programs are not identified in the survey. As a result, it is not possible to analyze the effects of potential policies net of most of the existing programs. Familias and Seguro beneficiaries cannot be identified in the microdata, and the Moratoria took place mostly after 2006. All of these programs are targeted to populations similar to the proposals discussed in this chapter. Only Jefes beneficiaries can be identified in the data set. The distortions in simulation results due to existing programs are potentially significant, especially the programs with high levels of coverage. At the beginning of 2008, the Moratoria, Familias, and Jefes all had a critical mass of beneficiaries. The case of the Moratoria is particularly troublesome for the simulations, as the data set used to build the baseline, from 2006, was collected before most beneficiaries of this programs began receiving their payments. To reduce the impact of these distortions, the baseline was updated to consider the likely impact of the Moratoria. In the case of Familias and Seguro, no corrections were possible. The adjusted baseline for 2007 is based on three assumptions. First, all monetary variables, including transfers, poverty lines, and others, were kept constant in real terms, based on 2006 prices. Second, elderly people in the sample were randomly assigned a Moratoria benefit, controlling for the gender and previous distribution of beneficiaries obtained from administrative sources. This includes 80 percent of women and men (minimum ages 60 and 65, respectively) without pension coverage, and 20 percent of women age 60 or above who already receive pensions, if there were no male pensioners in the household.1 Third, the extension in coverage went from highest to lowest quintiles. Argentina 65 Table 5.3. Adjusted Baseline Data Used for the Simulations Baseline Adjusted baseline Adjusted baseline (including moratoria) minus Jefes Extreme poverty 8.6 7.8 8.9 Poverty 26.4 25.0 25.8 Poverty gap 11.0 10.2 11.2 Gini 0.482 0.477 0.482 Spending (% of GDP) 8.2 9.1 8.9 Source: World Bank staff calculations based on EPH. As expected, the new baseline shows improvements in all welfare indicators relative to 2006, and the same indicators would deteriorate if Jefes were to be eliminated. Table 5.3 shows the relevant indicators for the original baseline, the adjusted baseline (that is, including the Moratoria), and the adjusted baseline without Jefes. It is interesting to note that Moratoria had a positive impact on all welfare indicators, as extreme poverty, total poverty, the poverty gap, and the Gini coefficient improve when the baseline is adjusted. However, most of these impacts would be fully compensated if Jefes were to be simply closed with no other program to replace it. The extreme poverty, poverty gap, and Gini indicators would be worse in a scenario with Moratoria but without Jefes than in the situation at the end of 2006. The Simulations Following the typology of transfers outlined earlier, this section presents four scenarios under which each transfer is simulated, applying a simple arithmetic model. The five alternatives are stylized and are general tools for illustrating how a specific policy might work in practice. In this sense, the analysis assesses the potential impact of (i) universal transfers for the unemployed, (ii) transfers to poor households, (iii) family allowances for households with children in the bottom two quintiles, and (iv) family allowances plus pensions for the elderly in the bottom two quintiles. Table 5.4 presents the basic characteristics of each scenario. Table 5.4. Simulated Scenarios Scenario Description Beneficiaries Transfer (Arg$) 1 Universal transfer for All unemployed workers, regardless Arg$400 per month the unemployed of their previous formality or informality status 2 Transfer to poor All households below the poverty line Arg$400 per month households 3 Noncontributory All children of families in the lowest Arg$155 per child per family allowances two quintiles month (on average, Arg$401 per family) 4 Noncontributory All children of families in the lowest Arg$155 per child per family allowances two quintiles plus month (on average, Arg$401 per family) + NCP All elderly in the lowest two quintiles Arg$270 per month not receiving pensions (equivalent to average Moratoria benefit) Source: World Bank staff. 66 A World Bank Country Study The potential impact of each scenario is assessed in comparison with the baseline along six indicators. The baseline scenario is the situation as of late 2006, adjusted by the impact of the Moratoria program. For the baseline and each scenario, six indicators are shown; four of them capture welfare, and the remaining two focus on the coverage and fiscal costs. The welfare indicators are extreme poverty, moderate poverty, the poverty gap, and inequality (as measured by the Gini coefficient). The other two indicators are the coverage of the lower two quintiles of the income distribution, and the total expenditure on income transfers (including all social security, Jefes, Familias, and Seguro programs). The following sections present the results of each scenario. The results of the simulations are presented in a hexagonal graph (figures 5.1 through 5.4), where each axis represents one indicator. Data have been normalized to the baseline, to show relative variations. Thus, a result of 0.83 for the extreme poverty indicator implies that, in such a scenario, extreme poverty incidence would be 17 percent lower than in the baseline. In the second panel of each figure the number of beneficiaries, average benefit, and value of the six indicators are also presented. A Universal Transfer for the Unemployed The first scenario is a transfer to all unemployed workers, regardless of their income levels. The benefit would allocate a transfer of Arg$400 per month to all unemployed workers. This would involve approximately 1 million beneficiaries. Figure 5.1. Simulation Results for a Transfer to Unemployed People Baseline Scenario I Beneficiaries - individuals 1,049,198 Benefit ARS 400 Extreme poverty Welfare indicators Annual expenditure in Extreme poverty 6% 0.83 Poverty Poverty 22% transfers 1.06 0.90 Poverty gap 9% Gini 0.464 1.54 0.87 Coverage quintiles 1 Poverty gap Program size and 2 0.97 Coverage quintiles 1 and 2 19% Gini % GDP in transfers 9.66% Diff with baseline (% of GDP) 0.55% Source: World Bank staff calculations based on EPH, and administrative data on expenditures; ANSES, MTESS, MDS. Argentina 67 The transfer to all unemployed workers would have marginal effects on welfare. Figure 5.1 shows the expected impacts of such a program. Thus, a transfer program for all unemployed workers would result in a decline in extreme poverty of 17 percent (from 1 to 0.83 in the figure), a reduction of 10 percent in poverty and of 13 percent in the poverty gap, and a reduction of 3 percent for the Gini coefficient. To achieve this, coverage of this program would be 54 percent larger among households in the first and second quintiles, and the total cost of income transfers (including current social security and social assistance programs) would increase by 6 percent, or more than half a percentage point of GDP. Poverty Relief Transfer A transfer to poor households would provide a flat benefit per month to all households with incomes below the official poverty line, regardless of household size. This benefit would reach 1.2 million beneficiaries and would provide a benefit level of Arg$400 per month. Considering that the poverty line is about Arg$280 per capita, not every household would move above the poverty line after receiving the benefit, although most would leave extreme poverty. Figure 5.2. Simulation Results for a Transfer to Poor Households Baseline Scenario II Beneficiaries - households 1,198,646 Benefit ARS 400 Welfare indicators Extreme poverty Extreme poverty 2% Poverty 15% Annual expenditure in Poverty Poverty gap 4% transfers 0.27 1.07 Gini 0.441 0.58 Program size 3.55 0.40 Coverage quintiles 1 0.93 Poverty gap Coverage quintiles 1 and 2 44% and 2 % GDP in transfers 9.77% Diff with baseline (% of GDP) 0.66% Gini Source: World Bank staff calculations based on EPH and administrative data on expenditures; ANSES, MTESS, MDS. 68 A World Bank Country Study A transfer targeted to the poor population would have a more significant redistributive impact than the first scenario. As shown in figure 5.2, a program for the poor would reduce extreme poverty by 73 percent and poverty by 42 percent. The poverty gap would fall from 10.2 percent to 4 percent, and the effect on inequality would be the most substantial of all scenarios, declining approximately 7 percent and reaching a Gini coefficient of 0.44. Coverage would substantially increase, by a factor of three, to reach 44 percent of first- and second-quintile households. However, the overall cost of this program would be very similar to that of the previous scenario, since the marginal expenditures would amount to 0.7 percent of GDP. Given these indicators, it is clear that this would be a much better targeted transfer than the previous one. Family Allowances for Households in the Lowest Quintiles Under the third scenario, each family in the bottom two quintiles would receive a transfer for each child under 18 years old. This policy, similar to the approach recently taken in Uruguay (see box 5.1), would provide Arg$155 per child, an amount that would result, on average, in nearly Arg$401 per month. Households that have already qualified for contributory family allowances--for example, those where the household head or spouse is formally employed and households in the third quintile or above--would not be eligible for the benefit. A benefit targeting children in the bottom two quintiles would have a notable impact on the bottom of the income distribution, at the lowest cost. Figure 5.3 shows that extreme poverty would drop by 66 percent, while poverty would decline by 21 percent. The poverty gap would also drop by 44 percent. The coverage of the bottom of the income distribution would more than double, to 38 percent. This policy would be less costly than the other scenarios, at 0.47 percent of GDP, a level of expenditure comparable to expenditures on noncontributory poverty relief in OECD countries. Figure 5.3. Simulation Results for a Transfer to Households with Children under 18 Beneficiaries - households 939,245 Baseline Scenario III Benefit per child ARS 155 Welfare indicators Extreme poverty Extreme poverty 3% Poverty 20% Annual expenditure in Poverty gap 6% Poverty transfers 0.34 Gini 0.457 1.05 0.79 Program size 3.07 0.56 Coverage quintiles 1 0.95 Poverty gap Coverage quintiles 1 and 2 38% and 2 % GDP in transfers 9.58% Diff with baseline (% of GDP) 0.47% Gini Source: World Bank staff calculations based on EPH and administrative data on expenditures; ANSES, MTESS, MDS. Argentina 69 Family Allowances Plus Pensions for Elderly People in the Bottom Quintiles As mentioned, most of the elderly population without pension coverage has been integrated into the Moratoria. Though available data make it impossible to know the exact number and characteristics of the elderly that continue to be excluded from the system, the third scenario assumes that those who did not join the program were the poorest in the distribution. Therefore, the scenario involves extending the Moratoria to those people 65 years of age or older in the first or second quintile. The benefit is set at Arg$270 per month, the average amount received by those who joined the Moratoria. In this last scenario, this benefit's impact was estimated in addition to the family allowances simulation of the third scenario, which would replace Jefes as the main income transfer for younger families. Because most of the elderly are already receiving benefits, the additional impact of this scenario would be minor. In fact, the results of the simulation indicate that all welfare indicators would have the same levels as in the previous scenario. In other words, providing universal pensions would not result in additional welfare for the population. On the other hand, coverage would be higher, at 47 percent of the lowest two quintiles, compared with 38 percent in the previous case, and the marginal spending in relation to the baseline would be 0.61 percent of GDP. What these results clearly indicate is that, given the already considerable impact of the Moratoria, further expansions of pension coverage would have little welfare impact but some effect on fiscal costs. Figure 5.4. Simulation Results for Family Allowances and a Transfer to the Elderly Baseline Scenario IV Beneficiaries Children 2,428,378 Elderly 295,378 Benefits Extreme poverty Children ARS 155 Elderly ARS 270 Annual expenditure in Welfare indicators 0.34 Poverty transfers 1.07 0.79 Extreme 4% Poverty 22% 0.56 Poverty gap 7% 3.81 Coverage quintiles 1 0.95 Gini 0.460 Poverty gap and 2 Program size Gini Coverage quintiles 1 and 2 47% % GDP in transfers 9.73% Diff with baseline (% of GDP) 0.61% Source: World Bank staff calculations based on EPH and administrative data on expenditures; ANSES, MTESS, MDS. 70 A World Bank Country Study Introducing Behavioral Responses As discussed in the consultations and survey showed a broad consensus that income transfer programs should be accompanied by conditions, or contraprestaciones. This requirement could involve conditionalities aimed at increasing investments in human capital, such as health or school attendance, similar to those involved with Familias, or work requirements, such as those for Jefes. There is also a strong interest in programs that involve tangible labor market outcomes, such as training, participation in employment services, or a guaranteed job. Behavioral simulation models provide the chance to analyze ex ante how these conditionalities might work, although with important limitations. Bourguignon, Ferreira, and Leite (2003) mentioned that the lack of these analyses for developing countries has two important causes, both of which seem to apply for Argentina. First, the cash element of redistribution may not be sufficient to modify labor supply significantly and to warrant this kind of analysis. Second, the estimation of structural econometric models of labor supply is complex because of the informality of a large part of the labor market. However, there are dimensions of household behavior other than labor supply that matter from a welfare point of view, and transfers and other public policies may affect that. This is the case in the demand for schooling, which is evaluated in the following. In this section, a behavioral simulation attempts to capture the behavioral effects of family allowances aimed at increasing adolescent school enrollment. Enrollment rates are almost universal for primary school in Argentina; therefore, conditionality on enrollment would produce no major effects in this indicator. In contrast, secondary school enrollment rates are far from being universal. The risk of youths leaving school increases with age in Argentina. In 2001, 4.4 percent of 13-year-olds did not attend school, but the proportion jumped to 8.1 percent for 14-year-olds and to 14 percent for 15-year-olds (World Bank 2008b). Therefore, conditionalities on secondary enrollments, aiming to change behavior for this age-group, could have an impact. Recent literature on youth dropouts asserted that, in the province of Buenos Aires, 52 percent of youth dropouts mention employment as the main reason for leaving. Poor academic performance (16 percent) and pregnancy (8 percent) are the next most important factors. Those who never started secondary school cite work, poor academic performance, and family problems as the critical constraints. As of late 2006, 14 percent of children aged 12 to 18 years were not attending school. Differentiating the sample of children who received family allowances and those who did not, the simulations show school attendance is 82 percent for those who did receive the benefit and 88 percent for those who did not. Figure 5.5 shows that the nonenrolled beneficiaries are concentrated between 12 and 18 years old. An ex ante evaluation implies simulating the number of children who would potentially change their school enrollment and employment status in response to the conditional family allowance. Following Ferreira and Leite (2003) on the specifications for estimating the behavioral response of a conditional family allowance, it is possible to obtain a counterfactual distribution of occupations based on the observed characteristics for each child. The details of the specifications are described in appendix 3. Argentina 71 Figure 5.5. No School Attendance of Beneficiaries, by Age .4 .3 Density .2 .1 0 5 10 15 20 Age Source: Authors' calculations based on EPH. The results indicate that some children (ages 12­18) would change their employment or education status. Children could move from not attending school to attending and working, or to only working, and could move from attending and working to only working. Table 5.5 summarizes the transition matrix. Table 5.5. Simulated Effect on School Attendance--Transition Matrix, Children Ages 12­18 Simulated (%) Not attending Attending school Attending Actual state school and working school Total All households Not attneding school 67.8 16.1 016.1 014.4 Attending school and working -- 78.0 022.0 003.5 Attending school -- -- 100.0 082.1 Total 09.7 05.1 085.2 100.0 Beneficiary households Not attending school 35.8 32.1 032.1 018.3 Attending school and working -- 60.5 039.5 005.0 Attending school -- -- 100.0 076.7 Total 06.6 08.9 084.6 100.0 Source: World Bank staff calculations based on EPH. Note: -- = no value. 72 A World Bank Country Study Family allowances would provide the strongest incentives for children to attend school exclusively. The first panel in table 5.5 shows the transition rates for all households with children between 12 and 18, and the second panel presents the results for beneficiaries. Looking at the first panel, almost 3 out of 10 children currently not enrolled in school would change their status to "attending school"--that means they would stop both working and not attending, or just not attending. In addition, 2 out of 10 children currently working and attending school would stop working. Box 5.1. Family Allowances in Uruguay The family allowances program in Uruguay was, in its origin, a contributory scheme, financed by workers' and employers' contributions and restricted to salaried formal workers. In 1980, a reform eliminated the earmarked contributions. Since 1999 the system of family allowances experienced three reforms that transformed it into a dual scheme in which noncontributory benefits are paid in parallel with contributory ones. According to Law 16.697, contributory family allowances can be claimed for children younger than 18 years old by salaried workers in the private sector, civil servants, retirees, and unemployment insurance beneficiaries. Benefits are paid per child, and there are two levels, depending on workers' income. Children of workers earning less than 6 BCP (base de prestaciones contributivas, an index used to adjust all social security variables, valued at US$1,636 in early 2007) receive an allowance of 16 percent of that value, while the children of workers earning between 6 and 10 BCP receive 8 percent of that value. Children of workers earning more than 10 BCP receive no benefit, with a few exceptions. Benefits are paid in the case of private sector employees, retirees, and unemployment insurance beneficiaries, or directly by the state, in the case of civil servants. Coverage of family allowances was extended to children of poor, informally employed families in 1999. In a departure from previous practice, Law 17.139 granted family allowances to children living in low-income households (less than three minimum salaries, later replaced by the BPC), if their father had been a beneficiary of the unemployment insurance and exhausted the six-month limit or if they are supported solely by their mother. Benefits were set at the same level of the contributory family allowance. Enrollment in this new family allowances program was slow until a reform introduced in late 2004 extended coverage to all children of low-income households, regardless of the employment status of their parents. As a result, by late 2007 nearly 215,000 children received this benefit. As part of the Plan de Equidad, a new major reform in the family allowances program was introduced in early 2008. During the first phase, the reform expects to reach 330,000 children by migrating beneficiaries from the old noncontributory system, plus those qualifying participants from the closing cash transfer program Ingreso Ciudadano. In a second phase (2009), authorities plan to add an extra 170,000 children, estimated to be eligible but not yet enrolled in the system. The new noncontributory family allowances program increases the amount of the benefit, introducing two payment scales that take into account the number of children in the household and their schooling level, the latter intending to reduce dropout rates at secondary schools. The impact of the new scheme on social indicators is expected to be significant. Once fully implemented, the program should result in a reduction of extreme poverty by nearly 20 percent, while total poverty would decline by 7 percent. In fiscal terms, the reform is expected to have a small impact in the first year, as the expanded benefits will be compensated for by the closing of Ingreso Ciudadano. During the second-phase expansion, the estimated additional cost could reach 0.3 percent of GDP. Argentina 73 The equal transition rates from children who are not attending school to attending school and working and just attending reflects the fact that most children are not participating in the labor market, despite not attending school. This particular characteristic of the Argentine youth labor market means that just a small number (if reinforcement works well) would have significant effects on enrollment. Most children not attending school are not working; therefore, there is no opportunity cost for them. And for those few children already working, the benefit amount represents half of their median earnings. On the other hand, in the case of those children working and attending school, the transfer amount is highly competitive. The median earnings for children who are working and studying reaches Arg$66 per month. Considering the group of children who change their status, the second panel shows that about a third of them would not meet the condition for receiving benefits. Although most of the beneficiaries are attending school, about 7 percent of the total beneficiaries would lose the benefit because of not moving from the status of not attending school. On the other hand, 12 percent of the total beneficiaries would change their behavior to meet the requirement. The small loss of beneficiaries means that there would not be major changes expected in the welfare impacts analyzed in the simulations. Implementation Challenges The simulation exercise masks the complexity of implementation and the political economy involved in any reform of existing programs. The simulations represent extremely simplified models of reality, where no implementation challenges, political economy considerations, or other obstacles are considered. Under these scenarios, perfect targeting, no errors or coverage gaps due to operational or design problems, and high efficiency in management are all assumed. Of course, in real-world politics there are many immediate and potential problems, which must be carefully considered before proposing any reform. Similar to welfare and fiscal impacts, transfer policy options differ widely in terms of implementation implications. A wide variety of mechanisms are needed to identify the target population, reach beneficiaries, provide benefits, and ensure that requirements are being met, among other actions. Some types of beneficiaries are easier to observe than others (for example, targeting based on indicators such as age and geographic location is much simpler than targeting based on income), and the whole process usually involves different government agencies, particularly in a federal system. Marginal improvements in targeting to improve welfare indicators can involve increasing implementation costs. This is the case of identification errors in targeting, in which increased targeting accuracy can raise costs. For example, poverty-targeted benefits have substantial information requirements and administrative costs for collection, information management, and verification. Similarly, targeting based on proxy indicator targeting can require detailed household data collection in order to identify beneficiaries. More sophisticated targeting mechanisms can require increased costs of implementation. In the case of family allowances, targeting costs can be reduced through the use of indicator targeting, rather than a combination of both indicator targeting and income testing. Transfers for poverty relief can be based on indicator targeting. Although the choice of the indicators used for targeting should be weighed against the cost of income testing, many of them are generally more reliable and easy to obtain. Family composition (for example, number of children or benefit status of elderly people) and labor market status are easily accessible 74 A World Bank Country Study through administrative data. Additionally, some indicators are typically more dynamic; as such, accurate weighting of each of them needs to be considered. In addition to the trade-offs surrounding targeting discussed earlier, the following implementation issues require attention: Institutions. Coherent institutional arrangements for program implementation are essential in any country, but they are particularly relevant in the federal context of Argentina. Institutional design needs to ensure coordination among actors at the national and provincial levels and to provide sound arrangements for program delivery, financing, monitoring, and evaluation. Limitations in reaching beneficiaries. Even in the best-designed programs, a number of potential beneficiaries never claim their benefits. This may sometimes serve as a form of self-targeting, as individuals or households who do not need the benefit do not take it. In other cases, transaction costs and administrative barriers, including inadequate access to information or resources to apply for the benefits, prevent people from applying or receiving the benefits. Information systems. Transfer programs are information intensive and require reliable beneficiary registries and coordination of databases to manage determination of eligibility, payments, and monitoring and controls. If conditionalities are incorporated, then an efficient system to verify compliance is needed. Labor market incentives. Labor market disincentive effects of transfer programs cannot be eliminated; however, they can be mitigated. Any transfer not related to work may create a disincentive. The challenge is not to eliminate these effects, but rather to reduce their relevance, in particular in comparison with potential positive effects of the program. OECD countries are increasingly incorporating labor market incentives into transfer programs to incentivize employment and ease the transition from welfare to work. Existing programs. Argentina has a strong base of existing social protection programs on which to build. Any new programs will need to take into account the existing national and provincial programs to ensure coordination of incentives and institutional arrangements. This chapter assessed the potential welfare impact and fiscal costs of policy options for income support. Although simplified and stylized, the scenarios used are based on the typology of views that emerged from the consultative part of this work. The results suggest that a noncontributory income support program is a feasible option for Argentina and would have desirable effects on poverty and inequality. However, as mentioned at the outset, such a program would be only a part of a comprehensive social protection system. The following chapter draws general conclusions based on the body of work included in this report. Notes 1 According to administrative data, 1.7 million people entered the Moratoria as of late 2007. A majority (90 percent) were women. Many beneficiaries were already receiving a survivors' pension, thus resulting in a duplication of benefits. CHAPTER 6 6. Reflections, Challenges, and Policy Considerations T his report considers the different dimensions of income support policies in Argentina, taking into consideration design, implementation, impact, and views of the population. Income support programs are a core component of social protection policy in Argentina. Together with other activities aimed at distribution of food and prescription drugs, as well as housing, communications, and, most important, community development programs, the policies represent the main tools the government has to prevent social shocks, ameliorate their impact if they occur, and reduce structural poverty. Traditionally, social protection policies in Argentina focused on formal workers, but this approach seems to be too narrow now. Argentina's social protection policies have operated for most of the twentieth century assuming that a fully functional labor market would be an effective tool to deliver services and transfers to the population. This assumption was proved wrong, or at least insufficient, with the 2001­02 crisis, as the population most affected by the crisis were either unemployed or informal workers and thus beyond the reach of traditional programs. Authorities responded with the rapid deployment of Plan Jefes, a large emergency income transfer program. The emergency response was effective, as it helped the country to overcome the worst of the crisis. Jefes provided benefits to nearly 2 million households at a time when poverty affected more than half of the population and unemployment reached record levels. The number of beneficiaries has slowly declined since 2003, and was at nearly one-third of its maximum value by early 2008. This reduction was the result of some beneficiaries reentering the formal labor market, others losing eligibility (in most cases, when children turned 18 years old), and the remainder transferring to the new programs (Seguro and Familias). Because the worst of the crisis has passed, the policy debate has shifted toward the longer-term and structural issues. The improvement in overall economic conditions since 2003 has resulted in a decline of unemployment, poverty, and inequality, while formal employment and real salaries have recovered to precrisis levels. These positive trends have generated some opportunities to start considering longer-term policy options, including the future role of income support policies. There is broad agreement among stakeholders and the population at large about the need for policies to alleviate structural poverty, including strategies to help individuals enter or reenter the labor market. There is little debate in Argentina about the need for these policies, and there are many concerns about how to design an effective scheme and, more importantly, how to ensure a successful, efficient, and transparent implementation. 75 76 A World Bank Country Study Concerns include the difficulties of coordinating policies and programs across multiple agencies and levels of government, how to respond to concrete demands from beneficiaries, and how to ensure that an adequate set of incentives and opportunities are available for beneficiaries to become self-sufficient within the labor market. Thus, the challenge is to design and fully implement a social protection system that is adequate in terms of coverage and benefits, integrated, and sustainable. The model of income transfer programs currently under implementation in Argentina has the potential to succeed in combining Seguro and Familias with more direct interventions at the community level to promote faster and stronger building of human and social capital. However, further work will be needed to ensure coverage of the vulnerable populations in the programs; ensure that the benefits they receive are adequate; avoid gaps and overlaps between programs, institutions, and government levels; and ensure that the system is fiscally and politically sustainable in the medium term. The question of coverage is probably the most challenging. Familias and Seguro could, as currently defined, cover all individuals and households in need of support. However, the implementation challenges are significant. Seguro is not only an income transfer, but also consists of a set of employment services that need to be set up at the local level throughout the country. There are also short-term fiscal restrictions. On the other hand, the pension system dramatically expanded its coverage within one year, but it is not clear whether this effort has reached all the poor elderly or whether further expansion will be necessary. The adequacy of Jefes benefits should also be considered, as they have been fixed since the introduction of the program in 2002, resulting in a real loss of more than 50 percent. The new programs provide higher benefits (around Arg$225, on average), but there have been no increases since their introduction, and no adjustment mechanism has been defined. This problem is similar for contributory benefits (for example, pensions, family allowances, and unemployment insurance), although these cases have involved ad hoc discretional adjustments. This approach does not seem consistent with a long-term view, as beneficiaries lack predictability about their incomes. Integration of programs is a central issue, because social protection programs in Argentina appear to function, in most cases, as independent systems with little interaction among them. This lack of program integration is seen across jurisdictions, as some provinces have schemes that replicate or overlap those at the national level, and also between national agencies and within specific programs. A clear example of insufficient integration in policy planning and implementation was the recent Moratoria, a program that granted retirement benefits to nearly 1.7 million older citizens without verifying whether they already had other benefits or transfers. Given Argentina's limited resources, it is critical to advance those efforts to improve coordination of policies and actions between the different actors at the national, provincial, and municipal levels. Fiscal sustainability has not been as critical as in the mid-1990s but is potentially fragile. After the pension reforms of the 1990s, fiscal sustainability became reasonably manageable, as pension expenditures began to decline. The introduction of Plan Jefes in 2002 and Familias and Seguro later on had a small impact, as their consolidated cost barely reached 1 percent of GDP in 2003, and their costs have fallen to half of that amount since then. The simulations presented in chapter 5 showed that an additional 0.5 percent of GDP could have notable impacts on welfare and coverage. On the other hand, the recent reforms in the pension system have created potential fiscal pressures, which may crowd Argentina 77 out other programs. The Moratoria program has an estimated cost of 1.5­2 percent of GDP in 2008, and while revenues have increased following the February 2007 reform, the medium-term impact of this reform should also be negative in fiscal terms. The combination of positive macroeconomic and social trends and the consensus across society about the need to advance structural policies has opened a window of opportunity for a long-term strategy for social protection. The 2002 crisis demonstrated the importance of having a social protection system in place that is able to respond effectively to short-term shocks. The response to the crisis was effective, but most policies and programs had to be designed and implemented from scratch, resulting in unnecessary implementation delays and problems. This experience has provided an important lesson for policy makers in Argentina. Rapid response to crises is very difficult to implement, unless the institutions, programs, and instruments are already in place to provide poverty alleviation to those in need, and to scale up programs where needed. CHAPTER 7 7. References ANSES (Administración Nacional de la Seguridad Social). 2007. "Informe de la Seguridad Social." Buenos Aires. Arriagada, A., and G. Hall. 2000. Managing Social Risks in Argentina. Mimeo, World Bank, Washington, DC. Barr, Nicholas. 2008. "Designing Poverty Relief Within and Integrated System: What have We Learned?" Keynote address delivered at the Second Conference on "Social Policy in Argentina: Towards the Bicentenary," April 9, 2008, Buenos Aires, Argentina. Barbeito, Alberto, Noemí Giosa-Zuazua, and Corina Rodriguez-Enriquez. 2005. "La cuestión social en la Argentina y el Plan Jefes y Jefas de Hogar Desocupados." OIT, Proyecto Crisis. Serie Documentos de Trabajo No. 3, World Bank, Buenos Aires. Bertranou, F., and D. Bonari. 2006. Protección Social en Argentina. Financiamiento, Cobertura y Desempeño. 1990­2003. Buenos Aires: MECON-OIT. Bertranou, F., and C. Grushka. 2002. "Beneficios Sociales y Pobreza en Argentina: Estudio del programa de pensiones no contributivas." In Pensiones no Contributivas y Asistenciales. Santiago: International Labour Organization (ILO). Bonvecchi, A. 2008. "Condiciones para Políticas Sociales Subnacionales Autónomas en Países Federales." Serie de Documentos de Trabajo sobre Políticas Sociales 3, World Bank, Buenos Aires. Bonvecchi, A., and C. Smulovitz. 2007. "Atender necesidades, crear oportunidades o garantizar derechos. Visiones sobre la politica social." Serie de Documentos de Trabajo sobre Políticas Sociales 1, World Bank, Buenos Aires. Bourguignon, F., F. Ferreira, and P. Leite. 2003. "Conditional Cash Transfers, Schooling, and Child Labor: Micro-Simulating Brazil's Bolsa Escola Program." World Bank Economic Review 17 (2). CEDLAS-MORI-BM. 2007. "Manual y documentación metodológica de la Encuesta sobre Programas Sociales." Documento técnico. Centro de Estudios Distributivos, Laborales y Sociales, UNLP; Equipos MORI Argentina; y Banco Mundial, área de Protección Social de América Latina. Cetrángolo, O., and J. P. Jiménez. 2003. "El gasto social y el Programa Jefes y Jefas de Hogar." ILO, Proyecto Crisis. Serie Documentos de Trabajo 4, World Bank, Buenos Aires. ------. 2004. "Las Relaciones entre Niveles de Gobierno en Argentina. Raíces Históricas, Instituciones y Conflictos Persistentes." Serie Gestión Pública No. 47, CEPAL, Santiago. 78 Argentina 79 Cruces, Guillermo, Helena Rovner, and Agustina Schijman. 2008. "Percepciones sobre los Planes Sociales en Argentina." Documento de trabajo sobre Políticas Sociales 2, World Bank, Buenos Aires. de Ferranti, D., G. Perry, I. Gill, and L. Serven. 2000 Securing our Future in a Global Economy. Washington, DC: World Bank. de Ferranti, D. G. Perry, D. Lederman, A. Valdes, and W. Foster (2005) Beyond the City: The Rural Contribution to Development. Washington, DC: World Bank. Dieguez, H., and A. Petrecolla. 1974. "La distribución funcional del ingresoy el sistema previsional en la Argentina. 1950­1972." Desarrollo Económico 14 (55). DAGPPS (Dirección de Análisis de Gasto Público y Programas Sociales). 2004. "Informe sobre los programas de empleo de ejecución provincial 2003." Documento de trabajo GP/15, Ministerio de Economía y Producción, Buenos Aires. ------. 2005. "Informe sobre los programas de empleo provinciales 2004." Documento de trabajo GP/16, Ministerio de Economía y Producción, Buenos Aires. ------. 2006. "Informe sobre los programas de empleo provinciales 2005." Documento de trabajo GP/17, Ministerio de Economía y Producción, Buenos Aires. ------. 2007. "Series de Gasto Público Consolidado por finalidad-función (1980­2006)." Ministerio de Economía y Producción, Buenos Aires. Dirección de Gastos Sociales Consolidados. 2001. "Informe sobre los programas de empleo provinciales." Documento de trabajo GP/9, Ministerio de Economía y Producción, Buenos Aires. ------. 2002. "Informe sobre los programas de empleo de ejecución provincial 2001." Documento de trabajo GP/13, Ministerio de Economía y Producción, Buenos Aires. ------. 2003. "Informe sobre los programas de empleo de ejecución provincial 2002." Documento de trabajo GP/14, Ministerio de Economía y Producción, Buenos Aires. Eissa, N., H. Kleven, and C. Kreiner. 2004. "Evaluation of Four Tax Reforms in the United States: Labor Supply and Welfare Effects for Single Mothers." NBER Working Paper 10935, National Bureau of Economic Research, Cambridge, MA. Ferreira, F., and P. Leite. 2003. "Meeting the Millennium Development Goals in Brazil: Can Microeconomic Simulations Help?" Economía (Spring). Fiszbein, A. 2005. "Beyond Truncated Welfare States: Quo Vadis Latin America?" Mimeo, World Bank, Washington, DC. Galasso, E., and M. Ravallion. 2004. "Social Protection in a Crisis: Argentina's Plan Jefes y Jefas." World Bank Economic Review 18 (3): 367­99. García, María Cristina, and Marcelo Langieri. 2003. "Análisis del Funcionamiento de los Consejos Consultivos del Plan Jefas y Jefes de Hogar." OIT, Proyecto Crisis. Serie Documentos de Trabajo No3, World Bank, Buenos Aires. Goldschmit, Ariela. 2008. "Income Support Policies in Argentina. Provincial Policies." Background paper, World Bank, Buenos Aires. INDEC (Instituto Nacional de Estadísticas y Censos). 2001. Press Reports. (http://www.indec.gov.ar/). ------. 2007. "Valorización mensual de la canasta básica alimentaria y de la canasta básica total--Aglomerado Gran Buenos Aires--Octubre de 2007." Comunicado de prensa, ISSN 0327-7968, Buenos Aires. Isuani, Ernesto. 1985. "Los orígenes conflictivos de la seguridad social argentina." En Biblioteca Política Argentina No.129, Centro Editor de América Latina, Buenos Aires. 80 A World Bank Country Study Jalan, J., and M. Ravallion. 1999. "Income Gains to the Poor from Workfare--Estimates for Argentina's Trabajar Program." Policy Research Working Paper 2149, World Bank, Washington, DC. Lindert, P. 2003. "Why the Welfare State Looks Like a Free Lunch." NBER Working Paper 9869, National Bureau of Economic Research, Cambridge, MA. Mesa Lago, C. 1978. Social Security in Latin America: Pressure Groups, Stratification, and Inequality. Pittsburgh: Pittsburgh University Press. MTESS (Ministry of Labor, Employment and Social Security). 2002. "Evaluación de los beneficiarios del Plan Jefes." Ministerio de Trabajo, Empleo y Seguridad Social, Mimeo. ------. 2003. "Plan Jefas y Jefes de Hogar Desocupados: Un año de gestión, Mayo 2002­ Mayo 2003." Ministerio de Trabajo, Empleo y Seguridad Social, Buenos Aires. ------. 2004. Segunda evaluación del Programa Jefes de Hogar--Resultados de la encuesta a beneficiarios. Buenos Aires: Ministerio de Trabajo, Empleo y Seguridad, Subsecretaría de Programación Técnica y Estudios Laborales. Monza, A., and C. Giacometti. 2003. "Los beneficiarios del Plan Jefes y Jefas de Hogar." Proyecto Cooperación OIT/Gobierno Argentino (MTSS): Enfrentando los retos al trabajo decente en la crisis argentina. Serie Documentos de Trabajo No 1. Moreno, Juan Martin (2007). "Se acuerda abuelo cuando las jubilaciones en Argentina eran bismarckianas?". Paper presented at the IX congress of AEPA, Cordoba, Argentina. Ortiz, Isabel. 2001. Social Protection in Asia and the Pacific. Manila: Asian Development Bank. Pautassi, Laura. 2003. "Plan Jefes y Jefas ¿Derecho social o beneficio sin derechos?" CELS. Pereyra, F. 2006. "Relevamiento documental sobre propuestas en materia de programas de transferencia de ingresos en Argentina." Background paper, World Bank, Buenos Aires. Ravallion, Martin, Emanuela Galasso, Teodoro Lazo, and Ernesto Philipp. 2001. "Do Workfare Participants Recover Quickly from Retrenchment?" Policy Research Working Paper 2672, World Bank, Washington, DC. Ravallion, Martin, Emanuela Galasso, and Agustín Salvia. 2001.: "Assisting the Transition from Workfare to Work: Argentina's Proempleo Experiment." Policy Research Working Paper 2738, World Bank, Washington, DC. Rofman, Rafael, and Leonardo Lucchetti. 2006. "Pension Systems in Latin America. Concepts and Measurements of Coverage." Social Protection Discussion Paper 0616, World Bank, Washington, DC. Schultess, Walter. 1990. "Sistema Nacional de Previsión Social. Su evolución y situación a fines de la década de los ochenta" BIRF Report ARG/88/055, Buenos Aires. World Bank. 2008a Informal Employment in Argentina: Causes and Consequences. Washington, DC: World Bank. ------. 2008b. Today's Youth--An Untapped Potential. Washington, DC: World Bank. Appendixes 81 82 APPENDIX A Appendix A: Methodology of the EPPS T he sample for the EPPS is representative of households on a national and regional level, in rural and urban areas, in mid-2007. The survey was developed according to a multiphase probabilistic design, with 2,500 face-to-face interviews with heads of households or spouses, living in residential households in Argentina. The survey opted to interview only heads of household or their spouses given that these respondents are the decision makers within the family. The questionnaire focused on capturing the perceptions of the general population on social programs, understood as income transfer programs. The EPPS questionnaire was designed around three main themes: knowledge and evaluation of current initiatives, broad normative perspectives about social programs, and perspectives and needs of the low-income population on several features of social policy. According to this goal, the questionnaire included a specific module for this population group. In order to gain national representation, the least-populated regions were included, as can be deduced from the comparison in the table of the number of homes in Argentina according to the National Population and Housing Census in 2001 and the number of surveys carried out in each region. In this way, the resulting sample is representative nationally and for each of the national regions considered (Metropolitan, Pampas, NEA, NOA, Cuyo, and Patagonia) (table A1.1). Likewise, the sample was designed to represent urban and rural areas, which results in a slight subrepresentation of the strata with the highest number of inhabitants, as shown in table A1.2. To control the imbalances brought about by the representation requirements of the sample, researchers constructed and used a series of adjustments to reach the empirical results shown in this document. Table A1.1 Distribution of Homes by Region and Surveys Carried Out REGION Households Households (%) Surveys Surveys (%) Metropolitan 03,557,488 035 0700 028 Pampas 03,545,541 035 0360 014 NEA 00,991,386 010 0360 014 NOA 00,839,646 008 0360 014 Cuyo 00,660,824 007 0360 014 Patagonia 00,479,049 005 0360 014 Total 10,073,934 100 2500 100 Source: National Population and Housing Census, INDEC 2001. 83 84 A World Bank Country Study Table A1.2 Distribution of Homes by Strata and Surveys Carried Out Locality with number Surveys Households STRATA of households Households Surveys (%) (%) 1 125,000 or more 04,955,978 049 1,025 041 2 25,000­124,999 01,585,227 016 0590 024 3 5,000­24,999 01,777,142 018 0360 014 4 500­4,999 01,228,709 012 0385 015 5 Less than 500 00,526,878 005 0140 006 Total 10,073,934 100 2500 100 Source: National Population and Housing Census, INDEC 2001. APPENDIX B Appendix B: Provincial Consultations: Process and Methodology T his appendix provides additional information on the methodology and approach of the provincial consultations undertaken by the UTDT team in partnership with the World Bank in 2007. Specifically, it covers the criteria used for selecting the provinces and interviewees, and the approach to constructing the "three visions" described here. Selection of Provinces and Key Informants The selection of the 10 provinces included in the study was based on three criteria aimed at identifying the provinces with one of the following characteristics: (i) the largest share of the population under the poverty line, (ii) conditions in which poverty indicators have worsened relative to the rest of the population, or (iii) less investment in social expenditures than the rest of the country. The first indicator used was the distance of the province from the national poverty rate. This measure was calculated as the percentage-point difference between the provincial poverty rate and the national average. Researchers selected those provinces whose share of the population under the poverty line was above the national average during at least 50 percent of the period between May 2001 and the second half of 2005. The second indicator was the recent trend in the poverty rate in each province. This indicator was based on the variation of the poverty rate in each of the EPH surveys from the value immediately preceding it in the series. The calculations were done at the provincial and national levels to compare the variation of each province. Researchers selected provinces where the number of people below the poverty line increased more or decreased less than at the national level. Finally, the third indicator used for selecting provinces was per capita social expenditures. Provincial data for 2001­04 were taken from the Direccion National de Coordinación Fiscal for analysis, and the provinces with the lowest level of social expenditures were selected. Based on these three indicators, 10 provinces were selected: Buenos Aires, Chaco, Chubut, Corrientes, Mendoza, Misiones, Rio Negro, Salta, Santa Fe, and Tucuman. The provinces were divided into three groups and covered by three research teams who undertook the field research. The first team, led by Carolina Garber, undertook the fieldwork in Chaco, Corrientes, Mendoza, and Santa Fe. A second team, led by Martin Lardone, worked in Mision, Salta, and Tucuman. Finally, the third group, coordinated by German Perez, did the fieldwork in Buenos Aires, Chubut, and Rio Negro. The research was under the overall guidance of Catalina Smulovitz and Alejandro Bonvecchi. 85 86 A World Bank Country Study The actors interviewed in each of the provinces were key informants selected for their direct involvement in the formulation or implementation of social policy in their respective provinces. Although their opinions are not statistically representative of the views of the population in their provinces, they provide insight into the views of stakeholders in the area. In each province, the teams tried to capture the opinions of public officials directly involved in the management of social policies, representatives of civil society, representatives of workers' and employers' associations, public officials at different levels of government (national, provincial, and municipal), and representatives of the opposition at the provincial level. In summary, the profiles sought in each province were as follows: Governor, minister, or principal political adviser to the provincial government Minister of social development, health, education, employment, or finance at the province level Representative from an employers' association, the chamber of commerce, or the most important or largest firm in the province Representative of Caritas, the Catholic Church, or another faith-based organization Political leader of the principal opposition party at the provincial level Mayor of an important urban center in the province Trade union leader or piquetero (association of unemployed), or representative of an NGO involved in social policy Methodology The process used for analyzing the interviews and defining the visions was inspired by Weber's approach of formulating "ideal types." The use of the ideal types is particularly helpful for organizing the results of the qualitative research, to identify common themes in the interviews, explore connections, and distinguish conceptual patterns. These patterns help filter out information that is less relevant for the focus of the study--in this case visions of social policy--and help organize the relevant information in a way that responds to the research question. The process of constructing the three visions used the following steps. First, researchers reviewed reports of the provincial interviews and classified the responses according to priorities, measures, resources, and challenges. Second, the interview responses were classified according to themes relevant to social policy, such as views about income transfers; the role of the nation, province, and municipality in the design of programs; management and financing of policies; the type of nongovernmental actors who should participate in the design and management of these policies; the types of policies and specific content that should be considered; the effects of decentralization on social policy; and so forth. These themes were then analyzed and categorized based on the type of actor and the socioeconomic conditions of the provinces. APPENDIX C Appendix C: Methodology of the Behavioral Simulation T he ex ante behavioral evaluation for the family allowance was modeled according to Ferreira and Leite's (2003) specifications for the Bolsa Escola program in Brazil. This approach imposes little structural specification on behavior in contrast to models based on demand for schooling for the intrahousehold labor allocation. Ferreira and Leite estimated a discrete occupational choice model--multinomial logit, which has three possible states: the child does not go to school and works full-time (inside or outside the household); the child goes to school and works part-time; or the child only goes to school. Then they used the estimated parameters to make predictions about the counterfactual occupational decisions of children. Because this procedure requires child earnings even for those children that do not work, a Mincerian earnings equation was estimated to obtain potential earnings. Because the income values accruing to each household are not symmetric across different occupational choices, standard estimation procedures for the multinomial logit are not valid. An identification assumption was made: children working outside the household and not attending school have zero domestic production. See Ferreira and Leite for methodological details. Some particularities of the relationship between children and the labor market require careful attention to the assumptions. About 19 percent of children who qualified for the family allowances do not attend school. However, only 35 percent of them are working outside the household and receive an income. This percentage decreases to 6 percent for those children who attend school. Therefore, the occupational choice model would include potential earnings for most of the target population. Despite this fact, the estimation parameters seem to be consistent with the assumptions. The Mincerian earnings model is estimated for the group ages 12­18: earnings and school attendance are negatively correlated. A dummy variable was included to identify those children who work and attend school. This simple specification of the earnings function reveals that a child working and studying will earn less income than a child who only works. If this coefficient is assumed to reflect fewer hours of work, then a child going to school works on average 35 percent less than a dropout. These results are similar to those obtained by the authors. Potential earnings for children who do not work (outside the household or at all) were obtained using the Mincerian regression; thus, the occupational choice estimation has all the necessary inputs. Considering that the reference category is "not attending school," the table presents the results for the two remaining states. The age reduces the probability to studying for both groups (working and not working), while years of 87 88 A World Bank Country Study education enhances it. Children's overage at school makes it possible. Another important result is that potential children's earnings are negatively associated with the probability of working and studying, but no statistical correlation is observed for those who only attend school. Table A3.1 Log Earnings Regression for Reported Earnings of Children Ages 12­18 Variable Coefficient Standard deviation P>t Years of education 0.035 0.012 0.003 Male 0.434 0.062 0.000 Age 0.179 0.023 0.000 Dummy (working and studying) -0.347 0.063 0.000 Intercept 1.931 0.360 0.000 Source: World Bank staff calculations based on EPH. 89 Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Endangered Forests and natural resources. We print World Bank Working Papers and Country Studies on postconsumer recycled paper, processed chlorine free. The World Bank has formally agreed to follow the recommended standards for paper usage set by Green Press Initiative--a nonprofit program supporting publishers in using fiber that is not sourced from Endangered Forests. For more information, visit www.greenpressinitiative.org. In 2008, the printing of these books on recycled paper saved the following: Trees* Solid Waste Water Net Greenhouse Gases Total Energy 355 16,663 129,550 31,256 247 mil. *40 feet in height and Pounds CO2 6­8 inches in Pounds Gallons BTUs Equivalent diameter 90 Argentina: Income Support Policies toward the Bicentennial is part of the World Bank Country Study series. These reports are published with the approval of the subject gov- ernment to communicate the results of the Bank's work on the economic and related conditions of member countries to governments and to the development community. Following the serious economic crisis in 2001­02, Argentina mobilized an unprecedented effort to provide income sup- port to the segment of the population most in need. Now, as growth has returned and social indicators recovered to pre- crisis levels, there is an opening to move from emergency support programs to a more comprehensive, long-term, and sustainable strategy for social protection. The challenge is to design and fully implement a social protection system that has adequate coverage and benefits and is integrated and fiscally and politically sustainable. The analysis contained in this book aimed to contribute to and inform the debate about the future of income support policies in Argentina, taking the views, values, and prefer- ences of the stakeholders and the population as starting points. The research included two innovative efforts to col- lect and understand the landscape of ideas regarding options for social protection circulating in Argentina: first, an exten- sive set of consultations with policy makers and practition- ers in social policy, mainly at the provincial level; and second, a national, representative opinion survey on the views and perceptions of the population regarding social policy and income support programs in particular. World Bank Country Studies are available individually or by subscription, both in print and online. ISBN 978-0-8213-8051-2 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org SKU 18051 E-mail: feedback@worldbank.org